University of Colorado, Boulder

CU Scholar
Civil Engineering Graduate Theses & Dissertations Civil, Environmental, and Architectural Engineering

Spring 1-1-2017

Measuring and Managing Construction Worker

Fatigue
Ulises Daniel Techera
University of Colorado at Boulder, [email protected]

Follow this and additional works at: https://scholar.colorado.edu/cven_gradetds

Part of the Civil Engineering Commons, and the Construction Engineering and Management
Commons

Recommended Citation
Techera, Ulises Daniel, "Measuring and Managing Construction Worker Fatigue" (2017). Civil Engineering Graduate Theses &
Dissertations. 98.
https://scholar.colorado.edu/cven_gradetds/98

This Dissertation is brought to you for free and open access by Civil, Environmental, and Architectural Engineering at CU Scholar. It has been accepted
for inclusion in Civil Engineering Graduate Theses & Dissertations by an authorized administrator of CU Scholar. For more information, please contact
[email protected].
 MEASURING AND MANAGING
CONSTRUCTION WORKER FATIGUE

by

ULISES D. TECHERA
B.S. Civil Eng. Polytechnic University of Catalonia, 2014
M.S. Civil Eng. Polytechnic University of Catalonia, 2014

A DISSERTATION
SUBMITTED TO THE DEPARTMENT OF CIVIL,
ENVIRONMENTAL, AND ARCHITECTURAL ENGINEERING
AND THE COMMITTEE ON GRADUATE STUDIES OF THE
UNIVERSITY OF COLORADO BOULDER
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
DEGREE OF DOCTOR OF PHILOSOPHY

April 2017

i
 This dissertation entitled:
Measuring and Managing Construction Worker Fatigue
Written by Ulises D. Techera
has been approved for the Department of Civil, Environmental, and Architectural Engineering

_____________________________
Professor Matthew R. Hallowell, Chair

_____________________________
Professor Ray L. Littlejohn

_____________________________
Professor Paul M. Goodrum

_____________________________
Professor Sathyanarayanan Rajendran

_____________________________
Professor Eric Marks

Date_______________________

The final copy of this dissertation has been examined by the signatories, and we find that both
the content and the form meet acceptable presentation standards of scholarly work in the above-
mentioned discipline

IRB protocol # 160354

ii
ABSTRACT
Ulises D. Techera (Ph.D., Civil Engineering)

Measuring and Managing Construction Worker Fatigue

Dissertation directed by: Professor. Matthew R. Hallowell

The construction industry presents a fatality rate three times greater than the average

considering all industries. Electrical transmission and distribution work presents one of the highest

fatality rates inside the construction industry. Such accidents come at a great economic and social

cost. Decades of accident causation research demonstrate that organizational and human factors,

rather than technical failures, are the principal causes of accidents. Fatigue showed to be a

significant trigger to human error, accident causation, and a bundle of other safety risks. This

dissertation represents the first research effort to meta-analyze the causes and consequences of

occupational fatigue and address the way in which fatigue can be identified, predicted and

managed for electrical transmission and distribution (TD) workers. Over the course of a year and

a half, a total of 343 TD workers, distributed across the US, participated in interviews, surveys,

and tests to accomplish the purpose of this dissertation. Additionally, a group of 52 general

construction (GC) workers also took part in the data collection process. The data was coded and

analyzed applying several statistical methods such as: Meta-analysis, Chi-square test, Proportion

test, Correspondence analysis, and Multiple Linear Regression. The results identified 9 of the

principal causes and 5 of the most relevant consequences of occupational fatigue together with

their relative impact. Furthermore, extreme temperatures and long shifts where identified as the

major causes for TD workers´ fatigue and loss of attention and slowing down were recognized as

significant consequences of fatigue among TD workers. Additionally, current fatigue identification

iii
and management techniques where documented. Lastly, fatigue predictive models for TD workers

and GC workers were created based on empirical data collected in the field. The level of

predictability of these models was low to medium, indicating that additional predictors need to be

identified. Fatigue predictors, as measured by two of the most reliable and valid tools to objectively

and subjectively assess fatigue, showed to vary between TD and GC workers. However, sleep

deprivation showed to be a common predictor. Future research should engage in the strengthening

of these models as well as the study of fatigue among other trades.

iv
DEDICATION

To my parents, Lucía Rocha and Daniel Techera, for showing me the value of obtaining

knowledge, for their example of hard work, and for their unconditional and tender love. To my

grandparents, Walter Techera, Irma Mederos, Sixto Rocha, and Dinorah Morales, for their

encouragement to achieve high goals. To my dear and beloved wife, Mary Techera, for her support,

help, patience, and love during this challenging journey. To my son, Walter Ulises Techera, who

filled my arms with love and always waited anxiously for my return. To my sons and daughters to

come, as a testimony that with hard work and the guidance of people who trust you, any goal can

be achieved in life. Additionally, to my beloved little sisters, Rafaela and Niza Techera, who by

looking up to me pushed me to reach farther. Last but not least, to my Heavenly Father, who

sustains me every day and watches over me.

Para mis padres, Lucía Rocha y Daniel Techera, por haberme mostrado la importancia de

obtener sabiduría, por su ejemplo de trabajo arduo, y por su amor tierno e incondicional. Para

mis abuelos, Walter Techera, Irma Mederos, Sixto Rocha, y Dinorah Morales, por darme aliento

para alcanzar elevadas metas. Para mi querida y amada esposa, Mary Techera, por su respaldo,

ayuda, paciencia, y amor durante este desafiante camino. Para mi hijo, Walter Ulises Techera,

quien llenó mis brazos de amor y siempre esperó ansioso por mi regreso a casa. Para mis futuros

hijos e hijas, como testimonio de que con esfuerzo y el apoyo de gente que confía en ti, cualquier

meta puede lograrse en la vida. Adicionalmente, para mis queridas hermanas, Rafaela y Niza

Techera, quienes al mirar hacia mí por un buen ejemplo me han animado a llegar más lejos. Por

último, para mi Padre Celestial, quien me sostiene y cuida cada día de mi vida.

v
ACKNOWLEDGEMENTS
I would like to express my deepest gratitude to my advisor Dr. Matthew R. Hallowell for

trusting me even before we met. For his exemplary guidance, his constant support and

encouragement, his words of wisdom delivered at the right time, and for always being by my side.

Also, I must express my gratitude to Dr. Ray Littlejohn for his friendship, constant advice,

and support during countless hours of statistical analyses. For lighting in me a passion for statistics

and for helping me understand fundamental concepts.

Moreover, I desire to acknowledge Dr. Sathyanarayanan Rajendran, who also went far

beyond his duties as a committee member and offered immense help to collect data. I would have

not accomplished such a populated and diverse sample without his help.

In addition, I must acknowledge Dr. Goodrum, who allowed me to use his computers to

collect data for several months, and for his exemplary teaching. I would have not achieved this

goal without his help.

Likewise, I wish to express my gratitude to Dr. Eric Marks, who helped me broaden the

perspective of my research and gave me important conceptual guidance when I needed it.

I also wish to acknowledge the industry leaders and workers who gave their time, effort

and support to collect the data that is the basis of this research. Their contribution makes this all

possible.

Last but not least, I appreciate the example, support, help, and friendship of all my fellow

students who were always present when I needed them the most.

vi
TABLE OF CONTENTS

ABSTRACT ................................................................................................................................... iii

DEDICATION ................................................................................................................................ v

ACKNOWLEDGEMENTS ........................................................................................................... vi

TABLE OF CONTENTS .............................................................................................................. vii

LIST OF TABLES ........................................................................................................................ xii

LIST OF FIGURES ..................................................................................................................... xiii

Chapter 1: Introduction ................................................................................................................... 1

1.1Background and Motivation: ................................................................................................. 1

1.2 Current needs for additional research: .................................................................................. 6

1.3 Dissertation organization ...................................................................................................... 7

1.4 Dissertation content and contributions.................................................................................. 7

1.5 References ........................................................................................................................... 12

Chapter 2: Causes and consequences of occupational fatigue: Meta-analysis and systems model.
....................................................................................................................................................... 14

2.1 Abstract: .............................................................................................................................. 14

2.2 Introduction: ........................................................................................................................ 15

2.3 Fatigue definition and classification: .................................................................................. 17

2.4 Causes of fatigue: ................................................................................................................ 18

2.4.1 Sleep deprivation ......................................................................................................... 19

2.4.2 Mental Exertion ........................................................................................................... 20

2.4.3 Muscular Exertion ........................................................................................................ 21

2.4.4 Work Load Characteristics .......................................................................................... 22

2.4.5 Overtime and Long Work Hours (LWH)..................................................................... 22

vii
 2.4.6 Incomplete recovery..................................................................................................... 24

2.4.7 At work environmental Factors ................................................................................... 25

2.4.8 Social environment at the workplace ........................................................................... 25

2.4.9 Emotional predisposition and distress ......................................................................... 26

2.5 Relationships among causes of fatigue ............................................................................... 28

2.5.1 Impacts of increased work load on other causal factors .............................................. 28

2.5.2 Impacts of overtime and LWH on other causal factors ............................................... 29

2.6 Consequences of fatigue: .................................................................................................... 30

2.6.1 Mood changes .............................................................................................................. 31

2.6.2 Cognitive degradation .................................................................................................. 32

2.6.3 Attention and concentration ......................................................................................... 33

2.6.4 Reaction time ............................................................................................................... 34

2.6.5 Physical degradation and pain ..................................................................................... 34

2.6.6 Illnesses ........................................................................................................................ 35

2.6.7 Human error and injuries ............................................................................................. 36

2.7 Relationships among outcomes of fatigue: ......................................................................... 37

2.7.1 Impacts of cognitive degradation on other outcomes .................................................. 38

2.7.2 Impacts of physical degradation on other outcomes .................................................... 38

2.7.3 Impacts of illnesses on other outcomes ....................................................................... 39

2.8 Systems model: ................................................................................................................... 39

2.9 Model limitations ................................................................................................................ 42

2.10 Meta-analysis approach .................................................................................................... 43

2.10.1 Studies selection......................................................................................................... 44

2.10.2 Studies coding ............................................................................................................ 45

2.10.3 Standardization .......................................................................................................... 46

viii
 2.10.4 Overall effect size calculation .................................................................................... 47

2.11 Results of the meta-analysis.............................................................................................. 49

2.12 Discussion of the meta-analysis: ....................................................................................... 53

2.13 Conclusions and recommendations: ................................................................................. 54

2.14 Acknowledgements ........................................................................................................... 56

2.15 References ......................................................................................................................... 56

Chapter 3: Fatigue Management in Electrical Transmission and Distribution Work ................. 63

3.1 Abstract ............................................................................................................................... 63

3.2 Introduction ......................................................................................................................... 63

3.3 Background ......................................................................................................................... 65

3.3.1 The importance of TD operations ................................................................................ 65

3.3.2 Safety in TD work ........................................................................................................ 66

3.3.3 The influence of fatigue on safety ............................................................................... 67

3.3.4 Causes and consequences of occupational fatigue....................................................... 68

3.4 Point of departure and contribution to knowledge .............................................................. 70

3.5 Research methods ............................................................................................................... 71

3.5.1 Development of questionnaires.................................................................................... 71

3.5.2 Respondent demographics ........................................................................................... 72

3.5.3 Data manipulation and statistical analysis: .................................................................. 73

3.6 Results and analysis ............................................................................................................ 74

3.6.1 Results from TD workers: ............................................................................................ 74

3.6.2 Results from TD supervisors: ...................................................................................... 86

3.7 Conclusion .......................................................................................................................... 88

3.8 Acknowledgements ............................................................................................................. 91

3.9 References ........................................................................................................................... 91

ix
Chapter 4: Measuring and Predicting Fatigue in Construction Workers: An Empirical
Field Study .................................................................................................................................... 96

4.1 Abstract ............................................................................................................................... 96

4.2 Introduction ......................................................................................................................... 97

4.3 Background ......................................................................................................................... 98

4.3.1 Potential Predictors of Fatigue ..................................................................................... 98

4.3.2 Sleep Deprivation......................................................................................................... 99

4.3.3 Incomplete Recovery ................................................................................................. 100

4.3.4 Workload.................................................................................................................... 100

4.3.5 Overtime and Long Shifts .......................................................................................... 100

4.3.6 Work Environment..................................................................................................... 101

4.3.7 Mental Exertion ......................................................................................................... 101

4.3.8 Muscular Exterior ...................................................................................................... 102

4.3.9 Social Work Environment.......................................................................................... 102

4.3.10 Personal Predisposition ............................................................................................ 102

4.4 Fatigue measurement tools ............................................................................................... 103

4.4.1 Subjective fatigue measurement ................................................................................ 104

4.4.2 Objective fatigue measurement.................................................................................. 108

4.5 Point of departure .............................................................................................................. 111

4.6 Methodology ..................................................................................................................... 111

4.6.1 Research tools development and selection ................................................................ 112

4.6.2 Sample Characteristics and Data Collection .............................................................. 114

4.6.3 Data Analysis ............................................................................................................. 115

4.7 Results and analysis .......................................................................................................... 117

4.7.1 Interactions ................................................................................................................. 125

x
 4.7.2 Limitations of the study ............................................................................................. 127

4.8 Conclusion ........................................................................................................................ 128

4.9 References ......................................................................................................................... 130

4.10 Appendix ......................................................................................................................... 136

Chapter 5: Conclusion................................................................................................................. 137

5.1 Contributions..................................................................................................................... 138

5.2 Suggestions for future research ......................................................................................... 142

5.2.1 Limitations of current data sets .................................................................................. 142

5.2.2 Future research based on findings.............................................................................. 143

5.2.3 Future research of fatigue in construction ................................................................. 144

5.3 Personal Reflection ........................................................................................................... 145

5.3.1 About research in the construction industry .............................................................. 145

5.3.2 About the journey to become a Ph.D. ........................................................................ 146

5.4 References ......................................................................................................................... 149

Bibliography ............................................................................................................................... 151

Appendix A: IRB Protocol.......................................................................................................... 162

Appendix B: Consent Form ........................................................................................................ 181

Appendix C: Activity Questionnaire .......................................................................................... 184

Appendix D: Swedish Occupational Fatigue Inventory ............................................................. 186

xi
LIST OF TABLES

Table 1 Causes of fatigue .............................................................................................................. 27

Table 2 Relationships among causes of fatigue ............................................................................ 30
Table 3 Consequences of fatigue .................................................................................................. 36
Table 4 Relationships among consequences of fatigue ................................................................ 39
Table 5 Effects of causes of fatigue .............................................................................................. 50
Table 6 Effects among causes of fatigue ...................................................................................... 50
Table 7 Effects of fatigue .............................................................................................................. 51
Table 8 Effects among consequences of fatigue ........................................................................... 51
Table 9 Effect Size of causes of fatigue........................................................................................ 69
Table 10 Effect Size of consequences of fatigue .......................................................................... 69
Table 11 TD workers and supervisors’ demographics ................................................................. 73
Table 12 Overall results from electrical workers .......................................................................... 75
Table 13 Percentage of accidents attributed to fatigue by worker ................................................ 76
Table 14 Self-reported level of fatigue ......................................................................................... 76
Table 15 One-way Chi-square test and proportion test results for underground workers ............ 77
Table 16 Chi-square test and proportion test results for substation workers ................................ 77
Table 17 Summary of results from TD supervisors ...................................................................... 88
Table 18 Potential predictors of fatigue ...................................................................................... 112
Table 19 MLR and Validation results for PVT data ................................................................... 119
Table 20 MLR and Validation results for SOFI data .................................................................. 119
Table 21 Predictors by fatigue measurement method and worker type ...................................... 123
Table 22 Descriptive statistics of continuous predictive variables ............................................. 125
Table 23 Frequencies of dichotomous predictive variables ........................................................ 125
Table 24 descriptive statistics of dependent variables ................................................................ 125
Table 25 Interaction’s analysis for a PVT model among TD workers ....................................... 126
Table 26 Interaction’s analysis for a SOFI model among TD workers ...................................... 126
Table 27 Descriptive statistics of continuous activity variables ................................................. 136
Table 28 Frequencies of dichotomous activity variables ............................................................ 136

xii
LIST OF FIGURES
Figure 1 Research questions and dissertation contributions ......................................................... 11
Figure 2: Systems model illustrating empirical literature on the relationships among causes and
effects of fatigue. .......................................................................................................................... 41
Figure 3: Systems model represents effect sizes (Cohen’s d values) for each relationship. ........ 52
Figure 4 Row principal normalization of consequences of fatigue and trade............................... 83
Figure 5 Symmetric normalization of consequences of fatigue and trade .................................... 84
Figure 6 Row principal normalization for identification techniques and trades ........................... 85
Figure 7 Symmetrical normalization for identification techniques and trades ............................. 86
Figure 8 Statistical Analysis ....................................................................................................... 117
Figure 9. Scatterplot matrix of Median RT predictors for GC workers...................................... 121
Figure 10 Dissertation contributions ........................................................................................... 141

xiii
Chapter 1: Introduction

1.1Background and Motivation:

The construction sector is one of the most hazardous and deadly of the industry (Sawacha

et al. 1999). According to the U.S. Bureau of Labor Statistics (2015) in 2014 there were a total of,

4,679 fatal work injuries, which equates to an all-industry and government jobs fatality rate of 3.3

fatalities for every 100,000 full-time equivalent workers (FEW). That same year, the construction

industry showed a fatality rate of 9.4 for every 100,000 FEW; thus, presenting a fatality rate nearly

3 times greater than that of the overall working population. Three of the ten most dangerous jobs

in the U.S. belong to the construction industry revealing its dangerous nature. From these 3

occupations, Electrical Transmission and Distribution workers (TD workers) experienced 19.2

fatalities for every 100,000 FEW during the year 2014. Furthermore, the average fatality rate for

this sector of the industry over the last 20 years is 26.1 fatalities every 100,000 FEW. These

statistics indicate that TD work is between 6 and 8 times more deadly than the average considering

all other professions (U.S. Bureau of Labor Statistics 2015). Unfortunately, construction related

accidents don’t only affect construction workers but also the public in general. In the U.K. on

average, every month a member of the public is killed and 1200 people are injured as a

consequence of a construction related accident (Sawacha et al. 1999).

Construction accidents come at a great economic cost for both the industry and society.

Considering direct costs (medical cost), indirect costs (wages and household productivity), and the

quality of life costs due to injury, the average construction fatality was estimated to cost $4 million.

Regarding non-fatal accidents, the average injury with days away from work was estimated to cost

1
$42,000, while restricted work and no-lost-work cases were estimated to cost $618 and $ 777

respectively (Waehrer et al. 2007). All these values exceed the equivalent values for the overall

private sector. According to data from 2002, the construction industry accounted for 5.2% of all

private industry employment but it was responsible for 15% of all private injury cost. All of which

indicates that the construction industry is disproportionally costly in regards to accidents and

fatalities. This translates to a yearly cost of approximately $11.5 billion (Waehrer et al. 2007).

There are several characteristics of construction work that contribute to its hazardous

nature. Construction work is constantly changing. The different stages of a project and the

uniqueness and complexity of each worksite require constant adaptation to a new environment and

a different way of performing activities. This jeopardizes the process of becoming familiar with

the environment to the point where most possible hazards are recognized (Buchholz et al. 1996).

Furthermore, this kind of job commonly requires working in awkward postures, in confined or

dangerous spaces, lifting heavy equipment or performing forceful exertions (Schneider and Susi

1994).

Accidents have shown to be caused either by unsafe human acts and/or an unsafe design

that generates a physical hazard (Kartam 1997). Furthermore, several decades of construction

safety research have revealed that organizational and human factors, rather than technical failures,

are the principal causes of accidents (Langford et al. 2000; Weick et al. 2008). This offers a

possible explanation for the fact that despite the numerous advances in technology and

communication during the past decades, which allowed for better equipment and training,

accidents and fatality in construction still occur at a disproportionate rate. All of this makes it

imperative to focus on the human aspect of accident causation.

2
 Fatigue plays a significant role in the causes of occupational injuries and fatalities,

particularly in high-energy situations. Fatigue can be defined as a decreased ability to perform

activities at the desired level due to lassitude or exhaustion of mental and/or physical strength

(Hallowell et al. 2010; Gander et al. 2011). When workers are fatigued, they experience

compromised alertness, judgement, reaction time, mental acuity, physical strength, and the

development of an uncooperative disposition (Gillberg and Åkerstedt 1994; Kajtna et al. 2011;

van der Linden et al. 2003a; Lorist et al. 2005; Scott et al. 2006; Yaggie and Armstrong 2004).

Such effects decrease a worker’s ability to complete their work safely due to the increased rate of

human error (Dembe et al. 2005). Thus, it is not surprising that many researchers include fatigue

as a salient factor in most accident causation models and theories (Craig 1992; Czeisler, et al.

1992 ; Lorist et al. 2000; van der Linden et al. 2003; Tixier et al. 2014). As described by Spurgeon

et al. (1997) there are mainly two reasons why fatigue compromises safety in an occupational

environment. First, fatigue diminishes the ability of an individual to perceive and react to new

information. (Grandjean 1979; Johnston et al. 1998; Lorist et al. 2000; Reiner and Krupinski 2011).

Second, fatigue diminishes the ability of an individual to perceive risk, and therefore,

subconsciously this individual assumes higher risk than he/she normally would under a non-

fatigued condition (Tixier et al. 2014). Along these lines, some authors have suggested that fatigue

may affect hazard recognition abilities given the fact that fatigued individuals assume higher risks;

however, as of now, no study has specifically studied the relationship between fatigue and hazard

recognition ability among either industry workers in general or construction workers.

3
 Despite the importance of fatigue in accident causation recognized by numerous

researchers, fatigue has remained as a latent, hidden variable which rarely appears in injury reports

and it is generally not mentioned until a major industry accident occurs (Harrington 2001; Reiner

and Krupinski 2011). Perhaps the familiarity with the phenomenon (fatigue) has made people

unaware of its real implications and effects. Furthermore, its dual nature presenting subjective and

objective qualities has contributed to the lack of a formal and universal definition of fatigue. This

lack of precision and formality has characterized fatigue related research until recent years.

The inability to properly define fatigue has inhibited the development of a cohesive list of

causes and consequences of fatigue. This has also compromised the ability to properly assess or

measure fatigue among individuals. Consequently, the current body of knowledge about fatigue

has remained disperse and disorganized. Researchers have focused on different aspects of fatigue,

thus achieving a deeper understanding of specific aspects, but no overall understanding of fatigue

is shared by the whole research community. This greatly diminishes the ability to properly manage

fatigue in the industry.

Early fatigue research focusses primarily on fatigue developed by patients of terminal or

chronic diseases. Consequently, original research addressed fatigue in clinical settings among

patients, and it studied mostly long term consequences of fatigue. In more recent years, fatigue

research has also focused on healthy and working populations, but this research focusses primarily

in the organizational, and managerial side of mitigating fatigue risk, giving place to good quality

products such as Fatigue Risk Management Systems (FRMS) and fatigue proofing techniques

4
(Dawson et al. 2012; Gander et al. 2011). But even this latter research is in need of a deeper

understanding of causes and consequences of fatigue among workers.

Currently, there is no study that addresses all principal causes and consequences of fatigue

in an occupational setting. Furthermore, there is no study that allows for comparison among

different causes and consequences of fatigue; therefore, it is difficult to establish an order of

importance among these causes and consequences. Lastly, very few studies address fatigue among

construction workers. These are trade specific studies and they reveal a need for further fatigue

research among other high risk construction trades. For example Chan (2011) arrived to the

conclusion that fatigue is the most critical aspect in accident causation among oil and gas

construction workers. Hallowell (2010) also demonstrated that fatigue plays a major role in

accident causation among rapid renewal highway construction. However, fatigue research among

construction workers is still in its infancy and further research is needed.

The understanding of some of the negative consequences of fatigue on health and the

industry has given place to an urgent need to assess fatigue levels among individuals. This need

was primarily addressed by developing subjective fatigue questionnaires. These surveys were first

developed to assess fatigue in clinical settings among unwell people. More recently, several

subjective fatigue questionnaires were developed to assess fatigue in the industry but none of them

focuses on the construction industry specifically. On the other hand, the technological advances of

the last decades have allowed for the creation of multiple innovative devices that measure fatigue

based on a physical aspect that changes under the effects of fatigue. Some of them base their

assessment on encephalography, others on neurocognitive behavior, others on pupilometry, and

5
others on oculometry, to mention a few. However, most of these devices lack validity and peer

reviewed research to support their performance. Additionally, these instruments that objectively

assess fatigue are usually very costly and time consuming, all of which limit their application in

construction and the industry in general (Dawson et al. 2014). Consequently, there is a need for a

valid and reliable way to measure and predict fatigue in construction workers.

The objective of this doctoral dissertation is to address all the aforementioned holes in the

current body of knowledge and further advance the knowledge of fatigue among the construction

industry with the ultimate goal of saving lives and improving productivity in the construction

industry.

1.2 Current needs for additional research:

In light of existing literature, there are several gaps in the current body of knowledge about

fatigue. Furthermore, even though fatigue has shown to greatly contribute to accident causation,

fatigue research in the construction industry is barely in its infancy. Additional knowledge about

the way in which fatigue affects construction workers and projects is needed. The foremost fatigue

research needs in the construction industry are summarized below.

• Properly define fatigue for the industry.

• Identify and summarize causes and consequences of fatigue in the industry.

• Objectively quantify the impact of causes and consequences of fatigue.

• Develop empirical research among different construction trades.

• Develop a valid and reliable way to measure and predict fatigue in construction workers.

Addressing these gaps in knowledge is the purpose of this dissertation.

6
1.3 Dissertation organization

The present document is organized into 5 chapters, the first chapter consists of the

introduction where the background, motivation, point of departure, and a summary of each

conducted study is presented. In addition, the introduction presents current industry problems and

how this dissertation will contribute to solve those problems. This document also contains 3 stand-

alone studies developed by the author which address the research needs presented above. These

documents are structured in a journal paper format with their corresponding abstract, introduction,

methodology, results, and conclusion. These 3 studies can be found in chapters 2 through 4. The

last chapter (Chapter 5) of this dissertation, presents a conclusion with a summary of the

contributions to knowledge achieved by these studies and a reflection of future steps on fatigue

related research in the construction industry.

1.4 Dissertation content and contributions

This section briefly explains the research needs addressed in and the knowledge

contribution obtained from each paper.

The first paper presented in this dissertation can be found in chapter 2 under the title:

“Causes and Consequences of Occupational Fatigue: Meta-analysis and Systems Model”. This

paper addresses the need for a formal definition of fatigue and the identification and quantification

of fatigue causes and consequences in an industry setting. The paper presents a definition of fatigue

that alludes to two well-identified dimensions of fatigue which are mental and muscular.

Additionally, this paper presents an exhaustive summary of the principal causes of fatigue and

consequences of fatigue that affect industry workers, organized as a systems model where the

7
volume of research behind each variable is presented. This analysis clearly identifies holes in the

current body of knowledge and it can be used for future research as a point of departure to decide

where further research is needed. Furthermore, this paper presents the first cohesive quantification

of the impact of causes and consequences of fatigue allowing for comparison among different

variables. This quantification was obtained as a result of a hefty meta-analysis of the existing

fatigue literature. This paper has been published in the Journal of Occupational and Environmental

Medicine.

The second paper presented in this document can be found in chapter 3 under the title

“Fatigue Management in Electrical Transmission and Distribution Work”. This paper is the result

of an empirical research project among TD workers whose jobs are among the 10 most deadly

professions in the country. This paper addresses the need for further fatigue research among at-

risk construction trades. The research questions that originated this study are as follows: What

causes fatigue among TD workers? What are the consequences of fatigue in TD work? How do

workers and supervisors identify fatigue? What do they do to manage fatigue? What things could

contribute to better management of fatigue in this trade? After an empirical study with the

collaboration of 143 electrical TD workers, the researchers discovered that according to the

opinion of TD workers, “Extreme Temperatures” and “Long Shifts” are the 2 most important and

significant causes of fatigue among TD workers followed by “Lack of Sleep”, and “Heavy Manual

Labor”. Among the consequences of fatigue, “Slowing Down” had a significant effect on fatigue

among workers followed by loss of attention and concentration which was also important. These

findings reveal that fatigue not only affects safety but also productivity in the construction industry.

Additionally, this paper presents the most common ways of identifying and mitigating fatigue

8
among TD workers which can be used for application in a FRMS for this sector of the industry.

These findings shed light on the impact of fatigue among TD workers and operations which can

be used to reduce accidents and fatalities in such a hazardous sector of the construction industry.

This journal paper will be submitted for review to the Journal of Construction Engineering and

Management.

The third and last paper presented in this dissertation can be found in chapter 4 under the

title “Predicting Fatigue in Construction Workers”. Previous research identified numerous

potential predictors of fatigue. Additionally, some of those were identified as especially important

to the onsetting of fatigue by TD workers. This last study, investigates the predictive character of

the potential aforementioned predictors. Two fatigue measurement tools were selected an objective

tool and a subjective tool, to register as many dimensions of fatigue as possible; a questionnaire

with the potential predictors was created; and a group of 253 US construction workers participated

in the study.

The alternative hypothesis of the study stated that fatigue changes could be predicted by

changes in some of the potential predictors identified by previous studies. Such a hypothesis turned

out to be true and models with explained variability between 10% and 50% were developed. There

results revealed the need to identify additional predictors that may be driving the percentage of

unexplained variability. Additionally, this study allowed for the identification of specific

predictors for two different types of workers: TD workers and general construction workers.

9
 The reader will notice the dependency among these above presented research studies. Each

study builds upon the knowledge obtained in the previous study, thus expanding the current body

of knowledge. The study presented in Chapter 2 identifies and quantifies the relative impact of

principal causes and consequences of fatigue per the existing literature. The study presented in

Chapter 3 explores which of those causes and consequences play a role in the onsetting of fatigue

among construction workers according to the opinion of construction workers. Lastly, the third

paper quantifies the predictability of these causes based on the empirical assessment of these

causes and the measurement of fatigue. Such work presents a unique and cohesive contribution to

the current body of knowledge about fatigue in the construction industry. Figure 1 summarizes the

research questions for this dissertation and the contribution in knowledge of each study.

10
What are the causes and consequences of occupational fatigue?

• Organized previous research into a

systems model.

Study • Identification of fatigue causes and

consequences
•
1 •
Representation of volume of research
Quantification of relative impact of
causes and consequences of fatigue
through Meta-Analysis
•
What are the causes and consequences of fatigue in construction?
How do construction workers identify fatigue?
How do they manage fatigue?
• First empirical fatigue study among
TD workers.

Study • Identification of fatigue:

➢ Causes
➢ Consequences
2 ➢ Identification techniques
➢ Mitigation techniques

Can we predict fatigue from its causal factors?

Study • Identification of fatigue predictors in

the construction industry.
•
3 Recognition of differences in
predictors by trade.

Figure 1 Research questions and dissertation contributions

11
1.5 References
Buchholz, B., Paquet, V., Punnett, L., Lee, D., and Moir, S. (1996). “PATH: A work sampling-
based approach to ergonomic job analysis for construction and other non-repetitive work.”
Applied Ergonomics, 27(3), 177–187.
Chan, M. (2011). “Fatigue: the most critical accident risk in oil and gas construction.”
Construction Management and Economics, 29(4), 341–353.
Craig, A. (1992). “Symptoms of acute and chronic fatigue.” Handbook of human performance,
Vol. 1: The physical environment; Vol. 2: Health and performance; Vol. 3: State and trait,
A. P. Smith and D. M. Jones, eds., Academic Press, San Diego, CA, US, 289–339.
Dawson, D., Chapman, J., and Thomas, M. J. W. (2012). “Fatigue-proofing: A new approach to
reducing fatigue-related risk using the principles of error management.” Sleep Medicine
Reviews, 16(2), 167–175.
Dawson, D., Searle, A. K., and Paterson, J. L. (2014). “Look before you (s)leep: Evaluating the
use of fatigue detection technologies within a fatigue risk management system for the road
transport industry.” Sleep Medicine Reviews, 18(2), 141–152.
Dembe, A. E., Erickson, J. B., Delbos, R. G., and Banks, S. M. (2005). “The impact of overtime
and long work hours on occupational injuries and illnesses: new evidence from the United
States.” Occupational and Environmental Medicine, 62(9), 588–597.
Gander, P., Hartley, L., Powell, D., Cabon, P., Hitchcock, E., Mills, A., and Popkin, S. (2011).
“Fatigue risk management: Organizational factors at the regulatory and industry/company
level.” Accident Analysis & Prevention, Advancing Fatigue and Safety Research, 43(2),
573–590.
Gillberg, M., and Åkerstedt, T. (1994). “Sleep restriction and SWS-suppression: effects on
daytime alertness and night-time recovery.” Journal of Sleep Research, 3(3), 144–151.
Grandjean, E. (1979). “Fatigue in industry.” British Journal of Industrial Medicine, 36(3), 175–
186.
Hallowell, M. R. (2010). “Worker fatigue: managing concerns in rapid renewal highway
construction projects.” Professional Safety, 55(12), 18–26.
Harrington, J. M. (2001). “Health effects of shift work and extended hours of work.” Occupational
and Environmental Medicine, 58(1), 68–72.
Johnston, R., Howard, M., Cawley, and Posse, G. (1998). “Effect of lower extremity muscular
fatigue on motor control performance.” Medicine & Science in Sports & Exercise®,
30(12), 1703–1707.
Kajtna, T., Štukovnik, V., and Grošelj, L. D. (2011). “Effect of acute sleep deprivation on
concentration and mood states with a controlled effect of experienced stress.” Zdravniški
Vestnik, 80(5).
Kartam, N. A. (1997). “Integrating Safety and Health Performance into Construction CPM.”
Journal of Construction Engineering and Management, 123(2), 121–126.
Langford, D., Rowlinson, S., and Sawacha, E. (2000). “Safety behaviour and safety management:
its influence on the attitudes of workers in the UK construction industry.” Engineering
Construction and Architectural Management, 7(2), 133–140.
van der Linden, D., Frese, M., and Meijman, T. F. (2003). “Mental fatigue and the control of
cognitive processes: effects on perseveration and planning.” Acta Psychologica, 113(1),
45–65.

12
Lorist, M. M., Boksem, M. A., and Ridderinkhof, K. R. (2005). “Impaired cognitive control and
reduced cingulate activity during mental fatigue.” Cognitive Brain Research, 24(2), 199–
205.
Lorist, M. M., Klein, M., Nieuwenhuis, S., de Jong, R., Mulder, G., and Meijman, T. F. (2000).
“Mental fatigue and task control: Planning and preparation.” Psychophysiology, 37(5),
614–625.
Reiner, B. I., and Krupinski, E. (2011). “Demystifying Occupational Stress and Fatigue Through
the Creation of an Adaptive End-User Profiling System.” Journal of Digital Imaging,
25(2), 201–205.
Sawacha, E., Naoum, S., and Fong, D. (1999). “Factors affecting safety performance on
construction sites.” International Journal of Project Management, 17(5), 309–315.
Schneider, S., and Susi, P. (1994). “Ergonomics and Construction: A Review of Potential Hazards
in New Construction.” American Industrial Hygiene Association Journal, 55(7), 635–649.
Scott, J. P., McNaughton, L. R., and Polman, R. C. (2006). “Effects of sleep deprivation and
exercise on cognitive, motor performance and mood.” Physiology & behavior, 87(2), 396–
408.
Spurgeon, A., Harrington, J. M., and Cooper, C. L. (1997). “Health and safety problems associated
with long working hours: a review of the current position.” Occupational and
Environmental Medicine, 54(6), 367–375.
Tixier, A. J.-P., Hallowell, M. R., Albert, A., van Boven, L., and Kleiner, B. M. (2014).
“Psychological antecedents of risk-taking behavior in construction.” Journal of
Construction Engineering and Management, ASCE(04014052), 1–10.
U.S. Bureau of Labor Statistics. (2015). “Census of Fatal Occupational Injuries Summary, 2014.”
<http://stats.bls.gov/news.release/cfoi.nr0.htm>.
Waehrer, G. M., Dong, X. S., Miller, T., Haile, E., and Men, Y. (2007). “Costs of occupational
injuries in construction in the United States.” Accident Analysis & Prevention, 39(6), 1258–
1266.
Weick, K. E., Sutcliffe, K. M., and Obstfeld, D. (2008). “Organizing for high reliability: Processes
of collective mindfulness.” Crisis management, 3(1), 81–123.
Yaggie, J., and Armstrong, W. J. (2004). “Effects of lower extremity fatigue on indices of
balance.” Balance, 312(322), 2004.

13
Chapter 2: Causes and consequences of occupational fatigue:
Meta-analysis and systems model.
Ulises Techera1, Matthew Hallowell2, Nathan Stambaugh3, and Ray Littlejohn4

2.1 Abstract:

Occupational fatigue was a latent variable until recent years when researchers began to

demonstrate the vast diversity of occupational and personal problems related to fatigue. The

negative consequences of fatigue highly impact our society, instigating productivity losses that

exceed $100 billion a year and injuries that total more than $50 billion yearly. Unfortunately,

fatigue is difficult to measure empirically and the plethora of causes and effects of fatigue are

described in literature published in fields as diverse as medicine, psychology, engineering, and

sociology. Researchers who wish to study fatigue and build upon the existing body of knowledge

find it difficult to access and codify the large, diverse, and scattered body of knowledge. To address

this issue, this study aims to codify the literature related to occupational fatigue into a single

systems model that: (1) identifies causal factors of fatigue; (2) evaluates interrelationships among

causal factors; (3) identifies outcomes of fatigue; and (4) evaluates the interrelationships among

fatigue outcomes. Furthermore, this study presents the first all-inclusive meta-analysis in this

domain by computing a single statistic comparable across studies. The body of knowledge was

cataloged and visually represented in the form of a systems model for the first time, allowing

1
PhD Student; Department of Civil, Environmental, and Architectural Engineering; University of Colorado at Boulder;
UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA; [email protected]
2
Beavers Endowed Professor of Construction Engineering; Department of Civil, Environmental, and Architectural
Engineering; University of Colorado at Boulder; UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA;
[email protected]
3
Undergraduate Research Assistant; Department of Civil, Environmental, and Architectural Engineering; University
of Colorado at Boulder; UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA;
[email protected]
4
W. Edwards Deming Professor of Management; Lockheed Martin Engineering Management Program; UCB 428;
1111 Engineering Drive; Boulder, CO 80309 USA; [email protected]

14
researchers and practitioners to easily access variables of interest and understand the dynamics of

fatigue and the effect of its causes and consequences. The model also clearly shows where future

research is needed to strategically address knowledge gaps.

2.2 Introduction:

Every day, more than 20% of the working population in the United States experiences

occupational fatigue (Kroenke and Price 1993), resulting in $136.4 billion in lost productivity and

healthcare costs each year (Ricci et al. 2007). Fatigue diminishes the quality of life of individuals

and weakens the immune system. Despite these important implications, fatigue is a common

occurrence, making it difficult for people to recognize its consequences and understand various

primary and secondary causes.

Unfortunately, despite its implications, fatigue is often uncovered only after a major

accident (Reiner and Krupinski 2011). For example, Harrington (2001) demonstrated how fatigue

played a role in the Three Mile Island, Chernobyl, and the Exxon Valdez oil spill events, as they

took place during the first hours of the morning after a night shift when the levels of fatigue among

workers were at their highest. The negative outcomes of fatigue are not surprising since fatigue

decreases the ability to process and react to new information and respond to hazards (Meijman and

Schaufeli 1996; Harrison and Horne 2000; Lorist et al. 2000).

Past literature has called for a deeper understanding of the causes and consequences of

occupational fatigue (Reiner and Krupinski 2011). In pursuit of this knowledge, researchers have

15
aimed to study specific causes or consequences of fatigue using rigorous experimental methods

and to validate each other’s research findings. Despite the relatively large and mature body of

knowledge in the field of occupational fatigue, there is no single resource that catalogs and

organizes empirical literature or that integrates the findings in a meaningful and comprehensive

manner. Thus, researchers may find it difficult to identify new theoretical points of departure and

cumbersome to perform high-quality literature reviews.

To address this need, we aimed to codify the literature related to occupational fatigue into

a systems model that: (1) identifies causal factors of fatigue; (2) evaluates interrelationships among

causal factors; (3) identifies outcomes of fatigue; and (4) evaluates the interrelationships among

fatigue outcomes. Although there are several hundred studies of fatigue, we built the model using

only empirical and validated findings.

Additionally, a meta-analysis was performed using published data and formal statistics.

Prior to the current paper, no formal or comprehensive meta-analysis had been performed on the

causes or consequences of occupational fatigue. The meta-analysis allowed us to use a single

statistic (Cohen’s d) to explain the direction and magnitude of relationships among causes and

consequences of fatigue. This value was computed based upon the effect sizes, sample sizes, and

variance reported in past empirical studies. The meta-analysis complements the systems model,

which focuses on illustrating the number of studies for each relationship by explaining the strength

of the relationships and their statistical significance. Consequently, a revised model was created to

illustrate validated effect sizes and directions of influence.

16
2.3 Fatigue definition and classification:

Before reviewing literature on the causes and consequences of fatigue, it is important to

provide a pithy definition of the various types of fatigue. In general, fatigue is a condition of the

individual that is recognized as a decreased ability to perform activities at the desired level due to

lassitude or exhaustion of mental and/or physical strength (Hallowell et al. 2010; Gander et al.

2011). Humans commonly identify this state when feeling tired or weary. The condition of fatigue

is experienced by every person (Aaronson et al. 1999) and, depending on its causes, can be

classified as acute or chronic (Piper 1989).

Acute and chronic fatigue are distinguished from one another by their principal causes and

long-term effects. Acute fatigue is experienced as a consequence of mental or bodily labor,

emotional stress, insufficient recovery, or a temporary illness (Aaronson et al. 1999; Maslach

2001; Hallowell et al. 2010). It is considered to be a normal regulatory response to adverse

condition and affects healthy people. Typically, acute fatigue can be relieved by quality rest, sleep,

appropriate diet, and exercise (Piper 1989; Jason et al. 2010). Chronic fatigue, on the other hand,

can manifest as a side-effect of a severe illnesses or treatments such as rheumatoid arthritis,

diabetes, multiple sclerosis, radiation exposure, and chemotherapy (Piper 1989; Gander et al. 2011;

Perry 2012). The symptoms of chronic fatigue are similar to those produced by acute fatigue except

that they are experienced constantly and cannot be alleviated by rest alone (Brown 1994). In most

occupational environments, the primary concern is acute fatigue, which is the focus of this paper.

In addition to the distinction between acute and chronic, fatigue can also be classified as

muscular or mental. Mental fatigue is associated with a decreased motivation to continue

17
performing a current activity and a sensation of weariness (Lee et al. 1991; Lorist et al. 2000).

Mental fatigue decreases the ability to process and respond to information, ultimately diminishing

competency, productivity, and error avoidance (Meijman and Schaufeli 1996; Harrison and Horne

2000; Lorist et al. 2000). Alternatively, muscular fatigue is a reduction in the physical ability to

exert a force or perform a task. It is often a result of either metabolic or neural decay over the

course of an extended physical activity. Muscular fatigue is most often caused by high intensity

work, long duration of work, or improper work posture. Blue collar workers who are often required

to perform heavy and repetitive tasks are more prone to muscular fatigue (Lipscomb et al. 2002).

Although muscular and mental fatigue can be distinguished from one another, they are often

experienced in concert to varying degrees and proportions. Thus, in literature and in professional

practice, fatigue is often modeled and discussed in a general form. When possible, we will

distinguish between muscular and mental fatigue for precision; otherwise, we will refer to fatigue

as general condition.

2.4 Causes of fatigue:

In order to build a comprehensive systems model of fatigue, implied causal factors were

identified, based on previous research, and the relationships among them were modeled in an

organized fashion. To increase the internal validity of the resulting systems model, our review is

limited to research-based studies that provide empirical evidence for conclusions and we focus our

attention on causal factors and relationships that have been validated by multiple studies.

Additionally, to reduce complexity and avoid repetition, factors that share similar root causes were

codified into single, fundamental causal factor that is aptly labeled. It should be noted that this

initial systems model includes all empirical evidence even if the data reported in the study did not

18
have sufficient detail to be included in the subsequent meta-analysis. This was done to illustrate

the focal areas of past fatigue research.

2.4.1 Sleep deprivation

Sleep deprivation is the most commonly discussed and arguably the most significant cause

of both muscular and mental fatigue. It affects between 15 and 20% of US adults, and is becoming

an increasing concern in our modern society due to various schedule irregularities and conflicting

priorities that compromise sleep (Webb and Agnew Jr 1975; Åkerstedt 1990; Bliwise et al. 1992).

Belenky et al. (2003) produced a comprehensive analysis of the fatigue-related

consequences of different sleep routines. In particular, they studied three different sleep restriction

conditions: a mild sleep restriction (7 hours of sleep per night), a moderate sleep restriction (5

hours), and a severe sleep restriction (3 hours or less). The results were compared to control group

which obtained 8 hours of sleep during the course of the study. They discovered that sleeping less

than 8 hours generated negative outcomes. In particular, those who experienced only mild or

moderate sleep deprivation showed a reduction in their mental and physical performance, which

stabilized after a few days but still was under the normal level of performance. Furthermore, those

who experienced severe sleep deprivation showed an ongoing degradation that didn’t stabilized

even after a few days. Several authors agree that a single night of partial sleep deprivation will

generate a reduction in performance that can last several days (Van Dongen and Dinges 2005; Lim

and Dinges 2008). Interestingly, Belenky et al. (2003) noted no significant improvements in

fatigue for workers who obtain more than 8 hours of sleep compared to the control group.

19
 Typically, sleep is compromised by early morning and night shifts that not only disrupt the

circadian rhythm of the individual but also affect their social and family life (Folkard and Tucker

2003). Such adverse effects are most prevalent the first night after a period of rest when typical

work shifts re-commence (Roach et al. 2004). In addition to fatigue, sleep disruption causes

impaired physical performance, decreased work satisfaction, increased stress, and interrupted

future sleep patterns (Harrison and Horne 2000; Philibert 2005). Unless this cycle is broken and

workers recover completely from the previous work day, the sleep deprivation process can become

chronic, ultimately resulting in serious long-term mental and physical health problems (Tepas and

Mahan 1989).

Sleep quantity and quality is important as the mind and body of human beings need rest to

avoid lassitude (Dawson and McCulloch 2005; Zhang and Liu 2008). Sleep deprivation is seen as

such a strong precursor of fatigue that Dawson and McCulloch (2005) claim that fatigue can be

accurately predicted by simply measuring the amount of sleep that a worker experiences in the 24-

48 hours prior to a shift. In fact, they argue that sleep patterns prior to work are a better indicator

of occupational fatigue than actual observations at work.

2.4.2 Mental Exertion

Some causes of fatigue can be purely mental. Mental exertion, defined as sustained

cognitive activity that requires extraordinary mental effort, is the principal cause of mental fatigue

(Lorist et al. 2005). Mental fatigue can be recognized by the individuals themselves as they

experience difficulty in focusing attention, making timely decisions, planning, and responding to

stimuli (van der Linden et al. 2003; 2003b; Boksem et al. 2005, 2006; van der Linden and Eling

20
2006). In addition to subjective indicators of mental fatigue, Lorist et al. (2005) used brain

scanning technology, such as the electro encephalogram, to directly measure mental lassitude and

unequivocally correlated mental exertion with mental fatigue.

2.4.3 Muscular Exertion

Muscular exertion, a purely physical occurrence, is a primary cause of localized fatigue

(Grandjean 1979; Brown 1994; Jason et al. 2010). The link between muscular exertion and

localized muscular fatigue has been studied in numerous ways. For example, fatigued is caused by

the repetitive contraction of a specific muscle through electrical impulses (dynamic exertion), or

by keeping a specific muscle contracted (static exertion) (Christensen et al. 1995; Edwards and

Lippold 1956).

Beyond the static and dynamic state of exertion, there is either a metabolic or central

nervous system degradation that explains the onset of muscular fatigue. For example, a fatigued

quadriceps muscle consumes much more energy than a relaxed muscle, although metabolic energy

production remains unchanged (Edwards et al. 1975). This results in exponential cumulative

fatigue generation.

Although mental and muscular exertion yield obvious and distinct impacts on mental and

muscular fatigue, respectively; other factors affect both mental and muscular capabilities. Some of

these factors interact producing a compounding effect on the development of general fatigue.

21
2.4.4 Work Load Characteristics

Assigned occupational work load, defined as the product of physical demand and time, has

been strongly linked to fatigue (Akerstedt et al. 2002). Ribet and Derriennic (1999) found that

heavy workloads negatively impact sleep and, thus, interfere with the recovery process

accumulating fatigue debt in the individual. In addition, high levels of required workload without

the option for lower-effort alternatives cause an effort-to-reward imbalance that can ultimately

lead to mental fatigue (Boksem et al. 2008). Hsiao and Simeonov (2001) and Gander et al. (2011)

found that time on task alone was inadequate for fatigue prediction as accurate predictions require

consideration of time on task, cognitive and physical demands, repetition, and scheduled breaks.

2.4.5 Overtime and Long Work Hours (LWH)

As defined by the Fair Labor Standards Act (1938), overtime in the US is the amount of

work time in a week that exceeds 40 hours. Overtime can be accomplished either by working more

than eight hours per day in a five-day-per-week schedule or by having fewer days off from work

per week.

The occurrence of overtime is a common problem of industrialized environments (Kodz et

al. 2003). More specifically, in Europe about 20% of the working population experiences at least

5 hours of overtime weekly and 13% of the full-time employees in Europe work at least 10 hours

of overtime a week (Pascal and Damien 2001). In the US the situation is similar where about 26%

of the male working population and 11% of the female working population work at least 10 hours

of overtime weekly. Additionally, the amount of overtime has been growing over the past 50 years,

22
especially among women due to their increasing presence in the workplace (Caruso 2006; Bureau

of Labor Statistics 2015).

The fatigue-related consequences of overtime are directly connected to the perceived

benefits and conditions under which overtime takes place. In this regard there are three different

types of overtime situations:

(1) Mandatory, low reward, low autonomy, and highly demanding overtime. In this case

the worker is forced to work overtime in an environment that the worker can’t control, required to

perform highly demanding activities, or does not feel that the increased payment or future time off

are sufficient compensation. This kind of overtime is associated with the most severe mental

fatigue, dissatisfaction with the job, job burnout, negative work/home interaction, and slow

recovery (Van Der Hulst and Geurts 2001; Beckers et al. 2008).

(2) Mandatory, high-reward overtime. The second kind of overtime is that which is

performed solely because the extra payment is desired. This kind of overtime encourages people

to work up to a point that exceeds their healthy state of functioning in order to earn more money,

causing relatively high levels of fatigue (Gander et al. 2011).

(3) Voluntary overtime. This overtime is not driven solely by the economic reward, but

also by personal fulfillment. When it is limited by the workers themselves, and it is performed

with a high level of autonomy, overtime does not appear to have significant negative

consequences, causing very low levels of both muscular and mental fatigue within natural

23
physiological limitations (Taris et al. 2007). Park et al. (2001) concluded that 20 hours of overtime

per week is the point after which fatigue recovery becomes extremely difficult, even when

overtime is voluntary.

Extended shifts can also cause fatigue. In contrast to overtime, which involves an

accumulated number of work-hours in a week, extended shifts or long working hours (LWH) refers

to single or compound shifts that involve more than eight hours of work in a twenty-four-hour

period. It is important to notice that LWH don’t necessarily imply overtime. Some schedules such

as 10 hour shifts for 4 days a week involve LWH but not in overtime.

LWH studies showed that memory efficiency decreases during cognitive activities when

uninterrupted time on task exceeds 3.5 hours (Meijman 1997). Once 3.5 hours of continuous work

has been accumulated, mental acuity can only be maintained at the expense of extra mental

exertion. The prolongation of such effort will cause mental fatigue (Meijman 1997). LWH also

causes fatigue by exposing the individual to other at work stressors such as noise, inadequate

lighting, extreme temperatures, and other environmental and social factors, when present, for an

additional period of time (Park et al. 2001; Dembe et al. 2005). The fatigue impact of LWH

depends on the activity performed. If the activity requires mental effort, it will generate mental

fatigue; otherwise, it will generate muscular fatigue or a combination of both.

2.4.6 Incomplete recovery

Recovery is the process of reversing the negative effects of mental and muscular exertion

to return to a pre-fatigued state. Acute fatigue occurs when there is inadequate time to rest and

24
recover from a work period (Jason et al. 2010; Swaen et al. 2003; Beurskens et al. 2000).

Incomplete recovery can result from inadequate time off from work, misclassifying non-recovery

time as recovery time (e.g., considering commute time as recovery), and inadequate use of the

given recovery periods by the worker such as substance abuse or physically and mentally

demanding recreational activities (Dawson and McCulloch 2005, Gander et al. 2011). Other

common threats to recovery time are irregular or split shifts in which a worker has between two

hours and four hours break before returning to work. Such shifts are usually insufficient to

effectively recover from fatigue because of common commute times.

2.4.7 At work environmental Factors

Noise, light intensity, vibration, and temperature are all environmental factors linked to

fatigue (Krause et al. 1997). Noise has been found to cause fatigue through over stimulation

(Landström and Lundström 1985; Landström and Löfstedt 1987). For example, Kjellberg et al.

(1998) showed that airplane mechanics and boat patrol crews experienced mental fatigue as a result

of high duration of exposure to low frequency noise. Additionally, extreme temperatures (i.e.,

below 35F and above 95F) increase the rate of muscular fatigue (Gonzalez-Alonso 1999; Zivin

and Neidell 2010). Similarly, Park and Gotoh (1993) found a significant positive correlation

between poor lightning and fatigue and Jiao et al. (2004) found a positive correlation between

vibration and fatigue.

2.4.8 Social environment at the workplace

Most of the working population spends one-third to one-half of their waking hours at work.

Relationships with co-workers, managers, and subordinates play a major role in the development

25
of mental fatigue (Bültmann et al. 2001). Some specific psychosocial experiences can exacerbate

mental fatigue. For example, workplace abuse, defined as daily harassment from co-workers or

abusive supervision for prolonged periods of time, can cause severe mental fatigue (Tepper 2000;

Zapf and Gross 2001). Similarly, Hardy et al. (1997) found that conflicts and adversarial situations

at work develop high levels of mental fatigue, even if they are essential attributes of the work type

(e.g., negotiations). Other social factors that strongly influence the onset of mental fatigue are the

perception of low autonomy, high emotional demands, and job insecurity (Taris et al. 2007).

2.4.9 Emotional predisposition and distress

Fatigue impacts everyone differently depending on their emotional predisposition and is

often exacerbated under distress. Emotional disposition pertains to the propensity of an individual

to experience negative emotions such as fear and sadness (Gibson et al. 2003). Bültmann et al.

(2001) performed a field study with 11,020 workers from 45 different companies, and discovered

that those who are more likely to experience negative emotions are more likely to encounter mental

fatigue than individuals who are more emotionally resilient. Furthermore, in a study of 16,139

employees from the public sector, Ala-Mursula et al. (2005) showed that the onset of fatigue is

accelerated under stressful conditions, especially for individuals who are emotionally vulnerable.

Table 1 summarizes these key causal factors of fatigue and provides the references

associated with each factor. Additionally, when possible, the impacts of each factor on mental,

muscular, and general fatigue are identified.

26
 Table 1 Causes of fatigue
Factor Factor Definition Impacted Reference over
Factor
Sleep Sleep deprivation refers to a loss in the GEN* Webb and Agnew Jr (1975);
Deprivation amount of consecutive hours of sleep. Tepas and Mahan (1989);
There are different degrees of sleep Åkerstedt (1990); Harrison and
deprivation, which Belenky et al. (2003) Horne (2000); Belenky et al.
classify as Mild SD (7h of time in time (2003); Folkard and Tucker
in bed), moderate SD (5h of TIB), severe (2003); Roach et al. (2004);
SD (3h of TIB) or total SD (no sleep at Dawson and McCulloch (2005);
all). Philibert (2005); Van Dongen and
Dinges (2005); Lim and Dinges
(2008); Zhang and Liu (2008).
Mental Sustained cognitive activity that requires MEN** Okogbaa et al. (1994); Meijman
Exertion extraordinary mental effort. (Lorist et al. (1997); Lorist et al. (2005);
2005; Meijman and Schaufeli 1996) Hallowell and others (2010).
Muscular Exhaustion of the muscle due to an MUS*** Edwards and Lippold (1956);
Exertion extent period of sustained tension or Edwards et al. (1975); Grandjean
repetitive activity. (1979);Brown (1994);
Christensen et al. (1995);
Johnston et al. (1998); Jason et al.
(2010);Yaggie and Armstorn
(2004); Jason et al. (2010).
Work Load High physical or mental demands at MEN Boksem and Tops (2008)
work. GEN Van Der Hulst and Geurts (2001);
Overtime As defined by (FLSA 1938) “overtime” MEN Meijman (1997)
& LWH is the amount of time worked that GEN Park et al. (2001); Van Der Hulst
exceeds 40h of work a week. and Geurts (2001); Dembe et al.
(2005); Taris et al. (2007);
Beckers et al. (2008)
Incomplete Recovery is the process of reverting or GEN Dawson and McCulloch (2005)
Recovery reversing the negative effects of job
demand to return to a pre-work state.
Work Noise, light intensity, vibration, and MEN Melamed and Bruhis (1996);
Environment temperature are all environmental Kjellberg et al. (1998)
factors linked to fatigue. GEN Krause et al. (1997); Gonzalez-
Alonso (1999); Ala-Mursula
(2005); Zivin and Neidell (2010).
Social The quality and characteristics of worker MEN Maslach and Jackson (1984);
Environment relationships with pear and supervisors, Hardy et al. (1997); Tepper
as well as the perceived freedom at work. (2000); Bültmann et al. (2001)
Emotional Emotional disposition pertains to the MEN Bültmann (2001); Gibson et al.
Predisposition level of fear, stress or overall attitude a (2003);
worker has towards a certain task or job
Notation: *MEN: Mental Fatigue; **GEN: General Fatigue; ***MUS: Muscular Fatigue

27
2.5 Relationships among causes of fatigue

Causal factors of fatigue can compound one another when they co-occur. Interestingly,

many of the interrelationships described in the literature relate to the organization of work by

management. For example, work load, overtime, and extended shifts tend to interact with each

other and with other causal factors. Thus, these managerial factors can be addressed through

mindful work organization.

2.5.1 Impacts of increased work load on other causal factors

As work load increases with additional or more demanding activities, the extra mental and

physical exertion can accelerate lassitude and, when extreme, it can lead to subsequent sleep

deprivation (Åkerstedt et al. 2002). These consequences can affect all types of workers; however,

women below 49 years of age who come from a high socioeconomic status are particularly

vulnerable (Van Der Hulst and Geurts 2001). Additionally, Van Der Hulst and Geurts (2001)

demonstrated that when workload is increased because of additional physical demands, the effects

could remain resident for up to a day after the exertion period.

With regards to mental fatigue, Boksem and Tops (2008) found that increased work load

is a significant mental stressor when the reward received by the worker is perceived as insufficient

to compensate for the extra effort required. Furthermore, Ribet and Derriennic (1999) discovered

that the pressure to hurry is the principal psychosocial occupational risk factor affecting sleep

deprivation.

28
2.5.2 Impacts of overtime and LWH on other causal factors

The synergistic effects of overtime and LWH magnify the effects of fatigue caused by other

factors. In fact, most of the aforementioned factors as discussed by the respective researchers in

the context of extended periods of work or exertion, implicitly suggesting that LWH or overtime

are pre-requisite factors. When the work requirements are mandated and workers have low

autonomy, lengthened work periods become especially problematic (Axelsson et al. 1998; Park et

al. 2001; Folkard and Tucker 2003; Dembe et al. 2005; van der Hulst et al. 2006).

When workers are required to work overtime or LWH, there is a natural reduction in

recovery because fewer hours are available to rest. On many occasions overtime and LWH are

combined with night or early morning shifts that disrupt circadian rhythms and interfere with

family life (Webb and Agnew Jr 1975; Åkerstedt 1990; Bliwise et al. 1992; Axelsson et al. 1998;

Folkard and Tucker 2003; Baulk et al. 2009). This affects the individual physically, emotionally,

and mentally, slowing down the recovery process (Åkerstedt 1990; Van Der Hulst and Geurts

2001; Folkard and Tucker 2003; Beckers et al. 2008; Charles, De, Wolff et al. 2013).

29
 Table 2 identifies and provides references for research of the interrelationships among the

causal factors of fatigue.

Table 2 Relationships among causes of fatigue

Factor Impacted References
Factor
Work Sleep Ribet and Derriennic (1999); Åkerstedt et al. (2002)
Load Deprivation
Mental Boksem and Tops (2008)
Exertion
Muscular Åkerstedt (2002).
Exertion
Overtime & Beckers et al. (2008)
LWH
Overtime Sleep Webb and Agnew Jr (1975); Åkerstedt (1990); Bliwise et al.
& LWH Deprivation (1992);Axelsson et al. (1998); Folkard and Tucker (2003); Baulk et al.
(2009)
Mental Meijman (1997); Park et al. (2001); Folkard and Tucker (2003); Dembe
Exertion et al. (2005)
Muscular Folkard and Tucker (2003); Dembe et al. (2005)
Exertion
Incomplete Åkerstedt (1990); Van Der Hulst and Geurts (2001); Folkard and
Recovery Tucker (2003); Beckers et al. (2008); Charles, De, Wolff et al. (2013).

2.6 Consequences of fatigue:

Fatigue research tends to focus on either causal factors or outcomes of fatigue. Causal

factors and outcomes are rarely modeled together. Here, implied causes and consequences of

fatigue are analyzed exclusively underlying their relationship with fatigue as the main variable of

interest. Occasionally, some causes and consequences of fatigue can be connected via mechanisms

other than fatigue, particularly with mental disorders and long-term health concerns. However, this

review did not include these secondary relationships.

Below is a comprehensive review of empirical literature that has linked occupational

fatigue with specific mental and physical consequences. Some of these consequences are subtle or

30
acute while others can manifest on delay or can be chronic. Understanding the interrelationships

among consequences of fatigue is critical to modeling the true impact that fatigue can have on an

individual.

2.6.1 Mood changes

Mood, related to short-term emotions and feelings, can be assessed by numerous variables

such as tension, anger, vigor, sadness, anxiety, and depression. Mood is typically measured

subjectively through the Profile of Mood State introduced by McNair et al. (1971). According to

the prevailing literature, fatigue results in strong negative changes to mood states, which can

change emotional stability, increase anxiety, and eventually lead to long-term depression (Pilcher

and Huffcutt 1996; Dinges et al. 1997; Harrison and Horne 2000; Sagaspe et al. 2006; Scott et al.

2006; Boonstra et al. 2007; Kajtna et al. 2011). In addition, a negative mood or emotional state has

been shown to decrease complex cognitive performance, clouding risk perception (Tixier et al.

2014).

Scott et al. (2006) studied the mood-related consequences of fatigue arising from 30 hours

of sleep deprivation and concluded that tension and anger were not affected but that feelings of

depression and sadness manifested quickly, even after only one missed sleep cycle. Mikulincer et

al. (1989) studied the timing of mood-related change arising from fatigue and found that fatigued

individuals experienced relatively high levels of anxiety between 0400 and 0800 hours and

relatively low levels between 1600 and 2000 hours. Interestingly, using a series of complex and

controlled experiments, Kajtna et al. (2011) found that the fatigue-induced mood changes

compromise an individual’s ability to concentrate on a stimulus.

31
2.6.2 Cognitive degradation

Cognitive functions, classified as complex or basic, are essential for effective completion

of work and for healthy life outside of work. Complex cognitive functions include, but are not

limited to, the ability to plan, perceive risk, and make decisions under uncertainty. Basic cognitive

functions include attention, vigilance, and response to stimuli (Jovanovic et al. 2012). When

studying the impact of fatigue on cognitive performance, researchers typically manipulate degrees

of sleep deprivation and measure performance with standardized cognitive activities (Angus and

Heslegrave 1985; Van Dongen and Dinges 2005; Lim and Dinges 2008; Zhang and Liu 2008)

Studies on complex cognitive performance involve sophisticated experimental methods

and the use of advanced technologies. Researchers have focused on the impacts of acute fatigue

on verbal fluency and communication, decision-making capability, creative thinking, planning,

executive control, and novelty performance. Overwhelmingly, data support that all of these

complex cognitive functions are significantly deteriorated even with modest fatigue levels

(Harrison and Horne 1999; 2000).

With the help of electroencephalographic (EEG) tools, functional neuroimaging, and

psychophysiological assessments, researchers have found that the Prefrontal Cortex (PFC), which

governs the executive functions of the brain, is the locus of deterioration when an individual is

fatigued (Jones and Harrison 2001; van der Linden et al. 2003). Unlike basic functions that can be

rote and automatic, complex tasks require uncompromised PFC functionality (Harrison and Horne

1999; Jones and Harrison 2001). Thus, the brain is highly susceptible to even modest fatigue levels

when executive functions are required to make complex decisions (Meijman 1997; Nilsson et al.

32
2005; Boonstra et al. 2007). In an occupational environment, this translates to a differential in

potential consequences depending on the tasks performed by different workers with the same

levels of fatigue.

In addition to complex cognitive functions, other more basic and automatic cognitive

functions can be compromised. One of the earliest and most comprehensive investigations of the

relationship between fatigue and cognitive performance found that a single night of sleep

deprivation decreased cognitive performance by 30% (Angus and Heslegrave 1985). Furthermore,

their findings showed that cognitive performance was reduced by an additional 30% during the

second 24-hour period of sleep deprivation and remained stable at 40% of the original cognitive

performance after 48 hours without sleep. This study was empirically validated by Van Dongen

and Dinges (2005); Lim and Dinges (2008); and Zhang and Liu (2008) who later replicated the

experiment with a more heterogeneous sample of participants and a more diverse set of cognitive

tasks and fatigue scenarios.

2.6.3 Attention and concentration

The pre-frontal cortex (PFC) helps an individual to concentrate attention on a task or

stimulus. As the functionality of the PFC is compromised by fatigue, both static and dynamic

functions are affected. Researchers have studied the impact of fatigue on concentration and

attention for many years, starting with Yoshitake's (1978) seminal work with 17,000 participants.

In this field study, three different dimensions of fatigue were identified: drowsiness and dullness,

inability to concentrate, and physical discomfort. Even though inability to concentrate was

experienced most strongly by workers with mentally demanding tasks, lassitude of concentration

33
and attention are experienced similarly by a fatigued individual regardless of causal factors. In

more recent years, fatigue has been studied in different naturalistic tasks such as driving (Brown

1994) and conflict tests (Fogt et al. 2010). As hypothesized, these studies demonstrated that fatigue

decreased processing of stimulus, concentration capacity, and made individuals more prone to

error.

2.6.4 Reaction time

When reacting to a stimulus, the human body must perceive, process, and physically react

to the stimulus. Previously, we discussed the reduced attention capacity resulting from fatigue. In

addition, reaction time, defined as the time between a stimulus onset and the response to such

stimulus, is highly compromised by fatigue.

The prevailing method for measuring reaction time involves using the psychomotor

vigilance test (PVT), which records the reaction time to a visual stimulus (Dinges and Powell

1985). Angus and Heslegrave (1985) used PVT to study the impact of fatigue on reaction time and

found that the reaction time of fatigued workers decreased by 24% when compared to well-rested

workers. Furthermore, higher levels of fatigue induced by a second night without sleep resulted in

a 57% reduction in reaction time. Similar studies found more modest but still significant decreases

in reaction time (Opstad et al. 1978; Naitoh 1981; Lorist et al. (2000).

2.6.5 Physical degradation and pain

The physical degradation caused by fatigue can be significant and pervasive. For example,

Johnston (1998) tested static and dynamic balance of fatigued individuals and found drastic

34
increases in the risk of falls compared to non-fatigued individual. In addition to reduced physical

stability, fatigue can result in reduced muscular strength causing slower movement and increased

number of errors (Grandjean 1979). The long-term impacts of physical degradation primarily

involve musculoskeletal disorders of the neck, shoulders, and back (Lipscomb et al. 2002).

Although acute, reduced balance was observed in fatigued subjects for at least 15 minutes after

prolonged exertion of cervical muscles, leading to longer-term localized muscular fatigue

(Gosselin et al. 2005; Yaggie and Armstrong 2004). The implication is that seemingly short-term

effects of acute fatigue can remain resident for longer periods of time than may be intuitive.

2.6.6 Illnesses

Fatigue has also shown to have significant long-term impacts on the human body. For

example, researchers have found evidence that night shift workers experience higher levels of

fatigue and, consequently, present gastrointestinal disorders, irregular bowel activity such as

constipation and diarrhea, bowel pain, and even ulcers at a rate that is nearly double their day-

worker counterparts (Segawa et al. 1987; Scott and La Dou 1994; Knutsson 2003). Furthermore,

Steenland and Fine (1996a), Bøggild and Knutsson (1999), and Knutsson (2003) found that shift

workers present a 40% higher risk of suffering cardiovascular disease or myocardial infarction

than day-time workers. Although gastrointestinal and cardiovascular conditions are the most

common occupational illnesses related to fatigue (Iwasaki et al. 1998), some postulate that fatigue

has greater consequences among those who are pregnant or have diabetes (Knutsson 2003).

35
2.6.7 Human error and injuries

As previously discussed, lassitude of cognitive functions decrease recall, reaction time,

ability to plan, and ability to respond to stimuli. Additionally, lassitude of the muscular system

decreases the ability to physically execute a plan. Thus, it is not surprising that researchers have

found evidence of a direct relationship between the degree of fatigue and rates of human error

(Craig 1992 ; Czeisler, et al. 1992 ; Lorist et al. 2000; van der Linden et al. 2003). Human error is

said to occur when an individual takes unintentional actions or is unable to execute planned

functions and an undesired outcome is realized.

Table 3 summarizes these key outcomes of fatigue. They are briefly described and

references are provided for each factor. Additionally, for consistency with Table 1, the impacts

of each factor on mental, muscular, and general fatigue are also distinguished.

Table 3 Consequences of fatigue

Fatigue Impacted Impacted Factor description References
Factor
MEN Cognitive Cognitive functions refer to both Opstad et al. (1978); Naitoh (1981);
Degradation complex and basic cognitive functions. Angus and Heslegrave (1985); Dinges
Complex cognitive activities are not and Powell (1985); Gillberg and
the result of a natural impulse or a Åkerstedt (1994); Lorist et al. (2000);
learned behavior as these activities van der Linden et al. (2003), (2003b);
require a high degree of processing Boksem et al. (2005); Lorist et al.
information in order to be performed (2005); Boksem et al. (2006); Tixier
(Boonstra et al. 2007). Basic cognitive et al. (2014).
functions have been defined as
attention, concentration, reaction time,
and well learned behavior. (Jovanović
et al. 2012)
Illnesses Illnesses refers to the affections that the Fredriksson et al. (1999);Van Der
individual can develop as a result of an Hulst and Geurts (2001); Amick III et
extended period of fatigue. Fatigue al. (2002); Kivimäki et al. (2002).
debilitates our immune system and
causes stress which opens the door for
health problems.

36
 Error or Error or Injuries refers to any kind of Lorist et al. (2000); van der Linden et
Injuries procedure or action that is not perform al. (2003), (2003b)
in the correct way, due to fatigue, and
that causes some kind of loss or
damage to the workers.
GEN Bad Mood Mood comprehends numerous Opstad et al. (1978); Mikulincer et al.
variables such as tension, anger, vigor, (1989); Pilcher and Huffcutt (1996);
anxiety, and depression. An important Dinges et al. (1997); Philibert (2005);
number of studies utilize the Profile of Sagaspe et al. (2006); Scott et al.
Mood State to assess mood. (2006); Boonstra et al. (2007); Kajtna
et al. (2011).
Cognitive See above Opstad et al. (1978); Yoshitake
Degradation (1978); Naitoh (1981); Angus and
Heslegrave (1985); Brown (1994);
Meijman (1997); Harrison and Horne
(1999), (2000); Lorist et al. (2000);
Jones and Harrison (2001); Lipscomb
et al. (2002); Belenky et al.(2003);
Van der Linden et al. (2003);
Gosselin et al. (2005); Nilsson et al.
(2005); Philibert (2005);Van Dongen
and Dinges (2005); Boonstra et al.
(2007); Lim and Dinges (2008);
Zhang and Liu (2008); Fogt et al.
(2010)
Physical Physical Degradation refers to any Yoshitake (1978); Grandjean (1979);
Degradation bodily symptom that impairs the Lipscomb et al. (2002); Nindl et al.
normal functioning of the body, such as (2002).
pain, somnolence, and lack of strength.
Illnesses See above. Segawa et al. (1987); Scott and LaDou
(1994); (Xu et al. (1994); Steenland
and Fine (1996); Krause et al. (1997);
Alfredsson et al. (1982); Iwasaki et al.
(1998); Bøggild and Knutsson (1999);
Hansen (2001); Nylén et al. (2001);
Åkerstedt et al. (2002); Liu and
Tanaka (2002); Knutsson (2003).
Error or See above Harrington (2001); Nylén et al.
Injuries (2001); Folkard and Tucker (2003);
Dembe et al. (2005).
MUS Physical See above. Grandjean (1979); Scott and LaDou
Degradation (1994); Johnston (1998); Yaggie and
Armstrong (2004).
Illnesses See above. Fredriksson et al. (1999)

2.7 Relationships among outcomes of fatigue:

Similar to section 2.5, there are some consequences of fatigue that exacerbate other

outcome factors, which, can magnify or perpetuate the effects of acute fatigue. These factors are

37
most often cited in medical, neuroscience, and psychology research because they are related to the

function (or dysfunction) of the human body.

2.7.1 Impacts of cognitive degradation on other outcomes

As cognition degrades, a plethora of negative outcomes arise. From a mental standpoint,

cognitive fatigue reduces complex and basic cognitive function, negatively affects mood, and

increases reaction time. From primary impacts of fatigue arise a host of secondary outcomes. For

example, weakened basic and complex cognitive functions increase human error, injuries, and

decrease the quality and productivity of work completed (Craig 1992; van der Linden et al. 2003;

Fogt et al. 2010). Also, mood changes often cause the affected worker to become more irritable,

thereby degrading workplace, social, and family relationships (Williams and Alliger 1994).

Finally, a delayed reaction time increases the rate of human error and increases safety risk

(Harrison and Horne 2000; Fogt et al. 2010). It should be noted that most authors agree that basic

cognitive functions are more resilient to fatigue than complex cognitive functions and mood.

However, even modest levels of fatigue can affect basic cognitive functions.

2.7.2 Impacts of physical degradation on other outcomes

After an intense or prolonged period of physical stress, human bodies experience muscular

fatigue, which is accompanied by a less noticeable, but equally important, reduction in motor

function (Grandjean 1979; Johnston 1998). The chemical, neural, and hormonal changes that take

place in human bodies when physically stressed is a natural defense that prevents further

degradation (Piper 1989; Brown 1994; Jason et al. 2010). However, these defenses slow down

physical reactions, decrease balance, and decrease coordination, making a worker prone to errors

and injuries (Grandjean 1979; van der Linden et al. 2003). Dramatic reductions in balance have

38
been shown to be a delayed consequence of fatigue that affects the worker even during rest periods

and social activity (Gosselin et al. 2005).

2.7.3 Impacts of illnesses on other outcomes

There are many ways that fatigue causes physical debilitation; however, it also causes long-

term impacts such as illnesses. When fatigue causes illness, the effects of the illness, in turn, often

reinforce fatigue thereby causing a downward cycle. Jason et al. (2010) found that between 5%

and 8% of the general population experience the fatigue-illness cycle for 1 to 5 months.

Table 4 Relationships among consequences of fatigue

Factor Impacted References
Factor Name
Cognitive Illnesses Brown (1994); Belenky et al. (2003).
performance Error or Tepas and Mahan (1989); Craig (1992); Czeeisler, et al. (1992);
Injuries Okogbaa et al. (1994); Meijman (1997); Harrison and Horne (1999),
(2000); Lorist et al. (2000); Jones and Harrison (2001); van der
Linden et al. (2003); Boonstra et al. (2007); Fogt et al. (2010).
Physical Illnesses Segawa et al. (1987); Scott and LaDou (1994); Knutsson (2003);
Degradation Lipscomb et al. (2002); Gosselin et al. (2005).
Error or Grandjean (1979); Craig (1992); Czeisler et al. (1992); Johnston et al.
Injuries (1989) Lorist et al. (2000); van der Linden et al. (2003); Yaggie and
Armstrong (2004).

2.8 Systems model:

After analyzing the causes and consequences of fatigue we aimed to represent the findings

in a systems model to illustrate the strengths and gaps in the current body of knowledge. Figure 2

illustrates the complexity of the relationships between causes and consequences of fatigue, where

fatigue is the mediating factor. The relationships among factors are represented by arrows and the

volume of research developed in that line of inquiry is represented by the thickness of the line.

Therefore, it becomes easy to distinguish mature areas of research from those where additional

research is needed. All the information presented in the model is a summary of the information

39
presented in Tables 1-4. It should be noted that the subsequent meta-analysis illustrates the strength

and direction of the effect sizes among these factors.

40
 CAUSES CONSEQUENCES

Emotional
Predisposition
Social
Environment

Cognitive
Mental Degradation
Work Exertion
Environment
MENTAL
Work FATIGUE
Error
Overtime Load Injury
LWH
Bad
Sleep
FATIGUE Mood Illness
Deprivation
es

MUSCULAR
Incomplete FATIGUE Physical
Recovery Degradation
Muscular
Exertion

Figure 2: Systems model illustrating empirical literature on the relationships among causes and effects of fatigue.
Notation: Line thickness corresponds directly to the number of studies devoted to each relationship.
Immediate relationships between causes, fatigue, and consequences.
Inter-relationships within causes or consequences of fatigue.

41
2.9 Model limitations

The systems model has numerous limitations. First, the model is built completely from

past research and the structure and content of the model was derived from what has been studied

in the past. Thus, the model reflects any limitations associated with the topics of investigation

to date. Second, the completeness of the model is commensurate with the completeness of the

literature review. Our analysis mainly considered research published in English in top

occupational safety and health, applied psychology, cognitive psychology, medical, industrial

systems, engineering, sleep, and neuroscience journals. To ensure completeness, we also cross-

referenced all references of past papers to ensure popular studies were not omitted. Nevertheless,

it is possible that some research was missed and that the model could be improved. Third, we

made the implicit assumption that the mechanics of fatigue are the same regardless of ethnic

background, culture, and occupation. The body of research affirms this assumption; however, it

remains just that: an assumption that has yet to be validated. Fourth, our review and the model

only include the direct causes and consequences of fatigue. There may be a plethora of factors

that indirectly affect fatigue or are an indirect outcome of fatigue through some mediating

variable. These interactions are out of the scope of this research. Fifth, the model does not

include direct connections between causal and outcome factors without fatigue as a mediator.

Future researchers may wish to study these direct relationships as a means to identify when

outcomes are falsely attributed to fatigue and when other mediating variables may be at play.

Finally, our descriptions of each causal factor and outcome of fatigue were abbreviated. There

is a wealth of information on research methods, nuances of findings, and implications that we

simply could not include within the scope of one paper. Rather, our aim was to produce a single-

source systems model of the many antecedents and consequences of fatigue and to acknowledge

42
the key research used to build a valid model. The meta-analysis presented in section 2.10 offers

a formal statistical analysis of published data in order to measure and model the effect sizes of

the relationships shown in Figure 2.

2.10 Meta-analysis approach

Since our main objective was to summarize the body of knowledge related to the causes

and consequences of occupational fatigue, we complemented the systems model with a formal

analysis of the magnitude and direction of the relationships. To achieve this objective we

performed a statistical meta-analysis. When formally performed, a meta-analysis allows one to

codify the effect sizes from multiple studies into a single statistic. Since effect sizes are

independent of the sample size and have a reportable standard error, they can be compared and

aggregated across multiple studies. Such analysis allowed us to identify the factors that have the

largest impact on fatigue and the factors that are most impacted by fatigue.

We used four basic steps in our meta-analysis as suggested by Lipsey and Wilson (2001)

and Field and Gillett (2010). The first step, studies selection, involved a comprehensive

literature review of studies that reported sufficient empirical data to calculate an effect size.

Additionally, in this step each study is evaluated according to certain criteria of quality and

methodology appropriates for their inclusion in the meta/analysis. Although each study

included in the systems model reported empirical evidence, many studies did not report

sufficient detail of their analysis and, thus, were not included in the meta-analysis. The second

step, studies coding, involved extracting the data needed to compute an effect size for each

43
relationship of interest. The result of this step is a matrix of effect sizes reported by all scientific

studies for each relationship. The third step, standardization, involves translating the effect sizes

between studies into common units so that they can be size aggregated. Finally, the fourth step,

overall effect size computation, involves computing a single effect size that combines the

standardized data from multiple studies. The details associated with each step are described

below.

2.10.1 Studies selection

In order to be included in the meta-analysis, a study must: (1) contain results about a

clear and specific relationship between a cause of fatigue and fatigue, or fatigue and a

consequence of fatigue; (2) involve a healthy working sample of subjects between 18 and 65

years of age and represent the working population (i.e., studies limited to athletes, children, sick,

or disabled persons were omitted); (3) include either an effect size or enough information to

compute an effect size; and (4) involve 5 participants or more. Of the 105 studies of occupational

fatigue that report empirical data, 23 met all four criteria.

It was also important also to ensure that only valid and reliable data were used in the analysis.

From the 23 studies a total of 64 effect sizes were obtained, including correlation coefficients

(r), Risk Ratios (RR), and Odds Ratios (OR). However, not all these effect sizes were acceptable

for the analysis because they were not accompanied by proper demographic statistics (e.g.,

sample size of the control group). This restriction yielded a total of 29 salient effect sizes.

44
2.10.2 Studies coding

The coding strategy to extract effect sizes from a study depended on the target aggregate

statistic. Since all of the variables of interest could be converted to a continuous scale and the

standard deviations could be pooled, Cohen’s d is the ideal statistic (Cohen 1988). Cohen’s d

yields information about the difference in means of two groups, measured in standard deviations.

We were able to use Cohen’s d because studies that reported the results of a questionnaire could

be as a relative continuous score and studies that reported objective measures such as reaction

time in a PVT were naturally continuous. When the raw data were available, the following

process was used.

̅̅̅̅
𝑋2 −𝑋̅̅̅̅
𝐶𝑜ℎ𝑒𝑛′ 𝑠 𝑑 = 𝑆 1
(1)
𝑃𝑜𝑜𝑙𝑒𝑑
Where 𝑋̅̅̅̅ ̅̅̅
1 is the expected value of the control group sample, 𝑋2 is the expected
value of the treatment group sample, and 𝑆𝑃𝑜𝑜𝑙𝑒𝑑 is the pooled standard
deviation.

We used the method introduced by Lipsey and Wilson (2001) to pool the standard error.

𝑛1 + 𝑛2 𝑑2
𝑆𝐸𝑑 = √ + (2)
𝑛1 𝑛2 2( 𝑛1 + 𝑛2 )
Where 𝑆𝐸𝑑 represents the standard error of the statistic d; 𝑛1 is the sample size
of the control group, 𝑛2 is the sample size of the treatment group and 𝑑 is the
effect size (Cohen’s d)

In order for a study to be included in our sample, the effect size and the standard

deviation had to be reported. Sometimes these variables are directly presented in a paper; other

times these variables can be extracted from the results of a t-test, an F-test, or an analysis of

variance (ANOVA), to mention a few. Since the effect size of small samples has been shown

to be biased by outliers, we used a sample size correction introduced by Hedges (1981). This

correction, shown in the equation below, reduces the relative impact of studies that report sample

45
sizes less than 20. When sample sizes are greater than 20, the adjustment is nullified. The

equation presented by Hedges (1981) is as follows:

3
𝑑′ = (1 − 4𝑁−9) 𝑑 (3)
Where d’ represents the unbiased effect size and N is the total sample (𝑛1 + 𝑛2 ).

2.10.3 Standardization

In some cases, effects sizes were reported by authors using a statistic other than Cohen’s

d (e.g., regression coefficient). These statistics were transformed into a Cohen’s d and a

corresponding confidence interval of 95% was created. When a correlation coefficient was

reported, Cohen’s d was calculated using the following formula:

2𝑟
𝑑𝑟 = √1−𝑟 2 (4)
Where 𝑑𝑟 represents the Cohen’s d effect size calculated from a correlation
coefficient (r).

The standard error was calculated as the standard error of a Fisher’s distribution and then

translated back with the hyperbolic function. This value was also transformed into a d effect

size.
1
𝑆𝐸𝑍𝑟 = (5)
√𝑛−3
Where 𝑆𝐸𝑍𝑟 represents the standard error of a correlation coefficient in a
Fisher’s distribution and n is the sample size.

For the case where Risk Ratios and Odds Ratios were reported with their standard error,

the statistics were standardized into a Cohen’s d using the following equation:

ln(𝑂𝑅)√3
𝑑𝑂𝑅 = (6)
𝜋
Where 𝑑𝑂𝑅 represents the Cohen’s d effect size calculated from an Odds Ratio
(OR).

46
 Ultimately, this process yielded all affect sizes expressed as a Cohen’s d with a
corresponding 95% confidence interval as suggested by Lipsey and Wilson (2001).

2.10.4 Overall effect size calculation

In this last step the effect sizes for multiple studies investigating the same relationship

were aggregated into an overall effect size. There are two basic ways of performing this

aggregation: the homogeneous effect meta-analysis and the random effect meta-analysis. The

homogeneous effect meta-analysis is only valid for exact replications of an experiment. Here,

there were few replications and far more studies that used a variety of methods and sample sizes.

Thus, we used the random effect meta-analysis in those situations. This method not only

accounts for variability within a sample due to chance, but also accounts for true variability

among samples from different studies. This ensured that each effect size was weighted according

to the quality of the study from which it was taken. Furthermore, random effect meta-analysis

doesn’t allow any study to dominate. The inverse variance weight method was employed and

each effect size (d) was weighted by the inverse of its variance. An overall effect for each

relationship was calculated using Eq. (7). In some cases, when only one study reported an effect

size for a relationship, the effect was converted to Cohen’s d and reported for such relationship.

∑ 𝑤𝑖 𝑑𝑖
𝐸𝑆𝑑 = ∑ 𝑊𝑖
(7)
Where 𝐸𝑆𝑑 represents the overall effect size, 𝑊𝑖 stands for each individual effect
size weight and 𝑑𝑖 denotes each individual effect size.

The standard error was computed using Eq. (8):

1
𝑆𝐸𝐸𝑆𝑑 = √∑ 𝑊 (8)
𝑖

47
 Where 𝑆𝐸𝐸𝑆𝑑 represents the overall effect size standard error and 𝑊𝑖 stands for
each individual effect size weight.

The equations presented above are sufficient for the fixed effect meta-analysis, however,

in order to compute a random effect meta-analysis three additional equations are required to

evaluate the homogeneity, random effect variance, and random effect weights using Eq. 9, 10,

and 11, respectively.

𝑄 = ∑ 𝑤𝑖 (𝑑𝑖 − 𝐸𝑆𝑑 )2 (9)

Where 𝑄 represents the homogeneity statistic, 𝑑𝑖 denotes each individual effect
size, and 𝐸𝑆𝑑 represents the overall effect size.

The homogeneity statistic is used to calculate the variance associated with true

differences among different studies, such variance receives the name of random effect

variance (t2) as shown in q. 10.

𝑄−𝑑𝑓𝑄
t2 = ∑ 𝑊2
(10)
∑ 𝑊𝑖 − 𝑖
∑ 𝑊𝑖
2
Where t represents the random effect variance, 𝑄 is the homogeneity statistic,
𝑑𝑓𝑄 represents the degrees of freedom of Q, and 𝑊𝑖 stands for each individual
effect size weight.

With this random effect variance, a new set of weights for each effect size can be calculated.

1
𝑊𝑖 ′ = (11)
𝑆𝐸𝑑2 + 𝑇 2

Where 𝑊𝑖′ represents the random effect weights for each effect size, 𝑆𝐸𝑑 is the

standard error of the statistic d, and t2 stands for the random effect variance.

With this new set of weights, the random effect sizes are calculating by applying Eq. (7) and

substituting 𝑊𝑖 by 𝑊𝑖 ′.

48
2.11 Results of the meta-analysis

The application of our meta-analysis method yielded the results summarized in Tables 5

through 8. Note that when only one study is reported on a specific relationship, the Cohen’s d

in the tables represents the effect size of that single study. In these tables, the column heading

“n.t” represents the sample size of the treatment group, “n.c” represents the sample size of the

control group, and 95% CI represents the upper and lower bounds of the 95% confidence

interval. The relationships derived from this meta-analysis are depicted in a revised systems

model in Figure 3. Here, the line thickness corresponds directly to Cohen’s d, the aggregated

effect size.

49
 Table 5 Effects of causes of fatigue
Relationship Source n.t. n.c. Cohen’s d 95% CI p-value
Gillberg and Akerstedt (1994)
Sleep Deprivation – General Fatigue 15 15 1.22 -0.86 3.98 0.125
Herscovitch and Broughton (1981)
Meijman (1997)
Mental Exertion – Mental Fatigue 49 45 0.91 0.47 1.34 0.000
Van der Linden et al. (2003)
Workload – General Fatigue Åkerstedt et al. (2002) 18828 18828 0.25 0.23 0.28 0.000
Overtime and LWH – Mental Fatigue Van Der Hulst and Geurts (2001) 325 325 0.43 0.11 0.75 0.000
Park et al. (2001)
Van Der Hulst and Geurts (2001)
Overtime and LWH – General Fatigue 19652 19864 0.49 0.16 0.82 0.002
Åkerstedt et al. (2002)
Beckers et al. (2008)
Environment – General Fatigue Gonzalez-Alonso (1999) 7 7 2.83 1.30 4.36 0.000
Maslach and Jackson (1984) 30 30
Social Environment – Mental Fatigue 0.39 0.04 0.74 0.015
Bültmann et al. (2001) 11020 11020
Note: n.t. = treatment group sample size; n.c.= control group sample sizew; 95% CI= 95% confidence interval.
One tailed p-values.

Table 6 Effects among causes of fatigue

Relationship Source n.t. n.c. Cohen’s d 95% CI p-value
Overtime – Incomplete recovery Van Der Hulst and Geurts (2001) 323 323 0.65 0.36 0.95 0.000
Note: n.t. = treatment group sample size; n.c.= control group sample sizew; 95% CI= 95% confidence interval.
One tailed p-values.

50
 Table 7 Effects of fatigue
Relationship Source n.t. n.c. Cohen’s d 95% CI p-value
Van der Linden et al. (2003)
Mental Fatigue - Cognitive Degradation 48 44 2.24 -0.41 4.90 0.049
Boksem et al. (2005)
Mental Fatigue – Illnesses Fredriksson et al. (1999) 484 484 0.33 0.26 0.40 0.000
Mental Fatigue – Error and Injury Van der Linden et al. (2003b) 36 32 0.32 -0.23 0.88 0.126
Scott et al. (2006)
General Fatigue - Bad Mood 15 15 0.95 0.02 1.88 0.023
Kajtna et al. (2011)
Gosselin et al. (2005)
General Fatigue - Cognitive Degradation 1039 1039 1.00 0.91 1.09 0.000
Philibert (2005)
Alfredsson et al (1982)
General Fatigue – Illnesses 594 1328 0.10 -0.04 0.23 0.079
Liu and Tanaka (2002)
Dembe et al. (2005) 1080
General Fatigue - Error and Injury 10805 0.64 -0.01 1.29 0.026
Gosselin et al. (2005) 4
Muscular Fatigue - Physical Degradation Yaggie and Armstorn (2004) 16 16 0.99 0.25 1.73 0.004
Muscular Fatigue - Illnesses Fredriksson et al. (1999) 232 232 0.41 0.00 0.82 0.000
Note: n.t. = treatment group sample size; n.c.= control group sample sizew; 95% CI= 95% confidence interval.
One tailed p-values.

Table 8 Effects among consequences of fatigue

Relationship Source n.t. n.c. Cohen’s d 95% CI p-value
Illnesses – Cognitive Degradation Vercoulen et al. (1994) 298 60 1.97 1.65 2.28 0.000
Note: n.t. = treatment group sample size; n.c.= control group sample sizew; 95% CI= 95% confidence interval.
One tailed p-values.

51
 CAUSES CONSEQUENCES

Social
Environment

Incomplete
Mental
Recovery
Exertion
MENTAL (2.24) Cognitive
Overtime FATIGUE Degradation
and LWH Work
Load
0.95 Bad Error or Illnesses
FATIGUE
Mood Injury
Work
Environment (0.64)
Sleep MUSCULAR
Deprivation FATIGUE

Physical
Degradation

Figure 3: Systems model represents effect sizes (Cohen’s d values) for each relationship.

Notation: Arrows indicate direction and line thickness corresponds directly to the magnitude of the effect size from the meta-analysis; (n) indicates
a 95% CI that includes zero or non-significance for such an effect size.

52
2.12 Discussion of the meta-analysis:

In a meta-analysis, an effect between 0 and 0.20 is considered small; between 0.20 and 0.50

is considered medium; and over 0.50, preferable closer to 0.80, is considered large (Cohen 1988).

Although most results achieved levels of high statistical significance, some were weak, either

because they didn’t achieved significance or because their confidence interval was too wide, even

including negative values. For example, the effect of sleep deprivation on fatigue was not

statistically significant and despite the large effect sizes reported, its confidence interval includes

the possibility of a null effect. This was observed because one study in this domain reported results

that were not statistically significant (Herscovitch and Broughton 1981). However, the other study

included in this relationship presented a large and significant effect (d=2.33; 95% C.I. = 0.97 to

3.68; p-value = 0.000) indicating an important influence of sleep deprivation in the development

of fatigue (Gillberg and Åkerstedt 1994). Intuitively some may think that a study that does not

report statistical significance should be discarded; however, a meta-analysis all past research

results must be considered for a relationship as long as they were obtained through high-quality

methods.

The largest significant effect, the relationship between environmental conditions at work

such as noise, vibration, and temperature and fatigue, was observed by Gonzalez-Alonzo (1999)

in a single study. Such a strong finding with little previous validation indicates the strong need for

scientific replication. Other medium effects include the impact of mental exertion, overtime, and

LWH on mental fatigue and the impact of social environment and workload on general fatigue.

53
 Only one study provided useful information about relationships among causes of fatigue

(Van Der Hulst and Geurts 2001), showing a strong relationship between overtime, LWH, and

incomplete recovery. This study confirms that overtime plays an important role as inhibitor of a

proper recovery but, again, should be replicated by future researchers to confirm the effects.

Most impacts of fatigue on physiology proved to be large. This was especially true for the

effect of general fatigue among cognitive degradation (d=1.00; 95% C.I. = 0.91 to 1.09) and

muscular fatigue among physical degradation (d=0.99; 95% C.I. = 0.25 to 1.73). This is not

surprising given the relatively large body of knowledge focused on this domain.

The effect sizes of the relationship between mental fatigue and illnesses showed to be

moderate and significant. However, this was not the case for general fatigue. A large and

significant effects can be observed between general fatigue and error. These results make sense

since errors are common consequences of acute fatigue. The influence of fatigue on mood has been

characterized as important with a large effect size, (d=0.95; 95% C.I. = 0.02 to 1.82) which agrees

with previous research.

2.13 Conclusions and recommendations:

The systems models presented in Figure 2 shows clearly that there is much less research

on the system of causal factors of fatigue than on the system of outcomes of fatigue. This may be

due to the fact that the outcomes of fatigue are more readily identifiable, and measurable than the

causal factors, which can be more difficult to empirically measure. Also, the psychological, health,

and medical research community that deals with the health and wellness aspects of fatigue is quite

54
large in comparison to the industrial systems, safety and health, and human factors research

community that often deals with the occupational causes of fatigue.

With respect to the causes of fatigue, Figure 2 clearly reveals some knowledge gaps that

may be useful for framing new research. For example, there is a wide range of research about the

way that fatigue arising from sleep deprivation impacts workers, and also about the way that

irregular and extended shifts impact workers. However, there is a clear need for additional research

about the way that workload impacts muscular exertion and the recovery process. Both have a

logical foundation that could be supported through rigorous research. These are simply a sample

of obvious gaps in the knowledge system. The reader may find additional connections upon which

to build a future study.

With respect to the consequences of fatigue, even though there is an abundant body of

literature addressing the relationship between general fatigue and mood changes, there is no

significant research that investigates the relationship between exclusively muscular or mental

fatigue and mood changes. Additionally, during our research we also found it strange that there is

no substantial research about the way in which mood changes affect work relationships, which

seems to be a reinforcing process. The relationship between mood and cognitive functions is ripe

for both confirmatory and seminal investigation.

The meta-analysis of the causes and consequences of fatigue illustrated in Figure 2 revealed

significant gaps in empirical evidence of fatigue-related factors. In fact, only 23 studies provided

sufficient empirical data. Furthermore, there is a dearth of replication among fatigue studies,

55
indicating a strong need for scientific validation. Future research is especially needed to confirm

proposed relationships among causal factors and among outcomes of fatigue. Since the medical

field has examined the relationship between fatigue and physiology, the impact of fatigue on other,

less direct outcomes like work performance, is needed.

This paper adds to the efforts of previous researchers to help formalize the study of fatigue

among the working population. It can serve as a point of departure for future research to evaluate

what has been done, where extra research is needed, and how to organize such research to

contribute to the enlargement of the current body of knowledge.

2.14 Acknowledgements

We would like to thank the ELECTRI International for supporting this research. Any

opinions, findings, and conclusions or recommendations expressed in this material are those of the

authors and do not necessarily reflect the views of ELECTRI International.

2.15 References

Åkerstedt, T. (1990). “Psychological and psychophysiological effects of shift work.”

Scandinavian Journal of Work, Environment & Health, 16, 67–73.
Åkerstedt, T., Fredlund, P., Gillberg, M., and Jansson, B. (2002a). “A prospective study of fatal
occupational accidents–relationship to sleeping difficulties and occupational factors.”
Journal of Sleep Research, 11(1), 69–71.
Åkerstedt, T., Fredlund, P., Gillberg, M., and Jansson, B. (2002b). “Work load and work hours in
relation to disturbed sleep and fatigue in a large representative sample.” Journal of
Psychosomatic Research, 53(1), 585–588.
Alfredsson, L., Karasek, R., and Theorell, T. (1982). “Myocardial infarction risk and psychosocial
work environment: An analysis of the male Swedish working force.” Social Science &
Medicine, 16(4), 463–467.
Amick III, B. C., McDonough, P., Chang, H., Rogers, W. H., Pieper, C. F., and Duncan, G. (2002).
“Relationship between all-cause mortality and cumulative working life course

56
 psychosocial and physical exposures in the United States labor market from 1968 to 1992.”
Psychosomatic Medicine, 64(3), 370–381.
Angus, R. G., and Heslegrave, R. J. (1985). “Effects of sleep loss on sustained cognitive
performance during a command and control simulation.” Behavior Research Methods,
Instruments, & Computers, 17(1), 55–67.
Axelsson, J., Kecklund, G., \AAkerstedt, T., and Lowden, A. (1998). “Effects of alternating 8-and
12-hour shifts on sleep, sleepiness, physical effort and performance.” Scandinavian journal
of work, environment & health, 62–68.
Baulk, S. D., Fletcher, A., Kandelaars, K. J., Dawson, D., and Roach, G. D. (2009). “A field study
of sleep and fatigue in a regular rotating 12-h shift system.” Applied ergonomics, 40(4),
694–698.
Beckers, D. G. J., Linden, D. van der, Smulders, P. G. W., Kompier, M. A. J., Taris, T. W., and
Geurts, S. A. E. (2008). “Voluntary or involuntary? Control over overtime and rewards for
overtime in relation to fatigue and work satisfaction.” Work & Stress, 22(1), 33–50.
Belenky, G., Wesensten, N. J., Thorne, D. R., Thomas, M. L., Sing, H. C., Redmond, D. P., Russo,
M. B., and Balkin, T. J. (2003). “Patterns of performance degradation and restoration
during sleep restriction and subsequent recovery: a sleep dose-response study.” Journal of
Sleep Research, 12(1), 1–12.
Bliwise, D. L., King, A. C., Harris, R. B., and Haskell, W. L. (1992). “Prevalence of self-reported
poor sleep in a healthy population aged 50–65.” Social Science & Medicine, 34(1), 49–55.
Bøggild, H., and Knutsson, A. (1999). “Shift work, risk factors and cardiovascular disease.”
Scandinavian Journal of Work, Environment & Health, 25(2), 85–99.
Boksem, M. A., Meijman, T. F., and Lorist, M. M. (2006). “Mental fatigue, motivation and action
monitoring.” Biological psychology, 72(2), 123–132.
Boksem, M. A. S., Meijman, T. F., and Lorist, M. M. (2005). “Effects of mental fatigue on
attention: An ERP study.” Cognitive Brain Research, 25(1), 107–116.
Boksem, M. A. S., and Tops, M. (2008). “Mental fatigue: Costs and benefits.” Brain Research
Reviews, 59(1), 125–139.
Boonstra, T. W., Stins, J. F., Daffertshofer, A., and Beek, P. J. (2007). “Effects of sleep deprivation
on neural functioning: an integrative review.” Cellular and Molecular Life Sciences, 64(7–
8), 934–946.
Brown, I. D. (1994). “Driver fatigue.” Human Factors: The Journal of the Human Factors and
Ergonomics Society, 36(2), 298–314.
Bültmann, U., Kant, I., Schröer, C., and Kasl, S. (2001). “The relationship between psychosocial
work characteristics and fatigue and psychological distress.” International Archives of
Occupational and Environmental Health, 75(4), 259–266.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences. Lawrence earbaum
Assuati. New York.
Craig, A. (1992). “Symptoms of acute and chronic fatigue.” Handbook of human performance,
Vol. 1: The physical environment; Vol. 2: Health and performance; Vol. 3: State and trait,
A. P. Smith and D. M. Jones, eds., Academic Press, San Diego, CA, US, 289–339.
Dawson, D., and McCulloch, K. (2005). “Managing fatigue: It’s about sleep.” Sleep Medicine
Reviews, 9(5), 365–380.
Dembe, A. E., Erickson, J. B., Delbos, R. G., and Banks, S. M. (2005). “The impact of overtime
and long work hours on occupational injuries and illnesses: new evidence from the United
States.” Occupational and Environmental Medicine, 62(9), 588–597.

57
Dinges, D. F., Pack, F., Williams, K., Gillen, K. A., Powell, J. W., Ott, G. E., Aptowicz, C., and
Pack, A. I. (1997). “Cumulative sleepiness, mood disturbance and psychomotor vigilance
performance decrements during aweek of sleep restricted to 4-5 hours per night.” Sleep:
Journal of Sleep Research & Sleep Medicine, 20(4), 267–277.
Edwards, R. G., and Lippold, O. C. J. (1956). “The relation between force and integrated electrical
activity in fatigued muscle.” The Journal of Physiology, 132(3), 677–681.
Edwards, R. H., Hill, D. K., and Jones, D. A. (1975). “Heat production and chemical changes
during isometric contractions of the human quadriceps muscle.” The Journal of Physiology,
251(2), 303–315.
Field, A. P., and Gillett, R. (2010). “How to do a meta-analysis.” British Journal of Mathematical
and Statistical Psychology, 63(3), 665–694.
Fogt, D. L., Kalns, J. E., and Michael, D. J. (2010). “A Comparison of Cognitive Performance
Decreases During Acute, Progressive Fatigue Arising From Different Concurrent
Stressors.” Military Medicine, 175(12), 939–944.
Folkard, S., and Tucker, P. (2003). “Shift work, safety and productivity.” Occupational medicine,
53(2), 95–101.
Gibson, A. S. C., Baden, D. A., Lambert, M. I., Lambert, E. V., Harley, Y. X. R., Hampson, D.,
Russell, V. A., and Noakes, T. D. (2003). “The Conscious Perception of the Sensation of
Fatigue.” Sports Medicine, 33(3), 167–176.
Gillberg, M., and Åkerstedt, T. (1994). “Sleep restriction and SWS-suppression: effects on
daytime alertness and night-time recovery.” Journal of Sleep Research, 3(3), 144–151.
Gosselin, A., De Koninck, J., and Campbell, K. B. (2005). “Total sleep deprivation and novelty
processing: implications for frontal lobe functioning.” Clinical Neurophysiology, 116(1),
211–222.
Grandjean, E. (1979). “Fatigue in industry.” British Journal of Industrial Medicine, 36(3), 175–
186.
Hallowell, M. R. (2010). “Worker fatigue: managing concerns in rapid renewal highway
construction projects.” Professional Safety, 55(12), 18–26.
Hansen, J. (2001). “Increased Breast Cancer Risk among Women Who Work Predominantly at
Night.” Epidemiology, 12(1), 74–77.
Hardy, G. E., Shapiro, D. A., and Borrill, C. S. (1997). “Fatigue in the workforce of national health
service trusts: Levels of symptomatology and links with minor psychiatric disorder,
demographic, occupational and work role factors.” Journal of Psychosomatic Research,
Workplace, 43(1), 83–92.
Harrington, J. M. (2001). “Health effects of shift work and extended hours of work.” Occupational
and Environmental Medicine, 58(1), 68–72.
Harrison, Y., and Horne, J. A. (1999). “One Night of Sleep Loss Impairs Innovative Thinking and
Flexible Decision Making.” Organizational Behavior and Human Decision Processes,
78(2), 128–145.
Harrison, Y., and Horne, J. A. (2000a). “The impact of sleep deprivation on decision making: A
review.” Journal of Experimental Psychology: Applied, 6(3), 236–249.
Harrison, Y., and Horne, J. A. (2000b). “The impact of sleep deprivation on decision making: A
review.” Journal of Experimental Psychology: Applied, 6(3), 236–249.
Hedges, L. V. (1981). “Distribution Theory for Glass’s Estimator of Effect size and Related
Estimators.” Journal of Educational and Behavioral Statistics, 6(2), 107–128.

58
Herscovitch, J., and Broughton, R. (1981). “Performance deficits following short-term partial sleep
deprivation and subsequent recovery oversleeping.” Canadian Journal of
Psychology/Revue canadienne de psychologie, 35(4), 309.
van der Hulst, M., and Geurts, S. (2001). “Associations between overtime and psychological health
in high and low reward jobs.” Work and Stress, 15(3), 227–240.
van der Hulst, M., Veldhoven, M. van, and Beckers, D. (2006). “Overtime and Need for Recovery
in Relation to Job Demands and Job Control.” Journal of Occupational Health, 48(1), 11–
19.
Iwasaki, K., Sasaki, T., Oka, T., and Hisanaga, N. (1998). “Effect of Working Hours on Biological
Functions related to Cardiovascular System among Salesmen in a Machinery
Manufacturing Company.” Industrial Health, 36(4), 361–367.
Jason, L. A., Evans, M., Brown, M., and Porter, N. (2010). “What is fatigue? Pathological and
nonpathological fatigue.” PM&R, 2(5), 327–331.
Johnston, M., Pollard, B., Manktelow, A., and others. (1989). “Stress and information processing
in 8 and 12-hour nursing shifts.” The pychologist.
Johnston, R., Howard, M., Cawley, and Posse, G. (1998). “Effect of lower extremity muscular
fatigue on motor control performance.” Medicine & Science in Sports & Exercise®,
30(12), 1703–1707.
Jones, K., and Harrison, Y. (2001). “Frontal lobe function, sleep loss and fragmented sleep.” Sleep
Medicine Reviews, 5(6), 463–475.
Jovanović, M., Sporiš, G., Šopar, J., Harasin, D., and Matika, D. (2012). “The Effects of Basic
Military Training on Shooting Tasks in Conditions of Sleep Deprivation.” Kinesiology,
44(1), 169–177.
Kajtna, T., Štukovnik, V., and Grošelj, L. D. (2011). “Effect of acute sleep deprivation on
concentration and mood states with a controlled effect of experienced stress.” Zdravniški
Vestnik, 80(5).
Kivimäki, M., Leino-Arjas, P., Luukkonen, R., Riihimäi, H., Vahtera, J., and Kirjonen, J. (2002).
“Work stress and risk of cardiovascular mortality: prospective cohort study of industrial
employees.” BMJ, 325(7369), 857.
Knutsson, A. (2003). “Health disorders of shift workers.” Occupational Medicine, 53(2), 103–108.
Krause, N., Lynch, J., Kaplan, G. A., Cohen, R. D., Goldberg, D. E., and Salonen, J. T. (1997).
“Predictors of disability retirement.” Scandinavian Journal of Work, Environment &
Health, 23(6), 403–413.
Kroenke K, and Price RK. (1993). “Symptoms in the community: Prevalence, classification, and
psychiatric comorbidity.” Archives of Internal Medicine, 153(21), 2474–2480.
Lim, J., and Dinges, D. F. (2008). “Sleep Deprivation and Vigilant Attention.” Annals of the New
York Academy of Sciences, 1129(1), 305–322.
van der Linden, D., Frese, M., and Meijman, T. F. (2003a). “Mental fatigue and the control of
cognitive processes: effects on perseveration and planning.” Acta Psychologica, 113(1),
45–65.
van der Linden, D. van der, Frese, M., and Sonnentag, S. (2003b). “The Impact of Mental Fatigue
on Exploration in a Complex Computer Task: Rigidity and Loss of Systematic Strategies.”
Human Factors: The Journal of the Human Factors and Ergonomics Society, 45(3), 483–
494.

59
Lipscomb, J. A., Trinkoff, A. M., Geiger-Brown, J., and Brady, B. (2002). “Work-schedule
characteristics and reported musculoskeletal disorders of registered nurses.” Scandinavian
journal of work, environment & health, 394–401.
Lipsey, M. W., and Wilson, D. B. (2001). Practical meta-analysis. Applied social research
methods series; vol 49., Sage Publications, Inc, Thousand Oaks, CA, US.
Liu, Y., and Tanaka, H. (2002). “Overtime work, insufficient sleep, and risk of non-fatal acute
myocardial infarction in Japanese men.” Occupational and Environmental Medicine,
59(7), 447–451.
Lorist, M. M., Boksem, M. A., and Ridderinkhof, K. R. (2005). “Impaired cognitive control and
reduced cingulate activity during mental fatigue.” Cognitive Brain Research, 24(2), 199–
205.
Lorist, M. M., Klein, M., Nieuwenhuis, S., de Jong, R., Mulder, G., and Meijman, T. F. (2000).
“Mental fatigue and task control: Planning and preparation.” Psychophysiology, 37(5),
614–625.
Maslach, C., and Jackson, S. E. (1984). “Burnout in organizational settings.” Applied social
psychology annual.
McNair, D. M., Lorr, M., and Droppleman, L. F. (1971). Profile of mood states. Univ.
Meijman, T. F. (1997). “Mental fatigue and the efficiency of information processing in relation to
work times.” International Journal of Industrial Ergonomics, 20(1), 31–38.
Meijman, T., and Schaufeli, W. (1996). “Psychische vermoeidheid en arbeid. Ontwikkelingen in
de A&O-psychologie.” PSYCHOLOOG-AMSTERDAM-, 31, 236–241.
Mikulincer, M., Babkoff, H., Caspy, T., and Sing, H. (1989). “The effects of 72 hours of sleep loss
on psychological variables.” British Journal of Psychology, 80(2), 145–162.
Naitoh, P. (1981). “Circadian cycles and restorative power of naps.” Biological rhythms, sleep and
shift work, 553–580.
Nilsson, J. P., Söderström, M., Karlsson, A. U., Lekander, M., \AAkerstedt, T., Lindroth, N. E.,
and Axelsson, J. (2005). “Less effective executive functioning after one night’s sleep
deprivation.” Journal of sleep research, 14(1), 1–6.
Nindl, B. C., Leone, C. D., Tharion, W. J., Johnson, R. F., Castellani, J. W., Patton, J. F., and
Montain, S. J. (2002). “Physical performance responses during 72 h of military operational
stress.” Medicine & Science in Sports & Exercise, 34(11), 1814–1822.
Nylén, L., Voss, M., and Floderus, B. (2001). “Mortality among women and men relative to
unemployment, part time work, overtime work, and extra work: a study based on data from
the Swedish twin registry.” Occupational and Environmental Medicine, 58(1), 52–57.
Okogbaa, O. G., Shell, R. L., and Filipusic, D. (1994). “On the investigation of the
neurophysiological correlates of knowledge worker mental fatigue using the EEG signal.”
Applied Ergonomics, 25(6), 355–365.
Opstad, P. K., Ekanger, R., Nummestad, M., and Raabe, N. (1978). “Performance, mood, and
clinical symptoms in men exposed to prolonged, severe physical work and sleep
deprivation.” Aviation, Space, and Environmental Medicine, 49, 1065–1073.
Park, J., Kim, Y., Chung, H. K., and Hisanaga, N. (2001). “Long working hours and subjective
fatigue symptoms.” Industrial health, 39(3), 250–254.
Philibert, I. (2005). “Sleep loss and performance in residents and nonphysicians: a meta-analytic
examination.” SLEEP-NEW YORK THEN WESTCHESTER-, 28(11), 1392.
Pilcher, J. J., and Huffcutt, A. J. (1996). “Effects of sleep deprivation on performance: a meta-
analysis.” Journal of Sleep Research & Sleep Medicine, 19(4), 318–326.

60
Piper, B. F. (1989). “Fatigue: current bases for practice.” Key Aspects of Comfort: Management of
Pain, Fatigue, and Nausea. New York: Springer Publishing Co, 187–98.
Reiner, B. I., and Krupinski, E. (2011). “Demystifying Occupational Stress and Fatigue Through
the Creation of an Adaptive End-User Profiling System.” Journal of Digital Imaging,
25(2), 201–205.
Ribet, C., and Derriennic, F. (1999). “Age, working conditions, and sleep disorders: A longitudinal
analysis in the French cohort E.S.T.E.V.” Sleep: Journal of Sleep Research & Sleep
Medicine, 22(4), 491–504.
Ricci, J. A., Chee, E., Lorandeau, A. L., and Berger, J. (2007). “Fatigue in the U.S. workforce:
prevalence and implications for lost productive work time.” Journal of Occupational and
Environmental Medicine / American College of Occupational and Environmental
Medicine, 49(1), 1–10.
Roach, G. D., Fletcher, A., and Dawson, D. (2004). “A Model to Predict Work-Related Fatigue
Based on Hours of Work.” Aviation, Space, and Environmental Medicine, 75(3), A61–
A69.
Sagaspe, P., Sanchez-Ortuno, M., Charles, A., Taillard, J., Valtat, C., Bioulac, B., and Philip, P.
(2006). “Effects of sleep deprivation on Color-Word, Emotional, and Specific Stroop
interference and on self-reported anxiety.” Brain and Cognition, 60(1), 76–87.
Scott, A. J., and LaDou, J. (1994). “Health and safety in shift workers.” Occupational Medicine,
960–971.
Scott, J. P., McNaughton, L. R., and Polman, R. C. (2006). “Effects of sleep deprivation and
exercise on cognitive, motor performance and mood.” Physiology & behavior, 87(2), 396–
408.
Segawa, K., Nakazawa, S., Tsukamoto, Y., Kurita, Y., Goto, H., Fukui, A., and Takano, K. (1987).
“Peptic ulcer is prevalent among shift workers.” Digestive Diseases and Sciences, 32(5),
449–453.
Taris, T. W., Beckers, D. G. J., Dahlgren, A., Geurts, S. a. E., and Tucker, P. (2007). “Overtime
work and well-being : Prevalence, conceptualization and effects of working overtime.” 40,
2, 21–40.
Tepas, D. I., and Mahan, R. P. (1989). “The many meanings of sleep.” Work & Stress, 3(1), 93–
102.
Tepper, B. J. (2000). “Consequences of abusive supervision.” Academy of management journal,
43(2), 178–190.
Tixier, A. J.-P., Hallowell, M. R., Albert, A., van Boven, L., and Kleiner, B. M. (2014).
“Psychological antecedents of risk-taking behavior in construction.” Journal of
Construction Engineering and Management, ASCE(04014052), 1–10.
Van Dongen, H. P. A., and Dinges, D. F. (2005). “Sleep, Circadian Rhythms, and Psychomotor
Vigilance.” Clinics in Sports Medicine, Sports Chronobiology: It’s About Time, 24(2),
237–249.
Xu, X., Ding, M., Li, B., and Christiani, D. C. (1994). “Association of rotating shiftwork with
preterm births and low birth weight among never smoking women textile workers in
China.” Occupational and Environmental Medicine, 51(7), 470–474.
Yaggie, J., and Armstrong, W. J. (2004). “Effects of lower extremity fatigue on indices of
balance.” Balance, 312(322), 2004.
Yoshitake, H. (1978). “Three Characteristic Patterns of Subjective Fatigue Symptoms.”
Ergonomics, 21(3), 231–233.

61
Zhang, N. Z., and Liu, H.-T. L. (2008). “Effects of sleep deprivation on cognitive functions.”
Neuroscience Bulletin, 24(1), 45–48.

62
Chapter 3: Fatigue Management in Electrical Transmission and
Distribution Work
Ulises Techera1, Matthew Hallowell2, and Ray Littlejohn3

3.1 Abstract

Objective: The electrical transmission and distribution power line (TD) sector presents a high

fatality risk. Fatigue greatly contributes to human error and injuries. The objective of this study is

to determine the fatigue risk factors in TD work.

Methods: A sample of 143 TD workers from the northeast coast of the U.S. participated in the

study. Standardized interviews were performed.

Results: Extreme temperatures (p-value 0.004) and long shifts (p-value < 0.001), were identified

by the workers as the principal contributors to their fatigue. Reduced work pace (p-value < 0.001),

and the loss of attention (p-value < 0.001) are the main consequences of fatigue observed.

Conclusions: The current study presents the basis for the development of a fatigue risk

management system in the TD sector. Further research with a larger and more diverse sample is

suggested.

3.2 Introduction

Most communities require a continuous supply of electrical power for the production of

basic goods, processing of water, and establishment of public safety(Balducci et al. 2003). Simply,

1
PhD Candidate; Department of Civil, Environmental, and Architectural Engineering; University of Colorado at
Boulder; UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA; [email protected]
2
Beavers Endowed Professor of Construction Engineering; Department of Civil, Environmental, and Architectural
Engineering; University of Colorado at Boulder; UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA;
[email protected]
3
W. Edwards Deming Professor of Management; Lockheed Martin Engineering Management Program; UCB 428;
1111 Engineering Drive; Boulder, CO 80309 USA; [email protected]

63
electrical power grids help to define the way we live. The delivery of electric power from its point

of generation to its point of use is conveyed through transmission and distribution (TD) power

lines and associated infrastructure that is constructed, operated, and maintained in the US by over

110,000 TD workers. These workers are exposed to high voltage, work at height, confined spaces,

and extremely harsh weather conditions. These risks translate to electrocutions, electrical shocks,

falls from high elevations, injuries from falling objects, overexertion, contusions, cuts, and

lacerations(U.S. Department of Labor 1996).

In comparison to the average occupation, TD construction, operation, and maintenance is

nearly an order of magnitude more dangerous. According to the U.S. Bureau of Labor Statistics

(2015)(U.S. Bureau of Labor Statistics 2015) in 2014 the all-industry fatality rate was 3.3 per

100,000 full-time equivalent workers (FEW) and the construction fatality rate was 9.4. That same

year, the TD subsector accounted for a rate of 19.2 fatalities per 100,000 FEW. During the past 20

years, the mean fatality rate for TD workers has been, on average, eight-times greater than the all-

industry average. These statistics set TD work among the 10 deadliest jobs in the country.

In 2014 the Occupational Safety and Health Administration (OSHA) created new

regulations for the TD sector (U.S. Department of Labor 2015), replacing regulations that were

nearly 40 years old. The new regulations address several new subjects such as proper training, job

briefings, fall protection, insulation, protection equipment, and minimum distances to live charge

sources. With this new set of rules, OSHA expects to save approximately 20 lives and prevent 118

serious injuries annually (U.S. Department of Labor 2015).

64
 Unfortunately, the regulations provide very little guidance for the detection and

management of worker fatigue. However, research has shown consistent evidence that fatigue

plays a significant and important role in accident causation (Craig 1992; Czeisler et al. 1992; van

der Linden et al. 2003a; Lorist et al. 2000; Tixier et al. 2014). In fact, fatigue has been shown to

trigger human error, which when combined with hazardous environments led to catastrophic

events like Three Mile Island, Chernobyl, and the Exxon Valdez oil spill (Harrington 2001). Given

that TD operations often involve working with live charges of over 500,000 volts, TD workers are

especially vulnerable to fatal accidents. The most significant causes of occupational fatigue are

lack of sleep, long shifts, involuntary overtime, high temperatures, and heavy workload(Techera

et al. 2016). These factors are ubiquitous in TD operations.

The objectives of this paper are to specifically study TD work and identify and describe

the: (1) specific causes of worker fatigue; (2) consequences of fatigue, (3) role of fatigue in injuries,

(4) methods for self-diagnosing fatigue and recognizing fatigue in co-workers; and (5) methods to

prevent or mitigate fatigue. Importantly, we also aim to understand the role that demographics play

in all of these aspects of TD worker fatigue. This is the first detailed study of occupational fatigue

in the TD sector.

3.3 Background

3.3.1 The importance of TD operations

The demand for electrical power has grown exponentially during the past 3 decades as the

use of technology has swiftly increased. In the United States, about 4 billion megawatts per hour

65
(mWh) of electrical power is produced yearly to meet demand. This production generates

approximately 388.1 billion dollars in revenue (U.S. Energy Information Administration 2016).

The ability to reliably deliver energy is a primary driver of the economy of developed

countries (Balducci et al. 2003). The public and private sectors have collaborated to ensure reliable

delivery of power to prevent the costly negative consequences of power outages, especially for the

industrial and commercial sectors. For example, the U.S. bulk power grid provides power for the

operation of lights, appliances, and electric motors with a reliability of 99.96% translating to 3.5

hours of power grid downtime per year (Balducci et al. 2003). This level of reliability is achieved

by the deployment of TD workers who construct and maintain the electrical grid. It is also these

workers who respond in emergency situations to restore power during outages caused by weather

events, disasters, and other disruptions.

3.3.2 Safety in TD work

Because of the regular exposure to electrical lines, TD workers are constantly in high-risk

environments. Contact with an energized line, whether through direct contact or arc flash, can

cause immediate death (Cadick et al. 2005). The transmission of a current through the human body

affects the nervous, muscular, cardiovascular, and pulmonary systems. It can cause external

damage such as burns and also internal tissue damage compromising organs and limbs (Gordon

and Cartelli 2009; Spies and Trohman 2006). Furthermore, due to electrical ignition, falls, and fire,

these damages can occur even when there is no direct contact with the electric source (Cadick et

al. 2005).

66
 Albert and Hallowell (2013) conducted the only notable research into TD safety. Their

study evaluated the efficacy and cost of safety measures in TD operations based on the opinion of

experts (Hallowell et al. 2011). The results showed that strict adherence to OSHA regulations and

the employer’s life-saving rules were perceived to be the most effective strategies to promote a

safe environment (91% consensus). This was followed by de-energizing the lines (86%) and the

use of insulated poles that prevent electric shock (81%). Despite the perceived effectiveness of de-

energizing lines, the cost is prohibitive due to disruptions in service. The implication is that TD

work must be performed on energized lines and that strict adherence to standard operating

procedures and the use of personal protective equipment (PPE) is the only feasible method of

controlling exposure. The result is a very dangerous work environment compounded by controls

that are susceptible to human error. Unfortunately, as will be discussed, fatigue compromises

cognitive and physical functions and ultimately causes human error. Thus, fatigue management is

critically important for the TD sector.

3.3.3 The influence of fatigue on safety

Fatigue plays a significant role in many occupational injuries and fatalities, particularly in

hazardous environments when concentration and skill are required to prevent contact with danger.

Fatigue can be defined as a decreased ability to perform activities at the desired level due to

lassitude or exhaustion of mental and/or physical strength (Gander et al. 2011; Hallowell 2010).

When workers are fatigued, they experience compromised alertness, judgement, reaction time,

mental acuity, physical strength, and the development of an uncooperative disposition (Gillberg

and Åkerstedt 1994; Kajtna et al. 2011; van der Linden et al. 2003a; Lorist et al. 2005; Scott et al.

2006; Yaggie and Armstrong 2004). Such effects decrease a worker’s ability to complete their

67
work safely due to the increased rate of human error (Dembe et al. 2005). Thus, it is not surprising

that many researchers include fatigue as a salient factor in most accident causation theories (Craig

1992; Czeisler et al. 1992; van der Linden et al. 2003a; Lorist et al. 2000, 2005; Scott et al. 2006;

Yaggie and Armstrong 2004).

As described by Spurgeon et al. (1997) there are two primary reasons why fatigue

compromises safety in an occupational environment. First, fatigue diminishes the ability of an

individual to perceive and react to new information. Mental fatigue impairs the processing of such

information, making it more difficult for an individual to understand incoming information and

exercise judgement (Lorist et al. 2000; Reiner and Krupinski 2011). Second, fatigue diminishes

the ability of an individual to accurately perceive and respond to risk, thereby clouding situational

awareness and compromising decision making (Tixier et al. 2014).

3.3.4 Causes and consequences of occupational fatigue

Since the focus of this paper is on identifying specific causes and consequences of fatigue

for TD workers, we have provided a brief review of the causes and consequences of occupational

fatigue in the general literature. Since this is a very broad field of study, we cannot include a

detailed description here. Rather, we refer the reader to Techera et al. (2016) for a comprehensive

review of these factors.

Researchers have found that sleep deprivation (Åkerstedt 1990; Bliwise et al. 1992;

Dawson and McCulloch 2005; Webb and Agnew Jr 1975; Zhang and Liu 2008), environmental

conditions such as noise (Kjellberg et al. 1998; Krause et al. 1997; Landström and Löfstedt 1987),

68
temperature (González-Alonso et al. 1999; Zivin and Neidell 2010),light (Park and Gotoh 1993),

and vibration (Jiao et al. 2004; Landström and Löfstedt 1987), muscular exertion (Brown 1994;

Grandjean 1979; Jason et al. 2010), overtime and long working hours (Beckers et al. 2008; Gander

et al. 2011; van der Hulst and Geurts 2001; Meijman 1997; Park et al. 2001), incomplete recovery

(Beurskens et al. 2000; Jason et al. 2010; Swaen et al. 2003), poor social environment (Bültmann

et al. 2001; Hardy et al. 1997; Tepper 2000; Zapf and Gross 2001), heavy workload (Åkerstedt et

al. 2002b; Boksem et al. 2008; Ribet and Derriennic 1999), and negative individual emotional

predisposition (Ala-Mursula et al. 2005; Bültmann et al. 2001) are primary drivers for occupational

fatigue. These causes are summarized in Table 9 together with their relative effect sizes(Techera

et al. 2016). Also, the main consequences of fatigue in a work environment are presented in Table

10.

Table 9 Effect Size of causes of fatigue

Causes of fatigue Cohen’s d (95% C.I.) p

Work Environment 2.83 (1.30 – 4.36) 0.000
Sleep Deprivation 1.22 (-0.86 – 3.29) 0.125
Mental Exertion 0.91 (0.47 – 1.34) 0.000
Overtime and Long Working Hours 0.49 (0.16 – 0.82) 0.002
Social Environment 0.39 (0.04 – 0.74) 0.015
Workload 0.25 (0.23 – 0.28) 0.000

Table 10 Effect Size of consequences of fatigue

Consequences of fatigue Cohen’s d (95% C.I.) p

Cognitive Degradation 100 (0.91 – 1.09) 0.000
Physical Degradation 0.99 (0.25 – 1.73) 0.004
Bad Mood 0.95 (0.02 – 1.88) 0.023
Error and Injury 0.64 (-0.01 – 1.29) 0.052
Illnesses 0.33 (0.26 – 0.40) 0.000

69
3.4 Point of departure and contribution to knowledge

This study deviates from and adds to the body of knowledge through the first investigation

of occupational fatigue in the TD sector and collection of data directly from workers. Most fatigue

research focuses on fatigue in general or in laboratory settings rather than in the context of specific

occupational environments. In general, TD workers are exposed to a different risk profile than

most workers because they are regularly assigned to extended shifts and overtime, are exposed to

the elements in often extreme weather conditions, are required to work at heights or in confined

spaces, and they perform much of their work in remote locations that are far away from home.

This work profile demands specific attention because the causes of fatigue may be different and

the consequences of fatigue can be dire. Also, within the context of their work environment, TD

workers must be able to self-diagnose fatigue and recognize fatigue in co-workers, especially when

working in remote locations where access to resources is limited. Therefore, we aimed to identify

the specific causes and consequences of worker fatigue for TD workers, the role of fatigue in past

incidents in TD work, methods for self-diagnosing fatigue and recognizing fatigue in co-workers,

and methods to prevent or mitigate fatigue in remote and often hostile conditions. As we present

our results we link to the existing body of knowledge to ultimately draw a comparison between

TD work and the general industry. Finally, we aim to investigate the role that demographics play

in these salient aspects of fatigue management for the first time (e.g., how age relates to the causes

of fatigue).

70
3.5 Research methods

Since there is a dearth of previous research into TD worker fatigue, we began with no

assumptions. Consequently, we did not develop and test a hypothesis; rather, we performed an

exploratory study and obtained data directly from a large population of TD workers in their field

with open-ended interview questions.

3.5.1 Development of questionnaires

Despite the exploratory nature of this study, we aimed to obtain quantitative answers to

allow for prioritization and statistical analyses. We used standardized survey interviews (SSI),

which allow for comparison among interviewees and for the production of quantitative descriptive

data(Fowler, Jr. and Mangione 1990). At the same time SSIs give the desired flexibility for

examination and further clarification needed in an exploratory process (Fowler, Jr. and Mangione

1990). Two brief SSI questionnaires were developed, one for workers and one for their supervisors.

We aimed to ensure that the questions were open-ended, clear, simple, and direct.

The questions are as follows:

Questions for TD workers:

1) What contributes to your fatigue?

2) How does fatigue affect your work performance?

3) How do you know when you or one of your coworkers is fatigued?

4) What do you do if you are fatigued?

5) What could you do differently in order to prevent fatigue?

6) Have you ever had an accident? If yes, were you fatigued at the time?

71
7) How fatigued do you feel on a daily basis from 1 to 10? —1 being not fatigued at all and 10

being extremely fatigued, to the point that you can’t continue working.

8) How fatigued do you feel towards the end of the work week (from 1 to 10)?

Questions for TD supervisors:

1) How do you recognize when your workers are fatigued?

2) What do you do to help them when they are fatigued?

3) How does fatigue impact workers’ performance?

4) What are the most difficult aspects of managing fatigue?

5) What do you think needs to change in order to help the workers not become fatigued?

6) How fatigued do you think the workers are on a daily basis from 1 to 10? —1 being not fatigued

at all and 10 being extremely fatigued, to the point that they can’t continue working.

7) How fatigued do you think the workers are towards the end of the work week (from 1 to 10)?

As one can see the two questionnaires are similar. Questions regarding the level of fatigue

were purposely asked in the manner to allow for comparisons between the perspectives of

supervisors and workers.

3.5.2 Respondent demographics

The sample population was comprised of 143 participants from the Northeastern United

States, divided into 123 workers and 20 field supervisors. These participants belonged to one of

the following categories: Underground operations, substation operations, or overhead operations.

The demographics of the participants is described in Table 11.

72
 Table 11 TD workers and supervisors’ demographics
Age Experience
TD
N µ (y/o) σ (yrs) µ (y/o) σ (yrs)
Workers, 77 41.5 11.0 14.4 11.8
Underground
Supervisors 12 48.6 15.1 27.8 13.8
Workers, 33 40.1 13.2 14.5 12.8
Substation
Supervisors 4 - - - -
Workers, 13 34.3 9.3 13.5 8.3
Overhead
Supervisors 4 48.5 8.3 28.0 9.3
Notation: (-)No demographic information was available for substation supervisors

The data were collected from workers in the summer months. During this time, 29% of the

participants were working extended shifts (about 13hs a day) due to an emergency situation caused

by abundant rain in the area. Additionally, during that period the weather was sunny with a mean

temperature of 79oF.

3.5.3 Data manipulation and statistical analysis:

Literal answers to each question were entered for every participant. During coding, some

answers such as “loss of concentration”, “distraction”, and “loss of focus” were combined into one

category with the name “loss of attention”. Similarly, “high temperatures”, “heat”, “cold”, and

“low temperatures” were codified as “extreme temperatures.” However, the original data were

always preserved. Once the data were coded, the frequency of each answer was calculated.

When evaluating the causes and consequences of fatigue, a one-way Chi-square analysis

for associations was performed in order to distinguish significant from non-significant

relationships and responses. The Chi-square analysis requires independence of data. Consequently,

for every question asked to a participant, we only considered the first answer for the Chi-square

analysis. Often, participants gave several answers to a given question. For example, for the

73
question: “what contributes to your fatigue?” most participants indicated several answers such as

“lack of sleep”, “high temperatures”, and “heavy work”. In order to comply with independence,

only the first answer of each participant was considered for the Chi-square analysis. Therefore, in

the case of the presented example, only the answer “lack of sleep” would be considered for Chi-

Square. The decision of choosing the first answer allows for the consideration of all participants

because some of them gave only one answer. The other answers were preserved for future analyses.

For variables that showed to have a significant association in the Chi-square analysis,

individual proportion tests were performed. Further statistical analyses were performed to test for

significant differences among distinct demographic groups. For instance, all variables were tested

for variation due to age and trade. When sample sizes of specific groups were too small for a Chi-

square test, Correspondence Analyses (CA) were performed to visually identify any possible

underlying association (Spencer 2013). The main purpose of a CA is to represent the complex

nominal data presented in a contingency table in a plot (low dimensional space) were each point

represents a variable or category (Spencer 2013). Variables with a similar distribution will appear

close to each other in a 2D representation (Spencer 2013). This would indicate a possible

association among such variables. Lastly, variables that showed similar behavior in the CA were

merged and a two-way Chi-square test of these combined subgroups was performed.

3.6 Results and analysis

3.6.1 Results from TD workers:

The univariate results of the interviews with TD workers are summarized in Table 12. The

results are specified for each major work type: underground utility work, substation work, and

74
overhead line work. The variables included in Table 12 are those which were mentioned by at least

9% of the workers in each category. Bonferroni correction for multiple tests was applied when

reporting significance. Table 13 summarizes the extent to which workers felt that fatigue had

played a role in their past incidents (note that this assessment was not performed among substation

workers due to an insufficient sample size). Lastly, Table 14 summarizes the self-reported levels

of fatigue experienced weekly and daily for each work type.

Table 12 Overall results from electrical workers

Summary table Underground Substation Overhead Total
Causes of fatigue n=76 n=33 n=13 n=122
Extreme Temperatures 63%* 82%* 31% 65%
Long Shifts 39%* 33% 62% 40%
Lack of Sleep 30% 52% 31% 36%
Heavy Manual Labor 24% 9% 22% 19%
Family Responsibilities 14% - - -
PPE 14% - 15% -
Double Shifts 12% 6% - -
Excessive Workload 12% 9% - -
Consequences of fatigue n=71 n=29 n=13 n=113
Loss of Attention 44% 69%* 15% 47%
Slow Down 52%* 17% 54% 43%
Rush up 17% - - -
Mistakes 14% - 8% -
Forgetfulness 10% - - -
Fatigue mitigation strategy n=77 n=31 n=13 n=121
Take a Break 57%* 58% 46% 56%
Coffee 34% 52%* 23% 37%
Water 45%* 10% 31% 35%
Eat 17% 32% 8% 20%
Power Nap 18% 13% 31% 18%
Fresh Air 17% 3% - -
Energy Drinks 13% 6% - -
Slow Down 10% 35% - -
Fatigue identification n=76 n=32 n=13 n=121
Body Language 28%* 31% 31% 29%
Slow Down 18% 56%* 23% 29%
Bad Language 18% - - -
Vocalize it 22% 9% 8% 17%
Different Behavior 14% 3% - -
Loss of Attention 9% 28% 8% 14%
Bad Mood 13% 41% - -
Drowsy 11% 3% - -

75
 Requested improvement n=62 n=25 n=12 n=99
Better Break Policy 26% 20% 8% 22%
Hiring 26% 8% 8% 19%
Inclement weather clause - 12% 33% -
Receive Water 13% - 17% -
(*) Significant. Note: According to Bonferroni correction, significance varies from <0.025 for “Causes of fatigue”
and “Fatigue mitigation strategies” to <0.05 for the remaining variables.

Table 13 Percentage of accidents attributed to fatigue by worker

Underground Overhead
Accidents and fatigue
n=68 n=12
No Accident 57% 17%
Accident not fatigue related 21% 50%
Accident fatigue related 22% 33%

Table 14 Self-reported level of fatigue

Self-reported level Underground Substation Overhead

of fatigue (1-10) * Mean Median Mode Mean Median Mode Mean Median Mode
Daily 5.3 5.0 5.0 4.2 4.0 5.0 3.7 3.5 5.0
Weekly 7.3 8.0 8.0 5.7 5.5 3.0 5.3 5.5 7.0
(*) 1 means not tired at all; 10 means so tired that you must stop working

One-way Chi-square and proportion test results:

After considering the first answer to each question from every electrical worker, only the

data obtained from underground workers and substation workers showed to be significant and

the results are presented here.

Tables 15 and 16 present a one-way Chi-square analysis of the data obtained from

underground and substation workers. The purpose of this analysis was to determine if the obtained

results represent significant underlying associations or if such results could have been obtained by

chance and chance alone. Once significantly different categories (different answers) were

identified, a proportion test was performed between each pair of variables to find the probability

76
of obtaining each specific difference. This final difference is described by a p-value, which is

drawn from a binomial distribution.

Table 15 One-way Chi-square test and proportion test results for underground workers
Categories: 9 𝜒 2 = 53.7 p-value: <0.000
Causes of fatigue
Expected Frequency: 8.4 Proportion Test (sig. p-val.<0.025)
Answers (n= 76) Actual Frequency 95% C.I. p-value
Extreme Temperatures 19 12.0 – 27.6 0.001
Long Shifts 20 12.8 – 28.6 <0.000
Categories: 6 𝜒 2 = 45.6 p-value: <0.000
Consequences of fatigue
Expected Frequency: 11.8 Proportion Test (sig. p-val.<0.05)
Answers: (n=71) Actual Frequency 95% C.I. p-value
Slow Down 30 21.7 – 38.7 <0.000
Categories: 7 𝜒 2 = 35.9 p-value: <0.000
Fatigue mitigation
Expected Frequency 11.0 Proportion Test (sig. p-val.<0.025)
Answers (n=77) Actual Frequency 95% C.I. p-value
Take a Break 24 16.2 – 32.9 <0.000
Water 19 12.0 – 27.6 0.022
Categories: 6 𝜒 2 = 15.1 p-value: 0.009
Fatigue identification
Expected Frequency: 12.7 Proportion Test (sig. p-val.<0.05)
Answers (n=76) Actual Frequency 95% C.I. p-value
Body Language 21 13.7 – 29.7 0.023
Categories: 4 𝜒 2 = 10.4 p-value: 0.017
Requested improvements
Expected Frequency: 15.5 Proportion Test (sig. p-val.<0.05)
Answers (n=62) Actual Frequency 95% C.I. p-value
Receive Water 3 0.6 – 8.4 <0.000

Table 16 Chi-square test and proportion test results for substation workers
Categories: 6 𝜒 2 = 18.8 p-value: <0.002
Causes of fatigue
Expected. F: 5.5 Proportion Test (sig. p-val.<0.05)
Answers (n=33) Frequency (1st answer) 95% C.I. p-value
Extreme temperatures 13 7.6 – 19.1 0.004
Categories: 5 𝜒 2 = 26.7 p-value: <0.000
Consequences of fatigue
Expected. F: 5.8 Proportion Test (sig. p-val.<0.05)
Answers (n=29) Frequency (1st answer) 95% C.I. p-value
Loss of attention 17 11.3 – 22.2 <0.000
Categories: 6 𝜒 2 = 27.9 p-value: <0.000
Fatigue mitigation
Expected. F: 5.2 Proportion Test (sig. p-val.<0.05)
Answers: (n=31) Frequency (1st answer) 95% C.I. p-value
Coffee 15 9.3 – 20.7 <0.000

77
 Categories: 5 𝜒 2 = 10.2 p-value: 0.037
Fatigue identification
Expected. F: 6.4 Proportion Test (sig. p-val.<0.05)
Answers (n=32) Frequency (1st answer) 95% C.I. p-value
Slow down 12 6.8 – 18.0 0.033

Analysis of Chi-square tests results

Perceived causes of TD worker fatigue:

Three main causes of fatigue were identified among all TD workers: “extreme

temperatures”, “long shifts”, and “lack of sleep”. These results confirm the importance of these

variables in the development of fatigue as described by several previous authors (Beckers et al.

2008; Belenky et al. 2003; González-Alonso et al. 1999; van der Hulst and Geurts 2001; Lim and

Dinges 2008; Van Dongen and Dinges 2005). Despite the fact that the results of this study indicate

“extreme temperatures” as a main contributor to fatigue, very little research has been done

regarding the influence of temperature in fatigue development. Consequently, future research

should address this gap in knowledge.

Perceived consequences of fatigue among TD workers:

Two main consequences of fatigue were identified across all sub-trades: (1) loss of

attention and (2) decreased productivity. Assuming that working slower affects productivity, then

it is appropriate to associate fatigue with an overall decrease in productivity. Also, lack of attention

induced by fatigue has been shown to compromise risk perception and decision making and

increase human error, all of which can contribute to accidents at work (Folkard and Tucker 2003;

Tixier et al. 2014). The perceived negative consequences identified here confirm some findings

from previous studies in other industries. For example, a study in the oil and gas construction

sector found that cognitive impairments, unsafe behavior; improper use of equipment, tools and

78
PPE, and even improper communication were outcomes of fatigue (n=321)(Chan 2011). Given the

consistency in findings, it is likely that the negative outcomes of fatigue in TD work are the same

as other industries while the margin for error is less with electrical work.

Methods to mitigate and manage TD worker fatigue:

In order to mitigate fatigue, workers take a break and drink water or coffee. It is important

to mention that for this analysis, the frequency of the variable “take a break” was determined only

by those answers that specifically mentioned “take a break”. Many other variables such as drink

water, drink a coffee, or take a power nap, also imply to take a break. However, these later variables

were considered as separate mitigation strategies. Despite the fact that power naps were the next

most mentioned strategy to mitigate fatigue among workers, this procedure is currently not allowed

by most employers. This situation may change in the future given the growing evidence that power

naps increase both productivity and safety (Purnell et al. 2002; Rosekind et al. 1995; Sallinen et

al. 1998).

Methods to identify fatigue in co-workers:

Identifying fatigue in co-workers is important to properly manage fatigue. Electrical TD

work is typically performed by crews of 2 to 6 workers who usually work together for a long period

of time. Consequently, TD workers are familiar with their coworkers and this allows them to

identify subtler changes in behavior produced by fatigue. Not surprisingly, “body language” was

found to be the most mentioned and significant (p-value: 0.023) way of recognizing a fatigued

coworker. Besides this form of identifying fatigue among coworkers, interviewees also noted a

slowing in work-pace as a sign of fatigue (p-value: 0.033). Not many researchers have investigated

79
the way in which individuals identify fatigue among coworkers. However, these findings confirm

suggestive results obtained by Lerman et al. (2012).

Suggested changes to better manage TD worker fatigue:

Even though the suggested fatigue mitigation methods were not found to be significant in

a Chi-square analysis, the suggestions are useful as a starting point for future research and policy

development. The most frequently mentioned suggestion were: (1) a better break policy that allows

for more frequent breaks where power naps are allowed (suggested by 22% of interviewees); (2)

increasing manpower (19%); (3) receiving water from the employer (10%); (4) an inclement

weather clause that restricts the work to only emergency tasks during extremely hot or cold days

(7 %). The perceptions were that suggested improvements would help mitigate the primary causes

of TD worker fatigue (i.e., extreme temperatures, long shifts, and work load).

Correspondence analysis results and interpretation

Correspondence analysis was used to investigate the underlying associations among all

variables, including demographics. That is, CA was used to investigate associations between

causes of fatigue, consequences, identification strategies, and mitigation strategies, age, and trade.

After performing CA some variables were consolidated and new two-way Chi-square analysis

were performed. These analyses were complemented by individual proportion tests to evaluate

significance of unique associations.

CA was used because it allowed for a visualization of the data structure, which enabled

comparisons among demographic groups (Spencer 2013). For example, we hypothesized that the

80
causes of fatigue are different for workers of varying ages or trades. Given a contingency table,

CA compares the observed frequency of responses to the expected random frequency. When a

significant deviation from the expected random frequencies is observed one can conclude that the

association is significant. Furthermore, similar profiles help researchers to identify variables that

behave similarly, which can be useful for dimension reduction (Greenacre 2007).

The concept of mass, inertia, and Chi-square distances are important in CA. The mass of a

row in a contingency table refers to the total frequency of a row divided by the addition of all row

frequencies. In the case of independent data (homogeneity), this number is equal to the sample

size. The concept of inertia refers to a weighted Chi-square distance, which in the case of the inertia

of a contingency table it is equal to the Chi-square statistic divided by the sample size. Lastly, Chi-

square distances are a weighted Euclidean distance, where the weight corresponds to the column

or row mass. Consequently, we can see a parallelism between the deviation of a specific frequency

from the expected frequency and the Chi-square distance (Greenacre 2007).

Correspondence analysis allows for the graphical representation of variables and categories

in a best-fitting plan where the Chi-square distances among variables and categories or between

variables and categories are represented. There are 3 possible representations: (1) Asymmetrical

Row Principal Normalization in which the row profiles are represented in their principal

dimension; (2) Asymmetrical Column Principal Normalization in which the column profiles are

represented in their principal dimension; and (3) Symmetrical Normalization in which both

profiles are represented in their principal dimension. In these 3 cases, the principal dimensions

are those which minimize the Chi-square distances of the individual profile elements to the average

81
profile. Consequently, those points further away from the origin represent a frequency different

from the expected frequency. Furthermore, two variables that appear close to each other manifest

a similar behavior. Lastly, in an Asymmetric Row Principal Normalization, if a category is closer

than the other categories to a specific variable, it is possible to deduce that the relative frequency

of such a variable is highest for that category. For example, if a specific contingency table has hair

color for rows and eye color for columns, row principal normalization will represent light color

hair close to light color eyes and further away from darker color eyes (Greenacre 2007).

It is important to underline that asymmetric normalizations give us information regarding

possible associations among variables or categories and between variables and categories for data

sets with high dimensionality such as this. For example, by looking at an asymmetric normalization

it is possible to consider the proximity of certain variables and the distances between variables and

categories. On the other hand, symmetric normalizations only allow us to make inferences within

variables or categories but not between them (Greenacre 2007).

Correspondence analysis for consequences of fatigue and trade:

From Figure 4 we can see that underground workers may be more prone to ‘slowing down’

when fatigued, compared to the other trades. This can be seen in Figure 4 because the graphical

representation for “slow down” is closer to the graphical representation for underground workers

than to the other trades. Likewise, the same graphic suggests that substation workers experience

‘lack of attention’ more than the other trades. Despite the low inertia represented by the raw

principal normalization plot (28.4%), a new Chi-square test showed a that these associations were

significant (𝜒 2 = 15.97, p-value = 0.003). This latter association may be explained by the fact that

82
substation workers spend most of their time in one location (the substation) while underground

and overhead workers are constantly changing from one location to another. Familiarity with the

environment may reduce alertness and attention, especially under the effects of fatigue. This new

hypothesis should be tested in future research. Figure 5 shows that some consequences of fatigue

have similar profiles given their proximity in the graphical representation. However, these

associations don’t add value to the analysis already described and for that reason they are not

explained.

o 1) Underground
o 2) Substation
o 3) Overhead

• 1) Slow Down
• 2) Loss of Attention
• 3) Rush up
• 4) Bad Mood
• 5) Judgment
• 6) Others

Figure 4 Row principal normalization of consequences of fatigue and trade

83
 o 1) Underground
o 2) Substation
o 3) Overhead

• 1) Slow Down
• 2) Loss of Attention
• 3) Rush up
• 4) Bad Mood
• 5) Judgment
• 6) Others

Figure 5 Symmetric normalization of consequences of fatigue and trade

Correspondence analysis for methods to identify fatigue and trade:

Figure 6 shows that underground workers identify fatigue among their coworkers by

changes in body language, the use of foul language, and specific mention of being tired. From the

same figure, it is possible to infer an association between substation workers and specific fatigue

identification techniques. In this case, we can see that these workers mostly identify fatigue by

noticing a slower pace of work and lower cooperation among their coworkers. All these

associations are interpreted from the graphical proximity of certain fatigue identification

techniques to specific trades. Similar to the previous case, the raw principal normalization

represented only 17.5% of the total inertia. Nonetheless, these associations were confirmed as

significant in a new Chi-square test (𝜒 2 = 13.64, p-value = 0.009).

84
 The associations among fatigue identification techniques described in the previous

paragraph can be visually recognized in Figure 7 by the proximity of the graphical representation

of specific fatigue identification techniques. For example, we see that identification techniques

1,2, and 3 are close to each other. This indicates that the column profiles of these variables may be

similar. Consequently, under that assumption we can say that when workers identify fatigue by a

change in body language they also identify fatigue by noticing the use of bad language and

manifestations of how tired their workers are.

o 1) Underground
o 2) Substation
o 3) Overhead

• 1) Body language
• 2) Vocalize it
• 3) Bad language
• 4) Slow down
• 5) Bad mood
• 6) Loss of attention
• 7) Others

Figure 6 Row principal normalization for identification techniques and trades

85
 o 1) Underground
o 2) Substation
o 3) Overhead

• 1) Body language
• 2) Vocalize it
• 3) Bad language
• 4) Slow down
• 5) Bad mood
• 6) Loss of attention
• 7) Others

Figure 7 Symmetrical normalization for identification techniques and trades

3.6.2 Results from TD supervisors:

Given the small sample of supervisors (20), no statistically significant results were obtained

from this group. However, the exploratory results obtained from this sample are presented because

they are informative and interesting when compared with worker perceptions.

Fatigue identification

Most supervisors agree that when operators are fatigued they work slower, which is also

the primary way in which supervisors identify fatigued workers. This is in stark contrast to how

workers identify fatigue in their peers as described above. Supervisors did mention changes in

mood and body language but to a lesser extent.

86
Action to help

According to supervisors, the most common action to help workers when they are fatigued

is to give them a break. Unfortunately, certain situations like emergency work opportunities for

breaks are practically limited. In such cases, supervisors try to assign additional workers to the

current crew. When that strategy is not possible, supervisors will redistribute the workload among

workers in order to avoid overexertion and provide water.

Consequences of fatigue at work

Supervisors felt that fatigue highly affects productivity and safety because workers are

slower and lose attention, which yields safety incidents and rework.

Difficult aspects of managing fatigue

Supervisors mentioned that the biggest challenge of managing fatigue is balancing the need

for completion of demanding tasks with a limited staff. Unlike other industries, TD workload

fluctuates rapidly and supervisors are often under pressure to completing several tasks in

emergency situations where public health and safety are at stake. In such situations, supervisors

are aware that their staff will be fatigued; however, there are few options to change workload,

hours, or number of workers.

Suggested improvements

Supervisors offered different potential improvements such as improving the current

workload-stuffing imbalance by hiring more employees and increasing labor budgets. While the

87
focus of fatigue improvements from the workers was on breaks and nutrition, supervisors focused

more on staffing imbalance.

Table 17 presents a summary of the results obtained from the supervisors together with the

percentage of supervisors that offered each answer. Although there were some differences in

responses between workers and supervisors, Table 17 reveals that there is significant alignment on

most topics. This suggests that supervisors are well connected with the work and the workers,

which may stem from the fact that many supervisors were TD workers earlier in their careers.

Table 17 Summary of results from TD supervisors

Fatigue identification (n=20) Action to help (n=20)
Body language 50% Give a break 75%
Slow down 50% Balance workload 60%
Bad mood 20% Give water 35%
Ask for breaks 15% Consequences of fatigue(n=12)
Requested improvement (n=12) Slow down 80%
Hiring 40%

3.7 Conclusion

Transmission and distribution power line managers, safety leaders, and union

representatives describe the sector as one with a comparatively high fluctuation of workload.

Unexpected events such as fires and inclement weather generate emergency work that requires

intense and prolonged work for a limited number of workers. On the other hand, quieter times

present the challenge of keeping the workforce active and vigilant. This fluctuation and

inconsistency makes TD work particularly difficult to manage using a single stable system. Thus,

fatigue management must be studied specifically in this industry, which has a different work

profile that typical construction work.

88
 Before this study, there was no research addressing fatigue for TD workers despite the fact

that it is one of the most dangerous industry sectors. The objective of this study was to identify

causes and consequences of fatigue, methods to identify fatigue in co-workers, and methods to

mitigate and manage fatigue. Importantly, the data for this study came directly from workers and

their field-level supervisors as to promote ecological validity and document the experiences

directly from the workforce.

The study revealed several characteristics of TD line work that contribute significantly to

fatigue. For example, extreme temperatures were identified as the principal contributors followed

by extended shifts. These results are especially important because the link between temperature

and fatigue has not been thoroughly studied in the past. We recommend future research in this area

as it will help practitioners to better design fatigue risk management systems.

The consequences of fatigue were similar to those documented for other industries. For

example, fatigue was perceived to have negative impacts to productivity and safety because of

decreased vigilance and increased rates of human error. Experimental research is suggested to

directly measure the impacts within the TD line work context. Surprisingly, there was no

statistically significant difference in the perceived levels of fatigue, the causes of fatigue, or the

consequences among different demographic groups. These causes and consequences largely were

similar across workers, which is counter to previous research conducted in laboratory settings.

This suggests either the laboratory experiments are not generalizable to the field or that worker

89
perceptions of their own fatigue levels are skewed to be relatively consistent across demographic

groups (e.g., old and young, new and experienced, etc.).

Although the results show that fatigue is perceived as a principle risk factor in their work

that impacts both productivity and safety, there were no formal fatigue risk management systems

in place. Furthermore, OSHA regulations do not emphasize the role that fatigue can play in safely

working around high or low voltage electrical lines. This is seen by the authors as an extraordinary

disconnect between research and practice that must be rectified. Improvements could be made by

implementing a FRMS like those observed in aviation, transportation, oil and gas, and mining

(Arnaldo et al. 2016; Chan 2011; Lerman et al. 2012; Marcus and Rosekind 2016). The findings

from this study serve as a foundation for such FRMS because specific risk factors, impacts, and

initial mitigation strategies have been identified and differentiated by specific TD trades. Previous

research has shown that FRMSs must be designed for specific work and that they are not one-size-

fits-all (Gander et al. 2011).

The authors offer some direct recommendations from this research. For example, the

effects of extreme temperatures can be mitigated through practical means such as using blowers

for underground work, supply water to workers on hot days, and providing regular breaks in warm,

quiet environments on cold days. Also implementing a regular break policy, in which power naps

are encouraged, will help to recover from physical and mental exertion. Work hours can be reduced

by reorganizing shifts even when staffing is limited. Lastly, an education program can be

implemented to help workers, supervisors, and their families to better understand what causes

fatigue, identify fatigue in their peers, and understand how to obtain proper rest during and after

90
work. These recommendations are based largely upon the risk factors identified and the ideas

provided by the craft workers.

3.8 Acknowledgements

The present study was successful because of the collaboration and interest of safety leaders,

union representatives, supervisors, and workers who provided open and candid responses and who

accommodated the research team for in-person field visits.

3.9 References
Åkerstedt, T. (1990). “Psychological and psychophysiological effects of shift work.” Scandinavian
Journal of Work, Environment & Health, 16, 67–73.
Åkerstedt, T., Fredlund, P., Gillberg, M., and Jansson, B. (2002). “Work load and work hours in
relation to disturbed sleep and fatigue in a large representative sample.” Journal of
Psychosomatic Research, 53(1), 585–588.
Ala-Mursula, L., Vahtera, J., Linna, A., Pentti, J., and Kivimäki, M. (2005). “Employee worktime
control moderates the effects of job strain and effort-reward imbalance on sickness
absence: the 10-town study.” Journal of Epidemiology and Community Health, 59(10),
851–857.
Arnaldo, R. M., Comendador, F. G., Sanz, L. P., and Sanz, A. R. (2016). “Friendly Fatigue Alert
Mobile Apps to Help Aviation Workers Prevent, Identify and Manage Alertness and
Fatigue.” Advances in Safety Management and Human Factors, Advances in Intelligent
Systems and Computing, P. Arezes, ed., Springer International Publishing, 421–432.
Balducci, P. J., Weimar, M. R., DeSteese, J. G., Schienbein, L. A., and Roop, J. M. (2003).
“Electric Power Interruption Cost Estimates for Individual Industries, Sectors, and the U.S.
Economy.” Proceedings from theTwenty-Fifth Industrial Energy Technology Conference,
Texas A & M University, Houston,TX., 109–118.
Beckers, D. G. J., Linden, D. van der, Smulders, P. G. W., Kompier, M. A. J., Taris, T. W., and
Geurts, S. A. E. (2008). “Voluntary or involuntary? Control over overtime and rewards for
overtime in relation to fatigue and work satisfaction.” Work & Stress, 22(1), 33–50.
Belenky, G., Wesensten, N. J., Thorne, D. R., Thomas, M. L., Sing, H. C., Redmond, D. P., Russo,
M. B., and Balkin, T. J. (2003). “Patterns of performance degradation and restoration
during sleep restriction and subsequent recovery: a sleep dose-response study.” Journal of
Sleep Research, 12(1), 1–12.
Beurskens, A. J. H. M., Bültmann, U., Kant, Ij., Vercoulen, J. H. M. M., Bleijenberg, G., and
Swaen, G. M. H. (2000). “Fatigue among working people: validity of a questionnaire
measure.” Occupational and Environmental Medicine, 57(5), 353–357.
Bliwise, D. L., King, A. C., Harris, R. B., and Haskell, W. L. (1992). “Prevalence of self-reported
poor sleep in a healthy population aged 50–65.” Social Science & Medicine, 34(1), 49–55.

91
Boksem, M. A. S., Tops, M., Kostermans, E., and De Cremer, D. (2008). “Sensitivity to
punishment and reward omission: Evidence from error-related ERP components.”
Biological Psychology, 79(2), 185–192.
Brown, I. D. (1994). “Driver fatigue.” Human Factors: The Journal of the Human Factors and
Ergonomics Society, 36(2), 298–314.
Bültmann, U., Kant, I., Schröer, C., and Kasl, S. (2001). “The relationship between psychosocial
work characteristics and fatigue and psychological distress.” International Archives of
Occupational and Environmental Health, 75(4), 259–266.
Cadick, J., Capelli-Schellpfeffer, M., and Neitzel, D. (2005). Electrical safety handbook. McGraw-
Hill:, New York, Chicago, San Francisco, Lisbon, London, Madrid, Mexico City, Milan,
New Delhi, San Juan, Seoul, Singapore, Sydney, Toronto.
Chan, M. (2011). “Fatigue: the most critical accident risk in oil and gas construction.”
Construction Management and Economics, 29(4), 341–353.
Craig, A. (1992). “Symptoms of acute and chronic fatigue.” Handbook of human performance,
Vol. 1: The physical environment; Vol. 2: Health and performance; Vol. 3: State and trait,
A. P. Smith and D. M. Jones, eds., Academic Press, San Diego, CA, US, 289–339.
Czeisler, C. A., Dumont, M., Duffy, J. F., Steinberg, J. D., Richardson, G. S., Brown, E. N.,
Sánchez, R., Ríos, C. D., and Ronda, J. M. (1992). “Association of sleep-wake habits in
older people with changes in output of circadian pacemaker.” The Lancet, Originally
published as Volume 2, Issue 8825, 340(8825), 933–936.
Dawson, D., and McCulloch, K. (2005). “Managing fatigue: It’s about sleep.” Sleep Medicine
Reviews, 9(5), 365–380.
Dembe, A. E., Erickson, J. B., Delbos, R. G., and Banks, S. M. (2005). “The impact of overtime
and long work hours on occupational injuries and illnesses: new evidence from the United
States.” Occupational and Environmental Medicine, 62(9), 588–597.
Folkard, S., and Tucker, P. (2003). “Shift work, safety and productivity.” Occupational medicine,
53(2), 95–101.
Fowler, Jr., F. J., and Mangione, T. W. (1990). Standardized Survey Interviewing: Minimizing
Interviewer-Related Error. SAGE.
Gander, P., Hartley, L., Powell, D., Cabon, P., Hitchcock, E., Mills, A., and Popkin, S. (2011).
“Fatigue risk management: Organizational factors at the regulatory and industry/company
level.” Accident Analysis & Prevention, Advancing Fatigue and Safety Research, 43(2),
573–590.
Gillberg, M., and Åkerstedt, T. (1994). “Sleep restriction and SWS-suppression: effects on
daytime alertness and night-time recovery.” Journal of Sleep Research, 3(3), 144–151.
González-Alonso, J., Teller, C., Andersen, S. L., Jensen, F. B., Hyldig, T., and Nielsen, B. (1999).
“Influence of body temperature on the development of fatigue during prolonged exercise
in the heat.” Journal of Applied Physiology, 86(3), 1032–1039.
Gordon, L. B., and Cartelli, L. (2009). “A complete electrical hazard classification system and its
application.” Electrical Safety Workshop, 2009. IEEE IAS, 1–12.
Grandjean, E. (1979). “Fatigue in industry.” British Journal of Industrial Medicine, 36(3), 175–
186.
Greenacre, M. (2007). Correspondence Analysis in Practice, Second Edition. CRC Press.
Hallowell, M., Esmaeili, B., and Chinowsky, P. (2011). “Safety risk interactions among highway
construction work tasks.” Construction Management and Economics, 29(4), 417–429.

92
Hallowell, M. R. (2010). “Worker fatigue: managing concerns in rapid renewal highway
construction projects.” Professional Safety, 55(12), 18–26.
Hardy, G. E., Shapiro, D. A., and Borrill, C. S. (1997). “Fatigue in the workforce of national health
service trusts: Levels of symptomatology and links with minor psychiatric disorder,
demographic, occupational and work role factors.” Journal of Psychosomatic Research,
Workplace, 43(1), 83–92.
Harrington, J. M. (2001). “Health effects of shift work and extended hours of work.” Occupational
and Environmental Medicine, 58(1), 68–72.
van der Hulst, M., and Geurts, S. (2001). “Associations between overtime and psychological health
in high and low reward jobs.” Work and Stress, 15(3), 227–240.
Jason, L. A., Evans, M., Brown, M., and Porter, N. (2010). “What is fatigue? Pathological and
nonpathological fatigue.” PM&R, 2(5), 327–331.
Jiao, K., Li, Z., Chen, M., Wang, C., and Qi, S. (2004). “Effect of different vibration frequencies
on heart rate variability and driving fatigue in healthy drivers.” International Archives of
Occupational and Environmental Health, 77(3), 205–212.
Kajtna, T., Štukovnik, V., and Grošelj, L. D. (2011). “Effect of acute sleep deprivation on
concentration and mood states with a controlled effect of experienced stress.” Zdravniški
Vestnik, 80(5).
Kjellberg, A., Muhr, P., and Skoldstrom, B. (1998). “Fatigue after work in noise - an
epidemiological survey study and three quasi-experimental field studies.” Noise & Health,
1(1), 47–55.
Krause, N., Lynch, J., Kaplan, G. A., Cohen, R. D., Goldberg, D. E., and Salonen, J. T. (1997).
“Predictors of disability retirement.” Scandinavian Journal of Work, Environment &
Health, 23(6), 403–413.
Landström, U., and Löfstedt, P. (1987). “Noise, vibration and changes in wakefulness during
helicopter flight.” Aviation, Space, and Environmental Medicine, 58(2), 109–118.
Lerman, S. E., Eskin, E., Flower, D. J., George, E. C., Gerson, B., Hartenbaum, N., Hursh, S. R.,
and Moore-Ede, M. (2012). “Fatigue Risk Management in the Workplace:” Journal of
Occupational and Environmental Medicine, 54(2), 231–258.
Lim, J., and Dinges, D. F. (2008). “Sleep Deprivation and Vigilant Attention.” Annals of the New
York Academy of Sciences, 1129(1), 305–322.
van der Linden, D., Frese, M., and Meijman, T. F. (2003). “Mental fatigue and the control of
cognitive processes: effects on perseveration and planning.” Acta Psychologica, 113(1),
45–65.
Lorist, M. M., Boksem, M. A., and Ridderinkhof, K. R. (2005). “Impaired cognitive control and
reduced cingulate activity during mental fatigue.” Cognitive Brain Research, 24(2), 199–
205.
Lorist, M. M., Klein, M., Nieuwenhuis, S., de Jong, R., Mulder, G., and Meijman, T. F. (2000).
“Mental fatigue and task control: Planning and preparation.” Psychophysiology, 37(5),
614–625.
Marcus, J. H., and Rosekind, M. R. (2016). “Fatigue in transportation: NTSB investigations and
safety recommendations.” Injury Prevention, injuryprev-2015-041791.
Meijman, T. F. (1997). “Mental fatigue and the efficiency of information processing in relation to
work times.” International Journal of Industrial Ergonomics, 20(1), 31–38.
Park, J., Kim, Y., Chung, H. K., and Hisanaga, N. (2001). “Long working hours and subjective
fatigue symptoms.” Industrial health, 39(3), 250–254.

93
Park, S., and Gotoh, S. (1993). “Research on office environment evaluation and worker fatigue.”
Journal of Thermal Biology, 18(5–6), 565–570.
Purnell, M. T., Feyer, A.-M., and Herbison, G. P. (2002). “The impact of a nap opportunity during
the night shift on the performance and alertness of 12-h shift workers.” Journal of sleep
research, 11(3), 219–227.
Reiner, B. I., and Krupinski, E. (2011). “Demystifying Occupational Stress and Fatigue Through
the Creation of an Adaptive End-User Profiling System.” Journal of Digital Imaging,
25(2), 201–205.
Ribet, C., and Derriennic, F. (1999). “Age, working conditions, and sleep disorders: A longitudinal
analysis in the French cohort E.S.T.E.V.” Sleep: Journal of Sleep Research & Sleep
Medicine, 22(4), 491–504.
Rosekind, M. R., Smith, R. M., Miller, D. L., Co, E. L., Gregory, K. B., Webbon, L. L., Gander,
P. H., and Lebacqz, J. V. (1995). “Alertness management: strategic naps in operational
settings.” Journal of Sleep Research, 4(s2), 62–66.
Sallinen, M., Härmä, m., Akerstedt, t., Rosa, R., and Lillqvist, O. (1998). “Promoting alertness
with a short nap during a night shift.” Journal of Sleep Research, 7(4), 240–247.
Scott, J. P., McNaughton, L. R., and Polman, R. C. (2006). “Effects of sleep deprivation and
exercise on cognitive, motor performance and mood.” Physiology & behavior, 87(2), 396–
408.
Spencer, N. H. (2013). Essentials of Multivariate Data Analysis. CRC Press.
Spies, C., and Trohman, R. G. (2006). “Narrative Review: Electrocution and Life-Threatening
Electrical Injuries.” Annals of Internal Medicine, 145(7), 531–537.
Spurgeon, A., Harrington, J. M., and Cooper, C. L. (1997). “Health and safety problems associated
with long working hours: a review of the current position.” Occupational and
Environmental Medicine, 54(6), 367–375.
Swaen, G. M. H., Amelsvoort, L. G. P. M. van, Bültmann, U., and Kant, I. J. (2003). “Fatigue as
a risk factor for being injured in an occupational accident: results from the Maastricht
Cohort Study.” Occupational and Environmental Medicine, 60(suppl 1), i88–i92.
Techera, U., Hallowell, M., Stambaugh, N., and Littlejohn, R. (2016). “Causes and consequences
of occupational fatigue: Meta-analysis and systems model.” Journal of Occupational and
Environmental Medicine, 58(10), 961-973.
Tepper, B. J. (2000). “Consequences of abusive supervision.” Academy of management journal,
43(2), 178–190.
Tixier, A. J.-P., Hallowell, M. R., Albert, A., van Boven, L., and Kleiner, B. M. (2014).
“Psychological antecedents of risk-taking behavior in construction.” Journal of
Construction Engineering and Management, ASCE(04014052), 1–10.
U.S. Bureau of Labor Statistics. (2015). “Census of Fatal Occupational Injuries Summary, 2014.”
<http://stats.bls.gov/news.release/cfoi.nr0.htm>.
U.S. Department of Labor. (1996). “OSHA Priorities - Power Transmission and Distribution in
Construction.” <https://www.osha.gov/archive/oshinfo/priorities/power.html>.
U.S. Department of Labor. (2015). “Electric Power Generation, Transmission and Distribution
Standard.” <https://www.osha.gov/dsg/power_generation/index.html>.
U.S. Energy Information Administration. (2016). “Electricity data browser - Net generation for
electric power.” <http://www.eia.gov/electricity/data.cfm>.

94
Van Dongen, H. P. A., and Dinges, D. F. (2005). “Sleep, Circadian Rhythms, and Psychomotor
Vigilance.” Clinics in Sports Medicine, Sports Chronobiology: It’s About Time, 24(2),
237–249.
Webb, W. B., and Agnew Jr, H. W. (1975). “Are we chronically sleep deprived?” Bulletin of the
Psychonomic Society, 6(1), 47–48.
Yaggie, J., and Armstrong, W. J. (2004). “Effects of lower extremity fatigue on indices of
balance.” Balance, 312(322), 2004.
Zapf, D., and Gross, C. (2001). “Conflict escalation and coping with workplace bullying: A
replication and extension.” European Journal of Work and Organizational Psychology,
10(4), 497–522.
Zhang, N. Z., and Liu, H.-T. L. (2008). “Effects of sleep deprivation on cognitive functions.”
Neuroscience Bulletin, 24(1), 45–48.
Zivin, J. G., and Neidell, M. J. (2010). Temperature and the Allocation of Time: Implications for
Climate Change. Working Paper, National Bureau of Economic Research.

95
Chapter 4: Measuring and Predicting Fatigue in Construction
Workers: An Empirical Field Study
Ulises Techera1, Matthew Hallowell2, Ray Littlejohn3, and Sathyanarayanan Rajendran4

4.1 Abstract

The upsurge in commitment to safety from industry leaders and the implementation of

better equipment and training have generated a 67% decline in recordable incident rates during the

past two decades. Fatalities, however, have plateaued for the last decade and, in the past three

years, have even increased. Current research indicates that human factors like fatigue play a major

role in accident causation and fatality occurrence. To manage fatigue, it is essential to determine

how fatigue can be objectively measured and predicted. The current body of knowledge suggests

several potential predictors based upon laboratory experiments; however, no causal factors of

fatigue have been validated in the context of construction work. The present study tested the

hypothesis that a set of objective variables can predict variability in construction worker fatigue.

A total of 252 US construction workers participated in the study. The results indicate that the

amount of sleep and rest obtained in the past 24 h, the number and length of previous shifts, the

temperature and level of noise at the worksite, and personal somatic characteristics can explain

between 9% and 50% of fatigue variance. Future research should address the discovery of

additional fatigue predictors that would allow for a more accurate prediction of fatigue.

1
PhD Candidate; Department of Civil, Environmental, and Architectural Engineering; University of Colorado at
Boulder; UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA; [email protected]
2
Beavers Endowed Professor of Construction Engineering; Department of Civil, Environmental, and Architectural
Engineering; University of Colorado at Boulder; UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA;
[email protected]
3
W. Edwards Deming Professor of Management; Lockheed Martin Engineering Management Program; University
of Colorado at Boulder, UCB 428; 1111 Engineering Drive; Boulder, CO 80309 USA; [email protected]
4
Associate Professor and Program Director of Safety and Health Management Program; Central Washington
University; 400 E. University Way, Ellensburg, WA 98926; [email protected]

96
4.2 Introduction

It is not news that construction is one of the most hazardous industries (Sawacha et al.

1999). In 2015 the construction accounted for a fatality rate of 9.4 for every 100,000 full-time

equivalent workers, nearly three-times greater than that of the overall working population (U.S.

Bureau of Labor Statistics 2015). Accidents have shown to be caused by unsafe human acts and

unsafe conditions (Kartam 1997).

Several decades of construction safety research have revealed that organizational and

human factors, rather than technical failures, are the principal causes of accidents (Langford et al.

2000; Weick et al. 2008). Such research also classifies fatigue as a principal risk factor. In fact,

Chan et al. (2011) carried out an exploratory analysis of the main safety risks in oil and gas

construction, in which 78% of stakeholders (n=351) acknowledged fatigue as the main risk factor

to accidents and the trigger to several other risks. Similarly, Hallowell (2010) also identified

fatigue as a main concern in rapid renewal highway construction.

Fatigue can be defined as a decreased ability to perform activities at the desired level due

to lassitude or exhaustion of mental or physical strength (Gander et al. 2011; Hallowell 2010).

These are manifestations of the physiological deterioration that takes place under the effects of

fatigue. For example, fatigue compromises the function of the prefrontal cortex thus affecting basic

and complex cognitive functions like the ability to concentrate and assimilate new information,

plan, communicate, and react to new stimuli (Angus and Heslegrave 1985; Harrison and Horne

1999, 2000a; Lorist et al. 2000; Van Dongen and Dinges 2005; Zhang and Liu 2008). Because of

97
physiological degradation, the ability of an individual to work safely and efficiently can be

severely compromised by fatigue.

To properly manage fatigue it is essential to detect early signs of onset among workers,

thus allowing for a corrective intervention before unsafe levels have been reached (Dawson et al.

2014). The identification of fatigue predictors would enable meaningful ways of prioritizing

management strategies. The results from past fatigue research suggests potential predictors such

as the amount of sleep obtained, the length of the shift, and the physical or mental intensity of a

job, and others. However, these potential factors, designated causal factors, have not been

validated as predictive in an occupational setting as they are mainly based on laboratory results

(Van Dongen 2004). To address this knowledge gap, the hypothesis that a set of objective

variables predicts variability in fatigue was tested specifically with construction workers in high-

risk tasks.

4.3 Background

4.3.1 Potential Predictors of Fatigue

Previous researchers have studied numerous variables that influence the onset of fatigue.

While this study doesn’t address the many cognitive and physical ways in which these factors

cause fatigue, a brief introduction to these variables is provided. For a detailed analysis of fatigue

causing factors, together with their relative importance and impact, the reader should refer to

Techera et al. (2016), who performed a recent meta-analysis of the causes and consequences of

fatigue. The salient content from this previous study is abstracted here to provide context.

98
4.3.2 Sleep Deprivation

Among fatigue causal factors, sleep deprivation has been the most widely studied and

several researchers affirm that it is the principal cause for fatigue (Belenky et al. 2003; Dawson

and McCulloch 2005; Folkard and Tucker 2003; Lim and Dinges 2008; Van Dongen and Dinges

2005). In fact, Dawson and McCulloch (2005) proposed a model to determine fitness for duty

based on previous sleep patterns. The prior sleep wake model (PSWM) indicates that if the time

between waking up and the end of a worker’s shift is longer than the amount of sleep obtained

during the past 48hs, then such a worker presents a significantly high likelihood of causing a

fatigue-related error. Furthermore, the same study concluded that a worker needs to obtain at least

5 h of sleep in the 24 h prior to work and 12 h of sleep in the 48 h before work to meet minimum

requirements for fitness.

The PSWM was recently used in a longitudinal study with 347 train condcutors using the

Fatigue Audit InterDyne (FAID) software developed by Dawson and Fletcher (2001). Although

they obtained a strong positive correlation between violations of prescriptive thresholds (sleep

deprivation) and fatigue levels, the methodological approach of the study is questionable because

of construct validity. The FAID software used to measure fatigue only considers the shift schedule

for the past 7 days. Consequently, the software delivers higher fatigue scores when a worker

experiences shorter rest periods, and the correlation between rest period and sleep deprivation is

obvious. Further, the factor of sleep deprivation is considered in isolation, although it well-

documented that there are many, interconnected causal factors.

99
4.3.3 Incomplete Recovery

Incomplete recovery, understood as the inability to fully recover from a fatiguing episode,

has been shown to contribute to fatigue accumulation (Beurskens et al. 2000; Jason et al. 2010;

Swaen et al. 2003). Although related, incomplete recovery is typically considered independently

of shift length because it is also strongly impacted by a worker’s personal life and rest habits

(Dawson and McCulloch 2005; Gander et al. 2011). The link between fatigue and recovery time

has traditionally been studied across shifts with different characteristics (i.e. split shifts, rotating

shifts). It has yet to be validated as predictive.

4.3.4 Workload

Defined as high mental or physical demands at work, workload is another potentially

obvious cause of fatigue important to consider in an industry setting. The degree to which workload

contributes to fatigue depends on the extent to which a worker perceives the reward for completing

a work-task in the demanded period of time as worth it (Boksem et al. 2008). The assessment of

this effort to reward balance is usually assessed subjectively. Further, this factor includes the

cognitive and physical demands, repetition, and schedule breaks (Gander et al. 2011; Hsiao and

Simeonov 2001). Workload has typically been studied as an isolated proxy for fatigue but not as a

predictor.

4.3.5 Overtime and Long Shifts

The time that a person spends at work exacerbates fatigue directly and indirectly. Long

shifts and overtime are a direct fatigue stressor by limiting the available time for recovery (Gander

et al. 2011). Additionally, overtime and long shifts expose the worker to other stressors, inherit to

100
the work task, for additional time (Park et al. 2001; Dembe et al. 2005). The relationship between

overtime or long shifts and fatigue are also indirectly related to fatigue through mediators such as

sleep deprivation (Åkerstedt et al. 2002b).

4.3.6 Work Environment

It seems intuitive that work environmental factors, such as the level of noise, lighting,

vibration, and temperature in the worksite can determine the type and intensity of workers’ fatigue.

However, despite the apparent obvious link, little research has been done to validate the impacts

of such factors. Furthermore, most studies conducted were performed in a controlled environments

rather than in the field where the true intensity of these factors is manifest (Jiao et al. 2004;

Kjellberg et al. 1998; Krause et al. 1997; Landström and Löfstedt 1987; Park and Gotoh 1993;

Zivin and Neidell 2010).

4.3.7 Mental Exertion

Mental exertion is conceptualized as sustained cognitive activity that requires extraordinary

mental effort, mental exertion is a fatigue driver (Lorist et al. 2005). In laboratory settings, mental

exertion has been shown to negatively impact the functioning of the pre-frontal cortex, which

regulates executive functions. As the mental demand required for a task increases, so does the rate

of onset of cognitive fatigue (Boksem et al. 2006, 2005; van der Linden et al. 2003a; van der

Linden and Eling 2006; van der Linden et al. 2003b). For instance, Boksem et al. (2005) performed

a laboratory experiment to assess the impact of mentally demanding activities on fatigue. In this

experiment fatigue was measured as changes in brain activity (EEG) and by the subjective self-

101
assessment of the participants. The results showed a strong correlation between time on task and

fatigue increments.

4.3.8 Muscular Exterior

Fatigue can be caused by both mental and physical exertion. The latter often relates to the

physical demands requirement of repetitive muscular effort or sustained muscular tension (Brown

1994; Christensen et al. 1995; Edwards and Lippold 1956; Grandjean 1979; Jason et al. 2010).

Christensen et al. (1995) demonstrated that the frequency and power of electromyographic (EMG)

signals decrease after dynamic or static muscular exertion, testifying of muscular fatigue.

Consequently, muscular exertion can be considered as a predictor of fatigue and EMG signals

present a possible scale to measure muscular fatigue. Thus, muscular exertion is yet another

potential predictor of fatigue that may relate to other factors.

4.3.9 Social Work Environment

Most people spend between 25 and 40% of their time at work. Thus, the quality of the

relationships with managers, co-workers, and subordinates becomes an important factor in the

development of mental fatigue (Bültmann et al. 2001). For example, supervisors’ abusive behavior

were linked to emotional exhaustion in a subjective assessment among 249 employees in varied

workplaces (Yagil 2006). Although difficult to measure objectively, this factor may be a predictor.

4.3.10 Personal Predisposition

It is important to consider individual differences while measuring fatigue. The source for

such differences correspond to somatic or psychic characteristics of everyone, which can be

102
temporal or permanent. For example, a somatic contributor to fatigue could be sleep apnea, which

inhibits normal rest thus generating a fatigue debt that accumulates overtime (Chervin 2000). On

the other hand, the negative or positive emotional state of an individual also plays a role in the

development of mental fatigue (Bültmann et al. 2001). In general, the impact of personal

predisposition on fatigue has been tested by comparing scores between questionnaires that measure

personal characteristics (independent variables) and subjective fatigue scales (dependent variable).

The reviewed causal factors of fatigue all offer opportunities for prediction and objective

measurement. In theory, objectifying and concurrently measuring these factors should produce the

best predictive model based upon the current state of knowledge. Thus, we hypothesize that each

of the reviewed factors offers some predictive power for fatigue. If so, a method could be created

to objectively measure the potential for fatigue before work begins, thereby improving ability to

manage and control its onset.

4.4 Fatigue measurement tools

In order to validate predictors, fatigue must be measured as a dependent (predictand)

variable using state-of-the-art methods. There is a rich and extensive body of literature on fatigue

measurement. However, given the complex physiological and psychosocial nature of fatigue, there

is no consensus on how it should be measured. Consequently, researchers have relied on proxies

for fatigue, often considering it as a product of its causes (e.g., workload) and effects (e.g., moving

slowly). Such an approach assumes that the measurement of these variables determines the state

of the underlying construct, which is problematic for scientific inquiry as fatigue itself cannot be

103
measured. Thus, as a dependent variable, the prevailing subjective and objective methods of

measuring symptoms of fatigue will be used.

There are two main ways of assessing fatigue: subjectively and objectively. The subjective

assessment of fatigue is founded on the perception of the symptoms that people experience when

they are fatigued. Such symptoms include sleepiness, lack of energy, and physical impairments

among others. The apparent inability to measure these sub constructs directly resulted in their

subjective assessments as the best means to measure fatigue. More recently, new technologies

allowed for the direct assessment of specific physical variables that change under the effects of

fatigue, such as neuronal activity or cardiorespiratory metrics. These technologies have enabled

objective measures of fatigue.

4.4.1 Subjective fatigue measurement

Subjective instruments were initially developed in clinical settings to monitor patients with

a terminal disease or who were undergoing long and aggressive medical treatment like

chemotherapy. These patients experience acute fatigue for prolonged periods of time as a side

effect of their condition (Piper et al. 1998; Shapiro et al. 2002; Smets et al. 1995; Taylor et al.

2000). Later, some of these scales were used to assess fatigue in healthy working populations,

revealing some incompatibilities between clinical and industry settings. Consequently, additional

scales were later developed specifically for healthy working individuals (Bültmann et al. 2000;

Dainoff et al. 1981; Mota and Pimenta 2006).

Subjective fatigue assessment tools are self-administering questionnaires or scales that

inquire about the manifestation or the intensity of specific fatigue consequences. For instance,

104
assuming drowsiness as a consequence of fatigue, a subjective questionnaire may include a

question related to the perception of drowsiness. The quality of these subjective instruments is

largely determined by the ability of the subject to recognize and value the magnitude of his or her

own drowsiness. Thus, the subjective nature of fatigue questionnaires yield scientific limitations

because of reliability and construct validity.

Reliability or internal consistency refers to the degree to which different participants who

experience the same level of fatigue will give the same answer to a specific question, thus allowing

for comparisons across individuals (Santos 1999). This factor is determined by the inter-variable

correlation across all elements of the questionnaire. Cronbach’s alpha is the most common statistic

to perform such an assessment (Cronbach 1951). This coefficient fluctuates between 0 and 1 and

values equal or greater than 0.7 indicate good reliability (Nunnally 1978). An elevated alpha value

indicates that all the items considered measure the same underlying construct.

Alternatively, construct validity is determined by compliance with convergent validity and

discriminant validity. Convergent validity is the measure of the extent to which a questionnaire

measures that which is intended to measure (Vries et al. 2003). The assessment of convergent

validity relies on the comparison between the instrument being assessed, and an assumed perfect

measure of the underlying construct. Specifically referring to fatigue, such perfect instrument

doesn’t exist. Therefore, the assessment of convergent validity is as good as the ability of the

referent instrument to truly measure fatigue. Discriminant validity investigates the degree to which

an instrument doesn’t measure that which it is not intended to measure. Researchers usually

perform this analysis by using groups with expected significant differences in fatigue.

105
 Multiple studies address the characteristics of the available fatigue assessment tools. It is

not in the scope of this study to reproduce such analyses; however, these past studies were

considered when selecting the most reliable and valid subjective fatigue assessment tool for

validation. A brief review of the salient subjective fatigue measurement tools is provided.

Need for Recovery Scale (NRS). The NRS is an instrument designed to measure fatigued

caused by work shift (Van Veldhoven and Broersen 2003). It presents 11 dichotomous items to

which a participant can answer “yes” or “no”. The scores are then transformed to a 100-point scale

that represents the final fatigue score of the individual. The internal consistency of the NRS was

measured as 0.87, making use of the internal correlation coefficient (ICC), a comparable statistic

to Cronbach’s alpha. These two coefficients are based on the same measurement model and on the

same definition of reliability (Bravo and Potvin 1991). Some of the items of this scale state: “I find

it difficult to relax at the end of a working day” or “I cannot really show any interest in other people

when I have just come home myself”. According to the character of these questions, the NRS was

designed to determine an overall level of fatigue that manifests regularly. Consequently, this scale

does not comply with the requirements of the study. The researchers are interested in a tool able

to assess subjective fatigue at a given point in time, rather than a period of time.

Fatigue Assessment Scale (FAS). FAS is a 10-item scale with 5 questions to measure

physical fatigue and 5 items to measure mental fatigue. The FAS was administered to a sample of

1893 volunteers obtaining a Cronbach’s Alpha of 0.87 . Similar to NRS, the FAS measures how

a person usually feels rather than how a person feels at the time of the assessment. For example,

one of the items states: “I have enough energy for everyday life” to which the responder chooses

106
an option from (1) never, to (5) always. The scale has been validated among large samples in cross-

comparisons with other fatigue measurement methods (Michielsen et al. 2004).

Swedish Occupational Fatigue Inventory (SOFI). The SOFI was developed from 95 verbal

expressions such as: drowsy, sweaty, passive, and uninterested. In this scale, participants indicate

to what extent they feel these sensations at a given point on a Likert scale. Initially, the

questionnaire was tested on 705 individuals belonging to 16 different occupations (Åhsberg et al.

1997b). Factor analysis and stepwise analysis reduced the initial poll to 25 variables organized in

5 different factors. Such factors were named: lack of energy, physical exertion, physical

discomfort, lack of motivation, and sleepiness (Åhsberg et al. 1997b). In its development, the SOFI

showed strong reliability and the ability to discriminate across different occupations (Åhsberg et

al. 1997b; Åhsberg 2000; Åhsberg et al. 2000b).

The discriminant and convergent validity of the SOFI was further explored among groups

with expected differences in fatigue dimensions. In 1998 a group of 20 men and 20 women

underwent 5 episodes of dynamic and static physical exertion (cycling and forward flexions) with

a 20-minute break in between sessions. The participants filled out the SOFI and the Category Ratio

Scale (CR-10) by Borg (1982) after every session. The results indicated high levels of both

physical exertion and physical discomfort after dynamic and static work respectively, thus

indicating good discriminant validity. Additionally, SOFI scores and CR-10 scores showed high

Pearson product moment correlation, demonstrating convergent validity. The dimension “lack of

energy” also scored high after both types of exercises. This suggests that the latter dimension

107
measures an overall level of fatigue. The rest of dimensions presented significantly lower scores

(Åhsberg and Gamberale 1998).

A similar protocol was followed to demonstrate discriminant validity for the remaining

SOFI dimensions. This third study consisted on a proof-reading task (2x90 min) and a vigilant task

(2x60 min) in which 20 men and 20 women took part. The results indicated significantly high

scores for lack of motivation, sleepiness, and also lack of energy, thus demonstrating the overall

construct validity of the scale (Åhsberg et al. 2000a).

More recently, the reliability and validity of the SOFI was reassessed on 597 workers,

belonging to 5 occupations with different workloads. A revised questionnaire with 4 items per

dimension sowed better psychometrics for the described sample. Based on this revised SOFI, new

reliability coefficients were calculated. An overall Cronbach’s alpha of 0.80 was found and the

reliability for each dimension was as follows: Lack of energy (0.92), Lack of motivation (0.92),

Sleepiness (0.89), Physical discomfort (0.92), and Physical exertion (0.87) (Åhsberg 2000). Since

its development, the SOFI has been widely used among working populations and it has been

translated into several languages, including Spanish (Åhsberg et al. 2000b; González Gutiérrez et

al. 2005; Leung et al. 2004; Sultanian et al. 2014). The strong and robust validation of the tool

and the fact that it measures fatigue at a single point in time made it ideal for the present study.

4.4.2 Objective fatigue measurement

Technological advances of the past few decades allowed for the development of tools that

can measure fatigue as a latent variable (Dawson et al. 2014). Researchers from several fields have

developed instruments to assess fatigue objectively based on the different observable variables that

108
change with fatigue. For example, researchers in neurobehavioral science developed tools based

on changes in cognitive ability, mood, and reaction time (RT) (Dinges et al. 1997). Performance

in all these factors decreases under the effects of fatigue due to a decay in the central and peripheral

nervous system (Balkin et al. 2004). Also, electroencephalography produced some products able

to measure fatigue using frequency change in brain activity considering Theta (3.5-7Hz), Alpha

(8-13Hz) and Beta (14-30Hz) waves’ change under the effects of fatigue (Lal et al. 2003; Lal and

Craig 2001; Okogbaa et al. 1994). Pupilometry, which measures visually-guided saccadic velocity

(SV), initial pupil diameter (PD), pupillary constriction latency (CL), and amplitude of pupil

constriction (CA) (Goldich et al. 2010); and oculometry, which measures the amplitude and

frequency of eyelid movement, have also produced devices to assess fatigue (Wierwille et al.

1994). Lastly, devices that identify posture changes and/or head nodding have also shown to be

able to alert of high levels of fatigue. Unfortunately these devices perceive fatigue usually when it

is already too late to prevent a fatigue related accident (Hartley and Arnold 2001; May and Baldwin

2009).

Objective fatigue assessment tools can be classified according to their intent in three

categories, (1) technologies that assess fitness for duty, (2) technologies that assess fatigue

continuously, and (3) technologies embedded in machinery or vehicles. The first 2 types of

technologies require interaction between the worker and the device, while the third type of

technology is non-intrusive.

Every fatigue-measuring technology presents advantages and disadvantages. Those that

assess fitness for duty require the worker to stop and dedicate time to the assessment; however,

109
these assessments usually take less than 10 minutes. Conversely, technologies that measure fatigue

continuously do not require workers to stop working for the assessment but require the participant

to carry a device, which can be intrusive. The third type of technology is attractive because it

doesnt require the worker to stop or to carry the device. However, these technologies only work in

very controlled environments (i.e. in trucks with drivers) and even in those environments their

performance is comparatively unreliable and inaccurate (Dawson et al. 2014). Since the goal of

this study was to test the predictive validity of candidate indicators, fitness for duty methods of

fatigue measurement that take place at specific times was preferred.

Fitness for duty technologies available in the market include: (1) The Psychomotor Vigilant

Test (PVT), which measures RT as an indirect assessment of fatigue; (2) The Occupational Safety

Performance Assessment Test (OSPAT), which is a computer based test that presents an

unpredictable tracking task that measures RT, sustained attention, and hand-eye coordination; (3)

The Online Continuous Performance Test (OCPT), which measures alertness and vigilance

through RT during an online task that presents multiple visual stimulus; (4) The Eye Check, which

assesses fatigue by measuring pupil diameter and constriction latency; (5) The Fitness Impairment

Tester (FIT), which uses eye tracking (saccadic velocity) and pupilometry (pupil diameter,

constriction amplitude, and latency) to assess fatigue; and (6) The Safety Scope, which measures

pupillary reflex and eye movement parameters to determine the level of fatigue of an individual

(Dawson et al. 2014). An exhaustive review of literature revealed spurious validity of most

aforementioned technologies. Furthermore, many of these studies were found in the quasi-

scientific commercial literature rather than in peer-reviewed journal articles. The extensive review

carried out by Dawson et al. (2014) reveal that the PVT is the only technology with strong evidence

110
of validation by independent researchers, laboratory studies, and field studies. Accordingly, it is

considered by most researchers as the eminent method to objectively measure fatigue (Dinges and

Powell 1985; Dorrian et al. 2005; Loh et al. 2004).

Traditionally the PVT was administered in a hand-held device known as PVT-192

(Ambulatory Monitoring, Inc., Ardsley, NY; Dinges & Powell, 1985). The principal challenge

with this technology is its extremely high cost. Recently, a new software ,PC-PVT was developed

to perform RT measures with the same functionality and reliability as the PVT-192 (Khitrov et al.

2013). Thus, the PC-PVT was the preferred objective method of measuring fatigue at single points

in time in the field.

4.5 Point of departure

Although there is a strong body of knowledge related to the causal factors of fatigue, most

previous studies were performed in a laboratory setting or in otherwise controlled environment.

Construction, however, is dynamic, diverse, and transient and the predictive validity of these

causal factors is unknown. The present study is the first to test the predictive validity of potential

fatigue indicators in an uncontrolled occupational setting.

4.6 Methodology

The overarching procedure to test the principal hypothesis consisted of three distinct

phases: (1) developing the experiment tools to assess potential predictors and measure fatigue; (2)

collecting data from construction workers in the field, (3) statistical hypothesis testing using

objective predictive statistics. These phases are described in detail below.

111
4.6.1 Research tools development and selection

Activity Questionnaire

Based on the causal factors of fatigue discussed in the literature review, an objective

questionnaire was developed. Each of the causal factors was converted into an objective question,

referred to as an activity variable. Collectively, these represented the activity questionnaire. Table

17 summarizes the causes, activity variables, and the corresponding questions. As one may note,

variables were continuous or dichotomous, which is relevant for statistical testing. Additionally,

demographic variables such as age, years of experience, position, and gender were included in the

Activity Questionnaire. In total, 26 variables were collected using the questionnaire. These

represented the independent predictor variables.

Table 18 Potential predictors of fatigue

Causes of fatigue Activity Variables Activity Questionnaire
Sleep 24h How many hours did you sleep during the past 24hs?
Sleep Deprivation
Sleep How many hours do you usually sleep?
Rest in 24h Besides sleeping, how many restful hours have you had in the
past 24hs?
Incomplete Individuals Do you have individual at home that require special attention?
Recovery Time off How many hours have you had off since your last shift?
Start time What is the start time of your shift today?
Rotating Do you have a rotating shift?
Workload Consecutive shifts How many previous consecutive shifts have you worked?
Overtime Overtime How many hours do you work a week?
& Long Shifts Last shift How long was your last shift?
Repetitive Are you performing repetitive tasks at work?
Vibration Are you exposed to vibration while working?
Noise Are you exposed to loud noises at work?
Work Environment
Light Is the work environment well lit?
Temperature What is the approximate temperature in your work
environment now?
Mental Exertion Mental demand 1 Are you performing mentally demanding tasks at work?
Muscular Exertion Physical demand 1 Are you performing physically demanding tasks at work?
Coworker Do you have a good relationship with your coworkers?
Social Environment
Supervisor Do you have a good relationship with your supervisor?
Caffeine Did you drink caffeine today? If yes, how much?
Self-Predisposition Condition 1 Do you have a condition that doesn’t allow you to rest well?
Condition 2 Do you get fatigued more easily than your coworkers?

112
Fatigue Measurement Tools

Subjective measurement of fatigue:

The current body of knowledge identified the SOFI as the most valid method of

subjectively measuring fatigue. The questionnaire includes 20 items such as: palpitations, worn

out, lack of interest, and others that participants rate on a 0 to 6 Likert scale, where 0 refers to ‘not

at all’ and 6 ‘to a very high degree’. The questionnaire is designed to produce higher scores when

participants are fatigued. For example, the questionnaire states ‘lack of energy’ as an item for

which a high answer (i.e. 6) would indicate ‘lack of energy to a very high degree.’ Because of

space limitations, the entire SOFI scale is not presented here. However, it can be found in its

entirety in (Åhsberg 2000).

Objective Measurement of Fatigue:

To objectively measure fatigue the PC-PVT was used. The implementation of the PC-PVT

requires a laptop computer with 2 GB of memory RAM, Windows 7 as the operating system, and

a gaming mouse with 1000hz response rate. To participate in a PC-PVT a participant sits in front

of a laptop computer with their dominant hand on the mouse. Before starting, the screen shows an

instructional message and when the participant clicks the mouse, the test starts. The participant

observes a black screen and, suddenly, a stimulus appears in the form of a red four-digit

millisecond counter that stops the count once the participant clicks the mouse. The counter displays

the RT of the individual for 500 ms and then disappears. This sequence repeats at random intervals

between 2 and 10 seconds. This test has been validated in a 5-min and a 10-min modality and the

shorter modality was selected (Khitrov et al. 2013).

113
 The PC-PVT stores the following relevant data: (1) time of the assessment; (2) number of

minor lapses (RT> 500ms); (3) number of major lapses (RT>1000ms); (4) number of stimulus

displayed; (4) minim RT; (5) maxim RT; (6) mean RT; and (7) median RT. The median RT was

considered as the dependent variable and most accurate objective measure of fatigue. The mean

can more easily be affected by lapses in RT, which are usually observed with the first stimulus due

to the inexperience of the construction worker with the PC-PVT protocol.

4.6.2 Sample Characteristics and Data Collection

When selecting participants for this study, the researchers desired to focus on high-risk

tasks where the impact of fatigue could be most severe. Among industry trades, electrical

transmission and distribution (TD) workers are known to have a disproportionately high fatality

rate. Specifically, the average TD worker’s fatality rate has been 3 times greater than the

construction industry average for the past 20 years (U.S. Bureau of Labor Statistics 2015). To add

breadth to the study and demonstrate generalizability, general construction workers who represent

several traditional trades, referred to here as GC workers, were also included but to a lesser extent.

The sample consisted of 252 construction workers from 6 different states across the US.

This group included 200 TD workers (age: µ = 39.0 y/o, σ = 10.9 yrs.; years of experience: µ =

11.4 yrs.; σ = 9.3 yrs.) and 52 GC workers (age: µ = 35.7 y/o, σ = 11.0 yrs.; years of experience:

µ = 7.1 yrs.; σ = 8.9 yrs.). Both groups were predominantly male: from TD workers, 193 men and

7 women; from GC workers, 50 men and 2 women; reflecting the actual demographic of the

industry. TD workers included ‘Lineman’ (n=110), ‘Control Room Operators’ (n=59), ‘Operators’

(n=14), ‘Foreman’ (n=13), and ‘Fuel Suppliers’ (n=4). Lineman perform overhead, underground,

or substation electrical work. Control Room Operators (CRO) are power plant workers who take

114
care of the correct functioning of the power plant and divide their time between the control room

and activities around the plant. Fuel Suppliers provide generators and equipment with the

appropriate fuel for their continuous operation. Lastly, Operators control equipment and Foremen

direct crews. These last two trades will be referred to as Electrical Operators (E. Operators) and

Electrical Foreman (E. Foreman) to distinguish them from trades with the same name among GC

workers. The group of GC workers was formed by ‘Carpenters’ (n=32), ‘Operators’ (n=7),

‘Laborers’ (n=6), ‘Foremen’ (n=3), ‘Electricians’ (n=2), and ‘Plumbers’ (n=2).

Data were collected at the beginning of each shift. Participants received an explanation of

the purpose of this research project and the data collection process. After, participants gave their

consent to participate in the study, completed the Activity Questionnaire, the SOFI, and took part

in the PC-PVT. The data were then transcribed to computer files for their posterior analysis.

4.6.3 Data Analysis

Multiple Linear Regression was selected as the initial method of analysis to test the

predictive validity of the activity variables. This method is generally used either to describe or to

predict a relationship (Pedhazur 1997). Despite the continuous nature of the underlying construct

(fatigue), it is important to ensure that the dependent variable used in the statistical analysis reflects

continuity. The study analyzed two dependent variables (1) PVT median RT and (2) SOFI scores.

The former dependent variable measures time, consequently it is a continuous variable; the latter

is an aggregate of Likert scales with continuous properties. The way in which the SOFI was

developed justifies the treatment of SOFI scores as continuous variables, and this is the way in

which these scores have been treated by all other researchers (Åhsberg 2000; González Gutiérrez

et al. 2005; Leung et al. 2004). Furthermore, the wide range of SOFI scores further justifies their

115
treatment as a continuous variable. However, the authors took an extra measure of precaution and

worked with the square root of the SOFI scores ‘SQRT(SOFI)’ which tested positive for normality,

thus justifying its treatment as a continuous variable (Lubke and Muthén 2004).

According to Miller and Kunce (1973) a minimum of 10 data points per variable of interest

may be sufficient to evaluate the predictive character of an independent variable. However, there

is no hard rule regarding this number because it depends entirely on the characteristics of the data

and the real correlation between the independent and dependent variables. In post-processing some

variables were removed from the analysis (e.g., gender and experience) given their low to null

variability.

The best model was selected after performing stepwise MLR (S) analysis, forward MLR

(F) analysis, backwards MLR (B) analysis, and confirming the significance of the regression

coefficients in a forced model. All statistical analyses were performed in SPSS Version 24. At the

end of this process two separate validations were performed. First, the validation of the model’s

fitness was performed, which examined compliance with (1) Linearity, (2) Homoscedasticity, (3)

Normality, and (4) Independence of the residuals (difference between predicted values and actual

values). Second, a k-fold cross validation was performed to document the variability of the model

when applied to a theoretically different population (Stone 1974). The k-fold cross validation

consists of randomly dividing the data set in k equal and constant groups and design k new models

using k-1 groups with a left-out group that changes every time. The results of this second validation

will provide information about the variability of the model when applied to a different sample.

116
 The analyses described above where applied to the overall sample and the TD and GC

groups as described in Figure 1, for both dependent variables (PVT and SOFI). K-fold validation

was applied to TD and GC groups separately as expected for a practical application of the model.

Transformations of dependent variables and elimination of outliers were performed to comply with

MLR validation requirements. The following section presents the results of such analyses.

TD
PVT/SOFI, MLR (S, F, B)
TD + GC Validation 1&2
PVT/SOFI
MLR (S, F, B)
Validation 1 GC
PVT/SOFI, MLR (S, F, B)
Validation 1&2

Figure 8 Statistical Analysis

4.7 Results and analysis

The results obtained from the MLR analyses and the k-fold validations are presented here

by dependent variable. First, Table 19 shows the result for the MLR analyses with median RT

values as the dependent variable. Second, Table 20 shows the results for the MLR models with

SQRT (SOFI) as the dependent variable. A respective 95% and a 90% significance level for

explained variability difference was applied to include and exclude a variable in any of the MLR

methods used.

In the case of RT as the dependent variable, when working with data obtained from TD

workers, a log-normal transformation of the PVT median values was applied to achieve normality

of the residuals. Such a transformation was not necessary when analyzing the GC data. The cross

validation for TD workers (n=200) consisted of a 10-fold cross validation with approximately 180

117
data strings per new sub model. A total of 7 data strings were not considered in the analysis due to

missing values in the data. Similarly, a 4-fold validation with approximately 37 data strings for

sub model was performed for GC worker’s data.

118
 Table 19 MLR and Validation results for PVT data
TD + GC workers TD workers GC workers
Validation
D.V. Median RT, n=236 D.V. Ln (Median RT), n=193 D.V. Median RT, n=47 Validation 𝝁 (𝝈)
𝝁 (𝝈)
R2 0.202 R2 0.092 0.095 (0.015) R2 0.198 0.200 (0.024)
Adjusted R2 0.185 Adjusted R2 0.077 0.080 (0.016) Adjusted R2 0.161 0.151 (0.026)
Std. Error 37.456 Std. Error 0.137 0.145 (0.002) Std. Error 35.285 35.330 (1.690)
Std. Std. Validation Std.
Predictors Bi Predictors Bi Predictors Bi Validation 𝝁 (𝝈)
Error Error 𝝁 (𝝈) Error
Intercept 303.664 16.481 Intercept 5.615 0.061 5.657 (0.017) Intercept 258.366 25.812 259.592 (11.941)
TD worker -29.607 6.467 Last Shift -0.012 0.005 -0.016 (0.002) Sleep 24 h 8.861 3.565 8.689 (1.446)
Last Shift -2.744 1.316 Noise 0.060 0.027 0.058 (0.005) Condition 1 -39.145 18.447 -38.680 (12.887)
Noise 13.091 7.388 Temperature 0.002 0.000 0.002 (0.000)
Temperature 0.372 0.120
Condition 2 19.200 7.132
Note: D.V.: dependent variable; Std.: standard deviation

Table 20 MLR and Validation results for SOFI data

TD + GC workers TD workers GC workers
D.V. SQRT(SOFI), n= 242 D.V. SQRT(SOFI), n= 180 Validation 𝝁 (𝝈) D.V. SQRT(SOFI), n= 42 Validation 𝝁 (𝝈)
R2 0.151 R2 0.161 0.162 (0.015) R 2
0.484 0.490 (0.060)
Adjusted R2 0.137 Adjusted R2 0.141 0.141 (0.015) Adjusted R2 0.458 0.454 (0.062)
Std. Error 1.972 Std. Error 1.201 1.201 (0.021) Std. Error 0.928 0.925 (0.009)
Std. Std. Std.
Predictors Bi Predictors Bi Validation 𝝁 (𝝈) Predictors Bi Validation 𝝁 (𝝈)
Error Error Error
Intercept 9.076 0.951 Intercept 7.651 0418 7.651 (0.131) Intercept 5.246 0.288 5.257 (0.055)
Age -0.031 0.012 E. Operator -0.998 0.348 -1.000 (0.060) Consecutive Shift 0.449 0.081 0.442 (0.042)
Sleep 24 h -0.317 0.081 Sleep 24 h -0.180 0.056 -0.180 (0.017) Condition 2 0.987 0.442 0.990 (0.265)
Supervisor -1.761 0.676 Rest -0.038 0.019 -0.038 (0.008)
Condition 2 1.523 0.382 Condition 2 1.065 0.268 1.067 (0.088)
Note: D.V.: dependent variable; Std.: standard deviation

119
 From the general PVT model including TD and GC workers, TD workers appear as a

predictor, indicating that there are differences in RT between the two groups of workers and that

these should be treated as separate groups. Furthermore, the larger number of TD workers biased

the overall model, because this model (TD + GC workers) shares most of the predictors with the

model that considers only TD workers (Last Shift, Noise, and Temperature).

Generally, the regression coefficients obtained in the models reflect the expected

relationship between the predictors and the predicted variable. For instance, the coefficients for

Consecutive Shifts, Noise, Temperature, and Condition 2 (becoming fatigued more easily than

coworkers) are positive in all models. This indicates that when considering each one of these

variables at the time, if all the other variables in the model are constant, then an increase in any of

these predictors would generate an increase in RT or SOFI scores, as expected. Likewise, the

coefficient for Supervisor and most of those for Sleep 24h are negative. This shows that a good

relationship with the supervisor and an increase in the amount of sleep obtained in the past 24h

would contribute to a lower RT or SOFI scores.

Nonetheless, some of the results are unexpected. For example, the coefficients for Last

Shift in two models resulted to be negative. This would indicate that the longer the previous shift

of the workers, the faster RT or lower SOFI score they would obtain. A possible explanation for

this could be that there is a slightly positive significant correlation between Last Shift and Time

Off (Pearson’s r = 0.171, p-value = 0.017). Consequently, one can infer that the longer the previous

shift, the longer time off the workers had, thus allowing for more recovery. However, Time off was

120
another variable in these MLR analyses and it was not found as a significant predictor. The full

discovery of this matter will require additional research.

Similarly, the results obtained from the analyses among GC workers, considering median

RT as a predictor, also suggest the need for further research with a bigger sample. In this case, the

coefficients for both predictors show counterintuitive signs. The coefficient for Sleep 24h is

positive, indicating that in general, the more a worker slept in the past 24h the slower RT observed.

Likewise, the coefficient of Condition 1 indicates that the few workers (n=4) that manifested to

have a condition that didn’t allow them to rest well, presented faster RT’s than the rest. Figure 9,

presents a Scatterplot matrix of the data showing that the counterintuitive relationship between the

predictors and the criterion measured, as manifested in the regression analysis, is the true nature

of the collected data.

Figure 9. Scatterplot matrix of Median RT predictors for GC workers

121
 To the extent to which fatigue can be measured by the PVT and SOFI, as justified by

previous researchers, activity variables showed to be able to predict fatigue (Åhsberg 2000;

Khitrov et al. 2013). However, the degree of predictability observed was low to medium. Overall,

the predictive models explain 10% to 50% of the variability in the dependent variables. The models

that predict RT median values, as measured by the PVT, showed a minimum explained variability

of 9% and a maximum of 20%, with Sleep 24h, Last Shift, Noise, Temperature, Condition 1, and

Condition 2 as predictors. The 10-fold validation performed among TD workers with median RT

as the dependent variable illustrates that in general, when applied to another theoretical sample,

the model will perform as expected, showing the same level of predictivity. On the other hand,

the models with SOFI scores as the dependent variable explain between 15% and 50% of

variability, with: Sleep 24h, Consecutive Shifts, Supervisor, and Condition 2 as predictors and the

4-fold cross validations indicates that the parameters of the model will be similar when applied to

a new but comparable population. Thus, the predictors considered in this study are better at

predicting SOFI scores than RT as measured by the PVT.

Overall, the models indicate an inability to predict between 80% and 90% of variability in

RT and between 50% and 84% inability to explain variability in SOFI scores. It is unknown to the

authors what specific variables could explain this variability and such a gap in knowledge should

be addressed in future research. However, regarding RT, previous researchers showed that in

addition to fatigue, age, sleep disorders, alcohol or stimulating substance consumption, and

physical fitness, influence RT (Hultsch et al. 2002; Kosinski 2008; Powell et al. 1999). This study

accounted for the effect of age, somatic disorders, and the ingestion of caffeine but not for physical

fitness. On the other hand, SOFI scores represent 5 dimensions of fatigue (lack of energy, lack of

motivation, sleepiness, physical discomfort, and physical exertion) and there might be a plethora

122
of variables that influence each of these categories which were not included in the activity

variables.

Furthermore, it is easy to see that SOFI and PVT models present 2 common predictors,

Sleep 24h and Condition 2, and the rest of predictors differ between models. The PVT presents

Last Shift, Noise, Temperature, and Condition 1 as predictors; while SOFI presents Consecutive

Shifts and Supervisor, as predictors. This indicates that PVT and SOFI measure different constructs

and that none of them represent an exhaustive measure of fatigue by themselves. Furthermore, a

correlation analysis revealed no significant correlation between PVT RT and SOFI scores

(Pearson’s r = 0.072, p-value = 0.254). The predictive model based on activity variables presented

in this study was able to represent both constructs thus offering a more complete measure of

fatigue. Table 21 represents a summary of causal factors, potential predictors, and actual predictors

by dependent variable and trade.

Table 21 Predictors by fatigue measurement method and worker type

PVT SOFI TD GC
Causes of fatigue Predictors
Predictors Predictors Predictors Predictors
Sleep Deprivation Sleep 24h X X X X
Incomplete Recovery Rest in 24h X
Workload Consecutive shifts X X
Overtime & Long Shifts Last shift X X
Noise X X
Work Environment
Temperature X X
Mental Exertion Mental demand 1
Muscular Exertion Physical demand 1
Social Environment Supervisor X
Condition 1 X X
Self-Predisposition
Condition 2 X X X X

The reader will notice that physical and mental demand didn’t appear as significant in any

model. The reason behind this is that such variables present almost null variability across

123
participants; indicating that these variables are present and play an obvious role in the development

of fatigue, but they are bad differentiators of fatigue.

Lastly, it is easy to see, from Table 21, that TD workers are more affected by work

environmental factors compared to GC workers. Most TD workers, perform their work activities

directly exposed to the elements all year around. For example, overhead lineman work on the pole

exposed to sun radiation, wind, rain, or snow. On the other hand, underground lineman work in

manholes which present a temperature approximately 10 ºF higher than outside, thus suffering the

effects of high temperatures. GC workers, however, spend most of the time working inside newly

constructed buildings once the infrastructure has been built. These differences may be the reason

behind the different predictors between these trades.

All the MLR analyses performed were validated for linearity, normality, and

homoscedasticity of the results. Furthermore, the data complies with independence of the residuals

as there was no time pattern in the data collection process. Lastly, all models passed checks for

outliers, leverage, and influence. In some cases, when specific data points compromised any of

these statistics, these points were removed from the data as it is typically recommended in

regression analysis.

Tables 22, 23, and 24 show the descriptive statistics of the predictors presented in the

models and the dependent variables. Section 4.10 of this study presents an appendix with the

descriptive statistics of all the activity variables considered.

124
 Table 22 Descriptive statistics of continuous predictive variables
Overall sample n = 249 TD workers n = 198 GC workers n = 51
Min Max 𝑋̅ S Min Max 𝑋̅ S Min Max 𝑋̅ S
Sleep 24 (h) 3.0 14.0 7.0 1.6 3.5 14.0 7.0 1.6 3.0 11.0 7.0 1.4
Consecutive shifts 0.0 12.0 2.3 1.9 0.0 12.0 2.1 1.9 0.0 6.0 3.0 1.7
Last shift (h) 4.0 16.0 10.0 1.9 4.0 16.0 10.2 2.0 5.0 12.0 9.0 1.5
Temperature (Fº) 17.0 110.0 60.5 20.9 17.0 110.0 59.9 22.4 42.0 80.0 63.0 11.6

Table 23 Frequencies of dichotomous predictive variables

Overall sample TD workers GC workers
Yes No Yes No Yes No
Noise 216 32 168 32 48 4
Supervisor 237 9 194 4 43 5
Condition1 34 214 30 170 4 44
Condition2 32 212 25 170 7 42

Table 24 descriptive statistics of dependent variables

Overall sample n = 252 TD workers n = 200 GC workers n = 52
Min Max 𝑋̅ S Min Max 𝑋̅ S Min Max 𝑋̅ S
Median RT 200 460 291 45 200 460 284 44 223 391 318 40
SOFI Score 0 76 22 18 0 71 21 17 0 76 27 20

4.7.1 Interactions

Possible interactions among all predictive variables where analyzed for every TD and GC

model with median RT and the square root of SOFI scores as the dependent variables. Only 2

double interactions were found to be significant and these are presented in Table 25 and Table 26

respectively.

125
 Table 25 Interaction’s analysis for a PVT model among TD workers
D.V. Ln (Median RT), n=193
R2 0.150
Adjusted R2 0.132
Std. Error 0.133
Std.
Predictors Bi
Error
Intercept 5.593 0.024
Last Shift_C -0.012 0.005
Noise 0.060 0.026
Temperature_C 0.002 0.000
Temp_Last Shift_C -0.001 0.000
Notation: D.V. = dependent variable; _C = Centered; Temp_Last Shift = Temperature x Last Shift

Table 26 Interaction’s analysis for a SOFI model among TD workers

D.V. L SQRT (SOFI), n=180
R2 0.192
Adjusted R2 0.169
Std. Error 1.181
Std.
Predictors Bi
Error
Intercept 7.590 0.412
E. Operator -1.002 0.342
Sleep 24 h -0.184 0.055
(Rest) -0.017 0.020
Condition 2 1.883 0.409
Rest_Condition 2 -0.131 0.050
Notation: D.V. = dependent variable; ( ) = Not significant; Rest_Condition 2 = Rest x Condition 2

After centering the predictive variables of the PVT model, by subtracting the variable mean

to each data point, no collinearity effect was detected. Table 25 indicates a significant interaction

between Temperature and Last Shift that increases the explained variability to 15%. Regarding the

SOFI model, the interaction between Rest and Condition 2 also offers a greater level of explained

variability at 19%. However, these interactions don’t dramatically change the predictability of the

previous models.

126
4.7.2 Limitations of the study

Despite minor limitations of this study, the authors recognize 3 main limitations. The first

limitation refers to the initial selection of potential predictors. Such a selection was driven by

previous research, therefore, only those variables recognized as causal factors of fatigue by other

researchers where included in this study. The results indicate, that such variables are able to explain

up to 50% of the variability in fatigue levels of the sample as measured by the SOFI and the PVT.

Thus, indicating that there may be other predictors, not considered in this study, able to add

explained variability to these models.

The second main limitation of this study refers to the extent to which the selected fatigue

measurement tools can register all fatigue’s dimensions. In theory, the study accounts for an

exhaustive measure of fatigue by measuring objective and subjective fatigue. However, it is

possible that other fatigue measurement tools may be able to register other dimensions of fatigue

not recognized by the SOFI or the PVT.

The third main limitation of this study refers to the volume of GC workers. When analyzing

the results of such a group some unexpected results were found regarding the sign of the predictors.

The use of a bigger sample would add more variability to the data and possibly correct biases.

Thus, offering more valid results. An overall reflection of this study is presented in the following

section.

127
4.8 Conclusion

The purpose of this study was to test the predictive validity of factors that have been found

to cause fatigue in laboratory settings, by measuring fatigue in-situ, fatigue in a high-risk industry.

Twenty-six variables representing nine fatigue causal factors were tested and the results revealed

low to medium predictive power, suggesting that laboratory results do not necessarily translate to

the field and that other factors not yet identified are likely at play.

This fact becomes evident while studying fatigue in physically active working conditions,

such as those experienced in the construction industry. Laboratory results may apply more directly

to those professions that require a level of activity similar to that experienced during laboratory

experiments. In the case of construction workers, achieving such a similitude between laboratory

conditions and the worksite is nearly impossible. Thus, questioning the external validity of

laboratory studies relative to actual field conditions.

Nonetheless, the models presented in this study offer an alternative to predict fatigue from

a few activity variables that is valid for the construction industry. Therefore, the researchers

conclude that construction workers’ fatigue can be predicted. Conversely, the authors sustain that

such models are not ready for their implementation in the industry as fatigue predictive tools. The

amount of error present in these models is still unacceptable for practical implementations.

Therefore, future research should strengthen the predictivity of these models by discovering new

predictors able to further explain fatigue variability.

128
 The discovery that predictors vary by trades is a valuable contribution to the industry.

Knowing that work environmental factors such as noise and temperature drive fatigue in TD

workers can influence decision making regarding the management of fatigue in this sector. For

instance, it will probably be more effective to buy ‘blowers’ to help underground workers cope

with a hot day than prevent them from working an extra day. However, such affirmation should

be studied in detail considering the characteristics of the work and job-site. This example is

presented only to illustrate the principle. Interestingly, the predictors obtained for RT among TD

workers coincide with those previously identified by TD workers as the main drivers of fatigue in

an exploratory study conducted by Techera et al (2017).

Additionally, the identification of actual fatigue predictors in this study can help the

industry implement Fatigue Risk Management System (FRMS). The multilayer system of defense

suggested by FRMS offers a general tactic to manage fatigue and the protocol followed in this

study should serve as a preliminary approach to craft a FRMS to a specific trade, thus increasing

the likelihood of its success.

Even though most predictive variables where objective (i.e. Sleep 24h and Consecutive

Shifts), these were better at forecasting subjective fatigue as measured by the SOFI than objective

fatigue as measured by the PVT. Thus, further validating the correlation between objective fatigue

causal factors and subjective levels of fatigue. The development of objective instruments to

measure fatigue offers a very attractive approach to fatigue measurement because these

instruments diminish the ability of an individual to manipulate the results. However, the results

129
obtained in this study indicate that objective and subjective fatigue measurement tools identify

different dimensions of fatigue and both should be considered to measure fatigue.

There is still no perfect or exhaustive method for measuring fatigue that is able to identify

all dimensions of fatigue. Consequently, future research should present a multi-instrumental

approach to fatigue measurement. Based on the results obtained in this study, a multidimensional

approach as the one presented here with activity variables, has the potential to predict fatigue levels

considering a wider range of fatigue dimensions than a specific instrument. Nonetheless, these

models need further development and validation before serving as fatigue predictive tools.

The research community’s understanding of sleep deprivation as a main driver of fatigue

is reinforced by this study. However, this study emphasizes the importance of further investigating

the impact of work environmental factors and social factors in the development of fatigue in the

field. Such areas of research have not been widely studied and there is a dearth of research in these

fields among construction workers.

4.9 References

Åhsberg, E. (2000). “Dimensions of fatigue in different working populations.” Scandinavian

Journal of Psychology, 41(3), 231–241.
Åhsberg, E., and Gamberale, F. (1998). “Perceived fatigue during physical work: an experimental
evaluation of a fatigue inventory.” International Journal of Industrial Ergonomics, 21(2),
117–131.
Åhsberg, E., Gamberale, F., and Gustafsson, K. (2000a). “Perceived fatigue after mental work: an
experimental evaluation of a fatigue inventory.” Ergonomics, 43(2), 252–268.
Åhsberg, E., Garnberale, F., and Kjellberg, A. (1997). “Perceived quality of fatigue during
different occupational tasks Development of a questionnaire.” International Journal of
Industrial Ergonomics, 20(2), 121–135.
Åhsberg, E., Kecklund, G., Åkerstedt, T., and Gamberale, F. (2000b). “Shiftwork and different
dimensions of fatigue.” International Journal of Industrial Ergonomics, 26(4), 457–465.

130
Åkerstedt, T., Fredlund, P., Gillberg, M., and Jansson, B. (2002). “Work load and work hours in
relation to disturbed sleep and fatigue in a large representative sample.” Journal of
Psychosomatic Research, 53(1), 585–588.
Angus, R. G., and Heslegrave, R. J. (1985). “Effects of sleep loss on sustained cognitive
performance during a command and control simulation.” Behavior Research Methods,
Instruments, & Computers, 17(1), 55–67.
Balkin, T. J., Bliese, P. D., Belenky, G., Sing, H., Thorne, D. R., Thomas, M., Redmond, D. P.,
Russo, M., and Wesensten, N. J. (2004). “Comparative utility of instruments for monitoring
sleepiness-related performance decrements in the operational environment.” Journal of
Sleep Research, 13(3), 219–227.
Belenky, G., Wesensten, N. J., Thorne, D. R., Thomas, M. L., Sing, H. C., Redmond, D. P., Russo,
M. B., and Balkin, T. J. (2003). “Patterns of performance degradation and restoration
during sleep restriction and subsequent recovery: a sleep dose-response study.” Journal of
Sleep Research, 12(1), 1–12.
Beurskens, A. J. H. M., Bültmann, U., Kant, Ij., Vercoulen, J. H. M. M., Bleijenberg, G., and
Swaen, G. M. H. (2000). “Fatigue among working people: validity of a questionnaire
measure.” Occupational and Environmental Medicine, 57(5), 353–357.
Boksem, M. A., Meijman, T. F., and Lorist, M. M. (2006). “Mental fatigue, motivation and action
monitoring.” Biological psychology, 72(2), 123–132.
Boksem, M. A. S., Meijman, T. F., and Lorist, M. M. (2005). “Effects of mental fatigue on
attention: An ERP study.” Cognitive Brain Research, 25(1), 107–116.
Boksem, M. A. S., Tops, M., Kostermans, E., and De Cremer, D. (2008). “Sensitivity to
punishment and reward omission: Evidence from error-related ERP components.”
Biological Psychology, 79(2), 185–192.
Borg, G. (1982). “A category scale with ratio properties for intermodal and interindividual
comparisons.” Psychophysical judgment and the process of perception, 25–34.
Bravo, G., and Potvin, L. (1991). “Estimating the reliability of continuous measures with
cronbach’s alpha or the intraclass correlation coefficient: Toward the integration of two
traditions.” Journal of Clinical Epidemiology, 44(4), 381–390.
Brown, I. D. (1994). “Driver fatigue.” Human Factors: The Journal of the Human Factors and
Ergonomics Society, 36(2), 298–314.
Bültmann, U., Kant, I., Schröer, C., and Kasl, S. (2001). “The relationship between psychosocial
work characteristics and fatigue and psychological distress.” International Archives of
Occupational and Environmental Health, 75(4), 259–266.
Bültmann, U., de Vries, M., M, J. H., Bleijenberg, G., M, H. M., and Kant, Ij. (2000).
“Measurement of prolonged fatigue in the working population: Determination of a cutoff
point for the Checklist Individual Strength.” Journal of Occupational Health Psychology,
5(4), 411–416.
Chan, M. (2011). “Fatigue: the most critical accident risk in oil and gas construction.”
Construction Management and Economics, 29(4), 341–353.
Chervin, R. D. (2000). “Sleepiness, Fatigue, Tiredness, and Lack of Energy in Obstructive Sleep
Apnea.” Chest, 118(2), 372–379.
Christensen, H., Søgaard, K., Jensen, B. R., Finsen, L., and Sjøgaard, G. (1995). “Intramuscular
and surface EMG power spectrum from dynamic and static contractions.” Journal of
Electromyography and Kinesiology, 5(1), 27–36.

131
Cronbach, L. J. (1951). “Coefficient alpha and the internal structure of tests.” Psychometrika,
16(3), 297–334.
Dainoff, M. J., Happ, A., and Crane, P. (1981). “Visual Fatigue and Occupational Stress in VDT
Operators.” Human Factors: The Journal of the Human Factors and Ergonomics Society,
23(4), 421–437.
Dawson, D., and Fletcher, A. (2001). “A quantitative model of work-related fatigue: background
and definition.” Ergonomics, 44(2), 144–163.
Dawson, D., and McCulloch, K. (2005). “Managing fatigue: It’s about sleep.” Sleep Medicine
Reviews, 9(5), 365–380.
Dawson, D., Searle, A. K., and Paterson, J. L. (2014). “Look before you (s)leep: Evaluating the
use of fatigue detection technologies within a fatigue risk management system for the road
transport industry.” Sleep Medicine Reviews, 18(2), 141–152.
Dembe, A. E., Erickson, J. B., Delbos, R. G., and Banks, S. M. (2005). “The impact of overtime
and long work hours on occupational injuries and illnesses: new evidence from the United
States.” Occupational and Environmental Medicine, 62(9), 588–597.
Dinges, D. F., Pack, F., Williams, K., Gillen, K. A., Powell, J. W., Ott, G. E., Aptowicz, C., and
Pack, A. I. (1997). “Cumulative sleepiness, mood disturbance and psychomotor vigilance
performance decrements during aweek of sleep restricted to 4-5 hours per night.” Sleep:
Journal of Sleep Research & Sleep Medicine, 20(4), 267–277.
Dinges, D. F., and Powell, J. W. (1985). “Microcomputer analyses of performance on a portable,
simple visual RT task during sustained operations.” Behavior Research Methods,
Instruments, & Computers, 17(6), 652–655.
Dorrian, J., Rogers, N. L., Dinges, D. F., and others. (2005). “Psychomotor vigilance performance:
Neurocognitive assay sensitive to sleep loss.” Marcel Dekker New York.
Edwards, R. G., and Lippold, O. C. J. (1956). “The relation between force and integrated electrical
activity in fatigued muscle.” The Journal of Physiology, 132(3), 677–681.
Folkard, S., and Tucker, P. (2003). “Shift work, safety and productivity.” Occupational medicine,
53(2), 95–101.
Gander, P., Hartley, L., Powell, D., Cabon, P., Hitchcock, E., Mills, A., and Popkin, S. (2011).
“Fatigue risk management: Organizational factors at the regulatory and industry/company
level.” Accident Analysis & Prevention, Advancing Fatigue and Safety Research, 43(2),
573–590.
Goldich, Y., Barkana, Y., Pras, E., Zadok, D., Hartstein, M., and Morad, Y. (2010). “The Effects
of Sleep Deprivation on Oculomotor Responses.” Current Eye Research, 35(12), 1135–
1141.
González Gutiérrez, J. L., Jiménez, B. M., Hernández, E. G., and López López, A. (2005).
“Spanish version of the Swedish Occupational Fatigue Inventory (SOFI): Factorial
replication, reliability and validity.” International Journal of Industrial Ergonomics, 35(8),
737–746.
Grandjean, E. (1979). “Fatigue in industry.” British Journal of Industrial Medicine, 36(3), 175–
186.
Hallowell, M. R. (2010). “Worker fatigue: managing concerns in rapid renewal highway
construction projects.” Professional Safety, 55(12), 18–26.
Harrison, Y., and Horne, J. A. (1999). “One Night of Sleep Loss Impairs Innovative Thinking and
Flexible Decision Making.” Organizational Behavior and Human Decision Processes,
78(2), 128–145.

132
Harrison, Y., and Horne, J. A. (2000). “The impact of sleep deprivation on decision making: A
review.” Journal of Experimental Psychology: Applied, 6(3), 236–249.
Hartley, L. R., and Arnold, P. (2001). “Managing fatigue in the road transport industry: an
occupational safety and health solution.” STRESS, WORKLOAD AND FATIGUE.
Hsiao, H., and Simeonov, P. (2001). “Preventing falls from roofs: a critical review.” Ergonomics,
44(5), 537–561.
Hultsch, D. F., MacDonald, S. W. S., and Dixon, R. A. (2002). “Variability in Reaction Time
Performance of Younger and Older Adults.” The Journals of Gerontology: Series B, 57(2),
P101–P115.
Jason, L. A., Evans, M., Brown, M., and Porter, N. (2010). “What is fatigue? Pathological and
nonpathological fatigue.” PM&R, 2(5), 327–331.
Jiao, K., Li, Z., Chen, M., Wang, C., and Qi, S. (2004). “Effect of different vibration frequencies
on heart rate variability and driving fatigue in healthy drivers.” International Archives of
Occupational and Environmental Health, 77(3), 205–212.
Kartam, N. A. (1997). “Integrating Safety and Health Performance into Construction CPM.”
Journal of Construction Engineering and Management, 123(2), 121–126.
Khitrov, M. Y., Laxminarayan, S., Thorsley, D., Ramakrishnan, S., Rajaraman, S., Wesensten, N.
J., and Reifman, J. (2013). “PC-PVT: A platform for psychomotor vigilance task testing,
analysis, and prediction.” Behavior Research Methods, 46(1), 140–147.
Kjellberg, A., Muhr, P., and Skoldstrom, B. (1998). “Fatigue after work in noise - an
epidemiological survey study and three quasi-experimental field studies.” Noise & Health,
1(1), 47–55.
Kosinski, R. J. (2008). “A literature review on reaction time.” Clemson University, 10.
Krause, N., Lynch, J., Kaplan, G. A., Cohen, R. D., Goldberg, D. E., and Salonen, J. T. (1997).
“Predictors of disability retirement.” Scandinavian Journal of Work, Environment &
Health, 23(6), 403–413.
Lal, S. K. L., and Craig, A. (2001). “A critical review of the psychophysiology of driver fatigue.”
Biological Psychology, 55(3), 173–194.
Lal, S. K. L., Craig, A., Boord, P., Kirkup, L., and Nguyen, H. (2003). “Development of an
algorithm for an EEG-based driver fatigue countermeasure.” Journal of Safety Research,
34(3), 321–328.
Landström, U., and Löfstedt, P. (1987). “Noise, vibration and changes in wakefulness during
helicopter flight.” Aviation, Space, and Environmental Medicine, 58(2), 109–118.
Leung, A. W. S., Chan, C. C. H., and He, J. (2004). “Structural stability and reliability of the
Swedish occupational fatigue inventory among Chinese VDT workers.” Applied
Ergonomics, The Occlusion Technique, 35(3), 233–241.
Lim, J., and Dinges, D. F. (2008). “Sleep Deprivation and Vigilant Attention.” Annals of the New
York Academy of Sciences, 1129(1), 305–322.
van der Linden, D., and Eling, P. (2006). “Mental fatigue disturbs local processing more than
global processing.” Psychological Research, 70(5), 395–402.
van der Linden, D., Frese, M., and Meijman, T. F. (2003a). “Mental fatigue and the control of
cognitive processes: effects on perseveration and planning.” Acta Psychologica, 113(1),
45–65.
van der Linden, D. van der, Frese, M., and Sonnentag, S. (2003b). “The Impact of Mental Fatigue
on Exploration in a Complex Computer Task: Rigidity and Loss of Systematic Strategies.”

133
 Human Factors: The Journal of the Human Factors and Ergonomics Society, 45(3), 483–
494.
Loh, S., Lamond, N., Dorrian, J., Roach, G., and Dawson, D. (2004). “The validity of psychomotor
vigilance tasks of less than 10-minute duration.” Behavior Research Methods, Instruments,
& Computers, 36(2), 339–346.
Lorist, M. M., Boksem, M. A., and Ridderinkhof, K. R. (2005). “Impaired cognitive control and
reduced cingulate activity during mental fatigue.” Cognitive Brain Research, 24(2), 199–
205.
Lorist, M. M., Klein, M., Nieuwenhuis, S., de Jong, R., Mulder, G., and Meijman, T. F. (2000).
“Mental fatigue and task control: Planning and preparation.” Psychophysiology, 37(5),
614–625.
Lubke, G. H., and Muthén, B. O. (2004). “Applying Multigroup Confirmatory Factor Models for
Continuous Outcomes to Likert Scale Data Complicates Meaningful Group Comparisons.”
Structural Equation Modeling: A Multidisciplinary Journal, 11(4), 514–534.
May, J. F., and Baldwin, C. L. (2009). “Driver fatigue: The importance of identifying causal factors
of fatigue when considering detection and countermeasure technologies.” Transportation
Research Part F: Traffic Psychology and Behaviour, 12(3), 218–224.
Michielsen, H. J., De Vries, J., Van Heck, G. L., R, J., and Sijtsma, K. (2004). “Examination of
the Dimensionality of Fatigue: The Construction of the Fatigue Assessment Scale (FAS).”
European Journal of Psychological Assessment, 20(1), 39–48.
Miller, D. E., and Kunce, J. T. (1973). “Prediction and Statistical Overkill Revisited.”
Measurement and Evaluation in Guidance.
Mota, D. D. C. F., and Pimenta, C. A. M. (2006). “Self-Report Instruments for Fatigue
Assessment: A Systematic Review.” Research and Theory for Nursing Practice, 20(1), 49–
78.
Nunnally, J. (1978). Psychometric methods. New York: McGraw-Hill.
Okogbaa, O. G., Shell, R. L., and Filipusic, D. (1994). “On the investigation of the
neurophysiological correlates of knowledge worker mental fatigue using the EEG signal.”
Applied Ergonomics, 25(6), 355–365.
Park, J., Kim, Y., Chung, H. K., and Hisanaga, N. (2001). “Long working hours and subjective
fatigue symptoms.” Industrial health, 39(3), 250–254.
Park, S., and Gotoh, S. (1993). “Research on office environment evaluation and worker fatigue.”
Journal of Thermal Biology, 18(5–6), 565–570.
Pedhazur, E. (1997). Multiple Regression in Behavioral Research: Explanation and Prediction.
Holt, Rinehart and Winston.
Piper, B. F., Dibble, S., Dodd, M., Weiss, M., Slaughter, R., and Paul, S. (1998). “The revised
Piper Fatigue Scale: psychometric evaluation in women with breast cancer.” Oncology
nursing forum, 25(4), 677–684.
Powell, N. B., Riley, R. W., Schechtman, K. B., Blumen, M. B., Dinges, D. F., and Guilleminault,
C. (1999). “A Comparative Model: Reaction Time Performance in Sleep-Disordered
Breathing Versus Alcohol-Impaired Controls.” The Laryngoscope, 109(10), 1648–1654.
Santos, J. R. A. (1999). “Cronbach’s alpha: A tool for assessing the reliability of scales.” Journal
of extension, 37(2), 1–5.
Sawacha, E., Naoum, S., and Fong, D. (1999). “Factors affecting safety performance on
construction sites.” International Journal of Project Management, 17(5), 309–315.

134
Shapiro, C. M., Flanigan, M., Fleming, J. A. E., Morehouse, R., Moscovitch, A., Plamondon, J.,
Reinish, L., and Devins, G. M. (2002). “Development of an adjective checklist to measure
five FACES of fatigue and sleepiness: Data from a national survey of insomniacs.” Journal
of Psychosomatic Research, 52(6), 467–473.
Smets, E. M. A., Garssen, B., Bonke, B., and De Haes, J. C. J. M. (1995). “The multidimensional
Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue.”
Journal of Psychosomatic Research, 39(3), 315–325.
Stone, M. (1974). “Cross-validatory choice and assessment of statistical predictions.” Journal of
the royal statistical society. Series B (Methodological), 111–147.
Sultanian, A. R., Motamedzade Torghabe, M., Shafii motlagh, M., Garkaz, A., and Mahdavi, N.
(2014). “Persian version of Swedish Occupational Fatigue Inventory )P-SOFI) : validity
and reliability.” Iran Occupational Health, 11(1), 34–43.
Swaen, G. M. H., Amelsvoort, L. G. P. M. van, Bültmann, U., and Kant, I. J. (2003). “Fatigue as
a risk factor for being injured in an occupational accident: results from the Maastricht
Cohort Study.” Occupational and Environmental Medicine, 60(suppl 1), i88–i92.
Taylor, R. R., Jason, L. A., and Torres, A. (2000). “Fatigue rating scales: an empirical
comparison.” Psychological Medicine, 30(04), 849–856.
Techera, U., Hallowell, M., Stambaugh, N., and Littlejohn, R. (2016). “Causes and consequences
of occupational fatigue: Meta-analysis and systems model.” Journal of Occupational and
Environmental Medicine, 58(10), 961–975.
Techera, U., Hallowell, M., and Littlejohn, R. (2017). “Fatigue Management for Electrical
Transmission and Distribution Construction” Journal of Construction Engineering and
Management, (Planned Submission: April 2017).
U.S. Bureau of Labor Statistics. (2015). “Census of Fatal Occupational Injuries Summary, 2014.”
<http://stats.bls.gov/news.release/cfoi.nr0.htm>.
Van Dongen, H. (2004). “Comparison of mathematical model predictions to experimental data of
fatigue and performance.” Aviation, space, and environmental medicine, 75(3), A15–A36.
Van Dongen, H. P. A., and Dinges, D. F. (2005). “Sleep, Circadian Rhythms, and Psychomotor
Vigilance.” Clinics in Sports Medicine, Sports Chronobiology: It’s About Time, 24(2),
237–249.
Van Veldhoven, M., and Broersen, S. (2003). “Measurement quality and validity of the ‘need for
recovery scale.’” Occupational and Environmental Medicine, 60(suppl 1), i3–i9.
Vries, J. D., Michielsen, H. J., and Heck, G. L. V. (2003). “Assessment of fatigue among working
people: a comparison of six questionnaires.” Occupational and Environmental Medicine,
60(suppl 1), i10–i15.
Wierwille, W. W., Wreggit, S. S., Kirn, C. L., Ellsworth, L. A., and Fairbanks, R. J. (1994).
“Research on vehicle-based driver status/performance monitoring; development,
validation, and refinement of algorithms for detection of driver drowsiness. Final report.”
(No. HS-808 247)
Yagil, D. (2006). “The Relationship of Abusive and Supportive Workplace Supervision to
Employee Burnout and Upward Influence Tactics.” Journal of Emotional Abuse, 6(1), 49–
65.
Zhang, N. Z., and Liu, H.-T. L. (2008). “Effects of sleep deprivation on cognitive functions.”
Neuroscience Bulletin, 24(1), 45–48.
Zivin, J. G., and Neidell, M. J. (2010). Temperature and the Allocation of Time: Implications for
Climate Change. Working Paper, National Bureau of Economic Research.

135
4.10 Appendix

Table 27 Descriptive statistics of continuous activity variables

Overall sample n = 249 TD workers n = 198 GC workers n = 51
Min Max 𝑋̅ S Min Max 𝑋̅ S Min Max 𝑋̅ S
Sleep 24 (h) 3.0 14.0 7.0 1.6 3.5 14.0 7.0 1.6 3.0 11.0 7.0 1.4
Sleep (h) 3.0 12.0 6.8 1.3 3.0 12.0 6.8 1.2 3.5 10.5 6.9 1.3
Rest (h) 0.0 20.0 4.4 4.5 0.0 20.0 4.7 4.9 0.0 11.0 3.0 2.3
Time off (h) 0.0 288.0 37.7 56.7 8.0 288.0 43.0 62.4 0.0 62.0 17.8 12.3
Consecutive shifts 0.0 12.0 2.3 1.9 0.0 12.0 2.1 1.9 0.0 6.0 3.0 1.7
Overtime (h) 0.0 44.0 7.1 8.4 0.0 44.0 7.4 9.0 0.0 20.0 6.1 5.3
Last shift (h) 4.0 16.0 10.0 1.9 4.0 16.0 10.2 2.0 5.0 12.0 9.0 1.5
Temperature (Fº) 17.0 110.0 60.5 20.9 17.0 110.0 59.9 22.4 42.0 80.0 63.0 11.6

Table 28 Frequencies of dichotomous activity variables

Overall sample TD workers GC workers
Yes No Yes No Yes No
Individuals 128 124 100 100 28 24
Rotating 77 174 74 126 3 48
Repetitive 217 35 167 33 50 2
Vibration 183 63 147 52 36 11
Noise 216 32 168 32 48 4
Light 223 25 176 24 47 1
Mental 201 51 161 39 40 12
Physical 200 52 155 45 45 7
Coworkers 247 1 199 1 48 1
Supervisor 237 9 194 4 43 5
Caffeine 195 45 162 33 33 12
Condition1 34 214 30 170 4 44
Condition2 32 212 25 170 7 42

136
Chapter 5: Conclusion

The effort of researchers, industry leaders, and policy makers to make construction a safer

environment has produced good fruits. A 67% reduction in recordable incident rates has been

observed during the past 2 decades (U.S. Bureau of Labor Statistics 2015). However, the

construction industry is still among the top most dangerous industries with a fatality rate nearly 3

times greater than the overall industry average. Furthermore, fatalities have plateaued for the last

decade, and they have even increased in the past 3 years (U.S. Bureau of Labor Statistics 2015).

Therefore, there is still much to do regarding safety in construction.

Safety becomes even more critical among sectors inside the construction industry such as

Transmission and Distribution Electrical Workers (TD workers). These workers present fatality

rates disproportionally high even compared to the overall construction industry. For instance, the

average TD workers’ fatality rate during the past 20 years is approximately 3 times that of the

overall construction industry (U.S. Bureau of Labor Statistics 2015). These statistics situate TD

workers among the 10 most dangerous jobs in the country.

Fortunately, past research sheds light on the way to accomplish a safer construction

industry. Kartam (1997) demonstrated that accidents are caused either by unsafe human acts and/or

an unsafe design that generates physical hazards. Additionally, Langford et al. (2000) and Weick

et al. (2008) showed that organizational and human factors, rather than technical failures, are the

principal causes of accidents.

137
 A large body of literature shows the impact of fatigue on human error and accident

causation (Craig 1992; James Yaggie 2010; van der Linden et al. 2003a; Lorist et al. 2000, 2005;

Scott et al. 2006). Furthermore, some researchers have recognized the negative impact that fatigue

has in the construction industry, identifying it as a main threat to safety and productivity (Chan

2011; Hallowell 2010). Nonetheless, such approaches study the impact of fatigue as a whole, based

on the opinion of stakeholders and extrapolating results from studies conducted in other industries

or in a laboratory setting. As of now, to the knowledge of the author, no study addresses the causes

and consequences of fatigue in the construction industry.

Lastly, the current body of knowledge about fatigue is disorganized and disperse. The

apparently fuzzy nature of the phenomenon, its multidimensional manifestation, and its

commonality have contributed to the inexistence of a universal definition of fatigue. Consequently,

a plethora of diverse fatigue measurement and management techniques exists with mostly little to

no validation. Such conditions make the research of fatigue in the construction industry even more

challenging. Several authors have identified a need for an organized body of knowledge that allows

for a better understanding of fatigue causes and consequences in occupational settings (Reiner and

Krupinski 2011).

5.1 Contributions

The research documents in this dissertation offer: (1) an organized framework to study

occupational fatigue in the industry as a whole and, more importantly, in the construction industry;

(2) new knowledge regarding fatigue causes, consequences, identification and management

techniques for the construction industry. The contributions in knowledge here presented are of

138
special importance to safety researchers, safety and industry leaders in the construction industry,

and policy makers.

The first study presented in this dissertation formalizes the study of fatigue in working

populations by organizing the disperse occupational fatigue literature and offering a clear overall

understanding of the phenomenon and its dimensionality. The existing body of knowledge was

organized in a single system’s model allowing for an identification and easy visualization of the

principal causes and consequences of occupational fatigue. Additionally, such a model illustrates

the link between these factors and a specific dimension of fatigue, together with a representation

of the volume of research performed for every link. This knowledge will help future researchers

to identify where further investigation is needed in order to achieve a higher understanding of

fatigue. Furthermore, now researchers and practitioners can access the principal causes and

consequences of occupational fatigue through a single document without addressing the more than

120 studies contemplated in this work.

In the same study a meta-analysis was conducted to quantify the relative impact of fatigue

causes and consequences. Such knowledge allows for comparisons among these factors helping

the research community and the industry make sound decisions on resource allocations to further

knowledge about specific fatigue causes and consequences or to decide on mitigation strategies.

The second study represents the first attempt to explore the presence and impact of fatigue

causes and consequences among construction workers in the field. The data was collected directly

from 143 TD workers thus providing excellent ecological validity. This fact is worth mentioning

139
because past fatigue literature has addressed the issue from a more theoretical point rather than

with actual empirical data. In addition to recognizing the specific causes and consequences of

fatigue that affect TD workers, according to the opinion of the participants, this study also

summarizes the strategies that construction workers use to identify and mitigate fatigue in the

worksite. Furthermore, this study also offers recommendations on the first steps to manage fatigue

in the TD sector based on empirical evidence. These results can be of great help to industry leaders

and policy makers who look forward to an effective way of managing fatigue in the TD sector,

which would potentially contribute to fewer accidents and higher productivity, among other

benefits.

Lastly, the third study presented in this dissertation offers an even more relevant

advancement in knowledge regarding fatigue management in the construction industry. To

properly manage fatigue it is important to understand its causal factors and how to detect or predict

fatigue impairments among workers (Dawson and McCulloch 2005). This study offers an

empirical work, performed among 253 US construction workers belonging to 6 different states

across the country. This diverse and populated sample adds strength to the obtained results.

This last study allows for the identification of fatigue predictive variables among two

different types of construction workers: TD workers and GC workers. Furthermore, this study

demonstrates (1) that fatigue predictors can vary across different trades and (2) that not one

available instrument to measure fatigue is exhaustive. Consequently, when developing fatigue

predictive models in working populations, there is a tradeoff between the predictors identified by

instruments and the quality of the prediction. These results are especially important for safety

140
researchers in the construction industry. Now researchers have an example of an approach to the

development of a fatigue predictive model that they can apply to enhance the current models.

Furthermore, the results presented in this study show that fatigue related results cannot be

generalized across different construction trades. Figure 10 summarizes the contributions of this

dissertation by study.

• Organized previous research into a

systems model.

Study • Identification of fatigue causes and

consequences
•
1 •
Representation of volume of research
Quantification of relative impact of
causes and consequences of fatigue
through Meta-Analysis

• First empirical fatigue study among

TD workers.

Study • Identification of fatigue:

➢ Causes
➢ Consequences
2 ➢ Identification techniques
➢ Mitigation techniques

Study • Identification of fatigue predictors in

the construction industry.

3 • Recognition of differences in
predictors by trade.

Figure 10 Dissertation contributions

141
5.2 Suggestions for future research

The research performed throughout this dissertation informed the author of several

opportunities for advancement in knowledge in this field of research. In the opinion of the author,

the next steps in this field of research should encompass: (1) the empirical quantification of fatigue

impact on accident causation; (2) the study of fatigue management among different construction

trades;(3) measure the efficacy of Fatigue Risk Management Systems (FRMS).

5.2.1 Limitations of current data sets

Despite the rigorous methodology followed in the research processes here presented, the

generalization of findings to a broader population will require further studies. The researcher is

content with the large sample of TD workers that participated in these projects (n = 343). However,

only 53 general construction workers took part in this research. Consequently, the author is

confident in the generalization of most results to other TD workers in the US, given that the sample

here considered was sufficient, diverse, and the data collection process was performed over the

course of a whole year, thus allowing for the consideration of diverse weather. Furthermore, the

samples considered were predominantly male while few participants were female. Therefore, these

results, a priori, should not be generalized to a group of female construction workers.

There are additional limitations to the data related to the conditions in which data collection

took place. Construction is a dynamic and fast paced environment in which workers have very

limited time for interaction with researchers. Therefore, the circumstances in which data is

collected vary from one worksite to another. However, these characteristics contributed to

142
ecological validity. Furthermore, this same reasoning applies to all in-field studies in the

construction industry.

5.2.2 Future research based on findings

The author has shown that additional fatigue predictors need to be identified to achieve

higher levels of predictability. Given the lack of an exhaustive fatigue measurement tool, a

multidimensional approach to fatigue measurement will give researchers the opportunity to

identify new predictors of fatigue.

Along those lines, the author also showed that fatigue predictors vary across different

trades, consequently, the multidimensional approach described above should be accompanied by

the inclusion of several diverse trades because these may experience fatigue accumulation in

different ways. This procedure will also help the discovery of new predictors. There are 2 other

trades that should receive priority treatment: Roofers and Steel workers, which are among the 10

most dangerous jobs in the country. Interestingly, in the third study presented in this dissertation

(Chapter 4), TD workers showed a significantly lower level of fatigue than GC workers. This was

an unexpected result because the researcher would expect the opposite given the higher injury and

fatality rate of TD workers compared to GC workers. Consequently, the implementation of the

protocols here presented to different trades will shed light on the importance of fatigue in accident

causation.

Additionally, the discovery of new predictors will also require the consideration of

additional prospective predictors. It is the opinion of the author, that a possible way to develop a

list of new potential predictors, would consist of conducting exploratory research such as that

143
presented in the second study of this dissertation (Chapter 3). This pragmatic approach allows for

open ended questions that would contribute with additional variables of interest, not foreseen by

the researchers. At the same time these answers can be coded for rigorous statistical analyses.

5.2.3 Future research of fatigue in construction

Surprisingly, despite decades of occupational fatigue research, to the knowledge of the

author, there are no longitudinal studies that directly investigate the impact of fatigue on injuries

and fatalities. There is a need for longitudinal studies able to track fatigue levels and accident

reports to empirically quantify the impact of fatigue in accident causation in the construction

industry. The studies presented in this dissertation underline the importance of addressing this

issue and even provide fatigue predictors in the construction industry that could be used to select

the most appropriate measurement tools. However, none of these studies followed up on worker’s

performance, safety record, and fatigue levels overtime. Such information may or may not justify

further research in this area in the construction industry.

Perhaps, for this reason, policy makers and industry leaders have not taken measures to

directly manage fatigue other than shift length limitations. It is the opinion of the author that efforts

to validate the impact of fatigue on accident causation and productivity loss should be

accompanied by equal attempts to investigate the efficacy of FRMS in preventing accidents and

improving productivity. This not only would allow for a more comprehensive approach to the

formal study of fatigue in the construction industry, but this would also enable true experiments

with control groups, thus allowing for direct comparison. It is in the best interest of industry

leaders to sponsor the implementation of FRMS suited for the construction industry and

longitudinal studies to quantify the impact of FRMS in accident reduction and productivity

144
improvement. Then, a comparison between positive effects of fatigue management techniques and

negative effects of fatigue in construction safety and productivity will be possible.

Additionally, recent studies indicate that the construction industry is experiencing a marked

shortage of craft workers and this situation will become even more critical in the near future

(Karimi et al. 2017). As a direct result, the workload for the available workforce increases and

these workers are required to work longer and more frequent shifts. All of these factors contribute

to the onset of fatigue. Thus, the author concludes that fatigue management, in the construction

industry, will become even more relevant in the upcoming years. As it was shown in this

dissertation, despite general fatigue causal factors such as sleep deprivation, there are specific

causes of fatigue that vary across trades. Consequently, it becomes vital to determine these causes

as early as possible to properly implement FRMS crafted to each trade, thus increasing the

probability of success in fatigue management.

5.3 Personal Reflection

5.3.1 About research in the construction industry

It is my opinion, that in general, safety research in the construction industry has

traditionally relied on a constructivist approached based on qualitative data obtained from experts

or on the extrapolation of studies conducted in other populations such as students or other

industries. Most current researchers are advocating for studies with strong ecological validity in

which the data is obtained from the construction workers themselves. As a researcher, it is my

145
responsibility and privilege to join this effort to further formalize research in construction safety

by developing studies with strong ecological validity and a rigorous scientific approach.

Several improvements in ergonomics and working conditions have positively impacted

safety in the construction industry. It is my opinion and that of other researchers that human factors

play a major role in accident causation. The management of human factors requires an awareness

of their influence on safety and a conscious personal effort to control them, if possible. Such a

control over our human factors will require a change of behavior. I believe that education is the

main avenue to overcome safety risks associated with human behavior. The personal conscious

effort to modify our conduct may have an important impact in accident mitigation.

Finally, other human factors besides fatigue need to be considered to understand the impact

that human factors have on safety in the construction industry. Some of those human factors

include but are not limited to: emotions, stress, cultural differences, and personal biases. While

conducting experiments to assess the impact of these human factors preliminary studies could be

conducted in laboratory settings, however, these studies should also be validated in the field to

ensure good ecological validity.

5.3.2 About the journey to become a Ph.D.

The path to the completion of this dissertation has been one of constant enthusiasm,

challenges, and growth. I have felt a gradual internal change and now I can say that I am not the

same individual that I was 3 years ago when I started this adventure. I have incorporated several

skills that open a new world of opportunities. Humbly, I would like to express that such skills

146
make me feel limitless. Not because consider myself better than anybody else, but because these

skills have given me the opportunity to acquire knowledge in any area of my interest. I have

developed the ability to: (1) critically see the world and recognize specific needs for additional

knowledge in order to solve a specific problem or to gain a deeper understanding; (2) obtain,

analyze, and synthesize the existing body of knowledge about a topic of interest to discover the

many angles in which such a topic has already been observed and recognize those perspectives yet

to be seen; (3) plan a scientific approach to the discovery of new knowledge; (4) execute such a

plan; and (5) clearly present the obtained new knowledge for the benefit of my fellowmen. The

development of these skills is my biggest takeaway in my journey to become a Ph.D. I happily

recognize that I would have never been able to develop these skills without the constant and patient

guidance of my advisor, Dr. Matthew R. Hallowell, to whom I am forever grateful.

The journey was full of challenging and rewarding tasks. Perhaps the most difficult task

was the one of engaging industry members in my research. Each organization has its own interests

and schedules and it is a difficult task to align interests and even more difficult to align schedules.

However, the industry is full of professionals who want to make a difference in their companies,

the industry, and the community. Thanks to those individuals, I was able to collect data from a

significant number of construction workers which was very rewarding. This allowed for studies

with excellent ecological validity and that is fulfilling because I know that the results of these

studies have direct applicability to the industry. Additionally, one of the things that excites me the

most about my research is that those advancements in knowledge that we obtain can potentially

help save lives.

147
 It is natural, in the process of learning, to look back and realize that some things could have

been done better, not because I was not careful enough, but simply because I didn’t have the

knowledge that I have now. In preparation to my comprehensive exam, I developed a research

plan, a protocol for its implementation, and documents to collect data. All of these were carefully

designed, however, mistakes still occurred. It is not relevant to mention the specific mistakes;

however, it may be enriching to share what I learned from those mistakes. I learned that despite

having a very good research plan, it is important to contemplate all the possible outcomes and plan

on a course of action for each possible alternative. By so doing, the researcher will always be able

to know what to do in order to achieve the desired objective despite changing circumstances.

Throughout my Ph.D. studies I learned some things that helped me complete my Ph.D.

goals and may be of use to future students. First, I found it very helpful to spend enough time on

a literature review of your topic of interest. In my case, my first study consisted in organizing the

current body of knowledge about fatigue and for that reason I spent 6 months on a deep literature

review and 6 more months working on a meta-analysis of such literature. After this, it was very

clear what things were missing in the body of knowledge and where further research was needed.

Probably most students will be able to reduce the amount of time if their field of interest possess

a more organized body of literature. Second, once you have a tight plan to accomplish your

research goals, present your plan to several people and learn from their feedback. I presented my

plan to many professors, students, and industry professionals that were helping me at the time. The

feedback that they provided was very enriching and allowed me to take better advantage of the

data collection process. Lastly, I would advise any student to expect changes in their research plan,

determine a plan of action for each foreseeable deviation from the plan and implement such a plan

148
of action early. This will allow students to know exactly when to stop collecting data given their

research objectives. These are simple tips but they helped me and they may help other students.

5.4 References

Chan, M. (2011). “Fatigue: the most critical accident risk in oil and gas construction.”
Construction Management and Economics, 29(4), 341–353.
Craig, A. (1992). “Symptoms of acute and chronic fatigue.” Handbook of human performance,
Vol. 1: The physical environment; Vol. 2: Health and performance; Vol. 3: State and trait,
A. P. Smith and D. M. Jones, eds., Academic Press, San Diego, CA, US, 289–339.
Dawson, D., and McCulloch, K. (2005). “Managing fatigue: It’s about sleep.” Sleep Medicine
Reviews, 9(5), 365–380.
Hallowell, M. R. (2010). “Worker fatigue: managing concerns in rapid renewal highway
construction projects.” Professional Safety, 55(12), 18–26.
James Yaggie, W. J. A. (2010). “Effects of Lower Extremity Fatigue on Indices of Balance.”
Human Kinetics Journals, <http://journals.humankinetics.com/jsr-back-
issues/jsrvolume13issue4november/effectsoflowerextremityfatigueonindicesofbalance>
(Aug. 10, 2015).
Karimi, H., Taylor, T. R. B., and Goodrum, P. M. (2017). “Analysis of the impact of craft labour
availability on North American construction project productivity and schedule
performance.” Construction Management and Economics, 0(0), 1–13.
Kartam, N. A. (1997). “Integrating Safety and Health Performance into Construction CPM.”
Journal of Construction Engineering and Management, 123(2), 121–126.
Langford, D., Rowlinson, S., and Sawacha, E. (2000). “Safety behaviour and safety management:
its influence on the attitudes of workers in the UK construction industry.” Engineering
Construction and Architectural Management, 7(2), 133–140.
van der Linden, D., Frese, M., and Meijman, T. F. (2003). “Mental fatigue and the control of
cognitive processes: effects on perseveration and planning.” Acta Psychologica, 113(1),
45–65.
Lorist, M. M., Boksem, M. A., and Ridderinkhof, K. R. (2005). “Impaired cognitive control and
reduced cingulate activity during mental fatigue.” Cognitive Brain Research, 24(2), 199–
205.
Lorist, M. M., Klein, M., Nieuwenhuis, S., de Jong, R., Mulder, G., and Meijman, T. F. (2000).
“Mental fatigue and task control: Planning and preparation.” Psychophysiology, 37(5),
614–625.
Reiner, B. I., and Krupinski, E. (2011). “Demystifying Occupational Stress and Fatigue Through
the Creation of an Adaptive End-User Profiling System.” Journal of Digital Imaging,
25(2), 201–205.
Scott, J. P., McNaughton, L. R., and Polman, R. C. (2006). “Effects of sleep deprivation and
exercise on cognitive, motor performance and mood.” Physiology & behavior, 87(2), 396–
408.
Techera, U., Hallowell, M., and Littlejohn, R. (2017). “Fatigue Management for Electrical
Transmission and Distribution Construction” Journal of Construction Engineering and
Management, (Planned Submission: April 2017).

149
U.S. Bureau of Labor Statistics. (2015). “Census of Fatal Occupational Injuries Summary, 2014.”
<http://stats.bls.gov/news.release/cfoi.nr0.htm>.
Weick, K. E., Sutcliffe, K. M., and Obstfeld, D. (2008). “Organizing for high reliability: Processes
of collective mindfulness.” Crisis management, 3(1), 81–123

150
Bibliography

Åhsberg, E. (2000). “Dimensions of fatigue in different working populations.” Scandinavian Journal

of Psychology, 41(3), 231–241.
Åhsberg, E., and Gamberale, F. (1998). “Perceived fatigue during physical work: an experimental
evaluation of a fatigue inventory.” International Journal of Industrial Ergonomics, 21(2), 117–
131.
Åhsberg, E., Gamberale, F., and Gustafsson, K. (2000a). “Perceived fatigue after mental work: an
experimental evaluation of a fatigue inventory.” Ergonomics, 43(2), 252–268.
Åhsberg, E., Garnberale, F., and Kjellberg, A. (1997). “Perceived quality of fatigue during different
occupational tasks Development of a questionnaire.” International Journal of Industrial
Ergonomics, 20(2), 121–135.
Åhsberg, E., Kecklund, G., Åkerstedt, T., and Gamberale, F. (2000b). “Shiftwork and different
dimensions of fatigue.” International Journal of Industrial Ergonomics, 26(4), 457–465.
Åkerstedt, T. (1990). “Psychological and psychophysiological effects of shift work.” Scandinavian
Journal of Work, Environment & Health, 16, 67–73.
Åkerstedt, T., Fredlund, P., Gillberg, M., and Jansson, B. (2002a). “A prospective study of fatal
occupational accidents–relationship to sleeping difficulties and occupational factors.” Journal
of Sleep Research, 11(1), 69–71.
Åkerstedt, T., Fredlund, P., Gillberg, M., and Jansson, B. (2002b). “Work load and work hours in
relation to disturbed sleep and fatigue in a large representative sample.” Journal of
Psychosomatic Research, 53(1), 585–588.
Ala-Mursula, L., Vahtera, J., Linna, A., Pentti, J., and Kivimäki, M. (2005). “Employee worktime
control moderates the effects of job strain and effort-reward imbalance on sickness absence:
the 10-town study.” Journal of Epidemiology and Community Health, 59(10), 851–857.
Albert, A., Hallowell, M. R., and Kleiner, B. M. (2014). “Experimental field testing of a real-time
construction hazard identification and transmission technique.” Construction Management and
Economics, 32(10), 1000–1016.
Alfredsson, L., Karasek, R., and Theorell, T. (1982). “Myocardial infarction risk and psychosocial
work environment: An analysis of the male Swedish working force.” Social Science &
Medicine, 16(4), 463–467.
Amick III, B. C., McDonough, P., Chang, H., Rogers, W. H., Pieper, C. F., and Duncan, G. (2002).
“Relationship between all-cause mortality and cumulative working life course psychosocial
and physical exposures in the United States labor market from 1968 to 1992.” Psychosomatic
Medicine, 64(3), 370–381.
Angus, R. G., and Heslegrave, R. J. (1985). “Effects of sleep loss on sustained cognitive performance
during a command and control simulation.” Behavior Research Methods, Instruments, &
Computers, 17(1), 55–67.
Arnaldo, R. M., Comendador, F. G., Sanz, L. P., and Sanz, A. R. (2016). “Friendly Fatigue Alert
Mobile Apps to Help Aviation Workers Prevent, Identify and Manage Alertness and Fatigue.”
Advances in Safety Management and Human Factors, Advances in Intelligent Systems and
Computing, P. Arezes, ed., Springer International Publishing, 421–432.
Axelsson, J., Kecklund, G., \AAkerstedt, T., and Lowden, A. (1998). “Effects of alternating 8-and 12-
hour shifts on sleep, sleepiness, physical effort and performance.” Scandinavian journal of
work, environment & health, 62–68.
Balducci, P. J., Weimar, M. R., DeSteese, J. G., Schienbein, L. A., and Roop, J. M. (2003). “Electric
Power Interruption Cost Estimates for Individual Industries, Sectors, and the U.S. Economy.”

151
 Proceedings from the Twenty-Fifth Industrial Energy Technology Conference, Texas A & M
University, Houston,TX., 109–118.
Balkin, T. J., Bliese, P. D., Belenky, G., Sing, H., Thorne, D. R., Thomas, M., Redmond, D. P., Russo,
M., and Wesensten, N. J. (2004). “Comparative utility of instruments for monitoring
sleepiness-related performance decrements in the operational environment.” Journal of Sleep
Research, 13(3), 219–227.
Baulk, S. D., Fletcher, A., Kandelaars, K. J., Dawson, D., and Roach, G. D. (2009). “A field study of
sleep and fatigue in a regular rotating 12-h shift system.” Applied ergonomics, 40(4), 694–698.
Beckers, D. G. J., Linden, D. van der, Smulders, P. G. W., Kompier, M. A. J., Taris, T. W., and Geurts,
S. A. E. (2008). “Voluntary or involuntary? Control over overtime and rewards for overtime
in relation to fatigue and work satisfaction.” Work & Stress, 22(1), 33–50.
Belenky, G., Wesensten, N. J., Thorne, D. R., Thomas, M. L., Sing, H. C., Redmond, D. P., Russo, M.
B., and Balkin, T. J. (2003). “Patterns of performance degradation and restoration during sleep
restriction and subsequent recovery: a sleep dose-response study.” Journal of Sleep Research,
12(1), 1–12.
Beurskens, A. J. H. M., Bültmann, U., Kant, Ij., Vercoulen, J. H. M. M., Bleijenberg, G., and Swaen,
G. M. H. (2000). “Fatigue among working people: validity of a questionnaire measure.”
Occupational and Environmental Medicine, 57(5), 353–357.
Bliwise, D. L., King, A. C., Harris, R. B., and Haskell, W. L. (1992). “Prevalence of self-reported poor
sleep in a healthy population aged 50–65.” Social Science & Medicine, 34(1), 49–55.
Bøggild, H., and Knutsson, A. (1999). “Shift work, risk factors and cardiovascular disease.”
Scandinavian Journal of Work, Environment & Health, 25(2), 85–99.
Boksem, M. A., Meijman, T. F., and Lorist, M. M. (2006). “Mental fatigue, motivation and action
monitoring.” Biological psychology, 72(2), 123–132.
Boksem, M. A. S., Meijman, T. F., and Lorist, M. M. (2005). “Effects of mental fatigue on attention:
An ERP study.” Cognitive Brain Research, 25(1), 107–116.
Boksem, M. A. S., and Tops, M. (2008). “Mental fatigue: Costs and benefits.” Brain Research Reviews,
59(1), 125–139.
Boksem, M. A. S., Tops, M., Kostermans, E., and De Cremer, D. (2008). “Sensitivity to punishment
and reward omission: Evidence from error-related ERP components.” Biological Psychology,
79(2), 185–192.
Boonstra, T. W., Stins, J. F., Daffertshofer, A., and Beek, P. J. (2007). “Effects of sleep deprivation on
neural functioning: an integrative review.” Cellular and Molecular Life Sciences, 64(7–8),
934–946.
Borg, G. (1982). “A category scale with ratio properties for intermodal and interindividual
comparisons.” Psychophysical judgment and the process of perception, 25–34.
Bravo, G., and Potvin, L. (1991). “Estimating the reliability of continuous measures with cronbach’s
alpha or the intraclass correlation coefficient: Toward the integration of two traditions.”
Journal of Clinical Epidemiology, 44(4), 381–390.
Brooks, G. P., and Barcikowski, R. S. (1995). “Precision Power Method for Selecting Regression
Sample Sizes.”
Brown, I. D. (1994). “Driver fatigue.” Human Factors: The Journal of the Human Factors and
Ergonomics Society, 36(2), 298–314.
Buchholz, B., Paquet, V., Punnett, L., Lee, D., and Moir, S. (1996). “PATH: A work sampling-based
approach to ergonomic job analysis for construction and other non-repetitive work.” Applied
Ergonomics, 27(3), 177–187.
Bültmann, U., Kant, I., Schröer, C., and Kasl, S. (2001). “The relationship between psychosocial work
characteristics and fatigue and psychological distress.” International Archives of Occupational
and Environmental Health, 75(4), 259–266.
152
Bültmann, U., de Vries, M., M, J. H., Bleijenberg, G., M, H. M., and Kant, Ij. (2000). “Measurement
of prolonged fatigue in the working population: Determination of a cutoff point for the
Checklist Individual Strength.” Journal of Occupational Health Psychology, 5(4), 411–416.
Cadick, J., Capelli-Schellpfeffer, M., and Neitzel, D. (2005). Electrical safety handbook. McGraw-
Hill:, New York, Chicago, San Francisco, Lisbon, London, Madrid, Mexico City, Milan, New
Delhi, San Juan, Seoul, Singapore, Sydney, Toronto.
Chan, M. (2011). “Fatigue: the most critical accident risk in oil and gas construction.” Construction
Management and Economics, 29(4), 341–353.
Chervin, R. D. (2000). “Sleepiness, Fatigue, Tiredness, and Lack of Energy in Obstructive Sleep
Apnea.” Chest, 118(2), 372–379.
Christensen, H., Søgaard, K., Jensen, B. R., Finsen, L., and Sjøgaard, G. (1995). “Intramuscular and
surface EMG power spectrum from dynamic and static contractions.” Journal of
Electromyography and Kinesiology, 5(1), 27–36.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences. Lawrence earbaum Assuati.
New York.
Craig, A. (1992). “Symptoms of acute and chronic fatigue.” Handbook of human performance, Vol. 1:
The physical environment; Vol. 2: Health and performance; Vol. 3: State and trait, A. P. Smith
and D. M. Jones, eds., Academic Press, San Diego, CA, US, 289–339.
Cronbach, L. J. (1951). “Coefficient alpha and the internal structure of tests.” Psychometrika, 16(3),
297–334.
Czeisler, C. A., Dumont, M., Duffy, J. F., Steinberg, J. D., Richardson, G. S., Brown, E. N., Sánchez,
R., Ríos, C. D., and Ronda, J. M. (1992). “Association of sleep-wake habits in older people
with changes in output of circadian pacemaker.” The Lancet, Originally published as Volume
2, Issue 8825, 340(8825), 933–936.
Dainoff, M. J., Happ, A., and Crane, P. (1981). “Visual Fatigue and Occupational Stress in VDT
Operators.” Human Factors: The Journal of the Human Factors and Ergonomics Society,
23(4), 421–437.
Darlington, R. B. (1990). Regression and linear models. McGraw-Hill New York.
Dawson, D., Chapman, J., and Thomas, M. J. W. (2012). “Fatigue-proofing: A new approach to
reducing fatigue-related risk using the principles of error management.” Sleep Medicine
Reviews, 16(2), 167–175.
Dawson, D., and Fletcher, A. (2001). “A quantitative model of work-related fatigue: background and
definition.” Ergonomics, 44(2), 144–163.
Dawson, D., and McCulloch, K. (2005). “Managing fatigue: It’s about sleep.” Sleep Medicine Reviews,
9(5), 365–380.
Dawson, D., Searle, A. K., and Paterson, J. L. (2014). “Look before you (s)leep: Evaluating the use of
fatigue detection technologies within a fatigue risk management system for the road transport
industry.” Sleep Medicine Reviews, 18(2), 141–152.
Dembe, A. E., Erickson, J. B., Delbos, R. G., and Banks, S. M. (2005). “The impact of overtime and
long work hours on occupational injuries and illnesses: new evidence from the United States.”
Occupational and Environmental Medicine, 62(9), 588–597.
Dinges, D. F., Pack, F., Williams, K., Gillen, K. A., Powell, J. W., Ott, G. E., Aptowicz, C., and Pack,
A. I. (1997). “Cumulative sleepiness, mood disturbance and psychomotor vigilance
performance decrements during aweek of sleep restricted to 4-5 hours per night.” Sleep:
Journal of Sleep Research & Sleep Medicine, 20(4), 267–277.
Dinges, D. F., and Powell, J. W. (1985). “Microcomputer analyses of performance on a portable,
simple visual RT task during sustained operations.” Behavior Research Methods, Instruments,
& Computers, 17(6), 652–655.

153
Dorrian, J., Lamond, N., Kozuchowski, K., and Dawson, D. (2008). “The driver vigilance telemetric
control system (DVTCS): Investigating sensitivity to experimentally induced sleep loss and
fatigue.” Behavior Research Methods, 40(4), 1016–1025.
Dorrian, J., Rogers, N. L., Dinges, D. F., and others. (2005). “Psychomotor vigilance performance:
Neurocognitive assay sensitive to sleep loss.” Marcel Dekker New York.
Edwards, R. G., and Lippold, O. C. J. (1956). “The relation between force and integrated electrical
activity in fatigued muscle.” The Journal of Physiology, 132(3), 677–681.
Edwards, R. H., Hill, D. K., and Jones, D. A. (1975). “Heat production and chemical changes during
isometric contractions of the human quadriceps muscle.” The Journal of Physiology, 251(2),
303–315.
Ferguson, S. A., Paech, G. M., Dorrian, J., Roach, G. D., and Jay, S. M. (2011). “Performance on a
simple response time task: Is sleep or work more important for miners?” Applied Ergonomics,
Special Section: Ergonomics, health and working time organization, 42(2), 210–213.
Field, A. P., and Gillett, R. (2010). “How to do a meta-analysis.” British Journal of Mathematical and
Statistical Psychology, 63(3), 665–694.
Fogt, D. L., Kalns, J. E., and Michael, D. J. (2010). “A Comparison of Cognitive Performance
Decreases During Acute, Progressive Fatigue Arising From Different Concurrent Stressors.”
Military Medicine, 175(12), 939–944.
Folkard, S., and Tucker, P. (2003). “Shift work, safety and productivity.” Occupational medicine,
53(2), 95–101.
Fowler, Jr., F. J., and Mangione, T. W. (1990). Standardized Survey Interviewing: Minimizing
Interviewer-Related Error. SAGE.
Gander, P., Hartley, L., Powell, D., Cabon, P., Hitchcock, E., Mills, A., and Popkin, S. (2011). “Fatigue
risk management: Organizational factors at the regulatory and industry/company level.”
Accident Analysis & Prevention, Advancing Fatigue and Safety Research, 43(2), 573–590.
Gibson, A. S. C., Baden, D. A., Lambert, M. I., Lambert, E. V., Harley, Y. X. R., Hampson, D., Russell,
V. A., and Noakes, T. D. (2003). “The Conscious Perception of the Sensation of Fatigue.”
Sports Medicine, 33(3), 167–176.
Gillberg, M., and Åkerstedt, T. (1994). “Sleep restriction and SWS-suppression: effects on daytime
alertness and night-time recovery.” Journal of Sleep Research, 3(3), 144–151.
Goldich, Y., Barkana, Y., Pras, E., Zadok, D., Hartstein, M., and Morad, Y. (2010). “The Effects of
Sleep Deprivation on Oculomotor Responses.” Current Eye Research, 35(12), 1135–1141.
González Gutiérrez, J. L., Jiménez, B. M., Hernández, E. G., and López López, A. (2005). “Spanish
version of the Swedish Occupational Fatigue Inventory (SOFI): Factorial replication, reliability
and validity.” International Journal of Industrial Ergonomics, 35(8), 737–746.
González-Alonso, J., Teller, C., Andersen, S. L., Jensen, F. B., Hyldig, T., and Nielsen, B. (1999).
“Influence of body temperature on the development of fatigue during prolonged exercise in the
heat.” Journal of Applied Physiology, 86(3), 1032–1039.
Gordon, L. B., and Cartelli, L. (2009). “A complete electrical hazard classification system and its
application.” Electrical Safety Workshop, 2009. IEEE IAS, 1–12.
Gosselin, A., De Koninck, J., and Campbell, K. B. (2005). “Total sleep deprivation and novelty
processing: implications for frontal lobe functioning.” Clinical Neurophysiology, 116(1), 211–
222.
Grandjean, E. (1979). “Fatigue in industry.” British Journal of Industrial Medicine, 36(3), 175–186.
Greenacre, M. (2007). Correspondence Analysis in Practice, Second Edition. CRC Press.
Gross, A. L. (1973). “How large should sample size be in a regression analysis?” Proceedings of the
Annual Convention of the American Psychological Association, 17–18.
Hallowell, M., Esmaeili, B., and Chinowsky, P. (2011). “Safety risk interactions among highway
construction work tasks.” Construction Management and Economics, 29(4), 417–429.
154
Hallowell, M. R. (2010). “Worker fatigue: managing concerns in rapid renewal highway construction
projects.” Professional Safety, 55(12), 18–26.
Hansen, J. (2001). “Increased Breast Cancer Risk among Women Who Work Predominantly at Night.”
Epidemiology, 12(1), 74–77.
Hardy, G. E., Shapiro, D. A., and Borrill, C. S. (1997). “Fatigue in the workforce of national health
service trusts: Levels of symptomatology and links with minor psychiatric disorder,
demographic, occupational and work role factors.” Journal of Psychosomatic Research,
Workplace, 43(1), 83–92.
Harrington, J. M. (2001). “Health effects of shift work and extended hours of work.” Occupational
and Environmental Medicine, 58(1), 68–72.
Harrison, Y., and Horne, J. A. (1999). “One Night of Sleep Loss Impairs Innovative Thinking and
Flexible Decision Making.” Organizational Behavior and Human Decision Processes, 78(2),
128–145.
Harrison, Y., and Horne, J. A. (2000a). “The impact of sleep deprivation on decision making: A
review.” Journal of Experimental Psychology: Applied, 6(3), 236–249.
Harrison, Y., and Horne, J. A. (2000b). “The impact of sleep deprivation on decision making: A
review.” Journal of Experimental Psychology: Applied, 6(3), 236–249.
Hartley, L. R., and Arnold, P. (2001). “Managing fatigue in the road transport industry: an occupational
safety and health solution.” Stress, workload and fatigue.
Hedges, L. V. (1981). “Distribution Theory for Glass’s Estimator of Effect size and Related
Estimators.” Journal of Educational and Behavioral Statistics, 6(2), 107–128.
Herscovitch, J., and Broughton, R. (1981). “Performance deficits following short-term partial sleep
deprivation and subsequent recovery oversleeping.” Canadian Journal of Psychology/Revue
Canadienne de psychologie, 35(4), 309.
Hsiao, H., and Simeonov, P. (2001). “Preventing falls from roofs: a critical review.” Ergonomics,
44(5), 537–561.
van der Hulst, M., and Geurts, S. (2001). “Associations between overtime and psychological health in
high and low reward jobs.” Work and Stress, 15(3), 227–240.
van der Hulst, M., Veldhoven, M. van, and Beckers, D. (2006). “Overtime and Need for Recovery in
Relation to Job Demands and Job Control.” Journal of Occupational Health, 48(1), 11–19.
Hultsch, D. F., MacDonald, S. W. S., and Dixon, R. A. (2002). “Variability in Reaction Time
Performance of Younger and Older Adults.” The Journals of Gerontology: Series B, 57(2),
P101–P115.
Iwasaki, K., Sasaki, T., Oka, T., and Hisanaga, N. (1998). “Effect of Working Hours on Biological
Functions related to Cardiovascular System among Salesmen in a Machinery Manufacturing
Company.” Industrial Health, 36(4), 361–367.
James Yaggie, W. J. A. (2010). “Effects of Lower Extremity Fatigue on Indices of Balance.” Human
Kinetics Journals, <http://journals.humankinetics.com/jsr-back-
issues/jsrvolume13issue4november/effectsoflowerextremityfatigueonindicesofbalance>
(Aug. 10, 2015).
Jason, L. A., Evans, M., Brown, M., and Porter, N. (2010). “What is fatigue? Pathological and
nonpathological fatigue.” PM&R, 2(5), 327–331.
Jiao, K., Li, Z., Chen, M., Wang, C., and Qi, S. (2004). “Effect of different vibration frequencies on
heart rate variability and driving fatigue in healthy drivers.” International Archives of
Occupational and Environmental Health, 77(3), 205–212.
Johnston, M., Pollard, B., Manktelow, A., and others. (1989). “Stress and information processing in 8
and 12-hour nursing shifts.” The pychologist.

155
Johnston, R., Howard, M., Cawley, and Posse, G. (1998). “Effect of lower extremity muscular fatigue
on motor control performance.” Medicine & Science in Sports & Exercise®, 30(12), 1703–
1707.
Jones, K., and Harrison, Y. (2001). “Frontal lobe function, sleep loss and fragmented sleep.” Sleep
Medicine Reviews, 5(6), 463–475.
Jovanović, M., Sporiš, G., Šopar, J., Harasin, D., and Matika, D. (2012). “The Effects of Basic Military
Training on Shooting Tasks in Conditions of Sleep Deprivation.” Kinesiology, 44(1), 169–177.
Kajtna, T., Štukovnik, V., and Grošelj, L. D. (2011). “Effect of acute sleep deprivation on concentration
and mood states with a controlled effect of experienced stress.” Zdravniški Vestnik, 80(5).
Karimi, H., Taylor, T. R. B., and Goodrum, P. M. (2017). “Analysis of the impact of craft labour
availability on North American construction project productivity and schedule performance.”
Construction Management and Economics, 0(0), 1–13.
Kartam, N. A. (1997). “Integrating Safety and Health Performance into Construction CPM.” Journal
of Construction Engineering and Management, 123(2), 121–126.
Kerlinger, F. N., and Pedhazur, E. J. (1973). “Multiple Regression in Behavioural Research. New
York: Holt, Rinehart and Winston.” Inc., l973.
Khitrov, M. Y., Laxminarayan, S., Thorsley, D., Ramakrishnan, S., Rajaraman, S., Wesensten, N. J.,
and Reifman, J. (2013). “PC-PVT: A platform for psychomotor vigilance task testing, analysis,
and prediction.” Behavior Research Methods, 46(1), 140–147.
Kivimäki, M., Leino-Arjas, P., Luukkonen, R., Riihimäi, H., Vahtera, J., and Kirjonen, J. (2002).
“Work stress and risk of cardiovascular mortality: prospective cohort study of industrial
employees.” BMJ, 325(7369), 857.
Kjellberg, A., Muhr, P., and Skoldstrom, B. (1998). “Fatigue after work in noise - an epidemiological
survey study and three quasi-experimental field studies.” Noise & Health, 1(1), 47–55.
Knofczynski, G. T., and Mundfrom, D. (2007). “Sample Sizes When Using Multiple Linear Regression
for Prediction.” Educational and Psychological Measurement.
Knutsson, A. (2003). “Health disorders of shift workers.” Occupational Medicine, 53(2), 103–108.
Kosinski, R. J. (2008). “A literature review on reaction time.” Clemson University, 10.
Krause, N., Lynch, J., Kaplan, G. A., Cohen, R. D., Goldberg, D. E., and Salonen, J. T. (1997).
“Predictors of disability retirement.” Scandinavian Journal of Work, Environment & Health,
23(6), 403–413.
Kroenke K, and Price RK. (1993). “Symptoms in the community: Prevalence, classification, and
psychiatric comorbidity.” Archives of Internal Medicine, 153(21), 2474–2480.
Lal, S. K. L., and Craig, A. (2001). “A critical review of the psychophysiology of driver fatigue.”
Biological Psychology, 55(3), 173–194.
Lal, S. K. L., Craig, A., Boord, P., Kirkup, L., and Nguyen, H. (2003). “Development of an algorithm
for an EEG-based driver fatigue countermeasure.” Journal of Safety Research, 34(3), 321–328.
Lamond, N., Dawson, D., and Roach, G. D. (2005). “Fatigue Assessment in the Field: Validation of a
Hand-Held Electronic Psychomotor Vigilance Task.” Aviation, Space, and Environmental
Medicine, 76(5), 486–489.
Landström, U., and Löfstedt, P. (1987). “Noise, vibration and changes in wakefulness during helicopter
flight.” Aviation, Space, and Environmental Medicine, 58(2), 109–118.
Langford, D., Rowlinson, S., and Sawacha, E. (2000). “Safety behaviour and safety management: its
influence on the attitudes of workers in the UK construction industry.” Engineering
Construction and Architectural Management, 7(2), 133–140.
Lerman, S. E., Eskin, E., Flower, D. J., George, E. C., Gerson, B., Hartenbaum, N., Hursh, S. R., and
Moore-Ede, M. (2012). “Fatigue Risk Management in the Workplace:” Journal of
Occupational and Environmental Medicine, 54(2), 231–258.

156
Leung, A. W. S., Chan, C. C. H., and He, J. (2004). “Structural stability and reliability of the Swedish
occupational fatigue inventory among Chinese VDT workers.” Applied Ergonomics, The
Occlusion Technique, 35(3), 233–241.
Lim, J., and Dinges, D. F. (2008). “Sleep Deprivation and Vigilant Attention.” Annals of the New York
Academy of Sciences, 1129(1), 305–322.
van der Linden, D., and Eling, P. (2006). “Mental fatigue disturbs local processing more than global
processing.” Psychological Research, 70(5), 395–402.
van der Linden, D., Frese, M., and Meijman, T. F. (2003a). “Mental fatigue and the control of cognitive
processes: effects on perseveration and planning.” Acta Psychologica, 113(1), 45–65.
van der Linden, D. van der, Frese, M., and Sonnentag, S. (2003b). “The Impact of Mental Fatigue on
Exploration in a Complex Computer Task: Rigidity and Loss of Systematic Strategies.” Human
Factors: The Journal of the Human Factors and Ergonomics Society, 45(3), 483–494.
Lipscomb, J. A., Trinkoff, A. M., Geiger-Brown, J., and Brady, B. (2002). “Work-schedule
characteristics and reported musculoskeletal disorders of registered nurses.” Scandinavian
journal of work, environment & health, 394–401.
Lipsey, M. W., and Wilson, D. B. (2001). Practical meta-analysis. Applied social research methods
series; vol 49., Sage Publications, Inc, Thousand Oaks, CA, US.
Liu, Y., and Tanaka, H. (2002). “Overtime work, insufficient sleep, and risk of non-fatal acute
myocardial infarction in Japanese men.” Occupational and Environmental Medicine, 59(7),
447–451.
Loh, S., Lamond, N., Dorrian, J., Roach, G., and Dawson, D. (2004). “The validity of psychomotor
vigilance tasks of less than 10-minute duration.” Behavior Research Methods, Instruments, &
Computers, 36(2), 339–346.
Lorist, M. M., Boksem, M. A., and Ridderinkhof, K. R. (2005). “Impaired cognitive control and
reduced cingulate activity during mental fatigue.” Cognitive Brain Research, 24(2), 199–205.
Lorist, M. M., Klein, M., Nieuwenhuis, S., de Jong, R., Mulder, G., and Meijman, T. F. (2000). “Mental
fatigue and task control: Planning and preparation.” Psychophysiology, 37(5), 614–625.
Lubke, G. H., and Muthén, B. O. (2004). “Applying Multigroup Confirmatory Factor Models for
Continuous Outcomes to Likert Scale Data Complicates Meaningful Group Comparisons.”
Structural Equation Modeling: A Multidisciplinary Journal, 11(4), 514–534.
Marcus, J. H., and Rosekind, M. R. (2016). “Fatigue in transportation: NTSB investigations and safety
recommendations.” Injury Prevention, injuryprev-2015-041791.
Maslach, C., and Jackson, S. E. (1984). “Burnout in organizational settings.” Applied social psychology
annual.
May, J. F., and Baldwin, C. L. (2009). “Driver fatigue: The importance of identifying causal factors of
fatigue when considering detection and countermeasure technologies.” Transportation
Research Part F: Traffic Psychology and Behaviour, 12(3), 218–224.
McNair, D. M., Lorr, M., and Droppleman, L. F. (1971). Profile of mood states. Univ.
Meijman, T. F. (1997). “Mental fatigue and the efficiency of information processing in relation to work
times.” International Journal of Industrial Ergonomics, 20(1), 31–38.
Meijman, T., and Schaufeli, W. (1996). “Psychische vermoeidheid en arbeid. Ontwikkelingen in de
A&O-psychologie.” Psycholoog-Amsterdam-, 31, 236–241.
Michielsen, H. J., De Vries, J., Van Heck, G. L., R, J., and Sijtsma, K. (2004). “Examination of the
Dimensionality of Fatigue: The Construction of the Fatigue Assessment Scale (FAS).”
European Journal of Psychological Assessment, 20(1), 39–48.
Mikulincer, M., Babkoff, H., Caspy, T., and Sing, H. (1989). “The effects of 72 hours of sleep loss on
psychological variables.” British Journal of Psychology, 80(2), 145–162.
Miller, D. E., and Kunce, J. T. (1973). “Prediction and Statistical Overkill Revisited.” Measurement
and Evaluation in Guidance.
157
Mota, D. D. C. F., and Pimenta, C. A. M. (2006). “Self-Report Instruments for Fatigue Assessment: A
Systematic Review.” Research and Theory for Nursing Practice, 20(1), 49–78.
Naitoh, P. (1981). “Circadian cycles and restorative power of naps.” Biological rhythms, sleep and
shift work, 553–580.
Nilsson, J. P., Söderström, M., Karlsson, A. U., Lekander, M., \AAkerstedt, T., Lindroth, N. E., and
Axelsson, J. (2005). “Less effective executive functioning after one night’s sleep deprivation.”
Journal of sleep research, 14(1), 1–6.
Nindl, B. C., Leone, C. D., Tharion, W. J., Johnson, R. F., Castellani, J. W., Patton, J. F., and Montain,
S. J. (2002). “Physical performance responses during 72 h of military operational stress.”
Medicine & Science in Sports & Exercise, 34(11), 1814–1822.
Nunnally, J. (1978). Psychometric methods. New York: McGraw-Hill.
Nylén, L., Voss, M., and Floderus, B. (2001). “Mortality among women and men relative to
unemployment, part time work, overtime work, and extra work: a study based on data from the
Swedish twin registry.” Occupational and Environmental Medicine, 58(1), 52–57.
Okogbaa, O. G., Shell, R. L., and Filipusic, D. (1994). “On the investigation of the neurophysiological
correlates of knowledge worker mental fatigue using the EEG signal.” Applied Ergonomics,
25(6), 355–365.
Opstad, P. K., Ekanger, R., Nummestad, M., and Raabe, N. (1978). “Performance, mood, and clinical
symptoms in men exposed to prolonged, severe physical work and sleep deprivation.” Aviation,
Space, and Environmental Medicine, 49, 1065–1073.
Park, J., Kim, Y., Chung, H. K., and Hisanaga, N. (2001). “Long working hours and subjective fatigue
symptoms.” Industrial health, 39(3), 250–254.
Park, S., and Gotoh, S. (1993). “Research on office environment evaluation and worker fatigue.”
Journal of Thermal Biology, 18(5–6), 565–570.
Pedhazur, E. (1997). Multiple Regression in Behavioral Research: Explanation and Prediction. Holt,
Rinehart and Winston.
Pedhazur, E. J., and Schmelkin, L. P. (2013). Measurement, Design, and Analysis: An Integrated
Approach. Psychology Press.
Philibert, I. (2005). “Sleep loss and performance in residents and nonphysicians: a meta-analytic
examination.” SLEEP-NEW YORK THEN WESTCHESTER-, 28(11), 1392.
Pilcher, J. J., Band, D., Odle-Dusseau, H. N., and Muth, E. R. (2007). “Human Performance Under
Sustained Operations and Acute Sleep Deprivation Conditions: Toward a Model of Controlled
Attention.” Aviation, Space, and Environmental Medicine, 78(5), B15–B24.
Pilcher, J. J., and Huffcutt, A. J. (1996). “Effects of sleep deprivation on performance: a meta-
analysis.” Journal of Sleep Research & Sleep Medicine, 19(4), 318–326.
Piper, B. F. (1989). “Fatigue: current bases for practice.” Key Aspects of Comfort: Management of
Pain, Fatigue, and Nausea. New York: Springer Publishing Co, 187–98.
Piper, B. F., Dibble, S., Dodd, M., Weiss, M., Slaughter, R., and Paul, S. (1998). “The revised Piper
Fatigue Scale: psychometric evaluation in women with breast cancer.” Oncology nursing
forum, 25(4), 677–684.
Powell, N. B., Riley, R. W., Schechtman, K. B., Blumen, M. B., Dinges, D. F., and Guilleminault, C.
(1999). “A Comparative Model: Reaction Time Performance in Sleep-Disordered Breathing
Versus Alcohol-Impaired Controls.” The Laryngoscope, 109(10), 1648–1654.
Purnell, M. T., Feyer, A.-M., and Herbison, G. P. (2002). “The impact of a nap opportunity during the
night shift on the performance and alertness of 12-h shift workers.” Journal of sleep research,
11(3), 219–227.
Rajaraman, S., Ramakrishnan, S., Thorsley, D., Wesensten, N. J., Balkin, T. J., and Reifman, J. (2012).
“A new metric for quantifying performance impairment on the psychomotor vigilance test.”
Journal of Sleep Research, 21(6), 659–674.
158
Reiner, B. I., and Krupinski, E. (2011). “Demystifying Occupational Stress and Fatigue Through the
Creation of an Adaptive End-User Profiling System.” Journal of Digital Imaging, 25(2), 201–
205.
Ribet, C., and Derriennic, F. (1999). “Age, working conditions, and sleep disorders: A longitudinal
analysis in the French cohort E.S.T.E.V.” Sleep: Journal of Sleep Research & Sleep Medicine,
22(4), 491–504.
Ricci, J. A., Chee, E., Lorandeau, A. L., and Berger, J. (2007). “Fatigue in the U.S. workforce:
prevalence and implications for lost productive work time.” Journal of Occupational and
Environmental Medicine / American College of Occupational and Environmental Medicine,
49(1), 1–10.
Roach, G. D., Fletcher, A., and Dawson, D. (2004). “A Model to Predict Work-Related Fatigue Based
on Hours of Work.” Aviation, Space, and Environmental Medicine, 75(3), A61–A69.
Rosekind, M. R., Smith, R. M., Miller, D. L., Co, E. L., Gregory, K. B., Webbon, L. L., Gander, P. H.,
and Lebacqz, J. V. (1995). “Alertness management: strategic naps in operational settings.”
Journal of Sleep Research, 4(s2), 62–66.
Rupp, T. L., Wesensten, N. J., and Balkin, T. J. (2012). “Trait-Like Vulnerability to Total and Partial
Sleep Loss.” Sleep, 35(8), 1163–1172.
Sagaspe, P., Sanchez-Ortuno, M., Charles, A., Taillard, J., Valtat, C., Bioulac, B., and Philip, P. (2006).
“Effects of sleep deprivation on Color-Word, Emotional, and Specific Stroop interference and
on self-reported anxiety.” Brain and Cognition, 60(1), 76–87.
Sallinen, M., HÄRMÄ, M., \AAKERSTEDT, T., Rosa, R., and Lillqvist, O. (1998). “Promoting
alertness with a short nap during a night shift.” Journal of Sleep Research, 7(4), 240–247.
Santos, J. R. A. (1999). “Cronbach’s alpha: A tool for assessing the reliability of scales.” Journal of
extension, 37(2), 1–5.
Sawacha, E., Naoum, S., and Fong, D. (1999). “Factors affecting safety performance on construction
sites.” International Journal of Project Management, 17(5), 309–315.
Schneider, S., and Susi, P. (1994). “Ergonomics and Construction: A Review of Potential Hazards in
New Construction.” American Industrial Hygiene Association Journal, 55(7), 635–649.
Scott, A. J., and LaDou, J. (1994). “Health and safety in shift workers.” Occupational Medicine, 960–
971.
Scott, J. P., McNaughton, L. R., and Polman, R. C. (2006). “Effects of sleep deprivation and exercise
on cognitive, motor performance and mood.” Physiology & behavior, 87(2), 396–408.
Segawa, K., Nakazawa, S., Tsukamoto, Y., Kurita, Y., Goto, H., Fukui, A., and Takano, K. (1987).
“Peptic ulcer is prevalent among shift workers.” Digestive Diseases and Sciences, 32(5), 449–
453.
Shapiro, C. M., Flanigan, M., Fleming, J. A. E., Morehouse, R., Moscovitch, A., Plamondon, J.,
Reinish, L., and Devins, G. M. (2002). “Development of an adjective checklist to measure five
FACES of fatigue and sleepiness: Data from a national survey of insomniacs.” Journal of
Psychosomatic Research, 52(6), 467–473.
Smets, E. M. A., Garssen, B., Bonke, B., and De Haes, J. C. J. M. (1995). “The multidimensional
Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue.” Journal of
Psychosomatic Research, 39(3), 315–325.
Spencer, N. H. (2013). Essentials of Multivariate Data Analysis. CRC Press.
Spies, C., and Trohman, R. G. (2006). “Narrative Review: Electrocution and Life-Threatening
Electrical Injuries.” Annals of Internal Medicine, 145(7), 531–537.
Spurgeon, A., Harrington, J. M., and Cooper, C. L. (1997). “Health and safety problems associated
with long working hours: a review of the current position.” Occupational and Environmental
Medicine, 54(6), 367–375.

159
Stone, M. (1974). “Cross-validatory choice and assessment of statistical predictions.” Journal of the
royal statistical society. Series B (Methodological), 111–147.
Sultanian, A. R., Motamedzade Torghabe, M., Shafii motlagh, M., Garkaz, A., and Mahdavi, N.
(2014). “Persian version of Swedish Occupational Fatigue Inventory )P-SOFI) : validity and
reliability.” Iran Occupational Health, 11(1), 34–43.
Swaen, G. M. H., Amelsvoort, L. G. P. M. van, Bültmann, U., and Kant, I. J. (2003). “Fatigue as a risk
factor for being injured in an occupational accident: results from the Maastricht Cohort Study.”
Occupational and Environmental Medicine, 60(suppl 1), i88–i92.
Taris, T. W., Beckers, D. G. J., Dahlgren, A., Geurts, S. a. E., and Tucker, P. (2007). “Overtime work
and well-being : Prevalence, conceptualization and effects of working overtime.” 40, 2, 21–40.
Taylor, R. R., Jason, L. A., and Torres, A. (2000). “Fatigue rating scales: an empirical comparison.”
Psychological Medicine, 30(04), 849–856.
Techera, U., Hallowell, M., and Littlejohn, R. (2017). “Fatigue Management for Electrical
Transmission and Distribution Construction” Journal of Construction Engineering and
Management, (Planned Submission: April 2017).
Techera, U., Hallowell, M., Stambaugh, N., and Littlejohn, R. (2016). “Causes and consequences of
occupational fatigue: Meta-analysis and systems model.” Journal of Occupational and
Environmental Medicine, 58(10), 961–975.
Tepas, D. I., and Mahan, R. P. (1989). “The many meanings of sleep.” Work & Stress, 3(1), 93–102.
Tepper, B. J. (2000). “Consequences of abusive supervision.” Academy of management journal, 43(2),
178–190.
Thorne, D. R., Johnson, D. E., Redmond, D. P., Sing, H. C., Belenky, G., and Shapiro, J. M. (2005).
“The Walter Reed palm-held psychomotor vigilance test.” Behavior Research Methods, 37(1),
111–118.
Tixier, A. J.-P., Hallowell, M. R., Albert, A., van Boven, L., and Kleiner, B. M. (2014). “Psychological
antecedents of risk-taking behavior in construction.” Journal of Construction Engineering and
Management, ASCE(04014052), 1–10.
U.S. Bureau of Labor Statistics. (2015). “Census of Fatal Occupational Injuries Summary, 2014.”
<http://stats.bls.gov/news.release/cfoi.nr0.htm>.
U.S. Department of Labor. (1996). “OSHA Priorities - Power Transmission and Distribution in
Construction.” <https://www.osha.gov/archive/oshinfo/priorities/power.html>.
U.S. Department of Labor. (2015). “Electric Power Generation, Transmission and Distribution
Standard.” <https://www.osha.gov/dsg/power_generation/index.html>.
U.S. Energy Information Administration. (2016). “Electricity data browser - Net generation for electric
power.” <http://www.eia.gov/electricity/data.cfm>.
Van Dongen, H. (2004). “Comparison of mathematical model predictions to experimental data of
fatigue and performance.” Aviation, space, and environmental medicine, 75(3), A15–A36.
Van Dongen, H. P. A., and Dinges, D. F. (2005). “Sleep, Circadian Rhythms, and Psychomotor
Vigilance.” Clinics in Sports Medicine, Sports Chronobiology: It’s About Time, 24(2), 237–
249.
Van Dongen, H. P., Maislin, G., Mullington, J. M., and Dinges, D. F. (2003). “The cumulative cost of
additional wakefulness: dose-response effects on neurobehavioral functions and sleep
physiology from chronic sleep restriction and total sleep deprivation.” SLEEP-NEW YORK
THEN WESTCHESTER-, 26(2), 117–129.
Van Veldhoven, M., and Broersen, S. (2003). “Measurement quality and validity of the ‘need for
recovery scale.’” Occupational and Environmental Medicine, 60(suppl 1), i3–i9.

160
Vries, J. D., Michielsen, H. J., and Heck, G. L. V. (2003). “Assessment of fatigue among working
people: a comparison of six questionnaires.” Occupational and Environmental Medicine,
60(suppl 1), i10–i15.
Waehrer, G. M., Dong, X. S., Miller, T., Haile, E., and Men, Y. (2007). “Costs of occupational injuries
in construction in the United States.” Accident Analysis & Prevention, 39(6), 1258–1266.
Webb, W. B., and Agnew Jr, H. W. (1975). “Are we chronically sleep deprived?” Bulletin of the
Psychonomic Society, 6(1), 47–48.
Weick, K. E., Sutcliffe, K. M., and Obstfeld, D. (2008). “Organizing for high reliability: Processes of
collective mindfulness.” Crisis management, 3(1), 81–123.
Wierwille, W. W., Wreggit, S. S., Kirn, C. L., Ellsworth, L. A., and Fairbanks, R. J. (1994). “Research
on vehicle-based driver status/performance monitoring; development, validation, and
refinement of algorithms for detection of driver drowsiness. Final report.” (No. HS-808 247)
Xu, X., Ding, M., Li, B., and Christiani, D. C. (1994). “Association of rotating shiftwork with
preterm births and low birth weight among never smoking women textile workers in China.”
Occupational and Environmental Medicine, 51(7), 470-474.

161
Appendix A: IRB Protocol
TITLE: Study of the influence of fatigue on hazard recognition and development of a fatigue
predictive tool

PROTOCOL VERSION DATE: 1st May 2016

VERSION: 1

PRINCIPAL INVESTIGATOR (PI):

Name: Ulises Techera

Address: 3095 Blue Sky Circle, #305. Erie, CO. 80516
Telephone: 801 718 1228
Email: [email protected]
KEY PERSONNEL
Name: Ulises Techera
Role: Principal Investigator

Name: Dr. Matthew Hallowell

Role in Project: Faculty Supervisor, Co-Investigator.

Name: Dr. Ray Littlejohn

Role in Project: Co-Investigator

OBJECTIVES
The current study has two specific primary objectives:

First, we aim to develop a fatigue predictive tool. Some authors claim that the level of fatigue in
an individual can be predicted based on the level of activity experienced during the past 24h-48h
(Dawson and McCulloch, 2005). Previous research in the construction industry suggests that the
amount of sleep obtained within such a period of time, the number of hours worked in the previous
shift, the rest period between shifts, the type of shift, the intensity of the workload, the time on
task, and the temperature are variables which could predict the level of fatigue in an subject
(Techera et al. 2016). This hypothesis will be tested by comparison to an objective measure of
fatigue. By so doing, a fatigue risk predictive tool will be created.

Second, we aim to study the influence of fatigue among the hazard recognition ability of
individuals. We will test the hypothesis that fatigue diminishes the ability of an individual to
recognize hazards. The data for this test will be obtained by asking individuals to recognize hazards
present on a set of pictures, right after assessing their level of fatigue. The researchers expect to

162
find a correlation between the level of fatigue in the individuals and their hazard recognition
ability.

The findings of this research will have their application in the construction industry in general, and
more specifically among Transmission and Distribution power line workers (TD workers).
BACKGROUND AND SIGNIFICANCE
In comparison to the average occupation, electrical TD construction, operation, and maintenance
is nearly an order of magnitude more dangerous. According to the U.S. Bureau of Labor Statistics
(2015) in 2014 there were a total of, 4,679 fatal work injuries, which equates to an all-industry
fatality rate of 3.3 fatalities every 100,000 full-time equivalent workers (FEW). That same year,
the construction industry showed a fatality rate of 9.4 every 100,000 FEW and electrical TD
workers experience a rate of 19.2 fatalities every 100,000 FEW. The mean fatality rate for
electrical TD workers for the past 20 years is 26.1 per 100,000 workers, over three times greater
than the construction fatality rate and eight times greater than the all-industry average. These
statistics set electrical TD operations among the 10 most deadly jobs in the country.

Fatigue plays a significant role in the causes of occupational injuries and fatalities, particularly in
high-energy situations. When workers are fatigued, they experience compromised alertness,
judgement, reaction time, mental acuity, physical strength, and the development of an
uncooperative disposition (Gillberg and Åkerstedt 1994; Kajtna et al. 2011; van der Linden et al.
2003a; Lorist et al. 2005; Scott et al. 2006; Yaggie and Armstrong 2004). Such effects decrease a
worker’s ability to complete their work safely due to the increased rate of human error (Dembe et
al. 2005). Currently, there are two ways of measuring fatigue, subjective questionnaires or devices
that measure fatigue objectively. Subjective measures of fatigue are prone to easy manipulation.
The majority of current technologies, that objectively measure fatigue, lack proper validation and
are very costly (Dawson et al. 2014). These characteristics, diminish the applicability of these tools
in the industry. The development of an objective fatigue risk predictive questionnaire will allow
for a cheap and easy evaluation of the level of fatigue among individuals, thus alerting supervisors
of a dangerous fatigue related situation and preventing accidents and fatalities.

Furthermore, a few studies suggest that fatigue diminishes the ability of an individual to perceive
risk, therefore, subconsciously individuals assume higher risk than they would assume under a
non-fatigued condition (Spurgeon et al. 1997; Tixier et al. 2014). However, none of these two past
studies focused on hazard recognition with fatigue as the only treatment variable. For this reason,
further research is needed to confirm these suggestions. The objective evaluation of the impact of
fatigue among the hazard recognition ability of workers will represent and important and
meaningful contribution to the current body of knowledge about fatigue and to the industry.
PRELIMINARY STUDIES
This study will be the first effort to create an objective fatigue risk predicting tool specifically
designed for construction workers. Additionally, this study will confirm or deny the predicting
character of the variables presented above in regards to fatigue, thus building upon the research
developed by Dawson and McCulloch (2005).

Additionally, the current research is the first effort to address hazard recognition as a function of
fatigue solely. The results of this study will either confirm or refute past research.
163
RESEARCH STUDY DESIGN
According to the research objectives, the study design is organized in the following way:

1) Fatigue risk predicting tool

According to the first objective of creating a fatigue risk predictive tool, the study is designed in 3
different sections which are, first Pilot Testing (PT), second “Instrument Development Phase
(IDP), and third “Instrument Validation Phase” (IVP). The data collecting process will take
approximately 9 months.

The point of departure consists of identifying variables that previous research theorize predicts
fatigue. These variables have been previously studied separately; however, non-study has
considered all variables together as predictors of fatigue. The current stage of the research process
focuses on obtaining the regression parameters that will describe the predictive ability of each
variable. To do that, the researchers will assess the state or value of each activity variable (the
seven variables presented above) through an activity questionnaire, and analyze the relationship
of those variables with a true measure of fatigue, which will be taken by a reaction time test (PC-
PVT). These steps will be described in detail in the “Procedure” section of this document.

Sample size calculation:

In order to accomplish this first objective presented above, the researchers seek to determine the
predictive validity of the variables presented above. The statistical method chosen to investigate
the predictive validity is Multiple Linear Regression (MLR). This method is generally used either
to describe or to predict a relationship (Gross 1973; Kerlinger and Pedhazur 1973; Pedhazur 1997).
While using this method the minimum sample size varies depending on several factors such as the
nature of each variable and the true correlation between independent variables and the dependent
variable. Values as low as 10 data points per independent variable of interest have been shown to
produce acceptable results (Miller and Kunce 1973). On the other hand some authors state that a
minimum sample size of 30 data points per independent variable of interest are needed to obtain
acceptable results (Pedhazur and Schmelkin 2013).

In this experiment, the sample size for the pilot test will be calculated based on the
recommendation given by Miller and Kunce (1973) of 10 data points per variable of interest. For
the second phase of this experiment (instrument development), we will base our initial decision of
minimum sample size in the work produced by Knofczynski and Mundfrom (2007). These latter
researchers utilized more than 23, 000,000 computed generated samples to calculate minimum
sample sizes while using MLR for prediction. According to that study in order to guarantee that
95% of the correlation coefficients between the predictive variables and the dependent variable
(fatigue) exceed 0.92 (good predicting level) and in order to be able to detect a squared population
multiple correlation coefficient (ρ2) equal or greater than 0.25 a ratio of approximately 30 data
points per variable of interest should be obtained.

1.a) Pilot Testing:

This first data collection effort has two purposes. The first purpose is the one of ensuring the correct
functioning of the research tools. These tools are: (1) activity questionnaire that inquiries about

164
the participant’s level of activity during the previous 24-48hs, a subjective fatigue questionnaire;
(2) software that performs a reaction time test; and (3) a hazard recognition test based on pictures.
The second purpose consist of acquiring enough data to more accurately calculate sample sizes
needed for subsequent phases. The pilot testing phase is expected to last approximately 1 month
where a sample of 70 graduate students from the University of Colorado at Boulder will voluntarily
take part in the experiment. The sample size for this pilot test was calculated based on the fact that
there are 7 principal variables of interest and using the method presented in the previous
subsection. The reader will notice that the activity questionnaire presents more than 7 variables;
the additional variables, besides the principal 7 variables, were included for exploratory purposes
given the fact that they experience changes with fatigue, but our main focus is to determine the
predictive ability of the 7 principal variables of interest. These variables are as follows: the amount
of sleep obtained within the previous 24hs, the number of hours worked in the previous shift, the
rest period between shifts, the type of shift, the intensity of the workload, the time on task, and the
temperature.

During this phase, the level of activity during the 24-48hs previous to the testing time will be
assessed, by the activity questionnaire, as well as the participant’s level of fatigue (with the
reaction time test) both objectively and subjectively (with a subjective fatigue questionnaire). A
correlation will then be drawn between the level of activity of individuals and their level of fatigue.
The results will give information for a more accurate sample calculation for the subsequent stages
of data collection.

1.b) Instrument Development:

This second phase has the objective of collecting data among the population of interest which is
construction workers (more specifically, electrical transmission and distribution (TD) line
workers) in order to obtain the fatigue predictive variables and weights that fit this population.

In instrument development research the more participants the better. However, in reality the
obtained sample size usually is far smaller than the desired sample size. The target sample size
will be more accurately calculated after the pilot test; however, an approximate number of 300
construction workers will be established as the target population. This sample size was calculated
on the basis that 30 data points per variable are needed to obtain good results from regression
analyses and there are a total of 7 main variables plus 17 other possible fatigue predictive variables
(30 X 7 = 210) (Knofczynski and Mundfrom 2007). An additional sample of 90 workers is
considered which will allow for a stronger analysis or the possible inclusion of other 3 additional
variables. These variables are presented in an objective activity questionnaire that will be described
in the “Procedures” section of this document, however, the reader can refer to appendix A for such
a questionnaire. The results’ characteristics and the data analysis agree with those of the previous
section.
1.c) Instrument Validation:
The main objective of this phase of the research project is validating a fatigue risk objective
predictive tool developed in the previous step with a population of construction workers. This
phase will be sub-divided between two stages which are (1) validation among TD workers and (2)
validation among construction workers from a variety of trades to test generalizability.

165
The sample size calculated for this stage consists of at least the same number of participants that
took part of the previous phase, per group where the questionnaire is to be validated. This indicates
that the questionnaire will try to be validated among 300 TD workers and 300 construction workers
in other sectors (e.g., concrete finishers, carpenters, plumbers). This validation procedure is
expected to last about 5 months. The timeline will include: One month and a half to collect data
among TD line workers; one month and a half to collect data among other construction workers;
and one month for data review and analysis; and one month to coordinate visits with the different
contractors.

By the end of this data collection effort, the researchers will have collected fatigue scores obtained
by a new objective fatigue measurement/predictive tool and objective fatigue levels obtained with
the PVT, which is the current gold standard to measure fatigue. The comparison among the fatigue
scores and the true measure of fatigue will validate the researcher’s tool.

2) Hypothesis testing of the influence of fatigue on hazard recognition ability.

The main objective of this second phase of the research project consists of obtaining an answer for
the following question: Does fatigue affect the hazard recognition ability of workers? To obtain
such an answer, the researchers propose the alternative hypothesis that individuals will higher
levels of fatigue have lower levels of hazard recognition.

This phase of the experiment will be incorporated to the stages 1.a, 1.b, and if needed 1.c. 70
graduate students and 900 construction workers will take part in this experiment. It is important to
mention that the sample size needed to test this hypothesis is smaller than the sample size needed
to develop a fatigue predictive tool because the experiment will involve fewer variables of interest
(Brooks and Barcikowski 1995; Darlington 1990; Gross 1973; Pedhazur 1997). The total duration
of this phase of the research project is expected to last approximately 9 months.

During this stage of the data collection process the hazard recognition ability of individuals will
be measured and compared to their objective level of fatigue. A correlation between these two
variables will be investigated. In order to measure hazard recognition ability a questionnaire with
pictures will be presented to the participants. The pictures will be randomly pulled from a portfolio
of pictures. Each picture contains a specific number of hazards identified by a panel of experts.
Participants will be invited to recognize as many hazards as possible and the results from different
participants with the same picture but different level of fatigue will be compared.

166
Research design flow chart:

Pilot Testing 210

H1. testing data collection 1 Graduate Students
1 Month

Instrument development 720

H1 testing data collection 2 TD line workers
2.5 Months

Instrument Validation 720 TD line workers

H1 testing data collection 3 720 Construction workers
5 Months

Final Data Analysis

3.5 Months
Articles Redaction

FUNDING
None
ABOUT THE SUBJECTS
➢ A total of 970 subjects are expected to participate in the study. All these are volunteers older
than 18 years of age which will voluntarily read and sign the consent form in order to take part
in the study.
➢ From these volunteers, 70 are expected to be graduate students from the University of Colorado
at Boulder, 600 are expected to be TD workers from the east coast of the United States, and
the remaining 300 are expected to be construction workers from Colorado (U.S.).
➢ Besides these groups of participants, a panel of 5 members integrated by safety experts will be
created to assist with stage 2 of the data collection process. This panel will consist of Professor
Matthew R. Hallowell (co-author) and four graduate student experts in construction safety.
More details about this panel of safety experts will be provided in the “Procedures” section of
this document.

167
 Subject Population(s) Number to be enrolled in each group
Graduate Students 70
TD line workers 600
Construction workers 300
Safety Experts panel 5

➢ The researchers anticipate that approximately 70% of the planed number of volunteers will
actually complete the experiment.
➢ Among the graduate students the researchers expect to find a slightly higher number of men
participating in the pilot test due to the demographics of engineering students. We also expect
to find a diverse group of ethnicities given the diversity of students in this campus.

Among the TD workers sample the researchers expect to find a predominant white male
population give the demographic characteristics of the TD workers in the east coast of the
United States.

Among the more general construction workers sample the researchers expect to find a
predominant Hispanic male population due to the demographic characteristics of the industry.

➢ List the inclusion criteria - characteristics that must be met for individuals to be enrolled in
study.
o Older than 18 years old. Having a full proficiency in English. Graduate student or
construction worker.
➢ List the exclusion criteria - characteristics that will exclude individuals from the study.
o Younger than 18 years old. Non-English speaking individuals. Non graduate students
or construction workers.
The objectives of this research project are directed to the construction industry and for this reason
no minor or non-construction worker or graduate student will be considered for the study.
VULNERABLE POPULATIONS
➢ What vulnerable populations will be considered for this study?
o None

➢ Describe the additional safeguards that are included to protect their rights and welfare.
o N/A

➢ How will coercion be avoided in this population?

o All participation in this research process will be voluntary and anonymous. The data
collected will be dis-attached from any participant’s personal identifier. No reward is
offered to graduate students and construction workers will not receive any additional
material compensation. In the same way, there is no sanction for not participating in
the study for either group. Consequently the motivation to participate in this study is
purely personal and has not external incentive or disincentive.

168
RECRUITMENT METHODS
o Graduate students in Engineering at CU Boulder will be recruited via email. The email will briefly
explain the objective of the research and will ask for the voluntary participation in the experiment.
The consent form will be attached to the email and those who desire to participate in the study will
bring the consent form signed to the collecting data session. After that participants will receive an
email with dates and times to take part of the experiment and these will enroll in one of those dates.
The day previous to their testing an email will be sent to them to remind them of the coming testing
session. No incentive or punishment are presented to the student in order to take part of the study.
This guarantees the absence of undue-influence or coercion to force students to participate in this
study.

o During the collecting data process you will interact with the principal investigator, Ulises Techera.
This data collection procedure will take place at the conference room of the Civil, Environmental,
and Architectural Engineering department of the Applied Sciences Engineering School, at the
University of Colorado at Boulder (1111 Engineering Drive 428 UCB / ECCOT441, Boulder , CO.
80309-0428). The Pilot Testing phase of this research will take place during the month of May and
June of 2016.

o TD line workers and construction workers will also be free to participate and in order to do so they
must sign the consent agreement beforehand. The presentation of the research study and solicitation
for participation will occur during a regular day of work. Once again, participants won’t receive
any material benefit or punishment for participating in the study. Consequently, their participation
is entirely voluntary. Participants will be allowed by their supervisors to participate in the study or
to continue working if that is what they prefer. It is in the best interest for the supervisor to have an
operator continually working, however, no punishment will be given to any worker for their
voluntary effort of participating in this study. Consequently, there is no undue-influence or coercion
to participate in the study.

o During the collecting data process you will interact with the principal investigator, Ulises Techera.
This data collection procedure will take place at the beginning of the shift, during the meeting with
supervisors, at the company’s facilities, and additionally in a quiet and spacious vehicle, parked
near to where the operator will be working. The data collection process extends from June, 2016 to
January, 2017.

List recruitment methods/materials and attach a copy of each in eRA

1. Email communication
2. Personal invitation at the work site

COMPENSATION
o For graduate students no compensation will be offered other than the personal satisfaction
of contributing to a good cause.
o For construction workers (including TD workers) no monetary compensation will be
offered other than maintaining their salary amount after performing the aforementioned
tests.

169
CONSENT PROCESS
o Two different consent forms were created, one for graduate students which can be
found in Appendix C, and another for construction workers which can be found in
Appendix D.
o The group of graduate students will receive the consent form by email and they must
sign the consent form before taking part in the study.
o The group of TD workers and the rest of construction workers will receive a briefing
on the objectives of this research with a description of the experimental procedure
and an explanation of their role as participants. Following the briefing, they will be
invited to participate in the experiment and it will be indicated that no material reward
or punishment will be given by participating in the study. Furthermore, it will be
explained that the participation is entirely voluntary and the consent forms will be
handed out. Those workers that decide to sign the consent form will be admitted for
testing.

o All results will remain confidential and no results will be reported to employers or
participants until all data have been aggregated. At no single point in the data gathering
process identifiers will be collected (names, addresses, complete birth dates, etc.). The
data will be completely anonymous and results will be presented as an aggregate, thus
responses of any participant cannot be traced back to one particular individual.

PROCESS TO DOCUMENT CONSENT IN WRITING

o As mentioned above, consent will be obtained via 2 different documents depending on the
characteristics of the sample. Volunteers participating in the Pilot Testing experiment will
complete the consent form addressed to graduate students. On the other hand, construction
workers will complete the consent form developed for construction workers.

o In every case consent forms must be delivered to the researcher in order for volunteers to
take part in the experiment. Participants can abandon the experiment at any time.

PROCEDURES
Before describing procedures of the 3 different sections mentioned above it is convenient to go
over the research tools that will be employed in such phases.

Research tools
The research tools that will be used to accomplish this first overall objective are: a questionnaire
developed by the researchers (activity questionnaire), a subjective fatigue assessment
questionnaire specifically designed to assess fatigue among a working population, and finally a
device to objectively assess the level of fatigue of an individual.

The survey developed by the researchers is a self-administered questionnaire that presents 4

biographical questions (age, years of experience in the sector, position, and gender), 24 objective
questions about the participant’s level of activity during the previous 24-48hs, and 4 subjective
questions. These questions inquire about the treatment variables which fatigue predictive ability

170
wants to be determined. This questionnaire will be handed out to volunteers who will complete it
in approximately 5 min. For more details on this questionnaire please refer to Appendix A.

The subjective assessment of fatigue will be performed by the Swedish Occupational Fatigue
Inventory (SOFI-20) (Åhsberg et al. 1997), which is a self-administered questionnaire that has
been widely used to assess fatigue among a working population (Åhsberg 2000; González
Gutiérrez et al. 2005).

The questionnaire presents 20 questions rated on a Likert scale from 0 to 6. It inquires about 5
different dimensions of fatigue which have shown to behave differently depending on the type of
job. The five dimensions mentioned before are: 1) Lack of energy; 2) Physical exertion; 3) Physical
discomfort; 4) Lack of motivation; 5) Sleepiness. The internal consistency of this questionnaire
was evaluated in a sample of 597 volunteers and the results showed an overall Cronbach’s alpha
of 0.80, indicated good internal consistency (Åhsberg 2000). For further information about the
SOFI please refer to Appendix B.

The third and last tool used in this phase of the research project is a device developed to objectively
obtain an indirect measure of fatigue, such a device is the Psychomotor Vigilance Test (PVT). The
PVT is a reliable and validated technology that has been widely used in research during several
decades to obtain an objective measure of fatigue. The PVT is a technology that assesses the
Neuro-behavioral performance of an individual by measuring its reaction time to a stimulus. This
method is considered the current gold standard to measure fatigue (Dinges and Powell 1985;
Dorrian et al. 2005; Loh et al. 2004). The PVT-192 is the most commonly used and widely
validated tool to objectively measure fatigue and it has been extensively validated in its 2 different
modalities, a 5 minute version of the test (Ferguson et al. 2011; Lamond et al. 2005; Thorne et al.
2005), and a 10 minute version of the test (Balkin et al. 2004; Dinges et al. 1997; Dinges and
Powell 1985; Dorrian et al. 2005, 2008; Pilcher et al. 2007; Van Dongen et al. 2003).

The PVT-192 (Ambulatory Monitoring, Inc., Ardsley, NY; Dinges & Powell, 1985), is the current
“gold standard” to measure simple reaction time. The device consists of a hand held apparatus
with a small screen (approximately 2” x 0.5”) and two buttons at the bottom. The device shows a
visual stimulus which consists of a four digit millisecond counter that stops the count once the
participant presses a button. At that time, the reaction time remains on the screen for 500ms and
then it disappears. These stimuli appear after a random interval of time denominated “Inter-
stimulus Interval” (ISI), which varies between 2 seconds and 10 seconds. The device stores the
data for each session, then, this data can be transferred to a PC for analysis. Some of the most
common analyses of the data are: the number of lapses experienced by an individual (more than
500ms without response after a stimulus) and the average reaction time (Khitrov et al. 2013).

Recently, a new software “PC-PVT” was developed to perform simple reaction time measures
among individuals. The principal objective of the developers of this software was to develop a
PVT with the same reliability and functionality as the PVT-192 (Khitrov et al. 2013). The
mentioned software has achieved these objectives offering a reliable, valid, more familiar and more
economical way of measuring simple reaction time.

171
The PC-PVT is a software that requires Windows XP or a later version to run, Windows 7 being
the preferred operating system (OS) for providing support for modern timing hardware. A PC is
not specifically designed to measure reaction time. For this reason a PC presents 2 main sources
of RT data degradation. One is a reduced accuracy in the measurement due to the multitasking
nature of a PC, and second, is a systematic delay in response detection introduced by the hardware
and /or by the software. Nonetheless, the PC-PVT, when in operation, raises its own priority level
to have the OS dedicate as much of the central processing unit (CPU) time as possible to the PVT
session (Khitrov et al. 2013) .

The PC-PVT test requires the PC-PVT software, a PC, and a mouse. The test can be administered
in 2 different modalities of 5 minutes or 10 minutes. The test consists of sitting in front of a
computer with one hand on the mouse and pressing a button on the mouse as soon as possible after
receiving a stimulus. The screen will show a black background and the stimulus will appear as a 5
digit millisecond counter in red that stops the count when the individual presses the button. The
RT remains on the screen for 500ms and then it disappears. The ISI is distributed randomly
between 2 seconds and 10 seconds, as in the PVT-192. The software allows for the immediate
storage of data by individual and by session in the PC right after a session finishes. This represents
an important advantage in comparison to the PVT-192 which requires the transmission of the
stored data in the device to a PC for statistical analysis with its consequent risks of corruption. The
summary statistics of the PC-PVT currently display: number of lapses, mean RT, speed, mean of
the fastest and lowest 10% of RTs and RT divergence (Rajaraman et al. 2012). The software even
provides an interesting and useful feature which consists of individualized predicting algorithms
capable of predicting reaction times for an individual up to 24hrs into the future. This feature can
then be used after the first session when the software matches your performance to an empirical
database to predict your future RTs. However, when an individual participates in multiple sessions
the software refines the predictive algorithm by assigning higher weights to the participant’s data
and lower weights to the data coming from the empirical database, using a Bayesian approach
(Khitrov et al. 2013).

The PC-PVT was validated in a wide range of PCs and laptops with different operating systems,
screens, and mouses to contemplate the variety of devices used by researchers. The validation was
against the PVT-192, using the RTBox as the true measure of delay. The RTBox, was developed
at the University of Southern California (Los Angeles, CA; Li et al., 2010) and consists of a closed
circuit capable of measuring the time between the onset of the stimulus and the response with sub-
millisecond precision (0.1ms). The RTBox allowed for three different comparisons: 1) The RTBox
against the PVT-192, 2) The RTBox against the PC-PVT with a standard mouse, and 3) The
RTBox against the PC-PVT with a gaming mouse with a scanning frequency of 1000Hz. Each
comparison took place in 5 different sessions of 5 minutes each where the RT varied from the
fastest humanly possible (~ 160 ms) to ~ 2000 ms. The results indicated a delay for the PVT-192
characterized by a mean of 3.4ms and a SD = 0.8 ms. For the PC-PVT with a gaming mouse the
mean = 7.8 ms with a SD = 1.0ms; and the PC-PVT with a standard mouse showed a delay
characterized by a mean = 35.7 ms and a SD = 2.6ms. The delay found in the PC-PVT with a
gaming mouse ( mean = 7.8, SD = 1.0) constitutes an error of 3% assuming a mean RT of 240 ms
with a SD = 29ms, which is common in fatigue research (Rupp et al. 2012). Consequently, the PC-
PVT with a gaming mouse ( 1ms mouse response) has been proven to be valid and reliable for RT
related research given its low margin of error (3%) which is comparable to the one presented by

172
the PVT-192 (1%) and its almost negligible variability (1ms) compared to the intra-subject RT
variability (~ 29ms) (Khitrov et al. 2013) . For this reason this research will utilize the 5minutes
PC-PVT protocol, with an appropriate gaming mouse to measure reaction time, and in this way
obtain an objective measure of fatigue.

In addition to the research tools presented above two additional tools will be employed during the
data collection process. The first additional tool consists of the Objective Fatigue Predictive
questionnaire that will be developed as part of this research process and later on validated as part
of this research effort. The second research tool used in this experiment will appear during the
second phase of this study. It consists of construction scenarios pictures which will display a
specific number of hazards. More details on these two tools will be presented in the following
subsections.

Name of Purpose (i.e. what data Time to Complete

instrument/tool/procedure is being collected?
Objective activity questionnaire Objective past level 5 minutes
of activity
SOFI Subjective level of 2 minutes
fatigue (score)
PC-PVT Objective level of 5 minutes
fatigue (RT)
Objective fatigue risk predictive Objective level of 4 minutes
questionnaire fatigue (score)
Construction site pictures Present Hazards 12 minutes

Now after describing the research tool we proceed with the description of each research phase to
accomplish our first research objective of developing a fatigue risk objective predicting tool.

1) Fatigue risk predicting tool

1.a) Pilot Testing:

The procedure of the pilot testing experiment is as follows: Participants will be invited to take part
in the experiment individually and after consenting by signing the consent form they will be given
both a day and a time convenient to them to take part in the experiment. The experiment will be
organized in groups of 10 volunteers. The 10 volunteers will start the testing protocol by receiving
an overall explanation of the testing procedure and an introduction to the testing tools (approximate
duration 3 min) where the anonymous character of the study will be described. Following the
introduction, the participants will complete the self-administered questionnaire (Appendix A),
which has a duration of approximately 5 minutes. Then, each participant will complete the self-
administered subjective fatigue questionnaire SOFI which has an approximate duration of 2
minutes. Following this questionnaire, each participant will perform the 5 minute PC-PVT test and
the data will be stored under a common identifier name that will appear on the other two previous
questionnaires (non-participant related). At this point the testing procedure will have finished. The

173
overall testing process will take approximately 20 minutes (considering 5 minutes for inter task
delay). This protocol will take place every 2 days allowing time to recruit additional volunteers.

Once the data are collected, the researchers will have obtained answers to the objective fatigue
questionnaire which consists of numerical values or Yes/No answers, subjective fatigue scores
from the SOFI, and objective fatigue scores interpolated from the RT data (e.g. average RT or
number of lapses). Regression techniques will be used to obtain weights that correlate the objective
fatigue questionnaire answers to the objective fatigue scores obtained from the PC-PVT. The
results from the SOFI will be used to compare fatigue scores with the objective PC-PVT results.

This pilot testing procedure will inform the researcher about the seemingly true number of
predictive variables, the functionality of the research tools, and the statistical power obtained with
the collected data. With this information, a more accurate minimum sample size for the next phases
can be calculated.

1.b) Instrument Development:

Regarding the testing procedure in this phase of the experiment, it will follow the order described
in the previous section utilizing the same tools and having the same duration. All workers will
have the opportunity to freely decide if they want to take part in this experiment after a detailed
explanation of the procedure and will be required to sign the consent form in order to participate.
The researchers will have the opportunity to interact with groups of 10 to 20 workers at the
beginning of the shift and smaller groups of workers throughout the working day. The researchers
assume a capability of testing 10 individuals simultaneously due to the limited number of available
laptops for this experiment (10 PC laptops). Given the situation, half of the worker groups will be
assessed during the beginning of their shift and the other half will be assessed during the middle
of their shift at their respective work locations. Companies participating in this research project
will allow workers to spend the necessary time for testing away from their usual activities without
receiving any salary reduction because of this procedure. Assuming that companies will allow the
participation of only one of their shifts a day (20 workers) this second data collection effort will
require approximately a month and a half (5 weeks of testing plus a week of travelling).

At the end of this second data collection cycle, the data will be analyzed by regression techniques
in order to obtain the weights that correlate the answers to the fatigue objective questionnaire to
the objective level of fatigue measured by the PC-PVT. Those variables that don’t correlate to the
results obtained by the PC-PVT will be discarded from the questionnaire and a new predictive
objective activity questionnaire will be developed including only those variables that appear to
have a predictive ability. After this, a fatigue scale that correlates to the PC-PVT results will be
developed. Consequently, a fatigue risk predictive tool will have been developed which provides
a score or level of fatigue as a result of the linear combination of the obtained weights and the
inputted answers. A comparison amongst the predicted levels of fatigue by the new questionnaire,
the PC-PVT scores and the SOFI will be performed in order to investigate the existence of a
correlation between the SOFI scores and the PC-PVT scores or the SOFI and the new objective
fatigue questionnaire scores. It is expected that this data analysis and the development of the new
predictive questionnaire will take approximately one month.

174
1.c) Instrument validation:
The validation process consists of testing the fatigue risk level prediction ability of the created
questionnaire. In order to accomplish this purpose participants will complete the questionnaire and
then take the 5 minute PC-PVT testing protocol on two occasions: one right after completing the
questionnaire and at a point in the last quarter of their shift (e.g., last 2 hours). This stage of the
research process will require following up with every participant during their shift. The first testing
procedure will take approximately 16 minutes. Because some questions on the initial survey will
have the same answers (e.g., how many hours of restful time did you get between shifts?), the
second test will take approximately 7 minutes.

If this last stage of the research process validates the objective fatigue risk predictive questionnaire
a new reliable and valid tool to accurately predict fatigue among construction workers will have
been created.

2) Hypothesis testing of the influence of fatigue on hazard recognition ability.

In order to test the hypothesis that fatigue affects hazard recognition a simple experiment is
designed. The experiment starts by using pictures of construction sites where hazards have already
been identified by an expert panel. A “hazard” is defined as anything that could injure, kill, or
make a person sick. These hazards can be classified as Motion, Mechanical, Electrical, Pressure,
Temperature, Chemical, Biological, Radiation, Sound, Gravity, or Motion (Albert et al. 2014)

Three pictures will be selected at random from a group of 6-9 pictures. Participants will receive
each picture in a separate sheet, with the mentioned definition of hazard, and enough space to write
down all the hazards that he or she can identify in the picture. The duration of this assessment is
expected to be approximately 12 minutes. This assessment will take place right after administering
the 5 min PC-PVT to the volunteer.

As a result of this phase, each picture will be evaluated by multiple participants who presented
different levels of fatigue. With that data, a correlation analysis will be performed to test the
alternative hypothesis that fatigue affects HR, with the null hypothesis that fatigue doesn’t affect
HR.

Visit # Procedures/Tools Location How much time the

visit will take
Pilot Test • Objective activity University of 1 month
questionnaire Colorado at
• SOFI Boulder
• PC-PVT
• Construction site
pictures
Instrument • Objective activity Power 1.5 months
Development questionnaire contractors from
• SOFI

175
 • PC-PVT the east coast of
Construction site the U.S.
pictures
Instrument • Objective fatigue Power 5 months
Validation risk predictive contractors from
questionnaire the east coast of
• PC-PVT the U.S.
• Construction site
pictures Construction
sites in Colorado

SPECIMEN MANAGEMENT
o N/A
DATA MANAGEMENT
o According to the HRB-211 the data obtain in this research process presents a risk level
of ____________

o All the data obtain during this research process will be initially delivered to the
researchers in one of the following two ways: a) Hard copy or b) Computer file. The
hard copy information will be obtained through the questionnaires administered to the
participants (Activity questionnaire, SOFI, Objective fatigue predictor questionnaire,
and the HR questionnaire). The computer files will be automatically saved in a
computer folder after each PVT session. The name of the computer file (numeric codes)
corresponding to a certain individual will be written on the hard copy questionnaires
filled out by the same individual in order to link both types of data from the same
participant. However, such a link or code has no identifying information that can trace
the data back to such an individual.

o The hard copy data will be scanned and stored as encrypted and password protected
files. At the end of each data collection day all computer files will be saved on 10
laptop computers and on a CUB server for remote access. In the laptop computers the
data will be protected by a password. Only the principal researcher will have access to
the storage files and the hard copy documents will be all shredded.

o Once the data collection period has finished the obtained data files will be transferred
to 3 desktop computers.. The data storage in laptops will be deleted and the data storage
in desktop computers will be password protected.

o It is very important to mention that none of the data collected is considered sensitive or
identifiable data.
WITHDRAWAL OF PARTICIPANTS
➢ Each participant is free to withdraw at any time during the data collection process with no
penalty. In case of withdrawal, the data obtained from that individual will be considered

176
 for analysis only if before withdrawing the volunteer was able to participate in all data
collection processes (questionnaires and PC-PVT) at least once.
➢ There is no need for replacement in case of withdrawal.
RISKS TO PARTICIPANT
➢ Participation will always take place in a safe environment inside a familiar room
(University or office workplace). Volunteers will have the opportunity to participate in a
safe environment free from distractions and recognizable hazards. Additionally, none of
the data acquisition procedures are invasive or present questions, states or images that
could put the participant at mental or emotional risk.

MANAGEMENT OF RISKS
➢ No potential risk was identified in the previous step.

POTENTIAL BENEFITS
➢ Benefits to the Participants: There is no proven benefit from the participation in the data
collection effort. However, presumably participants can benefit from the quiet time spent
participating in the experiment which will potentially reduce their workload by diminishing
the time that that participant will spend on their regular working activities. Additionally,
participants can potentially benefit from the exposure to the knowledge that fatigue is a risk
factor at work. This knowledge could influence the way that workers perform their activities
during the day, hopefully helping them develop lower levels of fatigue and improve their
safety.

➢ Benefits to the society/industry: As described before, fatigue has shown to be an important

contributor to accident causation. The first objective of this study is to provide a way to predict
high levels of fatigue before operators actually achieve those levels. This could be used to
prevent fatigue related accidents or incidents. This could potentially decrease the current
accidents and fatality rates of construction workers which are among the highest in the nation.
The second objective of this study is to test the influence of fatigue on hazard recognition
ability. The results of such a test will further inform the society/industry about the implications
of fatigue in safety. This knowledge will empower the industry/society to better prevent
accidents and fatalities.

PROVISIONS TO MONITOR THE DATA FOR THE SAFETY OF PARTICIPANTS

Participants, as mentioned above, will not be asked to disclose any personal information that
would link the responses back to them. Furthermore, access to the dataset will limited to the
key personnel (identified above). Each participant will be a random code number, however,
the code number has no significance to participants’ identity. All the data is stored in
password-protected computers of the key-personnel related with this project.
There is no way the data collected can be traced back to an individual and his/her responses.

MEDICAL CARE AND COMPENSATION FOR INJURY

N/A

177
COST TO PARTICIPANTS
There is no cost to the volunteers participating in this study.
DRUG ADMINISTRATION
No drugs will be administered at any stage of this experiment.
INVESTIGATIONAL DEVICES
No investigational devices will be used at any stage of this experiment.
MULTI-SITE STUDIES
N/A
SHARING OF RESULTS WITH PARTICIPANTS
➢ After developing the fatigue risk predictive tool (questionnaire) an effort to pilot test its use
in the industry will be organized by the collaborative parties (contractors). However, this
effort won’t directly depend on the researchers and therefore there are not current stated plans
for this procedure.

REFERENCES

Åhsberg, E. (2000). “Dimensions of fatigue in different working populations.” Scandinavian

Journal of Psychology, 41(3), 231–241.
Åhsberg, E., Garnberale, F., and Kjellberg, A. (1997). “Perceived quality of fatigue during
different occupational tasks Development of a questionnaire.” International Journal of
Industrial Ergonomics, 20(2), 121–135.
Balkin, T. J., Bliese, P. D., Belenky, G., Sing, H., Thorne, D. R., Thomas, M., Redmond, D. P.,
Russo, M., and Wesensten, N. J. (2004). “Comparative utility of instruments for monitoring
sleepiness-related performance decrements in the operational environment.” Journal of
Sleep Research, 13(3), 219–227.
Brooks, G. P., and Barcikowski, R. S. (1995). “Precision Power Method for Selecting Regression
Sample Sizes.”
Darlington, R. B. (1990). Regression and linear models. McGraw-Hill New York.
Dawson, D., Searle, A. K., and Paterson, J. L. (2014). “Look before you (s)leep: Evaluating the
use of fatigue detection technologies within a fatigue risk management system for the road
transport industry.” Sleep Medicine Reviews, 18(2), 141–152.
Dembe, A. E., Erickson, J. B., Delbos, R. G., and Banks, S. M. (2005). “The impact of overtime
and long work hours on occupational injuries and illnesses: new evidence from the United
States.” Occupational and Environmental Medicine, 62(9), 588–597.
Dinges, D. F., Pack, F., Williams, K., Gillen, K. A., Powell, J. W., Ott, G. E., Aptowicz, C., and
Pack, A. I. (1997). “Cumulative sleepiness, mood disturbance and psychomotor vigilance
performance decrements during aweek of sleep restricted to 4-5 hours per night.” Sleep:
Journal of Sleep Research & Sleep Medicine, 20(4), 267–277.
Dinges, D. F., and Powell, J. W. (1985). “Microcomputer analyses of performance on a portable,
simple visual RT task during sustained operations.” Behavior Research Methods,
Instruments, & Computers, 17(6), 652–655.
Dorrian, J., Lamond, N., Kozuchowski, K., and Dawson, D. (2008). “The driver vigilance
telemetric control system (DVTCS): Investigating sensitivity to experimentally induced
sleep loss and fatigue.” Behavior Research Methods, 40(4), 1016–1025.

178
Dorrian, J., Rogers, N. L., Dinges, D. F., and others. (2005). “Psychomotor vigilance performance:
Neurocognitive assay sensitive to sleep loss.” Marcel Dekker New York.
Ferguson, S. A., Paech, G. M., Dorrian, J., Roach, G. D., and Jay, S. M. (2011). “Performance on
a simple response time task: Is sleep or work more important for miners?” Applied
Ergonomics, Special Section: Ergonomics, health and working time organization, 42(2),
210–213.
Gillberg, M., and Åkerstedt, T. (1994). “Sleep restriction and SWS-suppression: effects on
daytime alertness and night-time recovery.” Journal of Sleep Research, 3(3), 144–151.
González Gutiérrez, J. L., Jiménez, B. M., Hernández, E. G., and López López, A. (2005).
“Spanish version of the Swedish Occupational Fatigue Inventory (SOFI): Factorial
replication, reliability and validity.” International Journal of Industrial Ergonomics, 35(8),
737–746.
Gross, A. L. (1973). “How large should sample size be in a regression analysis?” Proceedings of
the Annual Convention of the American Psychological Association, 17–18.
Kajtna, T., Štukovnik, V., and Grošelj, L. D. (2011). “Effect of acute sleep deprivation on
concentration and mood states with a controlled effect of experienced stress.” Zdravniški
Vestnik, 80(5).
Kerlinger, F. N., and Pedhazur, E. J. (1973). “Multiple Regression in Behavioural Research. New
York: Holt, Rinehart and Winston.” Inc., l973.
Khitrov, M. Y., Laxminarayan, S., Thorsley, D., Ramakrishnan, S., Rajaraman, S., Wesensten, N.
J., and Reifman, J. (2013). “PC-PVT: A platform for psychomotor vigilance task testing,
analysis, and prediction.” Behavior Research Methods, 46(1), 140–147.
Knofczynski, G. T., and Mundfrom, D. (2007). “Sample Sizes When Using Multiple Linear
Regression for Prediction.” Educational and Psychological Measurement.
Lamond, N., Dawson, D., and Roach, G. D. (2005). “Fatigue Assessment in the Field: Validation
of a Hand-Held Electronic Psychomotor Vigilance Task.” Aviation, Space, and
Environmental Medicine, 76(5), 486–489.
van der Linden, D., Frese, M., and Meijman, T. F. (2003). “Mental fatigue and the control of
cognitive processes: effects on perseveration and planning.” Acta Psychologica, 113(1),
45–65.

179
Loh, S., Lamond, N., Dorrian, J., Roach, G., and Dawson, D. (2004). “The validity of psychomotor
vigilance tasks of less than 10-minute duration.” Behavior Research Methods, Instruments,
& Computers, 36(2), 339–346.
Lorist, M. M., Boksem, M. A., and Ridderinkhof, K. R. (2005). “Impaired cognitive control and
reduced cingulate activity during mental fatigue.” Cognitive Brain Research, 24(2), 199–
205.
Miller, D. E., and Kunce, J. T. (1973). “Prediction and Statistical Overkill Revisited.”
Measurement and Evaluation in Guidance.
Pedhazur, E. (1997). Multiple Regression in Behavioral Research: Explanation and Prediction.
Holt, Rinehart and Winston.
Pedhazur, E. J., and Schmelkin, L. P. (2013). Measurement, Design, and Analysis: An Integrated
Approach. Psychology Press.
Pilcher, J. J., Band, D., Odle-Dusseau, H. N., and Muth, E. R. (2007). “Human Performance Under
Sustained Operations and Acute Sleep Deprivation Conditions: Toward a Model of
Controlled Attention.” Aviation, Space, and Environmental Medicine, 78(5), B15–B24.
Rajaraman, S., Ramakrishnan, S., Thorsley, D., Wesensten, N. J., Balkin, T. J., and Reifman, J.
(2012). “A new metric for quantifying performance impairment on the psychomotor
vigilance test.” Journal of Sleep Research, 21(6), 659–674.
Rupp, T. L., Wesensten, N. J., and Balkin, T. J. (2012). “Trait-Like Vulnerability to Total and
Partial Sleep Loss.” Sleep, 35(8), 1163–1172.
Scott, J. P., McNaughton, L. R., and Polman, R. C. (2006). “Effects of sleep deprivation and
exercise on cognitive, motor performance and mood.” Physiology & behavior, 87(2), 396–
408.
Spurgeon, A., Harrington, J. M., and Cooper, C. L. (1997). “Health and safety problems associated
with long working hours: a review of the current position.” Occupational and
Environmental Medicine, 54(6), 367–375.
Thorne, D. R., Johnson, D. E., Redmond, D. P., Sing, H. C., Belenky, G., and Shapiro, J. M. (2005).
“The Walter Reed palm-held psychomotor vigilance test.” Behavior Research Methods,
37(1), 111–118.
Tixier, A. J.-P., Hallowell, M. R., Albert, A., van Boven, L., and Kleiner, B. M. (2014).
“Psychological antecedents of risk-taking behavior in construction.” Journal of
Construction Engineering and Management, ASCE(4014052), 1–10.
U.S. Bureau of Labor Statistics. (2015). “Census of Fatal Occupational Injuries Summary, 2014.”
<http://stats.bls.gov/news.release/cfoi.nr0.htm>.
Van Dongen, H. P., Maislin, G., Mullington, J. M., and Dinges, D. F. (2003). “The cumulative
cost of additional wakefulness: dose-response effects on neurobehavioral functions and
sleep physiology from chronic sleep restriction and total sleep deprivation.” SLEEP-NEW
YORK THEN WESTCHESTER-, 26(2), 117–129.
Yaggie, J., and Armstrong, W. J. (2004). “Effects of lower extremity fatigue on indices of
balance.” Balance, 312(322), 2004

180
Appendix B: Consent Form
Title of research study: Study of the influence of fatigue on hazard recognition, the influence of
workload on fatigue, and development of a fatigue predictive tool.

Investigator: Ulises Techera

Why am I being invited to take part in a research study?

This research study focusses in the way that fatigue affects Transmission and Distribution power
line workers and more generally construction workers. For this reason you are the most
important factor of this research. Your opinion and experience matters the most in this research
project.

What should I know about a research study?

Someone will explain this research study to you.
Whether or not you take part is up to you.
You can choose not to take part.
You can agree to take part and later change your mind.
Your decision will not be held against you.
You can ask all the questions you want before you decide.
Who can I talk to?

If you have questions, concerns, or complaints, or think the research has hurt you, talk to the
research team at [email protected] (Ulises Techera , principal investigator),
[email protected] ( Matthew R. Hallowell, Co-Investigator/Faculty Advisor), or
at [email protected] (Ray Littlejohn, Co-Investigator).

This research has been reviewed and approved by an Institutional Review Board (“IRB”). You
may talk to them at (303) 735-3702 or [email protected] if:
Your questions, concerns, or complaints are not being answered by the research team.
You cannot reach the research team.
You want to talk to someone besides the research team.
You have questions about your rights as a research subject.
You want to get information or provide input about this research.
Why is this research being done?

The principal motivation behind this research project is the one of reducing injury and fatality
rates for one of the most dangerous professions in the U.S. (Transmission and Distribution power

181
line workers) and for construction workers in general. The fatality rate of these professions are 8
and 3 times the average of the whole working population in the country. Fatigue has shown to be
an important driver to accidents and for this reason it is the main focus of our research. After
conducting this research experiments we will be able to understand how fatigue affects hazard
recognition and we will have developed a tool to objectively predict the level of fatigue of
individuals, which can be used to prevent fatigue related accidents.

How long will the research last?

We expect that you will be in this research study for about 30 minutes.
How many people will be studied?
We expect about 480 people will be in this research study.
What happens if I say yes, I want to be in this research?

If you want to be part of this research you will participate in 2 data collection phases. Phase 1 will
take approximately 30 minutes. Phase 2 will take approximately 7 minutes
During Phase 1 you will:
Complete the Activity Questionnaire (AQ) which is a survey with 28 simple questions about your
level of activity. This will take approximately 5 minutes.
Complete the Swedish Occupational Fatigue Inventory (SOFI) which is a subjective fatigue
questionnaire with 20 questions. This will take approximately 2 minutes.
Complete the PC-PVT which is a reaction time test that takes place on a laptop computer and last
5 minutes.
The Hazard Recognition Questionnaire (HRQ) which is a questionnaire with 3 pictures and 3
questions for each picture. You will have 9 minutes to complete this questionnaire.
During Phase 2 you will:
Complete the PC-PVT once again.
Complete the NASA-TLX which is a simple survey with 6 questions that will be used to assess the
perceived workload of you activities. The completion of this questionnaire takes approximately 1
minute.
During the collecting data process you will interact with the principal investigator, Ulises Techera.
Phase 1 of the data collection procedure will take place at the beginning of the shift, during the
safety meeting with supervisors, at the company’s facilities. Phase 2 will take place in the work
truck later on your shift.
Your participation is entirely voluntarily and no compensation will be offered in return.
No material reward or punishment will be given by participating or not in the study.

What happens if I do not want to be in this research?

You can leave the research at any time and it will not be held against you.
What happens if I say yes, but I change my mind later?
You can leave the research at any time it will not be held against you.
Will being in this study help me anyway?
We cannot promise any benefits to you or others from you taking part in this research. However,
possible benefits include improving your awareness of the negative effects of fatigue in your life.

182
What happens to the information collected for the research?
Efforts will be made to limit the use and disclosure of your personal information, including
research study and medical records, to people who have a need to review this information. We
cannot promise complete secrecy. Organizations that may inspect and copy your information
include the IRB and other representatives of this organization.

The hard copies of the questionnaires that you may fill out today will be scanned and then
deleted. The computer files generated from this research will remain encrypted and password
protected in a desktop computer located in the University of Colorado at Boulder. Only the key
research personnel of this study will have access to those hard or soft files.

Signature Block for Capable Adult

Your signature documents your permission to take part in this research.

Signature of subject Date

Printed name of subject

Signature of person obtaining consent Date

Ulises Techera June 17th. 2016
Printed name of person obtaining consent IRB Approval Date

183
Appendix C: Activity Questionnaire
INSTRUCTIONS: Please answer each question. If you have questions, answer the researcher.

AGE: ___________________________________________
POSITION: ______________________________________
YEARS OF EXPERIENCE IN THIS SECTOR: _______________
GENDER: ________________________________________
Objective Questions:
1. How many hours did you sleep during the past 24hs? __________
2. How many hours do you usually sleep between 2 shifts? __________
3. Besides sleeping, how many restful hours have you had since your last shift? _________
(For example: the time you spend relaxing, watching TV, etc.)
4. Do you have individuals at home that require active care and attention?
5. How long was your last shift? __________
6. How many previous consecutive shifts have you worked? __________
7. How many hours have you had off since your last shift? __________
8. What is the start and end time of your shift today? START ________
END_________
9. If you are at work, how many hours have you been working during this shift?
__________
10. If you are at work, how frequent have your breaks been? (Example: every 2
hours)__________
11. On average how long was each break? __________
12. How long is your usual shift? __________
13. Do you have a rotating shift? If so, describe it.
E.g. One week in the morning, one week in the afternoon, and one week at night, repeat.
________________________________________________________________________
________________________________________________________________________
________________________
14. How many hours do you usually work a week? __________
15. Today, are you performing mentally demanding tasks at work?

184
 16. Today, are you performing physically demanding tasks at work?
17. Today, are you performing repetitive tasks at work?
18. Today, are you exposed to loud noises while working?
19. Today, are you exposed to vibration while working?
20. What is the approximate temperature in your work environment now? __________
21. Is the work environment well lit?
22. Did you ingest caffeine today? If YES, how much?
______________
23. Do you have a good relationship with your immediate supervisor?
24. Do you have a good relationship with your usual coworkers?
25. Do you have a condition that doesn’t allow you to rest well?

Not To a great
at all extent
Subjective Questions:
1 2 3 4 5 6 7 8 9 10
1. How mentally demanding is your job?

1 2 3 4 5 6 7 8 9 10
2. How physically demanding is your job?

1 2 3 4 5 6 7 8 9 10
3. How fatigued are you now?

4. Do you get fatigued more easily than your coworkers?

185
Appendix D: Swedish Occupational Fatigue Inventory
INSTRUCTIONS: Think about how you feel right now. To what extent do the expressions
below describe how you feel? For every expression, answer spontaneously, and mark the number
that corresponds to how you feel right now. The numbers vary between 0 (not at all) and 6 (to a
very high degree).

To a
very
Not at high
all degree
palpitations 0 1 2 3 4 5 6
lack of concern 0 1 2 3 4 5 6
worn out 0 1 2 3 4 5 6
tense muscles 0 1 2 3 4 5 6
falling asleep 0 1 2 3 4 5 6
numbness 0 1 2 3 4 5 6
sweaty 0 1 2 3 4 5 6
spent 0 1 2 3 4 5 6
drowsy 0 1 2 3 4 5 6
passive 0 1 2 3 4 5 6
stiff joints 0 1 2 3 4 5 6
indifferent 0 1 2 3 4 5 6
out of breath 0 1 2 3 4 5 6
yawning 0 1 2 3 4 5 6
drained 0 1 2 3 4 5 6
sleepy 0 1 2 3 4 5 6
overworked 0 1 2 3 4 5 6
aching 0 1 2 3 4 5 6
breathing 0 1 2 3 4 5 6
heavily
uninterested 0 1 2 3 4 5 6

Adapted from:
© The Swedish Occupational Fatigue Inventory-20
Arbetslivsinstitutet, E. Åhsberg, F. Gamberale, A. Kjellberg, 1998

186