Kamenov, 2017

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Int J Public Health

https://doi.org/10.1007/s00038-017-1047-5

ORIGINAL ARTICLE

What makes the difference in people’s lives when they have


a mental disorder?
Kaloyan Kamenov1,2 • Marı́a Cabello1,2 • Carolina Saskia Ballert3 •
Alarcos Cieza4,5,6 • Somnath Chatterji7 • Diego Rojas8 • Gloria Cerón8 •

Jerome Bickenbach3 • José Luis Ayuso-Mateos1,2,9 • Carla Sabariego5

Received: 22 November 2016 / Revised: 17 July 2017 / Accepted: 14 October 2017


Ó Swiss School of Public Health (SSPH+) 2017

Abstract also prominent factors. There was a huge overlap between


Objectives The objective of this study was to identify which the factors found relevant for mental and other non-com-
environmental factors are the most responsible for the dis- municable diseases, but a substantial variability depending
ability experienced by persons with mental disorders and on the intensity of difficulties in capacity.
whether they differ (1) from those in cardiovascular diseases, Conclusions This study challenges the appropriateness of
chronic respiratory conditions, diabetes, and cancer, and (2) disease-specific approaches and suggests that considering
depending on the capacity level-a proxy for the impact of intrinsic capacity levels is more informative than focusing on
health conditions on the health state of individuals. diagnosis alone when comparing needs and barriers that affect
Methods Nationally representative data from 12,265 adults the performance in daily life of specific groups of individuals.
in Chile collected in 2015 with the WHO Model Disability
Survey was analyzed. Keywords Environmental health  Mental health  Non-
Results The availability of personal assistance, frequency communicable disease  Disability, public health
of receiving personal assistance, and assistive devices for
mobility were the most important environmental factors
across mental and other non-communicable diseases. Per- Introduction
ception of discrimination and use of health services were
Mental disorders (MDs) are highly prevalent worldwide
and considered one of the five most burdensome non-
Alarcos Cieza and Somnath Chatterji: The views expressed in this communicable (NCD) conditions besides diabetes, cancer,
manuscript are those of the authors and do not necessarily represent cardiovascular disease (CVD) and chronic respiratory
views or policies of the World Health Organization.

& José Luis Ayuso-Mateos 6


Blindness and Deafness Prevention, Disability and
joseluis.ayuso@uam.es Rehabilitation (BDD), World Health Organization, Geneva,
Switzerland
1
Instituto de Salud Carlos III, Centro de Investigación 7
Department for Information, Evidence and Research, World
Biomédica en Red, CIBER, Madrid, Spain
Health Organization, Geneva, Switzerland
2
Department of Psychiatry, Universidad Autónoma de Madrid, 8
Department of Studies, Servicio Nacional de Discapacidad
Madrid, Spain
(Senadis), Ministerio de Desarrollo Social, Santiago, Chile
3
Swiss Paraplegic Research, Nottwil, Switzerland 9
Instituto de investigación de la Princesa, (IIS-IP), Hospital
4
Faculty of Social and Human Sciences, University of Universitario de la Princesa, C/Diego de León 62,
Southampton, Southampton, UK 28006 Madrid, Spain
5
Department of Medical Informatics, Biometry and
Epidemiology-IBE, Chair for Public Health and Health
Services Research, Research Unit for Biopsychosocial
Health, Ludwig-Maximilians-Universität München, Munich,
Germany

123
K. Kamenov et al.

diseases (Vos et al. 2013; World Health Organization however, of acknowledged importance across MDs—dis-
2014a). MDs have major social, human rights, and eco- crimination and use of health services. Studies have shown
nomic consequences. In terms of social determinants, there that stigmatization and discrimination is common among
is good evidence that MDs are distributed according to a people with MDs (Lewis et al. 2014; Thornicroft et al.
gradient of economic disadvantage across society (World 2010) and that a very low percentage of individuals with
Health Organization 2014a). Important systematic differ- MDs receive treatment, compared to other NCDs (Alonso
ences in mental health by gender, age, ethnicity, income, et al. 2007; Kessler et al. 2005; Wang et al. 2005). For
education, or geographic area of residence have been instance, across Europe 74% of those with MDs receive no
consistently reported (Campion et al. 2013; Patel and treatment compared to only 8% of people with diabetes
Kleinman 2003). On the other hand, MDs are associated (Alonso et al. 2007). The WHO has suggested that the scale
with human rights violations. People with MDs lack very of this ‘‘treatment gap’’ is disorder-specific and varies from
often basic human rights, such as shelter, food and cloth- 32% in schizophrenia to 78% in alcohol dependence
ing, and are discriminated in the fields of employment, (Kessler et al. 2005). Evidence on further EFs is still
education and housing. (World Health Organization seldom.
2014a). In terms of economic impact on society, MDs An important and direct source of broad data on EFs
cause significant socio-economic costs, estimated at US$ acting as barriers are comprehensive disability surveys that
2.5 trillion in 2010 and projected at 6.0 trillion US$ for go beyond the estimation of disability rates, such as the
2030 (Bloom et al. 2011). These costs are due to treatment WHO and World Bank (WB) Model Disability Survey
expenditures and costs associated to loss of income, and (MDS) (Loidl et al. 2016). MDS data of representative
indicate the enormous negative impact that MDs have on population samples offer an invaluable platform to study
individuals and society. built, social and attitudinal EFs most commonly associated
Persons with MDs experience high levels of disability in with MDs, to identify targets for public health interven-
their daily life (Alonso et al. 2013). The World Health tions. These data offer the possibility to determine whether
Organization (WHO) describes disability as the outcome of EFs associated with MDs have the same impact on other
the interaction between an individual with a health condi- major NCDs. Since most countries focus their public health
tion, and personal and environmental factors (EFs) (World interventions on major NCDs, this information would give
Health Organization 2001). Disability in this sense can be us a sense of the extent to which these interventions also
also referred to as performance and encompasses not only meet the needs of people with MDs. Using data from the
impairments in mental functions, such as lack of energy or implementation of the MDS as a national survey in Chile,
problems in regulating emotions, but also activities limi- this study will focus on the following objectives:
tations and restrictions in participation, such as problems in
1. to identify which factors of the environment have the
carrying out daily chores and getting a job (Cieza et al.
greatest impact on disability experienced by persons
2015). The extent of the disability a person experiences
with MDs;
varies greatly depending on the accessibility to good
2. to evaluate whether these factors are different from
quality treatment and other goods and services as well as
those most responsible for disability in persons with
on the built, political, social and attitudinal environment
CVD, chronic respiratory diseases, diabetes and
(Bostan et al. 2015; Sabariego et al. 2015a). The last
cancer, and
decades are associated with some dramatic shifts in the
3. to examine whether these factors of the environment
health and demographic profiles of populations (Vos et al.
have a different impact depending upon the capacity
2016). People are living longer and there is an increase of
level, used as a proxy of the impact of one or more
disabling chronic conditions that impact on their func-
health conditions on the health state of individuals.
tioning (Chatterji et al. 2015). Also, greater numbers of
people survive injury and illness but remain with important This study is important in the context of the global
limitations in functioning. In this sense, disability—how burden of disease and global human rights challenges that
well people live in terms of functioning in daily life—is as the MDs and other NCDs are posing. MDs are one of the
relevant as mortality—how long people live—and a major leading causes for years-lived-with-disability (YLDs)
public health priority. Health systems are challenged to worldwide responsible for more than 150 million YLDs
respond timely and efficiently to disability and not only to (Vos et al. 2016). Also, the world is facing a global human
mortality. rights emergency in mental health: people with mental
Data on the EFs that create or worsen the disability disorders very frequently do not have access to adequate
experienced by persons with MDs (World Health Organi- treatment, and institutional care—which is usually associ-
zation 2014c) is scarce, and tends to focus on single health ated with human rights violations including degrading
conditions and the impact of EFs in isolation. Two EFs are, treatment or living conditions—is still the only treatment

123
What makes the difference in people’s lives when they have a mental disorder?

option in several countries (World Health Organization randomly selected adult member of the household. The
2017). Disclosing factors of the environment impacting the present study used variables from three core MDS mod-
performance of persons with MDs and NCDs is therefore ules. Further details on the MDS design can be found
extremely important for laying the grounds for interven- elsewhere (Sabariego et al. 2015b).
tions and policy: only by broadly understanding the built, In the present study, we analyzed persons who had
political, social and attitudinal environment of persons with either a mental disorder (depression, anxiety, schizophre-
MDs and NCDs, concrete actions can be taken. nia, autism, bipolar disorder, alcohol use and drug use)
(N = 2699), a CVD (hypertension, heart or coronary dis-
ease, or heart attack) (N = 3679), diabetes (N = 1581),
Methods cancer (N = 273) or a chronic respiratory disease (chronic
bronchitis, emphysema, asthma or allergic respiratory dis-
Study design and participants ease) (N = 1244). Comorbidities were allowed between
these NCDs.
The current study is a secondary data analysis of the second
national disability survey of Chile (ENDISC II) carried out Variables
in 2015. ENDISC II aimed to determine the prevalence at
national level, to identify the main barriers and inequalities Performance and capacity
faced by people with disability and to provide evidence for
the further development of national regulations and poli- Performance targets the way people function in daily life in
cies, plans and programmes. The data collection was car- multiple functioning domains in the presence of health
ried out between June and September 2015, with a total of problems and taking into account all environmental barri-
12,015 households being surveyed with the same number ers or facilitators that constitute their real life setting.
of adults (aged 18 years and over) being interviewed. Performance is the dependent variable in the present study.
ENDISC II was representative at national, regional, and Capacity targets the level of functioning intrinsic to the
geographic (urban/rural) levels for the adult population individual in different functioning domains because of
([ 18 years). The sampling was based on the list of health problems and conditions. In this sense, capacity
households previously identified in the Survey of National reflects the intrinsic health state of an individual with one
Socioeconomic Characterization (Encuesta de Caracteri- or more health conditions and accounts for the presence of
zación Socioeconómica Nacional—Casen) in 2013 (Min- comorbidity.
isterio de Desarrollo Social 2016). A two-stage sampling Participants responded several questions regarding their
design stratified by communes and geographic area was performance and capacity in 17 functioning domains. All
used. The sample size of 12,196 households was calculated questions had five ordinal response options ranging from 1,
based on the national disability rate of 12.9% estimated in no problems, to 5, extreme problems. A single metrical
ENDISC I, assuming an absolute error of 0.4% and a rel- score, ranging from 0, no difficulties, to 100, extreme
ative error of 3.5%. In total, 11,981 households were difficulties, was created—using all questions of module
interviewed (98% response rate), resulting in a sample size 4000 for performance and all capacity questions of module
of 17,780 individuals (12,265 adults and 5515 children). 5000 for capacity—by the Chilean Statistics Bureau fol-
This study focuses only on the results for the adult lowing the WHO recommendations for the data analyses of
population. the MDS (Sabariego et al. 2015b).
ENDISC II has implemented the three core modules of
the full version of the MDS: environmental factors (module Environmental factors
3000), functioning (module 4000), and capacity and health
conditions (5000) (the MDS questionnaires are available Predictor variables were several EFs divided into (1) 12
at: http://www.who.int/disabilities/data/mds/en/). The hindering or facilitating factors of the general environment
MDS is a general population survey grounded in the like health facilities, places to socialize, transportation; (2)
International Classification of Functioning, Disability and frequency of use and need of personal assistance; (3) use of
Health (ICF) (World Health Organization 2001) and assistive devices and modifications related to mobility
operationalizes an advanced concept of disability mea- (crutches), seeing (glasses, lenses), hearing (hearing aid,
surement, focusing on performance, i.e. the degree of TV with subtitles), work (elevator), education (scanner,
execution of simple and complex actions modelled as the printer), home (door handles) and public spaces (adapted
outcome of the interaction between a health condition and public transport); (4) use of health care services or any
various environmental and personal factors. The MDS has rehabilitation service in the last 12 months; and (5) per-
a household and an individual questionnaire, filled out by a ception of discrimination in the last 12 months.

123
K. Kamenov et al.

Health conditions Descriptive statistics were obtained with SPSS, version


21 (IBM Corp 2012). RF and multiple regression analyses
Health conditions were assessed based on the Self-Ad- were performed in R Studio (Team 2015). For RF analyses,
ministered Comorbidity Questionnaire (SCQ) (Sangha the R function ‘cforest’ was used (package ‘party’) (Team
et al. 2003). This tool includes a list of country-specific 2015).
high prevalent or high priority health conditions. For each
condition, respondents were asked ‘‘1. Do you have
[DISEASE NAME]?’’ Results

Control variables Characteristics of the study population

Analysis was adjusted for age, sex, education and capacity. The majority of the sample was female (61% in CVD to
70% in cancer) and mean age ranged between 50 in MD
Statistical analyses and 61 years in CVD and diabetes (Table 1). Many par-
ticipants had more than one condition: 42% of individuals
Random forest (RF) analysis was used to identify which with MD, for example, had also a CVD.
EFs have the highest impact on performance. It is based on
the regression tree method and serves to identify variable Impact of environmental factors on performance
importance and rank predictors (Breiman 2001). The
responses of the single trees were averaged to obtain an The three EFs with highest VIM were the same for MD and
estimate of their importance in explaining variance in the the other NCD: use and frequency of use of personal
dependent variable (Hothorn et al. 2006). Control variables assistance as well as assistive devices for mobility
were forced in the model. The importance of predictors (Table 2). Further common EFs, among the ten highest,
was given by the variable importance measure (VIM) between MD and the NCDs were discrimination, hindrance
which represented the average of the frequency with which level of transportation and of shops or banks. Use of health
predictors were kept in the thousand regression trees. VIM care service was only highly ranked among the top ten EFs
provides unbiased rankings of the predictors according to for MDs. The starting model controlling for age, gender,
their association with the performance metric. Higher VIM level of education and capacity explained between 62 and
values indicate higher relevance of the variables in pre- 69% of the variance in performance across NCDs. The EFs
dicting performance. with the highest VIMs did not contribute much to the
The RF analysis was first carried out for five groups of additional explained variance, adding less than 3%.
NCDs—cancer, diabetes, CVD, respiratory diseases and
MDs. Additionally, the analyses were stratified by capacity Mild level of capacity difficulties
level. The levels of difficulties in capacity corresponded to
cut-off points previously set based on recommendations of There were not many common EFs across the five NCD
WHO for the MDS (Sabariego et al. 2015b). Persons with groups except the use of health services and use of personal
capacity scores [ 44.1 had severe difficulties in capacity, assistance (Table 3). Use of health services was the most
persons with capacity scores between 30 and 44.1 had important EF for MDs, but discrimination was not ranked
moderate, and individuals with capacity scores \ 30 had among the top ten EFs for MDs. Due to low number of
mild or no difficulties. As we were interested only in the people with cancer, the model had low power and only six
ranking of the importance of variables in a RF model, EFs were ranked. The starting model controlling for age,
sampling weights were not included. After the RF analyses gender, level of education and capacity explained between
were performed, multiple linear regressions were applied 14 and 36% of the variance in performance across the
repeatedly to determine the explained variance in perfor- NCDs. The EFs with the highest VIMs added only 2% to
mance by the independent variables with the highest VIMs. the additional explained variance.
Variables were included stepwise (in descending order of
importance) in a final model according to the VIM’s Moderate level of capacity difficulties
ranking in the RF analysis. The explained variance was
indicated with R2 and adjusted R2. The adjusted R2 was Discrimination was a common top ranked EF for all NCDs
used as a reference to assess the percentage of variance in (Table 4). Further common EFs across NCDs were use of
performance explained by the EFs with the highest personal assistance and assistive devices for seeing and
importance in the RF analysis. mobility. Use of health services was not a highly ranked EF

123
What makes the difference in people’s lives when they have a mental disorder?

Table 1 Characteristics of the sample; Chile, 2016


Mental disorders Cardiovascular Respiratory Diabetes Cancer
N = 2699 disorders N = 3679 disorders N = 1581 N = 273
N = 1244
N % N % N % N % N %

Females 1848 68 2276 61 807 64 1025 64 193 70


Educational level
No/primary school 916 34 1706 46.4 458 36.8 745 47.1 89 32.6
High school 1142 42.3 1434 39 484 38.9 616 39 121 44.3
University degree 639 23.7 537 14.6 301 24.2 220 13.9 63 23.1
Comorbidities
Mental disorders 1134 30.8 462 37.1 499 31.6 99 36.3
Cardiovascular disorders 1134 42 579 46.5 1060 67 120 44
Respiratory Disorders 462 17.1 579 15.7 234 14.8 47 17.2
Diabetes 499 18.5 1060 28.8 234 18.8 56 20.5
Cancer 99 3.7 120 3.3 47 3.8 56 3.5
Persons, who have felt discriminated 698 26 576 15.6 248 20 266 16.8 55 20.1
in the last 12 months
Persons, who have received health 2313 86 3316 90.1 1107 89 1476 93.4 264 96.7
care in the last 12 months
Persons, who have received rehabilitation 438 16.2 3197 87 204 16.4 222 14 67 24.5
service in the last 12 months
Mean SD Mean SD Mean SD Mean SD Mean SD
Age 50 17 61 15 53 19 61 14 57 16
Performance scorea 46.3 11.4 42.9 13.8 43.9 13.6 43.6 13.5 46.2 11.8
Capacity scoreb 40.4 15.1 37.3 16.2 38.7 16.4 38.1 16.4 40.9 16.1
a
Performance score: value ranges from 0 to 100, meaning the higher the score the greater the problems in the daily life performance
b
Capacity score: value range is from 0 to 100, meaning the higher the score the greater the difficulties experienced because of health related
decrements in functioning domains

for none of the NCDs. However, use of rehabilitation Discussion


services was highly ranked for both MDs and chronic
respiratory conditions. The starting model controlling for Using data of the implementation of the MDS as a national
age, gender, level of education and capacity explained up survey in Chile, we identified which factors of the envi-
to 15% of the variance in performance across the NCDs. ronment are the most responsible for the disability expe-
The EFs with the highest VIMs added less than 1% vari- rienced by persons with MDs and whether these factors are
ance. Only in the case of cancer, the EFs added 5% addi- different in the case of persons with four major NCDs—
tional variance. CVD, chronic respiratory diseases, diabetes and cancer. As
expected, discrimination and use of health services were
Severe level of capacity difficulties important factors for the overall performance in daily life
of people with MDs. EFs not commonly associated with
With the exception of diabetes, results were very similar MDs—such as personal assistance, use of assistive devices
for four NCDs (Table 5). Use and frequency of use of and the hindrance level of general environment—also had a
personal assistance as well as assistive devices for mobility considerable impact on performance. A large overlap
had the highest impact on performance. Use of health between the EFs found relevant for MDs and for other
services was not among the highly ranked EFs for any of NCDs was observed, which suggests that public health
the NCDs. The starting model explained between 33 and interventions developed for major NCDs might reach
48% of the variance in performance across the NCDs. The people with MDs as well. EFs most responsible for dis-
EFs with the highest VIMs contributed to the additional ability in MDs and in other NCDs differed considerably
explained variance with less than 2%. Only in the case of between persons with mild, moderate and severe levels of
cancer was there an additional variance (5%). difficulties in capacity. In this sense, public health actions

123
K. Kamenov et al.

Table 2 Results from the random forest and repeated linear regression models showing the environmental factors with the highest impact on
performance; Chile, 2016
Category Variable Mental disorders Cardiovascular Respiratory Cancer N = 273 Diabetes
N = 2699 disorders disorders N = 1581
N = 3679 N = 1244
VIM R2adj VIM R2adj VIM R2adj VIM R2adj VIM R2adj
(Rank) (Rank) (Rank) (Rank) (Rank)

Control Age 0.2 0.618 0.4 0.687 0.2 0.683 0.04 0.683 1.2 0.679
Sex 1.3 0.2 4.2 0.5 0,5
Level of education 0.7 0.8 1.9 2.3 0.5
Capacity 74.3 129.8 109.7 71.9 115.7
General EF Health facilities 0.1 (19) 0.57 0.3 (18) 0.679 0.8 (11) 0.678 0.1 (18) 0.632 0.4 (16) 0.678
Places to socialize 1.1 (4) 0.619 1.03 (8) 0.679 0.5 (14) 0.673 0.5 (11) 0.684 1.9 (7) 0.672
Shops/banks 0.6 (7) 0.619 1.1 (7) 0.682 0.5 (15) 0.671 1.1 (6) 0.689 3.04 (4) 0.675
Worship 0.1 (29) 0.567 0.3 (17) 0.68 0.3 (18) 0.682 0.4 (13) 0.669 0.5 (15) 0.677
Transportation 0.8 (5) 0.62 1.4 (4) 0.686 0.9 (7) 0.679 1.7 (5) 0.685 2.5 (6) 0.673
Dwelling 0.5 (10) 0.619 0.8 (9) 0.679 0.6 (13) 0.679 0.6 (10) 0.693 0.8 (12) 0.674
Terrain/climate 0.5 (11) 0.619 0.5 (14) 0.682 0.8 (12) 0.678 0.3 (14) 0.668 0.8 (11) 0.674
Lighting 0.4 (13) 0.619 0.4 (16) 0.682 0.08 (9) 0.677 0.2 (16) 0.669 0.1 (18) 0.677
Noise 0.3 (16) 0.571 0.3 (19) 0.679 0.09 (21) 0.591 1.1 (7) 0.69 0.1 (19) 0.677
Crowds 0.3 (15) 0.57 0.6 (12) 0.68 0.4 (16) 0.673 0.9 (8) 0.694 0.9 (10) 0.674
Workplace 0.4 (14) 0.571 0.2 (20) 0.632 0.09 (20) 0.589 0.2 (17) 0.634 0.09 (20) 0.626
Educational 0.03 (23) 0.567 0 (23) 0.632 - 0.02 0.59 0 (24) 0.617 0 (24) 0.624
institution (24)
Pers. assistance Personal Assistance 3,9 (1) 0.619 8.03 (1) 0.687 8.0 (1) 0.682 4.9 (3) 0.685 10.9 (1) 0.679
Frequency 2.9 (2) 0.619 7.06 (2) 0.687 6.2 (2) 0.681 11.2 (1) 0.686 8.8 (2) 0.679
Health care Health Service 0.7 (6) 0.621 0.1 (21) 0.632 - 0.03 0.59 0 (22) 0.617 0.2 (17) 0.678
(23)
Rehabilitation 0.09 (21) 0.567 0.4 (15) 0.682 0.3 (17) 0.673 0.1 (19) 0.628 0.09 (21) 0.626
Attitudes of Discrimination 0.5 (9) 0.619 1.2 (6) 0.684 0.9 (8) 0.678 0.7 (9) 0.693 2.6 (5) 0.673
others
Assistive Mobility 2.4 (3) 0.621 3.2 (3) 0.688 2.8 (3) 0.682 5.8 (2) 0.686 3.9 (3) 0.679
devices Seeing 0.6 (8) 0.621 0.5 (13) 0.682 1.3 (5) 0.682 0.4 (12) 0.684 0.5 (13) 0.674
Hearing 0.2 (17) 0.57 0.6 (11) 0.679 0.8 (10) 0.678 0.3 (15) 0.671 0.5 (14) 0.674
Work 0.07 (22) 0.567 0.05 (22) 0.632 0.002 (22) 0.591 0 (20) 0.623 0.05 (22) 0.625
Education - 0.01 0.567 0 (24) 0.632 0.1 (19) 0.683 0 (23) 0.617 0 (23) 0.624
(24)
Home 0.5 (12) 0.619 0.7 (10) 0.679 0.9 (6) 0.682 0.01 (21) 0.621 1.02 (9) 0.671
Public spaces 0.2 (18) 0.571 1.2 (5) 0.686 1.7 (4) 0.681 3.2 (4) 0.685 11 (8) 0.672
The most important EFs are marked in bold
VIM variable importance measures estimated with random forest regression, R2adj R2 adjusted showing the increase in explained variance
calculated with classical multiple linear regression analyses by adding the determinants stepwise in descending rank order into the model, EF
environmental factors, Control all models were controlled for age, gender, level of education and capacity. For gender the reference category was
male, for education—no/primary school)

tailored to MDs or other NCDs would gain in precision by This shows that continuous and coordinated care in the
considering the capacity of individuals when selecting community is a key element to prevent and overcome the
specific needs and barriers as targets. disability level associated with MDs, particularly for per-
Absence of personal assistance and assistive devices for sons with moderate to severe levels of capacity difficulties.
mobility was the most responsible factors for the disability While these findings sound intuitive for the rather ‘‘phys-
experienced both by persons with MDs and other NCDs. ical’’ NCDs, they are surprising for MDs. The strong

123
What makes the difference in people’s lives when they have a mental disorder?

Table 3 Results from a random forest analysis showing the ten environmental factors with the highest impact on performance in people with
mild levels of capacity difficulties; Chile, 2016
Rank Mental disorders N = 643 Cardiovascular disorders Respiratory disorders Cancer N = 69 Diabetes N = 500
N = 1230 N = 374
Environmental VIM Environmental VIM Environmental VIM Environmental VIM Environmental VIM
factor factor factor factor factor

1 Use of health 4.3 Assistive devices 1.4 Hindrance level 0.6 Assistive 1.07 Hindrance level 4.8
services for seeing of health devices for of shops/banks
facilities seeing
2 Hindrance level of 1.8 Hindrance level of 1.2 Hindrance level 0.5 Hindrance level 1.4 Hindrance level 2.8
public transport public transport of dwelling of shops/ of crowds
banks
3 Hindrance level of 0.9 Personal assistance 0.9 Hindrance level 0.5 Hindrance level 1.4 Discrimination 2.2
lighting in of public of public
surroundings transport transport
4 Hindrance level of 0.9 Frequency personal 0.8 Hindrance level 0.5 Use of 0.3 Personal 2.1
places to assistance of workplace rehabilitation assistance
socialize services
5 Assistive devices 0.4 Hindrance level of 0.6 Hindrance level 0.2 Discrimination 0.01 Hindrance level 1.1
for seeing lighting in of places to of health
surroundings worship facilities
6 Personal assistance 0.2 Assistive devices in 0.6 Personal 0.2 Hindrance level 0.001 Hindrance level 1.1
public spaces assistance of dwelling of places to
socialize
7 Hindrance level of 0.1 Discrimination 0.4 Hindrance level 0.1 Hindrance level 0.6
workplace of noise of
terrain/climate
8 Hindrance level of 0.1 Hindrance level of 0.3 Assistive 0.1 Frequency of 0.6
health facilities noise devices in personal
public spaces assistance
9 Assistive devices 0.1 Health service 0.3 Assistive 0.1 Use of health 0.6
for mobility devices for services
education
10 Hindrance level of 0.1 Hindrance level of 0.3 Hindrance level 0.1 Hindrance level 0.4
noise places to of shops/banks of public
socialize transport

research focus in MDs on clinical (severity of disease), health care was lower than in other NCDs but still very
psychological (personality, neuroticism), social (social high (86%) compared to published estimates showing that
adjustment, social support), cognitive and economic factors between 35.5 and 85.4% of serious cases in various
as determinants of disability in MDs, compared to the scant countries usually do not receive treatment (Demyttenaere
attention given to broader EFs so far (Harvey and Strassnig et al. 2004). This high percentage might be associated with
2012; Rytsala et al. 2006), may explain this. Another a high number of comorbidities in people with MDs, which
explanation may be the very frequent presence of comor- might lead to an easier access to care, or simply reflect the
bidities with ‘‘physical’’ conditions. broad coverage and accessibility of the Chilean’s health
This study, consistent with previous literature (Farrelly system (Missoni and Solimano 2010). It is important to
et al. 2014; Thornicroft et al. 2010), stresses the importance stress, however, that the question used to access use of
of discrimination as a determinant of the level of perfor- health care does not differentiate the kind of treatment
mance of individuals with NCDs and triggers the question received. In addition, the study collected self-reported
whether interventions targeting stigma towards MDs could information on the use of health services.
inform similar interventions for other NCDs. This study An important finding of this study is that the EFs most
also confirms the importance of use of health care services responsible for disability in MDs and NCDs differ con-
to MDs. The percentage of persons with MDs receiving siderably between persons with mild, moderate and severe

123
K. Kamenov et al.

Table 4 Results from a random forest analysis showing the ten environmental factors with the highest impact on performance in people with
moderate levels of capacity difficulties; Chile, 2016
Rank Mental disorders Cardiovascular disorders Respiratory disorders Cancer N = 86 Diabetes N = 504
N = 954 N = 1176 N = 381
Environmental VIM Environmental VIM Environmental VIM Environmental VIM Environmental VIM
factor factor factor factor factor

1 Discrimination 0.4 Use of personal 0.4 Hindrance level 0.9 Assistive devices 0.6 Hindrance level of 0.5
assistance of in public public transport
terrain/climate spaces
2 Hindrance level of 0.3 Frequency of 0.3 Assistive devices 0.8 Hindrance level 0.3 Use of personal 0.5
workplace personal for mobility of crowds assistance
assistance
3 Assistive devices 0.3 Discrimination 0.3 Discrimination 0.4 Discrimination 0.3 Frequency of 0.4
for mobility personal
assistance
4 Assistive devices 0.2 Assistive devices 0.3 Use of 0.2 Assistive devices 0.1 Discrimination 0.3
for seeing for mobility rehabilitation for mobility
5 Assistive devices 0.1 Assistive devices 0.2 Hindrance level 0.2 Hindrance level 0.1 Hindrance level of 0.3
for work for seeing of public of workplace shops/banks
transport
6 Hindrance level of 0.05 Hindrance level 0.2 Assistive devices 0.1 Hindrance level 0.05 Hindrance level of 0.1
noise of shops/banks in public spaces of health places to
facilities socialize
7 Hindrance level of 0.05 Hindrance level 0.2 Assistive devices 0.1 Assistive devices 0.01 Hindrance level of 0.1
places to of public for seeing for seeing crowds
worship transport
8 Hindrance level of 0.05 Hindrance level 0.1 Assistive devices 0.1 Hindrance level 0.01 Hindrance level of 0.1
shops/banks of for home of dwelling dwelling
terrain/climate
9 Hindrance level of 0.03 Assistive devices 0.1 Hindrance level 0.03 Assistive devices 0.05
public transport for work of noise for mobility
10 Hindrance level of 0.03 Use of 0.1 Use of health 0.02 Use of 0.05
lighting rehabilitation services rehabilitation

levels of difficulties in capacity. Some EFs, for instance use experience specific symptoms or functioning limitations
of health services, were relevant for persons with mild or associated with a health condition and report a diagnosis
moderate levels of capacity difficulties across MDs and without actually having it and vice versa. On the other
other NCDs, but not for severe cases, whereas personal hand, self-reported diagnostic tools are a cost-effective way
assistance was disclosed as a highly relevant factor for to obtain health status information in epidemiological
persons with moderate and severe difficulties in capacity, studies. Previous studies have showed that self-reported
but not for mild cases. Our results illustrate that consid- diagnoses were equally able to predict quality of life
ering capacity levels when comparing needs and barriers of problems in comparison with information collected in
specific groups of people is more accurate than focusing medical records (Olomu et al. 2012). The instrument used
solely on diagnosis. in ENDISC II has been proven especially useful in studies
This study has to be seen in the light of some limitations. based on general populations and in settings in which
First, the diagnosis of health conditions was not based on a medical records are not available. Second, we included
standardized diagnostic interview. ENDISC II used a self- people with comorbidities in the analyses. Evidence shows
reported diagnosis based on the Self-Administered that comorbidities across NCDs is a common phenomenon
Comorbidity Questionnaire (SCQ) (Sangha et al. 2003), (Prince et al. 2007) and reflect the real life experience of
which included a list of country-specific high prevalent or persons with MDs and other NCDs, so that excluding them
high priority health conditions and impairments. There are would introduce a selection bias. By basing our analysis on
certain pros and cons of using a self-reported diagnosis. intrinsic capacity, we accounted for the comorbidities in
The main problem is the lack of accuracy. People can the NCDs. Lastly, RF analysis, though a powerful tool for

123
What makes the difference in people’s lives when they have a mental disorder?

Table 5 Results from a random forest analysis showing the ten environmental factors with the highest impact on performance in people with
severe levels of capacity difficulties; Chile, 2016
Rank Mental disorders Cardiovascular disorders Respiratory disorders Cancer N = 118 Diabetes N = 576
N = 1101 N = 1271 N = 489
Environmental VIM Environmental VIM Environmental VIM Environmental VIM Environmental VIM
factor factor factor factor factor

1 Frequency of 1.8 Frequency of 2.1 Frequency of 2.9 Frequency of 4.8 Frequency of 1.8
personal personal personal personal personal
assistance assistance assistance assistance assistance
2 Assistive devices 1.5 Assistive devices 1.8 Assistive devices 1.5 Hindrance level 1.0 Assistive devices 1.7
for mobility for mobility for mobility of for mobility
terrain/climate
3 Use of personal 1.1 Use of personal 1.0 Use of personal 1.5 Use of personal 0.9 Hindrance level 1.3
assistance assistance assistance assistance of public
transport
4 Hindrance level of 0.8 Hindrance level of 0.9 Hindrance level 1.1 Assistive 0.8 Hindrance level 1.0
shops/banks shops/banks of shops/banks devices for of places to
mobility socialize
5 Hindrance level of 0.7 Hindrance level of 0.7 Discrimination 0.7 Hindrance level 0.6 Use of personal 0.9
places to places to of dwelling assistance
socialize socialize
6 Hindrance level of 0.7 Hindrance level of 0.7 Hindrance level 0.5 Hindrance level 0.4 Hindrance level 0.8
terrain/climate terrain/climate of places for of shops, of dwelling
worship banks
7 Hindrance level of 0.6 Hindrance level of 0.5 Hindrance level 0.5 Hindrance level 0.3 Hindrance level 0.6
public transport public transport of public of health of
transport facilities terrain/climate
8 Hindrance level of 0.3 Hindrance level of 0.6 Assistive devices 0.4 Discrimination 0.2 Hindrance level 0.5
dwelling dwelling for home of places to
worship
9 Hindrance level of 0.4 Hindrance level of 0.4 Hindrance level 0.3 Assistive 0.2 Hindrance level 0.4
lighting in places for of places to devices for of crowds
surroundings worship socialize seeing
10 Discrimination 0.3 Hindrance level of 0.4 Hindrance level 0.3 Hindrance level 0.04 Hindrance level 0.2
lighting in of dwelling of workplace of shops/banks
surroundings

ranking EFs for the level of performance of individuals Conclusions


does not provide information on the direction of the
association. Four key messages come out of the present study. First,
Two strengths should be as well mentioned. First, it is adopting an unbiased, comprehensive approach that takes
the first study to use a large general population sample to into account a range of EFs, encompassing the built,
explore the role of a wide range of EFs on the disability political, social and attitudinal environment, is very
level experienced by persons with MDs and other NCDs. important, not only to corroborate known determinants but
The study provides clear and reliable information about also to disclose other factors impacting the performance of
potential public health intervention targets to approach persons with MDs. Second, our results definitely show that
disability in MDs and other NCDs and can inform the persons with MD are in need, not only of emotional or
implementation of the WHO Global Disability Action Plan instrumental support, but also of physical support. Third,
2014–2021 (World Health Organization 2014b). Secondly, there is a complete overlap between MDs and NCDs in the
the present paper confirms the value of the implementation factors most responsible for the disability experienced in
of a comprehensive general population disability survey day-to-day life. Taken together, this suggests that public
like the MDS for generating evidence on MDs and NCDs. health interventions developed for CVD, chronic respira-
tory diseases, diabetes and cancer may well meet, at least

123
K. Kamenov et al.

partially, the needs of people with MDs. Finally, the large Kessler RC et al (2005) Prevalence and treatment of mental disorders,
overlap between MDs and NCDs contrasts with the dif- 1990 to 2003. N Engl J Med 352:2515–2523. https://doi.org/10.
1056/NEJMsa043266
ferences between persons with mild, moderate and severe Lewis TT, Williams DR, Tamene M, Clark CR (2014) Self-reported
levels of difficulties in capacity. This fact questions the experiences of discrimination and cardiovascular diseas. Curr
validity of disease-specific approaches and suggests that Cardiovasc Risk Rep 8:365. https://doi.org/10.1007/s12170-013-
looking at capacity levels when comparing needs and 0365-2
Loidl V, Oberhauser C, Ballert C, Coenen M, Cieza A, Sabariego C
barriers of specific groups is more informative than (2016) Which environmental factors have the highest impact on
focusing on diagnosis alone. the performance of people experiencing difficulties in capacity?
Int J Environ Res Public Health. https://doi.org/10.3390/
Acknowledgements The authors would like to thank Ms. Laura ijerph13040416
Marques for her assistance in the statistical analyses. Ministerio de Desarrollo Social (2016) Informe Metodológico: Marco
de Referencia Conceptual, Diseño del Instrumento y Medición
Compliance with ethical standards de la discapacidad en la población adulta (18 años y más). http://
www.senadis.gob.cl/pag/306/1570/publicaciones. Accessed 11
Conflict of interest The authors declare that they have no conflict of July 2017
interest. Missoni E, Solimano G (2010) Towards Universal health coverage:
the Chilean experience world health report. World Health
Organization
Olomu AB, Corser WD, Stommel M, Xie Y, Holmes-Rovner M
(2012) Do self-report and medical record comorbidity data
References predict longitudinal functional capacity and quality of life health
outcomes similarly? BMC Health Serv Res 12:398
Alonso J et al (2007) Population level of unmet need for mental Patel V, Kleinman A (2003) Poverty and common mental disorders in
healthcare in Europe. Br J Psychiatry 190:299–306. https://doi. developing countries. Bull World Health Organ 81:609–615
org/10.1192/bjp.bp.106.022004 Prince M, Patel V, Saxena S, Maj M, Maselko J, Phillips MR,
Alonso J, Chatterji S, He Y (2013) The burdens of mental disorders: Rahman A (2007) No health without mental health. Lancet
global perspectives from the WHO World Mental Health 370:859–877. https://doi.org/10.1016/S0140-6736(07)61238-0
Surveys. Cambridge University Press, Cambridge R Core Team (2015) R: a language and environment for statistical
Bloom DE, Cafiero E, Jané-Llopis E, Abrahams-Gessel S, Bloom LR, computing. R Foundation for Statistical Computing, Vienna.
Fathima S, Feigl AB, Gaziano T, Hamandi A, Mowafi M, http://www.R-project.org/
O’Farrell D (2011) The global economic burden of non- Rytsala HJ, Melartin TK, Leskela US, Lestela-Mielonen PS, Sokero
communicable diseases. World Economic Forum, Geneva TP, Isometsa ET (2006) Determinants of functional disability
Bostan C, Oberhauser C, Stucki G, Bickenbach J, Cieza A (2015) and social adjustment in major depressive disorder: a prospective
Which environmental factors are associated with lived health study. J Nerv Ment Dis 194:570–576. https://doi.org/10.1097/01.
when controlling for biological health?—A multilevel analysis. nmd.0000230394.21345.c4
BMC Public Health 15:508. https://doi.org/10.1186/s12889-015- Sabariego C et al (2015a) Determinants of psychosocial difficulties
1834-y experienced by persons with brain disorders: towards a ‘hori-
Breiman L (2001) Machine learning. Netherlands, Vol 45 zontal epidemiology’ approach. PLoS One 10:e0141322
Campion J, Bhugra D, Bailey S, Marmot M (2013) Inequality and Sabariego C et al (2015b) Measuring disability: Comparing the
mental disorders: opportunities for action. Lancet 382:183 impact of two data collection approaches on disability rates. Int J
Chatterji S, Byles J, Cutler D, Seeman T, Verdes E (2015) Health, Environ Res Public Health 12:10329–10351
functioning, and disability in older adults—present status and Sangha O, Stucki G, Liang MH, Fossel AH, Katz JN (2003) The self-
future implications. Lancet 385:563–575 administered comorbidity questionnaire: a new method to assess
Cieza A et al (2015) Understanding the impact of brain disorders: comorbidity for clinical and health services research. Arthritis
towards a ‘horizontal epidemiology’ of psychosocial difficulties Rheum 49:156–163. https://doi.org/10.1002/art.10993
and their determinants. PLoS One 10:e0136271. https://doi.org/ Thornicroft G, Rose D, Mehta N (2010) Discrimination against people
10.1371/journal.pone.0136271 with mental illness: what can psychiatrists do? Adv Psychiatr
Demyttenaere K et al (2004) Prevalence, severity, and unmet need for Treat 16(1):53–59. https://doi.org/10.1192/aptbp107004481
treatment of mental disorders in the World Health Organization Vos T et al (2013) Years lived with disability (YLDs) for 1160
World Mental Health Surveys. JAMA 291:2581–2590. https:// sequelae of 289 diseases and injuries 1990–2010: a systematic
doi.org/10.1001/jama.291.21.2581 analysis for the Global Burden of Disease Study 2010. Lancet
Farrelly S et al (2014) Anticipated and experienced discrimination 380:2163–2196
amongst people with schizophrenia, bipolar disorder and major Vos T et al (2016) Global, regional, and national incidence,
depressive disorder: a cross sectional study. BMC Psychiatry prevalence, and years lived with disability for 310 diseases and
14:157. https://doi.org/10.1186/1471-244X-14-157 injuries, 1990–2015: a systematic analysis for the Global Burden
Harvey PD, Strassnig M (2012) Predicting the severity of everyday of Disease Study 2015. Lancet 388:1545–1602
functional disability in people with schizophrenia: cognitive Wang PS, Lane M, Olfson M, Pincus HA, Wells KB, Kessler RC
deficits, functional capacity, symptoms, and health status. World (2005) Twelve-month use of mental health services in the United
Psychiatry 11:73–79 States: results from the National Comorbidity Survey Replica-
Hothorn T, Buhlmann P, Dudoit S, Molinaro A, van der Laan MJ tion. Arch Gen Psychiatry 62:629–640. https://doi.org/10.1001/
(2006) Survival ensembles. Biostatistics 7:355–373. https://doi. archpsyc.62.6.629
org/10.1093/biostatistics/kxj011 World Health Organization (2001) International classification of
IBM Corp (2012) IBM SPSS Statistics for Windows, Version 21.0. functioning, disability and health (ICF). World Health Organi-
IBM Corp, Armonk, NY zation, Geneva

123
What makes the difference in people’s lives when they have a mental disorder?

World Health Organization (2014a) Social determinants of mental World Health Organization (2017) Mental health, human rights &
health. World Health Organization, Geneva legislation. A global human rights emergency in mental health.
World Health Organization (2014b) WHO global disability action World Health Organization. http://www.who.int/mental_health/
plan 2014–2021. World Health Organization, Geneva policy/legislation/en/. Accessed 7 July 2017
World Health Organization (2014c) Global status report on noncom-
municable diseases 2014. World Health Organization, Geneva

123

You might also like