Epidemiological Approach For Malaria Control

Second edition
Epidemiological approach
TRAINING MODULE ON MALARIA CONTROL
for malaria control

guide for participantS
Epidemiological approach
for malaria control
MALARIA CONTROL
GUIDE FOR PARTICIPANTS

Second edition
TRAINING MODULE ON
WHO Library Cataloguing-in-Publication Data
Epidemiological approach for malaria control – 2nd ed.
2 v.
Contents: Guide for tutors – Guide for participants
First edition published as technical document WHO/CDS/CPE/SMT/2000.1
1.Malaria – prevention and control. 2.Malaria – epidemiology. 3.Teaching materials.
I.World Health Organization.
ISBN 978 92 4 150601 4 (NLM classification: WC 765)
The revision of this document was made possible through the

Russian Federation grant for malaria capacity development in Africa.
© World Health Organization 2013

All rights reserved. Publications of the World Health Organization are available on the WHO web site (www.who.int) or can be purchased
from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791
4857; e-mail: [email protected]). Requests for permission to reproduce or translate WHO publications – whether for sale or for non-
commercial distribution – should be addressed to WHO Press through the WHO web site (www.who.int/about/licensing/copyright_
form/en/index.html).
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever
on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or
concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may
not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the
World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names
of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication.
However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the
interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising
from its use.
Please consult the WHO Global Malaria Programme web site for the most up-to-date version of all documents (www.who.int/malaria).
Red blood cells: ©Ingram Publishing
Cover Design by Paprika-Annecy.com
Printed in Malta
iii
Contents
Foreword.............................................................................................................................................................v
Abbreviations.....................................................................................................................................................vi
Acknowledgements ..........................................................................................................................................vii
Development of the module ..............................................................................................................................ix
Introduction........................................................................................................................................................1
Part 1: Introduction to basic epidemiology and statistics

Learning Unit 1: Introduction to epidemiology...............................................................................................5
Learning Unit 2: Ratios, proportions and rates............................................................................................13
Learning Unit 3: Data presentation: tables, graphs and charts...................................................................21
Learning Unit 4: Measures of central tendency............................................................................................35
Learning Unit 5: Measures of variability and normal distribution................................................................43
Learning Unit 6: Assessing the accuracy of a test.......................................................................................51
Part 2: Applied malaria epidemiology

Learning Unit 7: Understanding malaria at regional and global levels........................................................57
Learning Unit 8: Determinants of malaria distribution.................................................................................61
Learning Unit 9: The life-cycle of the malaria parasite ................................................................................65
Learning Unit 10: The life-cycle of the vector ................................................................................................69
Learning Unit 11: Natural history of malaria in the human host ...................................................................73
Learning Unit 12: Intensity of malaria transmission......................................................................................79
Part 3: Malaria surveillance, monitoring and evaluation

Learning Unit 13: Overview of malaria programmes phases.........................................................................89
Learning Unit 14: Surveillance system............................................................................................................93
Learning Unit 15: Indicators for monitoring and evaluation of malaria control programmes ....................103
Learning Unit 16: Data collection methods...................................................................................................111
Learning Unit 17: Burden of malaria.............................................................................................................123
Part 4: Prevention and control of malaria epidemics

Learning Unit 18: Introduction to malaria epidemics...................................................................................131
Learning Unit 19: Early warning, early detection, notification and verification of a malaria epidemic ......135
Learning Unit 20: Prevention and early response to confirmed malaria epidemics....................................141
Learning Unit 21: Post-epidemic assessment and preparedness plan of action.........................................149
iv EPIDEMIOLOGICAL APPROACH FOR MALARIA CONTROL: GUIDE FOR PARTICIPANTS
Annex 1. Core surveillance indicators for malaria control...........................................................................153

Annex 2. Suggested register for community health workers, health posts and outpatient
departments of health centres and hospitals................................................................................159
Annex 3. Sheet for tallying outpatient attendance at health centres and hospitals....................................160
Annex 4. Daily and weekly records of outpatient attendance at health centres and hospitals...................161
Annex 5. Discharge register for inpatient departments of health centres and hospitals...........................162
Annex 6. Reports from health posts and community health workers to health facilities............................163
Annex 7. Reports from health facilities to the district level.........................................................................164
Annex 8. Line lists of inpatient malaria cases and deaths to be reported to district level in
low-transmission seetings.............................................................................................................166
Annex 9. Line lists of all confirmed malaria cases to be reported at district level in
low-transmission settings..............................................................................................................167
Annex 10. Example of supervisory checklist for malaria surveillance for countries with
high to moderate transmission......................................................................................................168
Annex 11. Example of questionnaire for malaria post-epidemic assessment...............................................169
v
Foreword
Malaria is a major global public health problem and a leading cause of morbidity and
mortality in many countries. Malaria caused an estimated 219 (range 154–289) million
cases and 660 000 (range 490 000–836 000) deaths in 2010. Approximately 80% of the
cases and 90% of the deaths occur in Africa while the remaining cases and deaths occur
mainly in the South-East Asia and Eastern Mediterranean Regions.1 For the most recent
figures on burden of malaria, search for the “World Malaria Report” available on WHO/
GMP websites (http://www.who.int/malaria/en/).
The World Health Assembly and Roll Back Malaria (RBM) targets for malaria control and
elimination are to achieve at least a 75% reduction in malaria incidence and deaths by 2015.
Elimination of malaria is defined as the reduction to zero of the incidence of locally acquired
infection by human malaria parasites in a defined geographical area as a result of deliberate
efforts. Elimination programmes require more technical malaria expertise than standard
malaria control programmes, and require by national expertise in malaria epidemiology and
entomology.
To achieve the objectives of malaria control and elimination programmes, appropriately
planned and targeted delivery of essential malaria interventions is critical, including: diagnos-
tic testing of all suspected malaria cases and prompt treatment of confirmed infections with
effective artemisinin-based combination therapy (ACT); chemoprevention of malaria in preg-
nant women (Intermittent preventive treatment during pregnancy – IPTp), infants (Intermit-
tent preventive treatment in infants – IPTi) and children (Seasonal malaria chemoprevention
– SMC), where appropriate; and application of appropriate vector control interventions, par-
ticularly the use of insecticide- treated nets (ITNs/LLINs) and indoor residual spraying (IRS).
This training module on the epidemiological approach for malaria control has been developed
to support the staff involved in the planning and management and in the monitoring and
evaluation of malaria control and elimination programmes.
WHO (2012). World malaria report. Geneva, World Health Organization. http://www.who.int/malaria/publications/world_
1.
malaria_report_2012/en/index.html
vi
Abbreviations
ABER Annual blood examination rate
ACT Artemisinin-based combination therapy
ANC Antenatal clinic
API Annual parasite index
CFR Case fatality rate
CSP Circumsporozoite protein
DALY Disability-adjusted life year
DHS Demographic and health survey
EIR Entomological inoculation rate
ELISA Enzyme-linked immunosorbent assay
GIS Geographical information system
HBI Human blood index
IEC Information, education and communication
IPT Intermittent preventive treatment
IRS Indoor residual spraying
ITN Insecticide-treated mosquito nets
IVC Integrated vector control
KAP Knowledge, attitudes and practices
LLIN Long-lasting insecticidal nets
MDA Mass-drug administration
MICS Multiple indicator cluster survey
MIS Malaria indicator survey
NGO nongovernmental organization
NMCP National malaria control programme
PCR Polymerase chain reaction
POPs Persistent organic pollutants
PPOA Preparedness plan of action
RBM Roll Back Malaria
RCT Randomized clinical trial
RDT Rapid diagnostic test
SPR Slide positivity rate
SMPH Summary measures of population health
TPR Test positivity rate
WHO World Health Organization
vii
Acknowledgements
This module was produced by the WHO Global Malaria Programme (GMP), with participa-
tion of current and former staff from WHO Headquarters and Regional Offices. WHO grate-
fully acknowledges the following experts who contributed to the development of this docu-
ment:
▶▶ M. Aregawi, N. Binkin, P.F. Beales, R.L Kouznetsov, F.A. Rio, and M.C. Thuriaux contrib-
uted to the development of the earlier WHO training module on Introduction to Basic Epide-
miology and Statistics, which is the basis for Learning Units 1, 2, 3, 4, 5 and 6 in Part I of this
module.
▶▶ L. Molineaux developed and field tested the content of the earlier WHO module on Applied
Malaria Epidemiology which is the basis for Part II of this module, and M. Aregawi, P.F.
Beales, A. Bosman, A. Haghdoost, S. Izadi, J.M. Jewsbury, R.L. Kousnetsov, A. Raeisi and
F.A. Rio gave their feedbacks for purpose of updating this Part. E. Renganathan contributed
to the earlier module.
▶▶ C. Delacollette, together with M. Aregawi and J. Sagbohan, developed the contents of the
trial version of the module on Prevention and Control of Malaria Epidemics which is the
basis for Part IV of this module.
▶▶ R. Majdzadeh led the updating of the current module and developed Learning Unit 7 of Part
I, Learning Units 8 and 9 in Part II and Part III on Malaria Surveillance, Monitoring and
Evaluation. B. Yazdizadeh contributed to the Learning Units 8 and 9 in Part II.
▶▶ The technical experts who guided the review and updating process of this module: A. A.
A. Adeel (King Saud University, Saudi Arabia), M. Sh. Al-Zedjali (Malaria Epidemiology,
Ministry of Health, Oman), A. Kondrashin (former WHO staff, Russian Federation),
B. Ayivi (National University Hospital, Benin), C. Hugo (ACT Malaria Foundation Inc,
Philippines), A. Baranova (Martzinovsky Institute of Medical Parasitology & Tropical
Medicine, Russian Federation), P. F. Beales (former WHO staff, United Kingdom), A. Beljaev
(Russian Medical Academy for Moscow, Russian Federation), S. Elbushra (University of
Gezira, Sudan), K. Kolaczinski (Malaria Consortium Africa, Uganda), S. Lutalo (Harare
Central Hospital, Zimbabwe), R. Majdzadeh (Tehran University of Medical Sciences, Iran),
E. M. Malik (Federal Ministry of Health, Sudan), P. S. Mapunda (Centre for Enhancement
of Effective Malaria Interventions, Tanzania), R. Mintcheva (Center of Infectious and
Parasitic Diseases, Bulgaria), O. Mokuolu (University of Ilorin Teaching Hospital, Nigeria),
E. Morozov (Martzinovsky Institute of Medical Parasitology & Tropical Medicine, Russian
Federation), A. Mwakilasa (Consultant, Tanzania), J. B. Ouedraogo (Direction Regionale
de l'Ouest, Burkina Faso), V. Sergiev (Martzinovsky Institute of Medical Parasitology &
Tropical Medicine, Russian Federation) and H. Vatandoost (Tehran University of Medical
Sciences, Iran).
▶▶ WHO staff who contributed to the technical content of the module over the period of its
development: H.Y. Atta, A. Bosman, K. Carter, K. Cham, F. Da Silveira, C. Delacollette,
G. A. Ki-Zerbo, M. Lynch, K. Mendis, B. Mulenda, R. Newman, P. Olumese, A. Rietveld,
viii EPIDEMIOLOGICAL APPROACH FOR MALARIA CONTROL: GUIDE FOR PARTICIPANTS
P. Ringwald, M. Robalo, A. Schapira, A. Teklehaimanot, M. Warsame, W. Were, R. Williams,

and G. Zamani.
▶▶ D. Chandramohan who reviewed the module as an independent expert.
WHO also thanks the participants, tutors and facilitators of several national and international
courses for their comments and suggestions during the field-testing which led to improvements
to the module.
The revision process was coordinated by M. Warsame; technical editing of the module was by
L.J. Martinez.
The revision of the module was made possible through the Russian Federation grant for malaria
capacity development in Africa.
ix
Development of the module

The content of the module is based on the current WHO guidelines and other evidence-based
technical documents.
This training module are arranged in four parts. The main framework of each part was
developed at different times, and has been evaluated and reviewed in various separate
education programmes. The decision to combine these four parts into a training module was
made through technical expert meetings held in 2008 and 2009.
The training module was updated under the guidance of technical experts representing
malaria training and academic institutions, malaria researchers, country programme manag-
ers, and WHO regional offices, who guided the process of reviewing and updating the module.
The process included the following steps:
▶▶ Three consultations of technical experts (7–9 April 2008; 14–16 October 2008 and 15–17
April 2009) were held to review the existing WHO training materials on the epidemiologi-
cal approach for malaria control and elimination, and to identify areas for update in view of
the development of new tools, technologies and strategies for malaria control and the chang-
ing disease epidemiology.
▶▶ Technical experts were commissioned to incorporate the recommended updates in the
module.
▶▶ The revised module was then reviewed for content and completeness by the technical
experts, the WHO technical staff and additional external experts in malaria epidemiology.
▶▶ The updated module was field-tested in several national and international courses.
▶▶ Based on feedback from field tests, and in consultation with technical experts, the text was
finalized for publication.
1
Introduction
The planning and implementation of a malaria control programme must be based on
epidemiological analysis and application of interventions suitable to specific local malaria
situations. Health workers and all stakeholders involved need to have a sound knowledge of
malaria epidemiology and prevention and control methods at national, district and peripheral
levels. The aim of this training module is to improve participants’ capacity in critical analysis
and synthesis of key determinants of malaria epidemiology, and their interactions, as the basis
for the selection of appropriate prevention and control interventions.
The module can be used for in-service training or as part of a basic course on malaria control.
For the latter purpose, it is recommended to deliver this module after the case management and
vector control modules have been covered. Prior knowledge of malaria control, including case
management and vector control options, would be beneficial.
The module is in two parts, the Guide for Participants and the Guide for Tutors. The Guide for
Participants covers basic concepts and information, and includes a series of exercises to be
carried out by the participants. The Guide for Tutors outlines the main points to be learnt, and
provides answers for the exercises which may be indicative in order to stimulate active learning.
The intended users of this training module

The module is designed for health professionals involved in the planning, implementing, and
monitoring of malaria control and elimination programmes. They include medical officers,
medical assistants, public health officers, environmental health officers, parasitologists and
entomologists involved in malaria control and working either with a national programme or
with an NGO.
Objectives
At the end of the training programme based on this module, participants should have acquired
the skills and competence necessary to:
▶▶ Describe the significance of malaria as a public health problem;
▶▶ Examine, analyse and interpret malaria data from routine health information systems, sur-
veillance and surveys;
▶▶ Explain the methods of acquiring evidence for a malaria control programme;
▶▶ Distinguish different stages of a malaria control programme, and the main strategies and
indicators for each stage;
▶▶ Analyse the malaria situation in a designated geographic area utilizing available information
from various data sources;
▶▶ Identify the appropriate control measures for specific epidemiological situations;
▶▶ Describe how to use early warning and detection systems, and to notify and verify malaria
epidemics;
2 EPIDEMIOLOGICAL APPROACH FOR MALARIA CONTROL: GUIDE FOR PARTICIPANTS
▶▶ Identify the most cost-effective malaria epidemic preventive and control options;
▶▶ Develop a preparedness plan of action for malaria epidemics.
The training approach used in this module emphasizes active involvement of participants
through a series of group exercises and discussions to stimulate active learning instead of
passive attendance at lectures given by a single person. The reasoning and deduction required
in the epidemiological approach makes the subject highly suitable for this training method,
but the success of the module will depend on active participation in the training activities
proposed. The module requires some basic knowledge of malaria case management,
parasitology, entomology, and vector control. However, the contents of the module are flexible
enough to allow the emphasis to be adjusted according to the specific training needs.
Use of the Guide for Participants

This Guide for Participants consists of instructional materials and exercises designed to enable
the participants to achieve the learning objectives of this module. The guide is divided into
21 Learning Units grouped in five parts: Part 1 – Introduction to basic epidemiology and
statistics; Part 2 – Applied malaria epidemiology; Part 3 – Malaria surveillance, monitoring
and evaluation; Part 4 – Prevention and control of malaria epidemics.
Each Learning Unit includes a series of exercises (and hints and partial solutions to some of
them) to be completed either individually or as a group, as stipulated by the tutor. The discussions
during small group work and during plenary sessions with the participation of facilitators and
tutors will assist the learning process. The exercises to be carried out in small groups aim to
stimulate discussions and exchange of experience between the participants (who will come
from different countries/areas with different experiences), the facilitators and the tutor. Ideally
participants need to acquire the skills and knowledge contained in each unit before progressing
to the next learning unit.
During the course, the Guide for Tutors will be available only to the tutor and facilitators. Upon
completion of the course/module, all participants should receive a copy of the Guide for Tutors
so that they can use it for further training and reference.
The module aims at developing an approach rather than to convey a body of facts (though
many facts may be conveyed in the process). Most factual information and details are referred
to relevant guidelines and other resource materials.
A single document cannot fully cover such a wide and dynamic subject as malaria epidemiology.
Prevention and control methods evolve over time. The module will be successful if it helps
the participants to understand the interactions between multiple factors influencing malaria
epidemiology thus preparing them to understand how new developments contribute to better
prevention and control approaches.
The Guide for Participants can also be used in conjunction with the Guide for Tutors for individual
study, but this module is best learned in group training.
INTRODUCTION 3
Evaluation
Judging whether or not the course was successful involves answering the following questions:
▶▶ How well did the participants learn?
▶▶ How did the participants view the training?
Evaluation of the participants

Progress and achievements are evaluated by the tutor, the facilitators, and by the participants
themselves. As well as general assessment during the group activities, a number of quizzes
and problems solve are used. The evaluation is intended to provide a helpful opportunity for
participants to measure their progress, and as a contribution to the learning process.
Whether this module is used for group training or individual learning, assessment of progress
made by the trainee in gaining skills and competence in the subject matter is essential. This can
be accomplished by means of a pre-test and a post-test, using a multiple-choice questionnaire
(MCQ). The pre-test will be given before the trainee reads the Guide for Participants. The post-
test will be administered after all the Learning Units have been completed. In MCQ tests,
each question is provided with a list of possible answers from which one must be selected (i.e.
considered to be correct). At the end of these sessions the tutor will analyse the results to identify
topics that were not fully understood. The tutor may also explain to individual participants
where mistakes were made and areas where improvement is needed.
The evaluation of the participant’s progress also includes assessment of classroom, practical and
field activities, degree of group participation, etc. including how the group work was presented
in plenary sessions, and the degree of clarity.
Evaluation of the training by the participants

The entire training activity, including the organization and content of the course, the suitability
of the learning methods, the quality of the teaching and training materials, and the competence
of the tutors and facilitators will be assessed by the participants. This will be done through
administration of a questionnaire, and at a plenary session after the post-test questionnaires
have been completed. This evaluation will take place at the end of the training period in order
to provide as much feedback from the participants as possible. All participants are encouraged
to make suggestions for improvement on the part of the tutor and facilitators as well as in the
content of the course and the training facilities. The objective of the plenary session is to ascertain
whether an issue(s) raised by one or more persons has the consensus of the whole group, and to
judge the importance of the issue(s) raised. Feedback provided through this exercise allows the
tutor to assess how well the training has been received and to propose modifications that seem
necessary for improving future programmes.
LEARNING UNIT 1
PART 1
Introduction to
epidemiology
EPIDEMIOLOGICAL APPROACH
Learning Objectives:
by the end, participants should be able to…
■ ■ Provide a definition of epidemiology
■ ■ Define descriptive studies and describe their purpose
■ ■ Describe the major types of descriptive studies and their
primary uses
■ ■ Describe the major types of analytic studies
■ ■ Provide a definition of random error, bias , confounding and
validity
5
Epidemiology may be defined as the study of the distribution and determinants of health-
related states or events (including disease) in human populations, and the application of this
study to the control of diseases and other health problems. The word epidemiology is derived
from the Greek words (epi) = among, (demos) = people, and (logos) = doctrine.
Different methods are used in carrying out an epidemiological investigation: surveillance
and descriptive studies are used to study distribution; analytic studies are used to study
determinants (causes, risk factors).
1.1 Types of epidemiological studies

1.1.1 Descriptive studies
Descriptive studies may be defined as studies that describe the patterns of disease occurrence
by time, place, and person.
Uses of descriptive studies

In health planning and administration, descriptive studies and the analysis of their results
allow planners and administrators to allocate resources efficiently.
They are also used to generate hypotheses, often providing first important clues about etiology.
Types of descriptive studies

▶▶ Case reports or case series
▶▶ These describe socio-demographic, behavioural and/or medical characteristics for one or
more persons with a similar diagnosis (example: characteristics of children admitted to a
hospital with cerebral malaria during a one-year period).
▶▶ They provide an important link between clinical medicine and epidemiology.
▶▶ They are often useful for generating hypotheses and examining new diseases. However,
conclusions about etiology or risk factors cannot be made without having undertaken
analytic studies (see below) to examine the expected frequency of exposure to the
etiological or risk factor in a group that does not have the illness under investigation.
▶▶ Ecological studies
▶▶ These may compare disease frequencies among different groups during the same period,
or compare disease frequencies in the same population at different points in time as a
function of a particular exposure. For instance Cambodia-Thailand, the increase over
time in the number of persons working as gem miners along the Cambodia-Thailand
border parallels the rise in P. falciparum malaria cases during the same time period (an
outcome).
▶▶ Ecological studies are usually quick and easy to perform, and can be undertaken with
already available information, but great care is needed to avoid reaching conclusions
based on spurious associations.
▶▶ Ecological studies cannot link exposure to outcome in a given individual.
▶▶ Descriptive studies constitute one of the first steps in outbreak investigation; and should
always be undertaken before initiating further analytic studies.
INTRODUCTION TO EPIDEMIOLOGY 7
1.1.2 Analytic studies U1

Analytic studies may be defined as studies used to test hypotheses concerning the relationship
between a suspected risk factor and an outcome, and to measure the magnitude of the
association effect, and its statistical significance. An analytic study always implies a comparison
among two or more groups.
There are two main types of analytic studies: observational and interventional.
Observational studies
▶▶ Most analytic studies fall in this category.
▶▶ There is no human intervention involved in assigning study groups; one simply observes
the relationship between exposure and disease.
▶▶ Observational studies are subject to many potential biases. Careful design and analysis
should help avoid many of these biases.
▶▶ There are three basic categories of observational studies: (i) cross-sectional studies (ii)
case-control studies and (iii) cohort studies.
▶▶ Cross-sectional studies (surveys)

▶▶ These examine the relationship between a disease, or other health-related characteristic,
and other variables of interest as they exist in a population at a given time. The presence
or absence (or the level) of a characteristic is examined in each member of the study
population or in a representative sample. These studies are used to obtain information
that is not routinely available from surveillance or case series.
▶▶ Cross-sectional studies provide no information on the temporal sequence of cause and
effect. In surveys examining the association between an exposure and an outcome, both
are measured simultaneously and it is often difficult to determine whether the exposure
preceded the outcome or vice versa.
▶▶ Surveys may simply describe characteristics or behaviours within a study population (e.g.
parasitaemia, use of insecticide-treated nets); or may be used to examine potential risk
factors (e.g. how those who receive vaccination differ from those who do not).
▶▶ In general, surveys measure the situation at a given moment, i.e. prevalence, rather than
the occurrence of new events, i.e. incidence (see Learning Unit 2).
▶▶ Surveys need very large sample sizes for the study of rare diseases or rare events. Surveys
are not recommended for the study of diseases with short duration.
▶▶ Case control studies

▶▶ Case-control studies proceed conceptually from outcome to exposure. They start with
groups affected with the outcome – in the case of a disease, the affected (“ill”) group and
the groups not affected (“well”), and retrospectively determine the rates of exposure to a
risk factor(s) for each group.
▶▶ In case-control studies, both exposure and disease are normally considered to have
occurred prior to enrolment in the study.
▶▶ As an example for case-control study, the rate of use of mosquito nets can be compared
between cases of malaria and healthy controls.
▶▶ Cohort studies1
Cohort studies proceed conceptually from exposure to outcome, starting with exposed and
unexposed groups and following them to see if the rates of occurrence of the outcome in the
two groups differ.
▶▶ Study groups are identified by exposure status prior to the occurrence of the outcome
of interest; both exposed and unexposed groups are then followed prospectively in
an identical manner until they develop the disease (outcome) under study, until the
study ends, or the subjects die or are lost to follow-up. Both cohorts should have similar
characteristics except for the exposure under investigation.
▶▶ Cohort studies differ from interventional studies in that the investigator does not determine
exposure status. This is determined by genetics or biology (sex, presence or absence of
genetic disease, etc.), subject’s choice (e.g. smoking behaviour, use of contraceptives,
sexual behaviours, food consumption) or other circumstances (e.g. rural versus urban,
socio-economic status).
▶▶ In some studies, called retrospective cohort studies, exposure and outcome both lie in the past
(before enrolment). The main conceptual element to remember is that the retrospective
cohort proceeds from exposure to disease.
Exercise 1.1
Match the type of study with one of the descriptions given:
a. Subjects are randomized into exposed and unexposed groups and followed
1. Descriptive study
over time to compare rates of disease development
b. Start with exposed and unexposed groups and determine if rates of
2. Analytic study
occurrence of outcome in the two groups differ
c. Start with cases of disease and unaffected controls and determine rates of
3. Case series
exposure to risk factors in each group
d. Examines the relationship between a disease and othe variables of interest in
4. Ecological study
a population at a particular point in time
e. Compares disease frequencies in a population at different points in time as a
5. Cross-sectional study
function of a particular exposure
6. Case-control study f. Describes characteristics of persons with a similar diagnosis
g. Test hypotheses concerning the relationship between suspected risk factor
7. Cohort study
and an outcome
8. Interventional study h. Describes patterns of disease occurrence by time, place, and person
Interventional studies
▶▶ The person conducting the study randomizes the subjects into exposed and unexposed
groups and follows them over time to compare their rates of disease (outcome)
development. Examples may include trials of the efficacy of a new drug compared with
the efficacy of the drug currently in use; or assessment of the efficacy of insecticide-treated
mosquito nets compared with non-treated nets.
A cohort can be defined as a designated group of people who have had a common experience vis-a-vis exposure, and are then
1.
followed up or traced over a period of time.

▶▶ Randomization helps ensure comparability of the groups and avoids many of the biases U1
inherent in non-interventional studies; for this reason interventional studies have been
considered as a widely accepted “gold-standard”.
▶▶ Interventional studies are nevertheless expensive. They may take a long time to carry out,
often present complex ethical problems,1 or may simply not be feasible (e.g. randomized
trials of the health benefits of breastfeeding). The results obtained may not be applicable
to routine programme conditions.
1.2
The logical sequence of epidemiological studies
In epidemiological research, the current state of knowledge often determines the most logical
study design. There is usually a progression from hypothesis-generating to hypothesis-testing
studies. For example, hypotheses are often generated by methods such as surveillance,
case reports, case series, or ecological studies. These hypotheses are then tested using data
from experience, from previous cross-sectional studies, from case-control studies, or from
retrospective cohort studies, which can be done relatively quickly and cheaply. If these studies
lend support to the hypothesis, a prospective cohort study may be undertaken. Finally, in some
situations, a randomized clinical trial may be appropriate.
The flowchart (Fig. 1.1) illustrates the application of the various types of primary studies. In all
types of study, hypothesis-setting must precede analysis.
STUDY
DESCRIPTIVE ANALYTICAL / EXPERIMENTAL
CASE REPORT
ECOLOGICAL OBSERVATIONAL INTERVENTIONAL
CASE SERIES
CASE CLINICAL
CROSS-SECTIONAL COHORT
CONTROL TRIAL
Figure 1.1 Types of primary epidemiological studies
1.2.1 Secondary research

In secondary research, the unit of study is the body of previous studies on the subject. If a
secondary study has been conducted with (i) a systematic search, (ii) appraisal of the gathered
studies and (iii) a systematic analysis, it is termed a ‘Systematic Review’. The statistical analysis
of the systematic review would be a ‘Meta-Analysis’. The evidence gathered is graded to assess
the strength of the evidence. Among all types of studies, including primary and secondary, the
results obtained from Systematic Reviews, especially on randomized clinical trials, provide the
highest level of evidence.
The ethical problems relative to epidemiological studies are addressed in International Ethical Guidelines for Epidemiological
1.
Studies prepared by Council for International Organizations of Medical Sciences (CIOMS) in collaboration with the World
Health Organization, Geneva, 2008.
1.3 Potential errors in epidemiological studies

1.3.1 Random error
Random error is the divergence, due to chance alone, of an observation on a sample1 from the
true population value, leading to lack of precision in the measurement of an association. There
are three major sources of random error: individual/biological variation, sampling error, and
measurement error.
Random error can be minimized but can never be completely eliminated since only a sample
of the population can be studied; individual variation always occurs and no measurement is
perfectly accurate. Random error can be reduced by careful measurement of exposure and
outcome, appropriate selection of study participants and enrolling a sample of sufficient size.
1.3.2 Systematic error (bias)

Bias occurs when there is a tendency to produce results that differ in a systematic manner from
the true values. A study with only small systematic bias is said to have high accuracy. Bias (or
systematic error) may lead to over- or underestimation of the strength of an association.2 The
sources of bias in epidemiology are many and over 30 specific types of bias have been identified.
The main biases are:
▶▶ Selection bias
▶▶ Information bias
▶▶ Bias due to confounding.
Selection bias
Selection bias occurs when there is a systematic difference between the characteristics of the
people enrolled for a study and the characteristics of the source population.
Information bias (also called measurement bias)

Information bias occurs when there are quality (accuracy) problems in the collection,
recording, coding or analysis of data among comparison groups. Interviewers might, for
example, interview the cases with more diligence than the controls, or a person with a disease
may recall previous exposures better than persons who are healthy (this type of bias is called
recall bias which is a form of information bias).
From the practical point of view, often there is not enough information to correct for selection
bias or information bias. It is best to think about possible sources of bias at the time of the study
design so that they can be minimized or avoided.
Confounding
In a study of the association between exposure to a cause (or risk factor or protecting factor)
and the occurrence of the disease, confounding can occur when another factor exists in the
1.
A sample is defined as a selected subset of a population. Sampling is the process of selecting a number of subjects from all the
subjects of a population.
2.
Statistical dependence among two or more variables, which are said to be associated if they occur together more frequently
than would be expected by chance. Statistical tests permit calculation of the association .
study population and is associated both with the disease and the initial factor being studied. U1
A problem arises if this second extraneous factor is unequally distributed among the exposure
subgroups. Confounding occurs when the effects of two protective or risk factors have not been
separated and it is therefore incorrectly concluded that the effect is due to one variable rather
than the other. For instance, in a study of the association between tobacco smoking and lung
cancer, age would be a confounding factor if the average ages of the non-smoking and smoking
groups in the study population were very different, since lung cancer incidence increases with
age.
Another example of confounding is shown in Figure 1.2. A study of the relationship between
income and malaria is illustrated by the top line of Figure 1.2. It is possible that income is
associated with the risk of malaria. However it is known that income is also associated with
the use of bednets to reduce the risk of malaria. The relationship between income and malaria
is thus affected by the relationship between bednets and income. In other words, bednets
confound the association between income and malaria.
EXPOSURE (income) DISEASE (malaria)
CONFOUNDING VARIABLE
(bednets)
Figure 1.2 Confounding – bednets, income and malaria
A confounding factor is an alternative explanation for presence of the outcome, other than the
exposure of interest, and this factor should not be an intermediate in the casual pathway from
exposure to the outcome. Supposing that income reduces the risk of malaria through the use
of bednets, the use of bednets is in the causal pathway of the effect of income on malaria, and
consequently the use of bednets should not be considered as a confounding factor.
Confounding can be controlled for in the analysis if appropriate information has been collected
during the study on potential confounding variables and if each factor is properly analysed and
interpreted.
Exercise 1.2
Match the type of potential study error with one of the examples given.
Study errors Examples
a. Studying the relationship between bednet use and occurrence of malaria if age
1. Random error is associated with both use of benets and susceptibility for symptomatic malaria
infection
b. Collecting information regarding malaria treatment through a household survey
2. Selection bias
conducted months after the treatment was received
3. Information bias c. Assessing outcomes of uncomplicated malaria among patients admitted to a hospital
d. Large divergence of a study result from the true value due to a small study sample
4. Confounding
size
1.4 Validity and reliability

Validity is an expression of high VALIDITY low
the degree to which a test is Measured values Measured values
capable of measuring what
high
it is intended to measure.
A study is valid if its results
RELIABILITY true value true value
correspond to the truth;
there should be no system- Measured values Measured values
atic error, and the random low
error should be as small
as possible. Figure 1.3 il-
true value true value
lustrates the relationship
between the true value and Figure 1.3 Validity and reliability
the measured values for low
and high validity and reliability (or repeatability). With low reliability but high validity the
measured values are spread out, but the average of the measured values is close to the true
value. High reliability does not
Validity high Validity low ensure validity since the meas-
urements may all be far from the
Reliability high true value.
Figure 1.4 shows the same con-
cept in a different graphical way:
the centre of the target corre-
Reliability low sponds to the true population
value; individual target shots are
Figure 1.4 Validity and reliability
individually measured from 5
samples in each example.
Exercise 1.3
A well done study with a small systematic error (minimal bias or confounding) but a large
random error could be described as having
a) High validity and low reliability
b) Low validity and high reliability
LEARNING UNIT 2
Ratios, proportions and rates
■ ■ Define the terms ratio, proportion and rate
■ ■ Differentiate between incidence rate and prevalence rate, and
give examples of their use
■ ■ Calculate rates, ratios, and proportions using appropriate
numerators, denominators and constants
■ ■ Apply the concepts of relative risk and risk differences
13
Depending on who is using the data and for what purposes, the data may be presented as raw
data, proportion, rate, and ratio.
2.1 Raw data

Raw data may be defined as the entire set of data collected in a study, before any rounding,
editing or statistical organization. They are of use primarily in helping health planners and
administrators determine health care needs. For example, a health planner may want to
know the number of cases of malaria seen in the last year so that an estimate can be made of
the quantity of antimalarial drugs to order for the next year. He/she may also want to know
approximately how many births take place each year so as to staff the obstetrics services
appropriately.
Any variable can be considered as one of two types: discrete or continuous.
Discrete variables have values that can fall into only a limited number of categories without
intermediate levels (e.g. sex – M/F; survival – dead or alive; exposure status – Yes/No; race,
marital status…). When the possible categories have a natural order of progression, the variables
are called ordinal (e.g. improvement in mobility, or level of current cigarette smoking – none,
light, moderate, heavy). Some quantitative data can also be discrete, e.g. parity (it is not possible
to have a fraction of a live birth).
Continuous variables can assume all possible values along a continuum within a specified range
(e.g. height, weight, blood pressure…). They are limited only by the accuracy and precision of
measurement.
2.2 Proportion
Proportion is also a measure that is of use primarily to planners and administrators. It is defined
as the number of events among all possible events, usually expressed as a percentage. The
formula is (x / y) k, where x is the number of individuals or events in a category and y is the total
number of events or individuals in the data set and k is a constant, in this case 100.
Example 1
Of the 120 cases of malaria admitted to hospital X last year, 80 were children. The proportion
(percentage) of children among the cases is (80 / 120) × 100 or 66.7%.
It may be useful for the hospital administrator to know that 67% (two-thirds) of malaria
hospitalizations occur in the paediatric age group and 33% (one-third) occur in adults. The
number of beds and staffing of various categories required to take care of malaria patients as
well as commodities (diagnostic testing and medicines) can be planned.
2.3 Rate
For the public health practitioner interested in determining who is at risk and monitoring
the success of prevention efforts, the most useful measure is a rate. Rates measure the relative
frequency of cases in a population during a specified period of time. The general formula is the
same as for proportions, i.e. (x / y) k, although here x, y, and k take on different meanings. Rates
measure incidence (new cases) within a specified period.
RATIO, PROPORTION AND RATE 15
An incidence rate1 is the occurrence of new cases of a disease within a defined population at
risk during a specified period of time. In this situation:
▶▶ x is the number of new cases in the defined population which had its onset during a specified
period of time U2
▶▶ y is the person time at risk. Typically the mid-year population of a defined geographic area is
used to determine person time at risk over a one-year period.
▶▶ k, a constant, depends on convention or is the value such that the smallest rate in the data set
has at least one digit to the left of the decimal point.
An attack rate is a variant of an incidence rate where shorter periods at risk (e.g. weekly or
monthly) are used as denominator; it is typically measured during an outbreak. In practice, the
attack rate will only differ from the incidence rate if there is a large proportion of persons in the
population who are not at risk (for instance, children who have been successfully vaccinated
against measles may be considered not to be at risk for the disease).
In a prevalence rate, x is the number of existing cases, new and old, in a defined population
during a specified period (period prevalence) or at a given point in time (point prevalence).
In reality prevalence is a proportion because it does not have time dimension. However, it is
also commonly called “prevalence rate”. Another form of prevalence is “period prevalence”
which can estimate old and new cases during a specific period of time. Whenever the term
“prevalence” is used without specification of “point” or “period”, point prevalence should be
assumed.
Example 1
In July, 3 new cases of malaria were detected in a village. There were already 10 people in the
village who had the disease, but two successfully completed a course of therapy during the
month and were considered cured. The population of the village was 2600. In this case:
• The incidence rate is:
(3 / 2600) × 1000 or 1.2 per 1000 persons per month2
• The period prevalence rate is:
[(3 + 10) / 2600] × 100 = 0.5%
• The point prevalence as of 31 July is:
[(3 + 10 – 2) / 2600] = 0.4%
1.
In more rigorous definition, the denominator of “incidence rate” is person-time (people in month, year, etc). Therefore, an
incidence rate shows how many cases (nominator) had occurred during in a certain person-time (denominator). In the same
literature, the incidence rate which is presented in this module is considered “cumulative incidence rate”. However, in the day
by day use of the incidence rate such distinction is not serious.
2.
Strictly speaking, the attack rate is the number of cases occurring during July in the population at risk (excluding those already
affected), i.e. [3 / (2600 – 10) × 1000] = 1.2 per 1000 per month, equal to the incidence rate. The Incidence Rate = [3 / (2600
total – (10 affected –2 cured and therefore sensitive again)] i.e. 3 / 2592. In practice, these requirements are often neglected
where they make little difference.
2.4 Ratio
A ratio is an expression of the relative frequency of the occurrence of some event compared to
some other event, for example, the ratio of male to female cases. Here, the formula is also (x / y) k,
where:
▶▶ x is the number of events or persons having a specified attribute
▶▶ y is the number of events or persons having an attribute different from those of the event
or person in x
▶▶ k is 1
In this situation, the ratio is often expressed as x:y, with y usually equal to 1 (y can be made
equal to 1 by dividing both x and y by y).
Example 1
If there are 15 male cases (x) and 5 female cases (y) of malaria, the male:female ratio can be
calculated as 3:1 by dividing both values by 5 (y).
Ratios are often used when it is difficult to ascertain the population denominator for a disease
or a condition correctly. One example is the abortion ratio, which is the number of abortions
divided by the number of live births during the same time period. The formula remains
(x / y) k with k determined either by convention or by the value that gives at least one digit to
the left of the decimal point.
2.5 Relative risk and risk difference

Rates for two or more groups (males/females, age categories, educational levels, presence or
absence of some behaviour) are often compared by dividing one by the other or by subtracting
one from the other.
If they are divided, the result is called a rate ratio or relative risk. The formula is simply:
rate a/rate b, where a is the incidence in the group exposed to the factor under investigation
and b the incidence rate in the group that is not exposed. The rate ratio or relative risk may be
used to identify possible causal risk factors and identify markers that may be useful in targeting
services. A ratio of 1 means that there is no difference in outcome between the exposed and the
unexposed groups (if the outcome is an incidence rate, this will be the same for both exposed
and unexposed groups). A ratio >1 suggests that the characteristic (exposure) is a risk factor; a
ratio of <1 suggests a protective effect.
Example 1
People who go into the forest have a malaria incidence rate of 10 / 1000 per month, while
people who do not go into the forest have a malaria incidence rate of 1 / 1000 per month. The
risk ratio is (10 / 1000) / (1 / 1000) = 10. Thus, people who go into the forest are 10 times
more likely to contract malaria than those who do not.
Example 2
People who use mosquito nets have a malaria incidence rate of 2 / 1000 per month; people
who do not use nets have a rate of 8 / 1000 for the same period. The ratio of the risks is
(2 / 1000) / (8 / 1000) = 0.25. Thus, those who use nets incur a lower rate of malaria incidence U2
than those who do not (this is called the protective effect and is calculated as 1– the relative
risk or 1– 0.25 = 0.75. This is roughly equivalent to saying that those who use bednets in these
circumstances will have 75% fewer episodes of malaria compared to those who do not use a
bednet.
Example 3
People who are illiterate have a malaria incidence rate of 9/1000, while those who are literate
have a rate of 3 /1000 for the same period. The ratio of the risks is 3. Thus, those who are
illiterate have three times more risk of malaria than those who are literate. Here, literacy is a
marker rather than a causal risk factor. Illiteracy does not cause malaria, but those who are
illiterate are at risk for other reasons, such as living conditions, occupation, etc.
Rates may also be compared by subtracting one from the other. The resulting value is known
as the (absolute) risk difference. This is calculated as: rate a – rate b. This represents
the absolute differences in risk between the exposed and the unexposed groups. If disease
incidence is the same in the exposed and in the unexposed group, the value of the absolute risk
difference will be zero. If there is a causal relationship between the characteristics being studied
and the outcome, the risk difference provides information on the amount of disease that could
be prevented if the characteristic could be eliminated.
Example 4
Those going into the forest have a malaria incidence rate of 10 / 1000 per month; those who do
not go into the forest have a malaria incidence rate of 1 / 1000 per month. The risk difference
is (10 / 1000 – 1 / 1000 = 9 / 1000). The absolute difference between the groups is 9 per 1000.
Because there is a presumably causal relationship, it could be concluded that if people stopped
going into the forest, the malaria rate would be reduced by as much as 9 / 1000, to 1 / 1000.
Caution is required in drawing such conclusions since often people will have more than one
characteristic (exposure) that puts them at risk for a disease; eliminating only one behaviour or
characteristic usually does not fully solve the problem. Caution is required if the characteristic
is a marker rather than a causative factor; changing a marker without changing the causal
factors associated with it is unlikely to result in a lower disease rate.
A note on rounding
The procedure for finding the last digit of a measure is called “rounding”. There are three general
rules for rounding:
Rule 1: if the digit beyond the last digit to be reported is less than 5, drop everything after the last
digit to be reported. Rounding to one decimal place, the number 5.3467 becomes 5.3.
Rule 2: if the digit after the last digit to be reported is greater than 5, increase the last digit to be
reported by one. The number 5.798 becomes 5.8 when rounding to one digit.
Rule 3: to prevent rounding bias, if the last significant digit is exactly 5, it is general practice to
round to the integer preceding the 5, and rounding up if this is an odd integer. Thus the number
3.55 (rounded to one digit) would be 3.6 (rounding up) and the number 6.450 would round to
6.4 (rounding down when rounding to one decimal).
It is also possible to round by taking the nearest whole number: 66.7% may be rounded to 67%.
Exercise 2.1
The following table 2.1 presents malaria morbidity data for Province X in Africa, which has
received a large number of immigrants in recent years:
Table 2.1 Number of malaria cases, Province X, 2001–2005
YEAR CASES POPULATION
2001 30 858 492 810

2002 36 602 585 540
2003 46 172 738 870
2004 56 439 891 280
2005 68 392 1 044 620
a. Describe in words the trend in the number of cases.

b. Calculate the incidence rate of malaria cases/100 population per year and describe the trend in
words.
c. Compare the trend in the number of cases and the trend in rates. How do you explain your
observations?
d. Which is the more appropriate measure to monitor changes over time in the area?
Exercise 2.2
The adjoining Province Z (population 169 250) had 15 233 cases of malaria in 2005.
a. Which province had the higher incidence rate in 2005?
b. In your opinion, which area should receive intensified control efforts and why?
Exercise 2.3
A survey among children aged < 5 years in Region A shows that 450 out of 950 children have
malaria parasites in their blood.
a. What is the parasite rate?
b. Is this the incidence rate or the prevalence? Explain your answer.
Exercise 2.4
In 2001, 49 140 malaria cases occurred among males and the remaining 23 250 occurred
among females.
a. What is the ratio of male:female cases? U2
b. What percentage of the total cases occurred in males? What percentage in females?
Exercise 2.5
At one of the health centres in Province X (Table 2.2), the age breakdown of malaria cases was
as follows:
Table 2.2 Age breakdown of malaria cases, Province X, 2006
AGE CASES % OF ALL CASES POPULATION AT RISK INCIDENCE RATE/100/year

0–11 months 71 2.4 1980 3.6
1–4 years 645 21.9 7920 8.1
5–14 years 698 23.7 12 300 5.7
≥ 15 years 1528 51.9 27 300 5.6
TOTAL 2942 100.0 49 500 (5.9)
a. Which age group accounts for the biggest percentage of all cases?
b. Which age group is at greatest risk of contracting malaria?
c. Why are the answers to questions a) and b) different?
Exercise 2.6
A study shows that the incidence rate of malaria is 10 / 1000 population per week among
Thai villagers who work as gem miners and go into the forests, whereas this rate is 2 / 1000
population per week among farmers from the same villages.
a. Calculate the relative risk of malaria among gem miners
b. Interpret your findings in words
c. Calculate the risk difference between the gem miners and the farmers
d. Interpret your findings in words
LEARNING UNIT 3
Data presentation:
tables, graphs and charts
■ ■ List the features of good tables, graphs, and charts
■ ■ Plot and label a series of tables, graphs and charts correctly
from raw data
■ ■ List the uses for semi-logarithmic presentation
21
3.1 Tables
A table may be defined as a set of data arranged in rows and columns designed to present the
frequency with which some event occurs in different categories or subdivisions of a variable, as
can be seen from Table 3.1.
3.1.1 Guidelines for developing tables

▶▶ Keep them simple: better 2 or 3 small tables than a single large table
▶▶ No more than 3 variables should be used in a table
▶▶ All tables should be self-explanatory
▶▶ Clear and concise title telling what, where, and when
▶▶ Rows and columns must be clearly labelled
▶▶ Units of measurement must be stated
▶▶ Codes, abbreviations, and symbols must be footnoted
▶▶ Totals must be shown
▶▶ If data are not original, their source must be footnoted
Rows and columns to be clearly labelled

Clear and concise title telling what, where and when
Table 3.1 Proportion of malaria cases in relation to total inpatients, State Z.,
2001–2005
YEAR All patients Malaria patients %
2001 136 289 16 946 12.4

2002 114 327 18 117 15.8
2003 101 050 13 821 13.7
2004 79 485 10 757 13.5
2005 76 403 11 533 15.1
Data collected from 24 district hospitals
Footnote to enhance clarity
3.2 Graphs
A graph may be defined as a method of showing quantitative data using a drawing on a
coordinate system. The most common form is a rectangular coordinate, with two sets of lines
at right angles to each other and divided into equal intervals. The x axis by convention is the
horizontal axis, and the y axis is the vertical one.
Graphs are used for continuous variables such as time, parasite counts, etc. Charts rather than
graphs are used for non-continuous variables such as sex or educational level.
The variable (age, year, etc.) is usually classified along the x axis; the y axis is the axis generally
used for measures of frequency. See Figure 3.1.
DATA PRESENTATION: TABLES, GRAPHS AND CHARTS 23
Label y axis:
frequency X41 Y7
(e.g. number,
percentage,
rate
equal
interval U3
equal Label x axis: method of classification

interval (e.g. time of on set of illness, in days;
age of cases, in years)
Figure 3.1 General graph with descriptive annotation
3.2.1 Guidelines for preparing graphs

▶▶ Keep them simple and do not try to put in too much information.
▶▶ Every graph should be self-explanatory.
▶▶ Avoid interrupting the axis (scale breaks) if at all possible.
▶▶ Title should be clear and concise.
▶▶ Axes clearly labelled.
▶▶ Units on the x and y axis clearly specified.
▶▶ Equal quantities must be represented by equal intervals on an axis; on the x axis, categories
covering 10 years, for example, should be twice as long as categories covering 5 years.
3.2.2 Types of graphs

The graph must have enough information to be clear without further explanation.
Charts
The most common forms are bar charts, pie charts and geographic coordinate charts.
Applications are given here for 620 patients, classified according to 4 age categories (<1, 1–5,
6–10 and 11–15 years). In the <1 group there were 250 patients; in the 1–5 group there were
325; in the 6–10 group there were 30; and in the 11–15 group 15 patients.
Pie charts
This is defined as a circular chart most frequently used to show percentage distributions, using
wedge-shaped portions proportionate to the size of the category. The convention is to start
at the 12:00 clock position and arrange “slices” anticlockwise in order of decreasing size. See
Figure 3.2.
d 5
5
e/pyrimethamine 12
r+lumefantrine 78
Example
The pie below shows the treatment profile of 100 paediatric malaria patients prior to admission
to hospital A in Province Z.
5
5
12
Not treated
Quinine
Sulfadoxine/pyrimethamine
Artemether+lumefantrine
78
Figure 3.2 Treatment of paediatric malaria patients prior to hospitalization, Hospital A in Province Z,
January–December 2001
Bar charts
Bar charts have cells, all of which have the some column width whatever the size of the category.
The bars may be arranged vertically or horizontally. By convention, there is always a space
between the bars. Bar charts are easier to use when categories are of unequal size; they must be
used if categories are not continuous (i.e. sex, marital status, etc.), as is the case in Figure 3.3.
Figure 3.4 is a bar chart of continuous data (age) with categories of unequal importance.
800
600
Cases
400
200
0
Town West North Central South
Zone of residence
Figure 3.3 Cases of malaria in hospitalized paediatric patients in hospital A, town B, January–March 2001
350
300
250
200
Cases
150 U3
100
50
0
5 1-5 6-10 11-13
Age in years
Figure 3.4 Age distribution of hospitalized paediatric patients in hospital A, town B, January–March 2001
Histogram
This may be defined as a bar graph of the frequency distribution of a continuous quantitative
variable in which the width of the bar is proportional to the unit of value of the variable on the
x axis and the height of the bar is proportional to the unit of value of frequency on the y axis. By
convention, there is no space between the bars, and no scale breaks are allowed on the y axis
(Fig. 3.5).
Histograms can be used to plot the number of cases or percentages on the y axis, but are
generally not used to plot rates.
45
40
Thousands of cases
35
30
25
20
15
10
5
0
Jun
Jul
Aug
Feb
Sep
Oct
Jun
Jul
Aug
Nov
Dec
Feb
Sep
Oct
Nov
Dec
Jan
Jan
Mar
Mar
Apr
Apr
May
May
2000 2001
Figure 3.5 Monthly distribution of malaria cases reported by PHC units and health centres, State Z,
2000–2001
Line graphs
A line graph may be defined as a graph of the frequency distribution of a continuous variable
created by plotting the frequency of a category on the y axis at the midpoint of the category on
the x axis. Values for each category are connected by a continuous line.
If a graph is to contain the frequency distribution by category for more than one group (e.g. the
frequency of cases over a 10-year period for males and females), it may be advisable to use line
graphs.
Line graphs may be used to plot number of cases and percentages; they are the method of choice
for plotting rates. See Figures 3.6 and 3.7.
60
50
40
Percentage
Percentage
30
20
10
0
1997 1998 1999 2000 2001
Year
District A State Z District B
Figure 3.6 Rate of positivity for malaria blood films examined at hospitals level, State Z, 1997–2001
Thousands of cases
Figure 3.7 Malaria cases and deaths officially notified in health facilities run by physicians, Country Y,
1998–2001
In Figure 3.7, it is difficult to show the variations in deaths over time. This can be easily shown
by using two vertical axises: the primary vertical axis for the malaria cases and the secondary
vertical axis for the malaria deaths as shown in Figure 3.8.
U3
Figure 3.8 Malaria cases and deaths officially notified in health facilities run by physicians, Country Y,
1998–2001
Geographic coordinate chart

This is a map where areas are shaded geographically according to the incidence or prevalence
rate of the disease considered. See Figures 3.9.
Belize Caribbean
Sea
Mexico
Honduras
Pacific
Ocean
El Salvador
No data 0 0 – 0.1 0.1–1 1–10 10–50 50–100 100
Figure 3.9 Distribution of confirmed malaria cases (per 1000 population), 2011 Guatemala
Semi-logarithmic graph
This may be defined as a graph in which the y axis is measured in logarithms of units and x axis
is measured in arithmetic units. These graphs are generally used to:
▶▶ Plot data when the range is too great to present meaningfully on an arithmetic graph.
▶▶ Examine relative rather than absolute changes over time.
If a line plotted on a semi-logarithmic graph is straight, it indicates a constant rate of change,
and the slope allows direct measurement of the rate of change. Two or more lines that follow
parallel paths have equal rates of change. See Figure 3.10, which represents a logarithmic plot
of the data shown in Figure 3.7 and shows more clearly the trends and relative rates of change
for cases and deaths.
The reason to use a log-scale is normally to obtain the relative change over time. However, this
is of course also the major limitation in the sense that a very steep increase in absolute counts
does not look very steep using a log-scale. One can thus easily underestimate the increase, or
decline.
1 000 000
100 000
Cases
10 000
1000
100
1988 1999 2000 2001
Year
Cases Deaths
Figure 3.10 Malaria cases and deaths officially notified in health facilities run by physicians, in country Y,
1998–2001
Note: With the generalization of computer graphics software programmes, charts, maps, graphs etc are increasingly prepared on
the computer; the use of semi-log paper is declining.
Box plot
In descriptive statistics, a box plot (also known as a box-and-whisker diagram or plot) is
a convenient way of graphically showing data through their descriptive statistics: sample
minimum, the lower quartile, the median, the upper quartile, and sample maximum. Note that
the box plots can be drawn vertically or horizontally, depending on whether you display the
descriptive statistics along a vertical or a horizontal axis.
Example: Admissions of malaria patients of Hospitals A and B are presented below). Prepare a
box plot for the data.
Months Hospital A Hospital B
January 321 465
February 235 399
March 202 345
April 304 456
May 621 746
June 600 802
July 590 845
August 431 578
September 381 503
October 142 478
November 98 389
December 96 390
The calculated sample minimum, lower quartile (25% percentile), median, upper quartile
(75% percentile), and sample maximum and box plot graph (Fig. 3.11) of malaria admissions in
Hospitals A and B are presented below.
Summary Range
Parameters
Hospital A Hospital B
Min 96 345
25th percentile 218.5 427.5
Median 321 478
U3
75th Percentile 510.5 662
Max 621 845
Admissions
Figure 3.11 Malaria inpatients in Hospitals A and B for 12 months period.
Venn diagram
Venn Diagrams are pictorial ways of representing interactions among sets to display informa-
tion that can be read easily. Each set of given information is designated a circle. Interactions
between the sets are shown in the circles' intersections; items common to both sets are found
in the intersection whereas other items are found outside the intersection. Figure 13.12 shows
case fatality rates (with respective sample size) among children with different clinical presen-
tations of severe malaria conditions admitted between 1989 to 1991 in Kilifi District hospital
in Kenya. The figure clearly presents the case fatality rates among children with the following
severe conditions:
▶▶ severe anaemia
▶▶ impaired consciousness
▶▶ acute respiratory distress
▶▶ severe anaemia and acute respiratory distress

▶▶ severe anaemia and impaired consciousness
▶▶ acute respiratory distress and impaired consciousness
▶▶ acute respiratory distress, severe anaemia and impaired consciousness
Acute respiratory
distress
24%
(21)
28% 16%
(36) 35% (49)
Impaired 7% (26)
consciousness 6% Severe
(219) 1%
(51) anaemia
(380)
Source: Marsh et al., NEJM 332, 1399-1404, 1995
Figure 3.12 Case fatality

Figure rate (sample
3.13. Case size)(sample
fatality rates among size)
Kenyan children
among Kenyanwith severewith
children malaria admitted
severe malaria in 2005
admied in 2005
Exercises
The government of a large Asian country started a national programme 5 years ago to reduce
the morbidity and mortality from 3 major childhood diseases. Each District Medical Office
(DMO) is now being asked to evaluate the effectiveness of this programme in their district
and to learn as much as possible about who remains at risk for developing these diseases. The
DMO of District W has decided that the best way to begin is to use the surveillance data from
the past 5 years to examine the trends in the incidence of these three diseases over that period.
The surveillance data provide the number of cases and deaths by month for each disease and
are broken down by broad age categories, including one for children <5 years of age.
The following exercises should be carried out in small groups.
Exercise 3.1
a. What are the advantages and disadvantages of using surveillance data to monitor trends for the 3
diseases?
b. What other sources of data might the DMO consider to gather information on trends in the 3
diseases?
Exercise 3.2
a. The DMO has the number of cases and deaths for children under 5 for each disease.
b. What other number(s) does the DMO need in order to monitor disease trends over a several year
period in an adequate manner? Where can the DMO obtain such numbers?
Exercise 3.3
The most recent national census was conducted in 2000, and no population estimates are
U3
available for the years 2001–2005. The population in 2000 in the <5 age group was 56 650.
The rate of natural increase of the population is 3.3% per year. How can the DMO estimate the
population aged <5 years in the district for each of the years 2001–2005?
Exercise 3.4
After estimating the mid-year population of children under 5 for each year between 2001 and
2005, the DMO develops a table containing data for the 3 diseases over the 5-year period (see
Table 3.1 below).
Divide each group into 3 smaller subgroups. Each subgroup should do one of the following:
a. Plot the trends for the incidence, mortality and case fatality for each of the 3 diseases.
b. Plot the trends for the incidence of each of the 3 diseases on the same graph.
c. Plot the trends for mortality from each of the 3 diseases on the same graph.
d. Plot the trends for case fatality rates in each of the 3 diseases on the same graph.
Describe the trends you have graphed to the rest of your group.
Exercise 3.5
Each subgroup will do one of the following:
a. For the disease assigned to your group, plot the age distribution from a hospital record review (Table
3.2).
b. Plot the seasonal distribution based on 5 years of surveillance data for the disease assigned to your
group (Table 3.3).
Exercise 3.6
Taking into account the graphs you have prepared, define the characteristics of the disease
assigned to your group. Describe in words the trends for incidence, mortality and case-fatality
as well as the age and seasonal distribution of the disease, and what types of actions or events
may have been responsible for the temporal trends observed. Prepare a summary to present to
the rest of the class. Presentations will be limited to 10 minutes per group.
Table 3.1 Incidence, Mortality, and Case-fatality rates for diseases A, B, and C,
District W, 2001–2005
Disease A
Population Cases in Deaths in

Year Cases/1000 Deaths/1000 Case-fatality %
aged <5 years <5 years <5 years
2001 58 520 10 241 205 175 3.5 2.0

2002 60 541 10 353 157 171 2.6 1.5
2003 62 446 10 616 131 170 2.1 1.2
2004 64 507 10 966 123 170 1.9 1.1
2005 66 635 11 261 113 169 1.7 1.0
Disease B
Population aged Cases in Deaths in

<5 years <5 year <5 year
2001 58 520 3113 152 53.2 2.6 4.9

2002 60 541 1604 85 26.5 1.4 5.3
2003 62 446 4571 219 73.2 3.5 4.8
2004 64 507 1251 71 19.4 1.1 5.7
2005 66 635 2259 113 33.9 1.7 5.0
Disease C
Population aged Cases in Deaths in

<5 years <5 years <5 years
2001 58 520 480 386 8.2 6.6 80.5

2002 60 541 454 394 7.5 6.5 86.7
2003 62 446 381 356 6.1 5.7 93.4
2004 64 507 348 329 5.4 5.1 94.4
2005 66 635 347 320 5.2 4.8 92.3
Table3.2 Age distribution for diseases A, B and C – District W, Hospital 1 January–

31 December 2005
Number of cases
Age group (in months)
Disease A Disease B Disease C
0–5 427 37 258

6–11 1063 296
12–23 2312 411
24–35 1239 223
36–59 647 115
Total 5688 1082 258
Table 3.3 Average monthly distribution for diseases A, B and C, 2001–2005
Percentage of cases (%)

Month
Disease A Disease B Disease C
January 3 13 8
February 7 10 7
March 15 9 6
April 18 7 8
U3
May 13 5 9
June 41 4 7
July 7 4 8
August 5 5 9
September 3 7 12
October 6 9 9
November 7 12 8
December 5 15 9
Total 100 100 100
LEARNING UNIT 4
Measures of central tendency
■ ■ Define the terms mean, median, and mode
■ ■ Describe the advantages and disadvantages of using the
mean versus the median
■ ■ Calculate means, medians and modes from individual and
from grouped data
35
With variables such as age, number of children, haemoglobin, and parasite counts, it is often
useful to develop a single value that is representative of the individual values in the group. These
single, representative values are known as measures of central tendency. These values not only
facilitate the description of a population, but also facilitate the comparison of populations.
▶▶ The most common measures of central tendency are the mean, the median, and the mode.
These measures can be calculated from individual data if the number of items in the data set
is small; if there are many items, these measures are calculated from grouped data.
4.1 Mean
The mean is the average of the values in the data set. It is calculated by taking the sum of the
individual values in the data set and dividing this sum by the number of values in the set. The
mean is the most commonly used measure of central tendency, in part because it is used in
other, more sophisticated statistical tests. Its major disadvantage is that it can be affected by the
presence in the set of a few extreme values, large or small.
Mathematically, the formula can be expressed as follows:
x= ∑ x / n
i
where: x is the arithmetic mean

∑ is the sum of the individual values
xi is each individual value in the set
n is the number of individual values in the set.
Example 1
Calculate the mean of a set of 5 values: 12, 15, 7, 13, 8
4.2 Median
The median is the value or point in a data set that divides the ranked values into 2 equal-sized
groups, one consisting of values smaller than the median, and the other consisting of values
greater than the median. If the data set is skewed (by one or more extreme values), the median
is a more representative measure of central tendency than the mean, because it is less influenced
by outliers. The median is calculated as follows:
▶▶ Order the values by rank (place the values in sequence, either in ascending or in descending
order)
▶▶ Identify the mid-point of the sequence. If there is an odd number of values identify the
middle number. If there is an even number of values identify the mid-point between the two
numbers in the middle of the sequence.
MEASURES OF CENTRAL TENDENCY 37
The general formula to identify the middle value is:
total number of values in sequence + 1

middle value =
2
▶▶ The number corresponding to this middle value is the median of the values in the data set.
Example 2
Determine the median of a data set containing an odd number of values (12, 15, 7, 11, 8).
▶▶ Rank values in ascending order: 7, 8, 11, 12, 15
▶▶ Determine the mid-point of the sequence (5 values +1) / 2 = 3 U4
▶▶ The median is therefore the 3rd value in the sequence
▶▶ The 3rd value is 11, therefore the median is 11
Example 3
Determine the median of a data set with an even number of values where the middle numbers
are different (12, 15, 18, 7, 13, 8).
▶▶ Rank values in ascending order: 7, 8, 12, 13, 15, 18
▶▶ Determine the midpoint of the sequence (6 values + 1) / 2 = 3.5
▶▶ The median is the value that lies halfway between the 3rd and 4th values
▶▶ The 3rd and 4th values are 12 and 13. The median is: (12+13) / 2 = 12.5
Example 4
Determine the median of a data set with an even number of values where the middle numbers
are the same (12, 15, 18, 7, 12, 8).
▶▶ Rank values in ascending order: 7, 8, 12, 12, 15, 18
▶▶ Determine the midpoint of the sequence (6 values + 1) / 2 = 3.5
▶▶ The median is the value that lies halfway between the 3rd and 4th values
▶▶ The 3rd and 4th values are 12 and 12. The median is: (12 + 12) / 2 = 12
4.3 Mode
The mode is the value in a set of data which occurs most frequently. It is identified by counting
the number of times a value occurs in the data set and determining which value occurs most
often. Sometimes a data set may have more than one mode.
Example 5
Determine the mode of the values 12, 15, 18, 7, 12, 8, 3, 19, 2.
The mode is 12 because this number appears twice while the others appear only once.
Example 6
Determine the mode of values 12, 15, 12, 3, 18, 7, 12, 8, 3, 15, 19, 3, 2.
This sequence has two modes, 3 and 12, since both numbers appear 3 times while the others
occur only once or twice.
Calculation of the mean, median and mode for grouped data

If there are many observations in a data set, or if the observations have been grouped together
for other purposes, the mean, median, and mode can be calculated using the frequency
distributions.
Example 7
Table 4.1 Haemoglobin levels in grams/100 ml among Amazonian Gold Miners, Brazil in 2000
Hb in grams fi Number of cases

8.0−9.9 6
10.0−11.9 23
12.0−13.9 9
14.0−15.9 1
16.0−17.9 2
TOTAL 41
Mean
In order to calculate the mean, the formula is:
x= ∑f x / n
i i
where: x is the arithmetic mean

∑ is the sum of the individual values
fi is the frequency of occurrence of an event
(i.e. the numbers in column 2 in the above table 4.1)
xi is the measurement units or the midpoints of the intervals
(see column 3 below)
n is the number of individual values in the data set.
In order to determine the mean haemoglobin value, it is necessary to add a 3rd and 4th columns
to the above table 4.1:
Table 4.2 Haemoglobin levels in grams/100 ml among Amazonian Gold Miners, Brazil, 2000
Class interval
Hb in grams
fi Cumulative
number of cases
xi Midpoint of interval f x
i × i – product
Number of cases shown of columns 2 & 4
8.0−9.9 6 6 8.95 53.70
10.0−11.9 23 29 10.95 251.85
12.0−13.9 9 38 12.95 116.55
14.0−15.9 1 39 14.95 14.95
16.0−17.9 2 41 16.95 33.90

TOTAL 41 470.95
Applying the formula,

∑fi xi = 470.95 and n = 41, hence
x= ∑f x / n = 470.5 g / 41 = 11.5 g
i i
Median
The median will occur in an interval. The steps in calculating a median for grouped data is to
determine the middle data value and then to find what interval this value occurs in. The median
is the midpoint of that interval. The middle data value is (total number of cases + 1) / 2. Here,
there are 41 cases and the middle case is the 21st (class interval 10.0 –11.9). It is now necessary to
find which haemoglobin value corresponds to the 21st observation. This is done by calculating
U4
the cumulative number of cases, starting with the lowest haemoglobin value:
The 21st case is included in the range 10.0–11.9 (cases 7–29). Next the following formula is
applied to determine the exact median value:
Median = L + (JW / f)
where:
L is the true lower limit of the class interval containing the median point
(in the present case, L = 10)
J is the number of cases in this interval below the midpoint case,
calculated as the number of cases below the midpoint (here = 21)
minus the cumulative number of cases up to (but not including) this
interval (here = 6), which is 15
W is the width of the class interval = 2
f is the number of cases in this class interval (in this case = 23)
Applying the formula,
Median = L + (JW / f) = 10 + (15 × 2 / 23) g/ml = 11.30 g/ml
Mode
The mode is the midpoint of the interval that occurs most frequently. Looking at the frequency
column of Table 4.2, we can see that the “10.0 – 11.9” interval occurs most often. So, the mode
is the midpoint of that interval, which is 10.95.
Exercise 4.1
Last month, 20 patients with malaria were admitted to hospital X. The age distribution is as
follows (in years): 4, 3, 3, 1, 2, 26, 64, 3, 2, 5, 7, 4, 22, 3, 1, 1, 12, 2, 3, 6.
a. What is the mode?
b. What is the median?
c. What is the mean?
d. Why are there differences between the median and the mean?
e. Which is a better measure of the age distribution in the population, and why?
Exercise 4.2
The duration of hospitalization for the 11 children admitted last month to hospital X with
cerebral malaria was as follows:
Child 1 3 days
Child 2 7 days
Child 3 8 days
Child 4 5 days
Child 5 4 days
Child 6 11 days
Child 7 6 days
Child 8 10 days
Child 9 5 days
Child 10 1 day
Child 11 4 days
a. What is the mode?

b. What is the median?
c. What is the mean?
d. Why are the median and the mean closer than in Exercise 1?
e. Which is a better measure of the distribution of length of stay of the children in this case? Why?
Exercise 4.3
The following data on parasite density were obtained for 200 consecutive patients seen in
clinic X during the first quarter of 2004:
Density: parasites/1000 WBC Frequency
1000−2999 20
3000−4999 70
5000−6999 80
7000−8999 25
9000−10999 5
Total 200
Calculate:
a. The mean parasite density
b. The median parasite density
c. The modal parasite density
Exercise 4.4
The parasite densities (asexual parasites per µl) of 23 malaria patients admitted to hospital X
last month were as follow:
Parasite density (per Parasite density (per
Patient number Patient number
µl) µl)
1 1069 13 22827
2 3941 14 23585
3 4514 15 29648
4 8048 16 32265
5 8793 17 32739
6 11654 18 38000 U4
7 12736 19 40684
8 13680 20 66211
9 17614 21 67231
10 18630 22 159091
11 21186 23 196500
12 22099
Calculate the mean and median parasite density and discuss the difference between the two
values.
If you know how to calculate the geometric mean, please also calculate this and discuss the
difference with the other participants and the tutor.
LEARNING UNIT 5
Measures of variability
and normal distribution
■ ■ Define the terms range, standard deviation and normal
distribution
■ ■ Describe the advantages and disadvantages of using range,
standard deviation and normal distribution
■ ■ Calculate a range and a standard deviation
■ ■ Calculate and interpret a chi-squared value
43
The measures of central tendency (mean, median and mode) are very useful for describing a
frequency distribution, but they do not indicate the spread of values that may have the same
central tendency. In making decisions for the management of tropical diseases, as in many
other public health fields, it is important to establish what is “normal”. The “normal” value is
a statistical concept and depends, to a great extent, on the distribution of the attribute in the
population. The extent of variability can be summarized through two measures:
▶▶ the range,
▶▶ the standard deviation.
5.1 Range
The range indicates the distance between the highest and the lowest value in the distribution.
Example 1
The range of 11 values 3, 4, 4, 5, 6, 6, 6, 7, 7, 8, 10 is 3 to 10.
Range can also be expressed as 10 –3 = 7
The range is simple to calculate and easy to understand, but the range tells only about two
values of a series of observations. An extremely high or low value may be due to a measurement
error. The range does not take into account variability of observations between the two
extreme values.
5.2 Standard deviation

The standard deviation is a measure that describes the scatter of observations around the mean.
If all the observations had the same value, the standard deviation would be zero; the further
apart from one another (and from the mean) the individual observations are, the greater is
the standard deviation. If the standard deviation of a sample is very small, the sample average
closely represents every individual value; a large standard deviation indicates that this is not so.
The steps in the calculation of the standard deviation are as follows:
▶▶ Calculate the difference between each observation and the mean (xi – x )
▶▶ Square each difference (xi – x )2
▶▶ Add the above squares and divide this sum of squares by the number of observation minus
1, i.e. (n – 1)
▶▶ Calculate the standard deviation by finding the square root of the number obtained in the
above steps in application of the formula:
MEASURES OF VARIABILITY AND NORMAL DISTRIBUTION 45
SD
where: xi each value

x the mean
(xi – x )2 the square of each difference
∑ the sum of
n number of observations
Note that the denominator is n – 1 rather than n. In practice, when n is reasonably large, it makes little difference.
However, for theoretical reasons, n – 1 is preferred.
Example 2
Calculate the standard deviation (SD) of a set of 11 values: 3, 4, 4, 5, 6, 6, 6, 7, 7, 8, 10 . U5
Following the above steps:
▶▶ Calculate the mean: 66 / 11 = 6
▶▶ Calculate the difference between the value of each observation and the mean:
Mean 6
Values xi 3 4 4 5 6 6 6 7 7 8 10
Difference from mean xi –6 –3 –2 –2 –1 0 0 0 1 1 2 4
Squared difference from mean 9 4 4 1 0 0 0 1 1 4 16
▶▶ Add the squares of differences

sum of squares = 9 + 4 + 4 + 1 + 0 + 0 + 0 + 1 + 1 + 4 + 16 = 40
▶▶ Divide by number of observations minus 1 (n –1), where: n = 11 = 40 / 10 = 4
▶▶ Calculate the standard deviation by finding the square root of the result:
SD =
A quicker calculation can be:
∑ x 2 − (∑ x) 2 ÷ n
SD =
n −1

where: ∑ x2 take each observation, square it, then sum the squares
(∑ x)2 sum the observations, then square the sum
Using the values of the previous example: 3, 4, 4, 5, 6, 6, 6, 7, 7, 8, 10 (n = 11)
▶▶ Square each observation and then sum the squares:
9 + 16 + 16 + 25 + 36 + 36 + 36 + 49 + 49 + 64 + 100 = 436
Thus ∑ x2 = 436
▶▶ Then sum the observations and square the sum:

(3 + 4 + 4 + 5 + 6 + 6 + 6 + 7 + 7 + 8 + 10)2 = (66)2 = 4356
Thus (∑ x2) = 4356
(∑ x) 2 4356
with = = 396
n 11
(436 − 396)
and SD = = 4 =2
10
The standard deviation is based on all observations; therefore it is better suited than the range
for describing the distribution.
5.3 The normal distribution

The standard deviation is especially applicable when the underlying distribution is close to
normal (Gaussian), i.e. symmetrically bell-shaped. This is often assumed to be the case for many
biological characteristics, such as height, weight and blood pressure. The normal distribution
has some useful characteristics and many statistical tests can be used if the observations follow
a normal distribution. Approximately two-thirds of the values under a normal distribution
curve fall within ± one standard deviation of the mean, and approximately 95% fall within
± two standard deviations of the mean (Fig. 5.1). However, many biological distributions in
parasitology and epidemiology do not follow a Gaussian (normal) curve.

Figure 5.1 The normal distribution curve
Standard deviation of a percentage

If a sample (of at least 30 subjects) yields a percentage of p%, it is possible to calculate
the standard deviation of this percentage in the population as follows:
p (100 – p)
SD=
n
Consider a sample of 100 persons of whom 80% are women and 20% are men. The standard
deviation of the percentage of women in the population is:
p (100 – p)
SD= with p = 80, 100 – p = 20, n = 100
n
here SD= p (100 – p) = 4

n
It would be expected that there are less than 5 chances in 100 that the proportion of women in
one sample is less than
80% minus (1.96 × 4) or 80% – 7.84% or 72.16%
or more than
80% plus (1.96 × 4) or 80% + 7.84% or 87.84%
If not using a calculator, 1.96 may be rounded up to 2 to simplify calculations.
Note: 95% of the observations fall within ± 2 times the standard deviation.
5.4 Test of association: the chi-square test U5

The chi-square (χ ) is a results statistics test used to show if results are as expected or unusual.
2
The test is commonly used to examine the null hypothesis.

The test is calculated as the sum of the squares of observed values minus expected values
divided by the expected values. The table below shows the distribution of two variables A and B
and how to calculate the chi-squared statistics to test for an association.
Note: the example is given for a 2 x 2 distribution with (2–1)(2–1) = 1 degree of freedom, for which the threshold
values of chi-squared statistics is 3.84 at 5% probability. For degrees of freedom greater than 1, see Table 5.1.
Observed values
Variable A
Present Absent Total
Variable B Present A B A+B
Absent C D C+D
Total A+C B+D A+B+C+D
▶▶ Establish the null hypothesis and calculate the expected frequencies (E) for each observed
(O) cell under the null hypothesis of independent (no association).
If the null hypothesis were not rejected, the expected values would have been as follows:
Expected values
Variable A
Variable B Present (A+C)(A+B) / N (B+D)(A+B) / N A+B
Absent (A+C)(C+D) / N (B+D)(C+D) / N C+D
Total A+C B+D N = A+B+C+D
▶▶ Determine the degree of freedom of the distribution (the freedom to choose frequencies in
the cells under the constraint of fixed marginal totals): this is equal to (number of columns
of data minus 1) × (number of rows of data minus 1).
▶▶ Calculate the chi-squared statistics
Chi-square = χ2 with a degree of freedom (df) = (row – 1)(column – 1)
▶▶ Compare the results with the theoretical distribution of chi-squared to determine significance.
If the calculated chi-square is greater than the tabulated value the null hypothesis can be
rejected at the corresponding level (5%, 10%) of significance. For 1 degree of freedom, the
value of χ2 corresponding to p = 0.05 is 3.84.
Examples 3
Association between recent forest exposure and malaria infection, State A
Exposure
Disease Present 50 11 61
Absent 16 41 57
Total 66 52 118
▶▶ Null hypothesis: absence of association between forest and malaria

▶▶ Expected values
The expected value for the first cell (exposure and disease both+) would be
Exposure
Disease Present 66 x 61 / 118 = 34.12 52 x 61 / 118 = 26.88 61
Absent 66 x 57 / 118 = 31.88 52 x 57 / 118 = 25.12 57
Total 66 52 118
▶▶ Number of degrees of freedom = 1

This table can be filled by difference:
Exposure
Disease Present 50 – 34.12 = 15.88 11 – 26.88 = –15.88 61
Absent 16 – 31.88 = 15.88 41 – 25.12 = –15.88 57
Total 66 52 118
The absolute value of the difference between expected and observed value is:
(50 – 34.12) = 15.88 for all cells: the square of this difference is 252.17
Chi-squared is the sum of the squared differences between observed and expected values/
observed value =
(252.17 / 34.12) + (252.17 / 26.88) + (252.17 / 31.88) + (252.17 / 25.12) = 34.73
This is greater than 3.84 and hence the null hypothesis is rejected and the association between
forest exposure and malaria infection is accepted.
Note: In practice, it is often sufficient to calculate (O – E)2/E for the smallest value of O. If, as is the case here, the
result is > 3.84, the chi-square test is positive and the distributions are significantly different.
Exercises
Calculation of the mean, standard deviation, range, and chi-squared
Exercise 5.1
The duration of hospitalization for 24 children admitted last month to hospital X with
pneumonia was as follows:
Child 1 6 days Child 13 10 days
Child 3 10 days Child 15 14 days U5
a. What is the range of values?

b. What is the mean duration of hospitalization?
c. What is the standard deviation?
d. Is range or standard deviation a better measure of the distribution in this case? Explain your answer.
Exercise 5.2
The following data on the pulse rate were taken on admission from 10 male patients
hospitalized in one week:
Male 1 83 beats/minute Male 6 59 beats/minute
a. What is the range of pulse rate values?

b. What is the mean?
c. What is the standard deviation?
Exercise 5.3
A sample of 200 population surveyed shows the following:
▶▶ Of 94 people with a positive blood slide, 34 regularly use an insecticide-treated bednet.
▶▶ Of 106 people with a negative blood slide, 80 regularly use an insecticide-treated bednet.
a. Tabulate the information
b. Is the distribution of “positive blood slides” and that of “bednet users” significantly different (p =
0.05)?
Table 5.1 Threshold values for chi-squared, p = 0.05 and p = 0.01
Degrees of freedom p = 0.05 p = 0.01

1 3.84 6.64
2 5.99 9.21
3 7.82 11.35
4 9.49 13.28
5 11.07 15.09
6 12.59 16.81
7 14.01 18.48
8 15.51 20.09
9 16.92 21.67
10 18.31 23.21
11 19.68 24.73
12 21.03 26.22
13 22.36 27.69
14 23.69 29.14
15 25.00 30.58
16 26.30 32.00
17 27.59 33.41
18 28.87 34.81
19 30.14 36.19
20 31.41 37.57
21 32.67 38.93
22 33.92 40.29
23 35.17 41.64
24 36.42 42.98
25 37.65 44.31
26 38.89 45.64
27 40.11 46.96
28 41.34 48.28
29 42.56 49.59
30 43.77 50.89
LEARNING UNIT 6
Assessing the accuracy of a test
■ ■ Define the terms sensitivity, specificity, positive predictive
value, negative predictive value, and describe their
importance to health practitioners and patients
■ ■ Describe the trade-offs between sensitivity and specificity
■ ■ Describe the effect of prevalence and incidence on positive
predictive value
■ ■ Calculate and interpret sensitivity, specificity, and positive
predictive value from sample data
51
An ideal laboratory test would detect all people who have a disease and at the same time identify
as normal all those who do not have the disease (Fig 6.1).

Figure 6.1 An ideal test should distinguish healthy and diseased
However, many biological parameters (such as haematocrit readings and blood glucose,) are
based on continuous data and the values between people with and those without a disease may
overlap (Fig 6.2).
true negatives true positives
false negative false positive

Figure 6.2 A schematic graph shows how cut-off point of a test can affect performance of a test
How well a laboratory test performs in the identification of individuals with or without the
disease can be assessed from the values in the following 2 x 2 table:
Disease present Disease absent
Test positive True positives (TP) (a) False positives (FP) (b) TOTAL POSITIVE
Test negative False negatives (FN) (c) True negatives (TN) (d) TOTAL NEGATIVE
From this table, it is possible to calculate the following values that summarize the performance
of the test:
People with the disease TP + FN (a + c)
Prevalence
Total population TN + TP + FN + FP (a + b + c + d)
People with the disease and a positive test TP (a)

Sensitivity
All people with the disease TP + FN (a + c)
People with a neg. test without the disease TN (d)

Specificity
All people without the disease FP + TN (b + d)
Positive predictive People with the disease and a positive test TP (a)
value: % of positives
with the disease All people with a positive test TP + FP (a + b)
Negative predictive People without the disease with a neg. test TN (d)
value: % of negative
without the disease All people with a negative test TN + FN (c + d)
ASSESSING THE ACCURACY OF A TEST 53
For most laboratory tests, the most critical items are:

▶▶ Sensitivity – does this test detect all or most persons with the disease?
▶▶ Specificity – does the test result exclude healthy individuals?
▶▶ Positive predictive value – among those testing positive, what is the probability that they are
patients with the disease?
The cut-off point can be moved down for what constitutes an abnormal value in order to
improve the sensitivity of a test, but by moving the cut-off point down, the rate of false positives
is increased. Conversely, if the cut-off point is moved up in order to provide greater specificity,
the sensitivity will decrease and the proportion of tests that are falsely negative will increase.
Where the cut-off point is set determines the proportion of false positives and negatives.
▶▶ To reduce the chance of having false negatives, in the case of conditions where early action is
important, the cut-off point should be set low (high sensitivity).
▶▶ In some cases, however, a positive test may entail other investigations or treatments that may
involve risk for the patient. In such cases, the chance of false positives should be reduced by
setting the cut-off point higher for higher specificity, even if this lowers the sensitivity of the test.
Sensitivity and specificity are independent of prevalence of a disease. However, positive predictive U6
value is dependent on prevalence, as shown in Figure 6.3:
Figure 6.3 Change of positive predictive value according to prevalence
Example
A new diagnostic test to determine the presence of malaria parasites has been tested on all
patients. The performance of the test has been assessed by comparison of the results with those
obtained by microscopy (the ‘Gold Standard’).
The results obtained are shown in the following table:
Microscopy results (Gold Standard)

TOTAL
Malaria parasites present Malaria parasites absent
Positive 215 16 231
New diagnostic test
Negative 15 114 129
TOTAL 230 130 360
Prevalence TP + FN / Total (215 + 15) / 360 = 64%

Sensitivity TP / (TP + FN) 215 / (215 + 15) = 93%
Specificity TN / (TN + FP) 114 / (114 + 16) = 88%
Positive predictive value TP / (TP + FP) 215 / (215 + 16) = 93%
Negative predictive value TN / (TN + FN) 114 / (114 + 15) = 88%
There is an interesting relationship between the prevalence of a condition and the positive
predictive value of the corresponding test. Consider what might have happened if everyone
coming to the outpatient clinic had been screened. There would have been more negative
slides, indicating that the latter population has less malaria than the previous one. This would
give the following:
Microscopy results (Gold Standard)
TOTAL
Malaria parasites present Malaria parasites absent
Positive 215 248 463
New diagnostic test
Negative 15 1822 1837
TOTAL 230 2070 2300
Prevalence TP + FN / Total (215 + 15) / 2300 = 64%

Sensitivity TP / (TP + FN) 215 / (215 + 15) = 93%
Specificity TN / (TN + FP) 1822 / (1822 + 248) = 88%
Positive predictive value TP / (TP + FP) 215 / (215 + 248) = 46%
Negative predictive value TN / (TN + FN) 1822 / (1822 + 15) = 99%
Sensitivity and specificity remained the same, but the positive predictive value and negative predictive
value changed. Thus, the higher the prevalence results, the higher the positive predictive value.
Conversely, the lower the prevalence results, the lower the positive predictive value.
Tests are sometimes used in sequence in order to maximize their sensitivity and specificity.
For example, the ELISA for HIV has high sensitivity but low specificity. When a positive result
is obtained, the ELISA must therefore be followed by a Western Blot test, which has higher
specificity. Repeating the test on a higher prevalence population also improves the positive
predictive value of the test.
ASSESSING THE ACCURACY OF A TEST 55
Exercises: sensitivity and specificity
A study was conducted to investigate how well the clinical signs (fever + rigors) of malaria
correlates with parasite positivity in a sample of adults from a high malaria endemic
province (2% prevalence) and from a low malaria endemic province (0.2% prevalence).
In both areas, 98% of those who have parasitaemia as diagnosed by a thick smear have
fever and rigors, but 1% of those who have negative parasitaemia on thick smear also have
fever and rigors.
Exercise 6.1
a. Using the thick smears as the gold standard, what is the sensitivity of fever + rigors for detection of
malaria?
b. What is the specificity of these signs?
Exercise 6.2
For a hypothetical population of 100 000 adults in the high malaria endemic area as indicated
above, calculate the following: U6
a. The number of persons who would receive treatment each week, assuming that all those with fever
and rigors would be treated
b. The number who would be unnecessarily treated
c. The positive predictive value
d. The negative predictive value
Exercise 6.3
State in words the meaning of the positive predictive value and negative predictive value that
you have calculated.
Exercise 6.4
Let us see what would happen in the areas of low prevalence.
For a hypothetical population of 100 000 adults in the low endemicity area, calculate the
following:
a. The number of persons who would receive treatment each week, assuming that all those with fever
and rigors would be treated
b. The number who would be unnecessarily treated
c. The positive predictive value
d. The negative predictive value
LEARNING UNIT 7
PART 2
Understanding malaria
at regional and global levels
■ ■ Describe the geographical distribution of malaria in the world
57
Introduction
The Learning Unit 7 of this Guide for Participants consists of a series of exercises to be complet-
ed either individually or in a group. The exercises aim to stimulate discussions and exchange of
experience between participants (who will come from different countries/areas with different
experience). The source materials and responses are in the Guide for Tutors which will be pro-
vided to the participants at the end of the course for future reference.
Figure 7.1 Countries or areas at risk of transmission, 2010
Exercise 7.1
a. Describe the burden of malaria in the world and explain its significance. See Figure 7.1.
Hint: what proportion of the world’s population is at risk of malaria? Malaria transmission
differs in intensity and regularity depending on local factors such as rainfall patterns,
proximity of mosquito breeding sites and mosquito species.
b. What are the differences between endemic and epidemic malaria? Is the pattern of malaria infection
different in the affected areas?
c. Discuss with group members the status of malaria in various parts of the world.
Exercise 7.2
List the factors which are different for malaria in Africa, South-East Asia and America. What are the
characteristics of each of these factors in different continents?
Table 7.1 lists some of these factors; participants may complete them.
UNDERSTANDING MALARIA AT REGIONAL AND GLOBAL LEVELS
Table 7.1 Distribution of malaria characteristics
AMERICAS SOUTH-EAST ASIA AFRICA

Parasitological species of malaria infections
Principal malaria vectors
Estimated proportion of population at malaria risk
Estimated contribution to the global burden of
clinical malaria cases
Exercise 7.3
a. Do differences in the distribution of malaria in various parts of the world affect the objectives of
national malaria programmes? Explain the programme objectives in different countries.
U7
LEARNING UNIT 8
Determinants of
malaria distribution
■ ■ Categorize the determinants of malaria distribution
■ ■ Describe each of these determinants
61
Introduction
This Learning Unit consists of a series of exercises to be completed either individually or
in a group. The exercises aim to stimulate discussions and exchange of experience between
participants (who will come from different countries/areas with different experience). The
source materials and responses are in the Guide for Tutors which will be provided to the
participants at the end of the course for future reference.
Parasite
Environmental factors Mosquito Global factors
Human
8.1 Determinants of malaria distribution

Exercise 8.1
Which determinants influence the distribution of malaria in the world?
In general, the factors affecting malaria distribution can be classified into the following three
groups:
1. Mosquito, parasite, human
2. Environmental factors
3. Global factors
8.2 Parasite
Exercise 8.2
How do the characteristics of the malaria parasite influence malaria distribution in the world?
8.3 Mosquito
Exercise 8.3
How do the characteristics of the Anopheles mosquito influence malaria distribution in the world?
8.4 Humans
Exercise 8.4
a. How can human characteristics influence malaria distribution in the world?
b. How do behavioural factors influence malaria distribution in different regions?
DETERMINANTS OF MALARIA DISTRIBUTION 63
8.5 Environmental factors

Exercise 8.5
a. How can climate determine the distribution of malaria? Give examples of climate change that have
led to a change in malaria transmission.
b. In what conditions does malaria transmission not take place in tropical regions?
c. Which species of malaria is most frequent in regions of cooler climate?
d. How can climate and change in behavioural patterns alter the rate of malaria transmission?
8.6 Global factors

Exercise 8.6
a. Is poverty is a contributing cause of malaria, or is malaria a cause of poverty? How can poverty
influence the distribution of malaria?
b. In addition to those mentioned above, what other factors can influence malaria transmission in a
region?
U8
LEARNING UNIT 9
The life-cycle of the malaria parasite
and its relation to pathogenesis, immune responses,
antimalarial drugs, and field measurements
■ ■ Describe the life-cycle of human malaria parasites
■ ■ Describe important characteristics of infection with each of
the 4 species of human malaria parasite
■ ■ Relate the parasite’s life-cycle to pathogenesis, immune
responses and potential vaccines
■ ■ Identify, in the parasite’s life-cycle, the points of impact of
the main antimalarial drugs
■ ■ Identify in the parasite’s life-cycle what can be measured,
using methods suitable for application in malaria control
programmes.
65
Introduction
The Learning Unit 9 of this Guide for participants consists of a series of exercises to be completed
either individually or in a group. The exercises aim to stimulate discussions and exchange of
experience between participants (who will come from different countries/areas with different
experience). The source materials and responses are in the Guide for Tutors which will be
provided to the participants at the end of the course for future reference.
9.1 The life-cycle of the malaria parasite

Exercise 9.1
a. Working in small groups, discuss ways of preparing a schematic representation of the life-cycle of the
malaria parasite. Consider the different viewpoints of your colleagues before preparing your own
diagram (each individual participant).
Study the diagram of the life-cycle of the malaria parasite provided by the tutor (Fig. 9.1). Compare
it with the one they have prepared and discuss any major differences with the facilitator.
b. Like all sexually reproducing organisms, the malaria parasite is diploid during part of its life-cycle
and haploid during the rest of the cycle. Identify in what parts of the cycle the parasite is diploid and
haploid, respectively. Discuss the biological significance of the haploid and diploid stages.
9.2 Characteristics of infection with each of the 4 species

of human malaria parasites
Exercise 9.2
Working in small groups, list the different characteristics of an infection with each of the 4
species of human malaria parasites.
Prepare:
▶▶ A table of time factors (i.e. prepatency, duration of incubation period, time of appearance of
gametocytes, and duration of asexual cycle in the blood) for each species
▶▶ A table of multiplication factors (number of sporozoites per oocyst, numbers of merozoites
per hepatic schizont and per blood schizont) for each species
▶▶ A short list of other important differences among the species
Compare the results of the exercise with the tables and list presented by the tutor.
Discuss the differences.
9.3 The life-cycle of the parasite, pathogenesis, and immunity

Exercise 9.3
a. On a copy of the diagram of the parasite’s life-cycle (Fig. 9.1), indicate the main ways in which the
parasite can produce disease. List the main mechanisms of pathogenesis. Compare with the tutor’s
Figure 9.2. Discuss differences and implications.
THE LIFE-CYCLE OF THE MALARIA PARASITE 67
b. On a copy of Figure 9.2, indicate which stages of the malaria parasite provoke naturally-acquired
immune responses (IRs) to the malaria parasite. For each category of IR, indicate by what parasite
life stage(s) it is stimulated, and in what part of the life-cycle it exerts its impact. Compare to the
tutor’s Figure 9.3. Discuss the differences and implications.
c. On a copy of Figure 9.3, indicate the expected points of impact of different types of potential malaria
vaccines. Compare with the tutor’s Figure 9.4. Discuss differences and implications.
9.4 The parasite’s life-cycle and antimalarial drugs

Exercise 9.4
On a copy of the diagram of the parasite’s life-cycle (Fig. 9.1), indicate the points of impact of
the main antimalarial drugs. Compare with the tutor’s Figure 9.5. Discuss differences and
implications.
9.5 Epidemiological measurements

Exercise 9.5
a. On another copy of the same diagram (Fig. 9.1) identify the possible epidemiological measurements.
Indicate what can be measured technically and state by what methods. Compare with the tutor’s
Figure 9.6. Discuss differences and implications.
b. What are the tools for diagnosis of human malaria cases at the field level?
c. What are their diagnostic performances?
U9
LEARNING UNIT 10
The life-cycle of the vector
and factors that influence it in relation
to malaria transmission
■ ■ Describe the life-cycle of the malaria vectors
■ ■ Describe the factors affecting the vector’s life-cycle in
relation to malaria transmission (including vector control)
■ ■ Relate vector collection methods to the vector’s life-cycle,
and entomological data to their epidemiological interpretation
69
Introduction
participants at the end of the course for future reference.
10.1 The vector’s life-cycle

Exercise 10.1
Working in groups, discuss ways of preparing a diagram of the life-cycle of the malaria vector. Consider
the different viewpoints of your colleagues before preparing your own diagram.
Study the diagram of the life-cycle of the malaria vector provided by the tutor (Fig. 10.1). Compare it
with the diagram your group have prepared and discuss any important differences with the facilitator.
10.2 Factors affecting the vector’s life-cycle in relation

to malaria transmission
Exercise 10.2
Working as a small group, list the major factors that have a direct effect on the vector’s life-cycle,
and then organize these factors into appropriate groups.
The following groups may be considered:
1. Numerical factors, e.g. duration of life-stages
2. Vector behaviour
3. Vector-parasite interactions, perhaps including possible density-dependent regulation in the
vector
4. Differences among vector species
5. Factors of the physical environment, perhaps subdivided according to type of effect on the
vector’s life-cycle, e.g. vector production, vector survival, human-vector contact, etc.
10.3 Vector control measures and their point of impact

Exercise 10.3
Group vector control methods according to their type, e.g. chemical, and to the stage of the vector’s life-
cycle to which they are applicable; also consider methods for reducing human-vector contact.
On a copy of the diagram of the life-cycle of the malaria vector (Fig. 10.1) indicate the points of impact
of the different control measures. Compare this diagram with the one presented by the tutor (Fig.10.2).
THE LIFE-CYCLE OF THE VECTOR 71
10.4 Efficacy of vector control measures

Exercise 10.4
Working in groups, discuss the following:
Do all control measures have the same effect on the transmission of malaria? If not, how and why do
they differ?
Consider the path followed by a vector in acquiring, incubating and transmitting the infection, and
the number of times the same individual vector passes through the various stages represented on the
diagram (Fig. 10.2).
Are conclusions about the expected efficacy of particular control methods affected by the assumption
that all the mosquitoes in a given population may or may not behave in the same way?
For instance a given amount of exophilic behaviour (outdoor resting after a blood-meal) may
be a function of the whole vector population or of only a part of that population (exophilic
subpopulation). Would that affect the efficacy of indoor residual spraying?
Similarly a given amount of zoophilic behavior (proportion of bites on animals) could be a
characteristic of the total population or concentrated only in a subpopulation; would that affect
the impact of a control measure on transmission?
10.5 Measurement methods

Exercise 10.5
On another copy of the same diagram of the vector’s life-cycle (Fig. 10.1), indicate the main
vector collection methods that are technically feasible in a control programme.
A number of variables can be measured in the samples collected, e.g. sex, species, number, age-
class of females, nulliparous or parous, infection rates, etc., and from these measurements
various indicators can be calculated, e.g. human-biting rate, sporozoite rate, survival rate, etc.
Information on the measurement of the variables and calculation of the indicators is provided
in the module Malaria Entomology and Vector Control.
Compare the diagram with the one presented by the tutor (Fig. 10.3).
U10
10.6 Interpretation of entomological data
Three problems of epidemiological interpretation of entomological data will be considered
here.
10.6.1 Representativeness
Representativeness of entomological measurements: relationship between the actual measure-
ment and the reality it is intended to measure.
The aim is to measure vector density and vector behaviour.
Exercise 10.6
a. Do the collection methods allow vector density and vector behaviour to be measured independently
of each other?
b. If the answer is no, do the methods measure the “true” density or only indicators of trend?
c. Is this significant for measuring the impact of vector control, e.g. entomological impact of the
introduction of insecticide-treated mosquito nets?
10.6.2 Ratio between sample size and population size

Exercise 10.7
a. Are entomological sampling fractions usually large, small, or very small?
b. Consider the following example. In an isolated village, during one whole night, 20 female
mosquitoes of species X are collected on two human baits. What other variables must be known to
calculate the mosquito population of that night in that village?
c. Attribute some plausible value to each of these variables, and calculate the mosquito population
size and the sampling fraction.
d. Is it common to find new cases of human infection without finding infected vectors? Can the sampling
fraction of the two measurements be compared? Starting from the example just calculated, does
the comparison explain the discrepancy?
10.6.3 Calculating survival from age-composition

Exercise 10.8
a. Name the two variables that determine age-composition of adult female vectors at a given point in
time.
b. What is being assumed if survival is calculated from age-composition at a given point in time?
c. What is being assumed if survival is calculated from average age-composition over a period of time,
e.g. a transmission season?
LEARNING UNIT 11
Natural history of malaria in the
human host and factors that affect it
■ ■ Describe the natural history of malaria in the human host
■ ■ Describe the factors that have an effect on the natural
history of malaria in the human host, including immunity and
malaria control measures
■ ■ Describe the role of improved diagnostic testing and
treatment services at intermediate and peripheral levels in
reducing morbidity and mortality due to malaria
73
Introduction
participants at the end of the course for future reference
11.1 Natural history of malaria in the human host

Exercise 11.1
Working in small groups, discuss ways of preparing a diagram of the natural history of malaria
in the human host. Prepare a flowchart from inoculation to death, including intermediate states
and reversible steps.
After considering the viewpoints of colleagues in the group, each participant should draw his/
her own diagram of the natural history of malaria in the human host.
Study the flowchart (Fig. 11.1) presented by the tutor. Compare it with the one prepared by the
group and discuss any important differences with the tutor or facilitator.
Exercise 11.2
Discuss the following three questions in the small group, draft answers, and compare them
with the tutor’s answers:
a. How rapidly do patients move from one state to another?
b. What are the implications for case-management?
c. Does the risk of disease and the risk of severe disease change in the course of time after inoculation?
Exercise 11.3
Not every case proceeds all the way from inoculation to death and different cases stop at different
intermediate states. Discuss in group what factors might affect the outcome of an inoculation, and
how to classify the factors in a meaningful table. Each participant should prepare a table. Note
that this section covers the identification and classification of factors; subsequent sections will
consider the different classes of factors, their possible importance and mode of action. Compare
the table with the one presented by the tutor (Table 11.1) and discuss any differences.
11.2 The inoculum’s intrinsic factors

Exercise 11.4
In small groups, discuss the following questions, then draft individual answers.
1. Number of sporozoites inoculated
a. How many sporozoites does a vector inoculate?
b. How can the number of sporozoites be measured?
c. Does the number of sporozoites affect the outcome?
NATURAL HISTORY OF MALARIA 75
2. Differences of “virulence” among parasite species

a. Why is P. falciparum the most pathogenic species? (see Learning Unit 10).
b. Do other species of Plasmodium cause lethal forms of malaria?
3. Differences of “virulence” within a parasite species or local parasite population
a. What observations suggest that there are such differences, within a species and within a local
population of a species?
Consider the following:

i. Some children who develop severe malaria, presumably after a recent inoculation, have
received many previous P. falciparum inoculations without developing severe malaria.
ii. Some asymptomatic parasitaemic children develop clinical malaria (uncomplicated
malaria or severe malaria) following superinfection; can that be explained by the
quantity of parasites added? Example: a child with an asymptomatic parasitaemia of
1000 parasitized RBC/µl, and a blood volume of 2000 ml, in which superinfection leads
to the successful maturation of 20 liver schizonts.
iii. If you conclude that the outcome – disease or no disease, uncomplicated malaria or
severe malaria – is related to the quality of the parasites rather than their quantity, can
you also conclude that the parasite’s variable intrinsic virulence is a sufficient explanation
of such differences of outcome? Discuss the answers and discrepancies with the tutor.
11.3 The human host’s intrinsic factors

11.3.1 Genetic traits and susceptibility
Exercise 11.5
List and discuss briefly examples of genetic traits that decrease or increase the human host’s
susceptibility to malaria.
11.3.2 Acquired immunity (active)

Exercise 11.6
a. What stimulates active immunity to malaria? Is it “all or nothing” or a gradual process? Is it species-
specific? Is it reversible? If so, how is it reversed? Discuss the questions in group, then draft individual
answers.
b. Draw a diagram of the effect(s) of active immunity on the natural history of malaria in the human U11
host. After discussing the subject in group, draw individual diagrams. Suggestion: on another copy
of the flowchart of malarial states (Fig. 11.1) draw, to the right of the flowchart, two additional,
modified, flowcharts corresponding to “moderate” and “high” levels of active immunity.
Compare the diagram to the tutor’s (Fig. 11.2). The diagram suggests that successive
inoculations are progressively less pathogenic and less dangerous. However, according to a
previous statement (see above), some children developing severe malaria, presumably after a
recent inoculation, have received many previous P. falciparum inoculations without developing
severe malaria. Is there a contradiction? Discuss.
c. Discuss, in relation to the diagram, the expected effects of different malaria control measures, such
as vector control, control of human-vector contact, treatment of cases, potential vaccines.
11.3.3 Acquired immunity (passive)

Exercise 11.7
a. How is passive immunity acquired? Is it reversible? If so, how is it reversed? After discussing in group,
answer individually.
b. Draw a diagram of the effect(s) of passive immunity. After discussing in group, draw individual
diagrams. Suggestion: on the copy of Figure 11.2 on which the effect(s) of active immunity have
already been indicated and add, to the left of the original flow-chart, a modified flowchart,
corresponding to passive immunity.
c. Compare the answers and diagrams with the tutor’s, in particular comparison with Figure 11.3.
11.4 Other human biological factors

Exercise 11.8
Discuss how pathogenesis is affected by:
a. Pregnancy
b. Nutritional status
c. Age per se (i.e. separately from its association with immunity)
11.5 Interaction between parasite diversity and host diversity

It appears that intrinsic parasite and host factors can explain only part of the variation in
outcome of host-parasite encounters, and that interaction may be important.
Exercise 11.9
a. Consider the parasite:
i. Does a parasite present one, a few, or many antigens to the host?
ii. Do different parasites of the same species present much antigenic diversity, or a little, or none?
b. Consider the human host:
i. Can every host acquire protective immunity against every antigen?
ii. Does the set of antigens to which a host can respond, i.e. can acquire protective immunity, differ
among hosts?
c. Do the answers to (a) or (b) suggest something about pathogenesis, other than the intrinsic virulence
of the parasite and the intrinsic susceptibility of the host?
NATURAL HISTORY OF MALARIA 77
11.6 Age-specific distribution of malaria

Exercise 11.10
Working in small groups, carry out the following exercise.
a. Suppose that exposure (to vectors) does not vary with age; show on a diagram the kinds of age-
specific distributions that would be expected for:
i. Malaria infection (parasitaemia)
ii. Malaria disease
iii. Malaria deaths
b. How would those distributions change if transmission increases or decreases? Show on a separate
diagram the effect of transmission changes.
c. Does exposure to vectors vary with age? If so, is this important?
How will the diagram appear in case of sudden exposure of a non-immune population to
intense transmission?
Compare the graph with the one provided by the tutor (Fig. 11.4) and discuss the differences.
11.7 Malaria mortality

In Figures 11.1 and 11.2 (presented by the tutor) the diagrams show that malaria (P. falciparum)
kills through severe malaria, i.e. clinical forms of malaria such as cerebral malaria, likely to be
lethal if left untreated. Is that the only way that malaria is responsible for death? Consider the risk
of death associated with some other diseases, e.g. pneumonia, measles, malnutrition. Among
children suffering from pneumonia, for example, some may suffer from uncomplicated malaria
at the same time. Is the risk of dying the same among the children suffering from pneumonia
plus uncomplicated malaria, as among the children suffering from pneumonia alone? More
generally, is the addition of uncomplicated malaria likely to affect the case fatality rate (CFR)
of some other diseases?
If the addition of uncomplicated malaria increases the CFR of some other diseases, the
additional deaths may be termed indirect malaria deaths, while the deaths due to severe malaria
may be termed direct malaria deaths, and the sum of the two may be termed total malaria
deaths. Participants should add these concepts to Figure 11.1, then compare the new diagram
to the tutor’s Figure 11.5 and discuss the differences.
Exercise 11.11
U11
a. If there is both direct and indirect mortality from malaria, is their relative magnitude of practical
importance, and, if so why?
b. What kind of data would allow estimation of their relative magnitude?
c. If such data are known to the participants, what do they show?
LEARNING UNIT 12
Intensity of malaria transmission
■ ■ Define the major parameters of intensity of transmission
which are used in malaria epidemiology
■ ■ Identify the relationships between the vectorial capacity,
the basic reproduction rate, the inoculation rate, and the
incidence and prevalence of malaria infection
■ ■ Describe the expected impact of mass drug administration
and/or vector control activities on malaria transmission at
different levels of endemicity
■ ■ Distinguish what models can or cannot contribute to the
planning of malaria control
■ ■ Indicate the epidemiological methods for measuring malaria
morbidity and mortality and how information can be collected
■ ■ Describe how the relationship between vectorial capacity
and other concepts of the intensity of malaria influence the
selection of control methods in different epidemiological
situations
79
12.1 Intensity of transmission

The concept of “intensity” can be applied to the transmission of the malarial infection, to the
disease, and to the related mortality.
Exercise 12.1
Name and define the different concepts of intensity that are applicable to malaria infection and
to its transmission.
12.2 Vectorial capacity and basic reproduction rate

12.2.1 Vectorial capacity (C)
Vectorial capacity C is defined as the potential number of secondary cases originating from
one primary case in one day, assuming that the population is, and remains, fully susceptible. In
malaria, where transmission is indirect, the anopheline mosquito’s circumstances determine
vectorial capacity. Note that the term refers to the combination of the components determining
the effectiveness of a local mosquito population to transmit malaria and, although the human
host is the beginning and end of the sequence, the term refers to the insect stages, not those
occurring in humans.
Understanding the concept of vectorial capacity helps in understanding the factors that
determine malaria transmission, and is therefore important for planning malaria control
strategies. The vectorial capacity is a measure of the efficiency of a local vector population in
potentially transmitting the parasite.
12.2.2 Basic reproduction rate (Ro)

The basic reproduction rate Ro is the potential number of secondary cases that could potentially
originate from one primary case during the whole period of the patient’s illness, assuming
that the community is and remains fully susceptible to malaria infection. Note that the basic
reproduction rate is a potential multiplication factor, not an actual secondary attack rate.
12.2.3 A revision of some definitions

Exercise 12.2
Working in small groups:
a. Revise and discuss the definitions of primary and secondary cases.
b. If a malaria patient has gametocytes in the blood for 14 days only (during which he/she can transmit
the infection), and this generates 5 new cases which can be added every day, then how many new
cases would be added to the community?
c. What is the vectorial capacity in the example above? What is the basic reproduction rate?
Note: The definition of vectorial capacity presented here includes the word ‘potential’. If a malaria patient
is present in the area, vectorial capacity is the number of cases that could potentially originate from this
person. Vectorial capacity is a measure of the potential for malaria transmission in an area, and it is one of the
characteristics and of the combined capabilities of anopheline mosquitoe population present in an area; malaria
itself may not be present in the area. In other words, vectorial capacity can be calculated for regions which have
anopheline mosquitoes but no malaria.
INTENSITY OF MALARIA 81
12.2.4 Derivation of a formula for vectorial capacity (C)

Exercise 12.3
Create a formula for vectorial capacity that considers the relation between the factors involved
in transmission by reviewing the events in malaria transmission from one patient to another
through mosquitoes.
Note: The following stages can be considered in malaria transmission:
▶▶ Stage 1: The primary case is bitten by a certain number of mosquitoes per day.
▶▶ Stage 2: A fraction of these vectors acquire the infection. Infection is transmitted by a
proportion of mosquitoes that have sucked blood. (Not all mosquitoes will become infected
by sucking blood from a malaria-infected patient and later produce sporozoites. Many
variables affect this process resulting in a situation where the probability of transmission of
infection to the insect is less than 100 %.)
▶▶ Stage 3: Some of the mosquitoes that have sucked blood survive the sporogonic cycle (the
parasite’s incubation period in the vector) and become infective.
▶▶ Stage 4: Some of the infective mosquitoes will live for a period of time during which they will
bite a number of humans.
▶▶ Stage 5: A fraction of humans bitten by infective (sporozoite-carrying) vectors acquire the
infection and go on to develop the disease – these are the secondary cases of malaria. It is
important to remember that the disease does not develop in all bitten persons.
12.2.5 Relationship between vectorial capacity and

basic reproduction rate
Exercise 12.4
a. Write an equation for the relation between vectorial capacity and basic reproduction rate variables
b. Define the possibility of recovery of a patient, or recovery rate, in a day "r".
c. Is there a relation between a patient’s recovery rate and the duration of disease?
d. Keeping in mind the answers to questions (b) and (c), enter "r" in the relation between vectorial
capacity and basic reproduction rate.
12.2.6 Derivation of a formula for basic reproduction rate (Ro)

Exercise 12.5
Based on the relationship between vectorial capacity (C) and basic reproduction rate (Ro),
write an equation for the basic reproduction rate Ro.
12.3 Vectorial capacity, basic reproduction rate, and control U12

of transmission
A number of deductions concerning malaria control can be made from the formulae for
vectorial capacity and basic reproduction rate.
Exercise 12.6
a. Identify for each control measure below the factor(s) affected.
Control measures Factors affected (among m, a, p, n, 1/r) a
Residual spraying
Space spraying
Source reduction
Larviciding
Reduction of human-vector contact
Insecticide-treated mosquito nets
Treatment of cases
a
m = vector density in relation to human; a = vector’s biting rate on humans; p = survival of the vector; n = duration of
sporogony cycle; 1/r = duration of infectivity (in humans)
b. On the basis of formulae for C and Ro evaluate the extent to which the variation in the different
factors listed above affects the magnitude of the vectorial capacity or the basic reproduction rate.
c. Calculate C, given m = 10, a = 0.5, p = 0.8, n = 10. Calculate the effect on C of reducing either m, a
or p by half. Why is a reduction of n not included in the exercise?
d. What can be concluded concerning the relative efficacy of the different control measures?
12.3.1 Identification of simplifying assumptions

Exercise 12.7
In deriving the formulae for vectorial capacity and basic reproduction rate, and in making
deductions about control, some simplifying assumptions were made. What are these
assumptions? Are they important, and if so, why? (see Learning Unit 11).
12.3.2 Measurement of intensity

Exercise 12.8
Working in small groups, consider successively the five kinds of “intensity” identified at the
start of this unit. For each one, go through the following exercise:
a. Can it be measured?
b. If so, how?
c. Discuss with your colleagues the measurements in terms of:
i. Technical feasibility
ii. Cost
iii. Reliability
d. Can the concept – especially those of vectorial capacity and basic reproduction rate – be
useful in the absence of the corresponding measurements?
12.3.3 Relationship between prevalence and incidence

Exercise 12.9
a. Discuss (in the group) the relationship between prevalence and incidence.
b. Express the relationship between prevalence and incidence as a formula.
c. Give a numerical example. For example: if the incidence rate is 200/1000/year, and a “case” lasts
on the average two months, what is the expected prevalence rate?
12.3.4 Qualitative relation between the different dimensions of

intensity of malarial infection and its transmission
Exercise 12.10
Working in small groups, construct a diagram in which the five dimensions of intensity
previously identified are connected by arrows indicating what determines what; you may need
to introduce some other factors, perhaps the duration of infection, immunity?
Discuss with your colleagues the relationships between the parameters of intensity of malaria
shown in the figure 12.1 provided by the tutor.
12.3.5 Quantitative relationship between prevalence and

vectorial capacity
To examine the relationship between vector capacity and prevalence of disease, Ronald Ross
divided the community into two groups – the ill group and the healthy group. In his model,
each day a number of the healthy acquire malaria and a number of ill individuals are cured
and return back to the healthy group. Note that drug treatment has no place in this model
and it is assumed that the ill individuals recover once the 1–5 year course of the disease is over
(recovery rate "r"). On the other hand, the healthy people become ill at a rate that is the product
of the prevalence (represented by "y") multiplied by vectorial capacity (represented by "C"); his
model can be represented as follows:
Negatives Positives
l-y y

On this basis Ross defined the mathematical model of malaria as follows:
y(t+1) = y(t) + y(t) C [1 – y(t)] – r y(t) FORMULA 1
where y = proportion of positives in the human population
1 – y = proportion of negatives in the human population
C = vectorial capacity (per time unit) U12
r = recovery rate (per time unit)
(t) means “at time t”
(t+1) means “at time t + 1 time unit”
The logic behind formula 1 is as follows: the prevalence at time (t+1) is equal to the prevalence
at time "t", plus the new cases occurring in the interval and minus the old cases recovering in
the interval.
Exercise 12.11
Working in small groups
a. Discuss the components of formula 1 and their relationship to each other.
If malaria in the region is in a state of equilibrium, i.e. the disease prevalence is neither rising
nor declining. Derive a formula (formula 2) for y as a function of C:
FORMULA 2

b. Using formula 2, draw the relationship between vectorial capacity (C) and disease prevalence (y)
in a graph. Note that in formula 2, prevalence is negatively related to recovery rate (r) and positively
related to vectorial capacity (C). Compare this graph with figure 12.2 provided by the tutor.
0
C
c. Assuming r=0.5, what are the changes in y for each unit change in C? Fill the following table:
C y
0.5
1.5
2.5
What should be y, if C=0.25? justify your answer.

d. Draw this graph in MS-Excel by simulating it for different rates of recovery and vectorial capacity.
e. Compare the graphs with figure12.2. Discuss various sections of the graph and try to interpret and
adjust it with different epidemiologic circumstances and conditions of malaria (study the ascending
slope and plateau of the curve).
f. Discuss the point of intersection of the graph with the X axis.
g. Having established the relationship between vector capacity (C) and basic reproduction rate Ro ,
prepare formula 2 and the graphs according to the basic reproduction rate Ro (discuss the graph
and equation obtained).
h. What would be y for different values of R0 in the following table:
R0 y
0.5
2

i. What should be the number of cases after four generation, if R0=2 and the primary number of cases
be 8? Re-calculate the number of cases for R0=0.5.
In summary, the following can be deduced from the Ross model:

1. There is a non-zero critical value of vectorial capacity below which malaria cannot
maintain itself in an area (C*). Vectorial capacity does not have to reach zero, nor does the
anopheline mosquito have to be completely eradicated in an area in order to eliminate
endemicity. This implies that malaria could disappear without the elimination of all
vectors.
2. There is no linear relationship between the prevalence of disease and vectorial capacity
above the critical value stated for vectorial capacity. Near the critical value minor
changes in vectorial capacity lead to major changes in prevalence or endemicity of
malaria in the region, i.e. malaria is unstable in the region. But a little above the critical
value, where the curve comes to plateau, even large changes in vectorial capacity will
not affect endemicity and do not displace it, i.e. malaria is stable in the region. Stable
malaria, in classifications of endemicity, is characterized by a high level of transmission
and prevalence of disease without marked seasonal changes, and regardless of
environmental events such as reduction or increase in rainfall in the region.
3. The danger of regional outbreaks is near the critical value where malaria is unstable.
Consequently, problems in the health sector and malaria control are more pronounced
in the area. In fact, it is near this critical value that vector density control has a significant
effect on malaria control and elimination.
4. The critical value for vectorial capacity that is represented by C* occurs when the
vectorial capacity and recovery rate ‘r’ are equal. It can be concluded that the lower the
recovery rate of malaria (or the longer lasting the infection), the higher is the vectorial U12
capacity. Thus the critical vectorial capacity below which the transmission cycle is
interrupted is lower for P. vivax than for P. falciparum.
Hint: R0=1= . Therefore, whenever C = r, R0 is at the situation that the epidemics cannot
occur. Therefore, when r=C, y will be zero. In this critical value we will have C*.
12.3.6 The value of the vectorial capacity below which malaria

transmission cannot be maintained
Exercise 12.12
Discuss in small groups why malaria transmission cannot maintain itself if C is below a certain
value and complete the following exercises.
a. What is the critical value of C below which malaria cannot maintain itself, i.e. C*?
b. What is the critical value of the vector density (represented as m*)?
c. Calculate m* for an efficient and an inefficient vector respectively, using the following numerical
values
r n a p m*
Efficient vector 0.01 10 0.5 0.9
Inefficient vector 0.01 10 0.1 0.5
d. Add on the graph (Fig. 12.2) the incidence rate, the entomological inoculation rate, and the
proportion seropositive as functions of C or Ro. Compare the group’s graph with the one provided
by the tutor (Fig. 12.3). Discuss the differences.
e. What does the graph imply in terms of selection of measurement methods to describe malaria
situations? Can a useful approximation of the vectorial capacity (or of the basic reproduction rate)
be obtained indirectly, i.e. without actually measuring them?
12.3.7 Effects of a reduction in prevalence or in vectorial capacity

Exercise 12.13
Working in small groups, discuss the effects of a reduction in prevalence or in vectorial capacity,
then complete the following exercise.
a. If the prevalence of infection (or the “parasite reservoir”) has been suddenly reduced to a low level by
a single mass drug administration, what would be expected to happen to the prevalence after that?
b. If the vectorial capacity has been suddenly and permanently reduced to a lower level, what would be
expected to happen to the prevalence after that?
Ross used the model represented by formulas 1 and 2 to explore these two questions, and
reached conclusions which are generally accepted today; participants could try to do the same.
Formula 1 is easy to calculate on a pocket calculator, especially if programmable.
To explore question (a), use C = 1 and r =0.5, and start at y (0) = 0.5; calculate y (1) .... y (5);
change arbitrarily y (5) to 0.1, calculate y (6), y (7), etc., and study the result – it may be helpful
to plot the values on a graph.
To explore question (b), use C = 1 and r =0.5, and start at y (0) = 0.5; calculate y (1) .... y (5);
reduce C from 1 to 2/3, then calculate y (6), y (7), etc. and study the result – again, it may be
helpful to plot the values on a graph.
Compare the group’s graphs with the ones provided by the tutor (Fig. 12.4). Discuss the
differences.
12.3.8 Is there a place for models in planning malaria control?

Exercise 12.14
a. Are any (or all) of the following statements correct?
i. Facts are preferable to models
ii. Models make questionable assumptions
iii. Models use questionable numerical values
iv. Models have failed – and will continue to fail – to make accurate predictions
b. If some (or all) of the statements are correct, is there any place left for the use of models in planning
malaria control?
12.4 Measuring the burden of malaria

Exercise 12.15
a. Of the five concepts applicable to the “intensity” of malaria in terms of infection, which ones are
applicable to malaria in terms of disease?
b. The burden of malaria is due to which circumstances?
c. Are there any other measurement(s) that could be applicable to the quantification of malarial
disease in a population?
d. How can the mortality from malaria be quantified?
U12
LEARNING UNIT 13
PART 3
Overview of stages in malaria
programme phases
■ ■ Compare and contrast the concepts of malaria control,
elimination and eradication
■ ■ Describe the different objectives for each phase of malaria
programme
89
13.1 Phases of a malaria programme

In 2012, malaria transmission occurred in 99 countries in tropical and subtropical zones,
spanning all continents of the world except Antarctica and Australia, with intensities of
transmission that vary from very low to extremely high.1 The objectives of malaria control
programmes range from reducing the disease burden and maintaining it at a reasonably low
level, to eliminating the disease from a defined geographical area and, ultimately to eradicating
the disease globally.
History shows that new goals and targets for global malaria control and elimination must be
realistic in order to avoid disappointment and disillusionment and the devastating implications
of disease resurgence, as experienced in the past when intensified control efforts were
interrupted. The prevailing state of health systems and the epidemiology of malaria must be
taken into account in setting realistic targets.
Exercise 13.1
a. Define malaria control, elimination, and eradication.
b. Match the levels of malaria transmission (1, 2, 3 and 4) with the appropriate phases of malaria
programme (a, b, c, and d) in the box below.
1. Areas where parasite rate (PR) has been reduced to < 1 malaria
a. Control
case/1000 population/ year
2. Areas where malaria parasite rate has been reduced to < 5% among
b. Elimination
fever cases in health facilities
3. Areas where transmission has been interrupted c. Pre-elimination
4. High to moderate transmission d. Prevention of re-introduction
c. Illustrate the phases of malaria programmes on a separate sheet. Compare your diagram with
Figure 13.1 provided by the tutor.
There are four different levels of malaria transmission as described in Table13.1.
Table 13.1 Different levels of malaria transmission
Characteristicsb
Transmission levela
Endemicity level Parasite prevalence
Low transmission Hypoendemic ≤ 10% in children aged 2–9 years
Moderate transmission Mesoendemic 11–50% in children aged 2–9 years
High transmission Hyperendemic Constantly > 50% in children aged 2–9 years
High transmission Holoendemic Constantly > 75% in infants aged 0–11 months
a. WHO 2012. Disease surveillance for malaria control: an operational manual. Geneva, World Health Organization. http://
www.who.int/malaria/surveillance_monitoring/operationalmanuals/en/index.html
b. WHO 1963. Terminology of malaria and of malaria eradication. Report of drafting committee. Geneva, World Health
Organization.
WHO 2012. World Malaria Report. Geneva, World Health Organization. http://www.who.int/malaria/publications/world_
1.
OVERVIEW OF STAGES IN MALARIA PROGRAMME PHASES 91
13.2 Transition from malaria control to elimination

Countries are engaged in different phases of malaria programme according to the malaria
transmission levels:
▶▶ High to moderate transmission areas: In areas where the geographical distribution of
malaria cases is widespread and intensity of transmission is high , the prime objective is to
reduce malaria mortality, morbidity and intensity of transmission. In these areas, the malaria
control programme should make all possible efforts to achieve universal coverage of cost-
effective preventive, diagnostic testing and treatment interventions. In addition, intensified
monitoring and surveillance, and cross-border collaboration are critical to reduce the burden
of malaria to consistently less than 5% parasite positivity rate among febrile patients with
suspected malaria at health facility level or among people of all ages with current fever or
history of fever in the past 24 hours identified in population based surveys.
▶▶ Low transmission areas: In some areas, natural conditions and/or control efforts have
reduced the transmission intensity to low levels and localized geographical areas, allowing
programme re-orientation towards elimination. The goal of the pre-elimination phase
is to reduce the malaria burden to an incidence rate of less than 1/1000 people at risk at a
sustainable level. This is achievable by perfecting the surveillance system and targeting of case
management and vector control operations, in residual and new active foci. A notification
system, and regulation of the private sector, must be established at this stage. In this phase,
the subsequent aim is to halt local malaria transmission through targeted interventions to
populations at risk in malaria foci.
▶▶ Areas where transmission has been interrupted: In these areas, the goal will be to
prevent re-establishment of local transmission from imported cases. Once the entire country
has been free from local malaria transmission for three consecutive years, the process for
certification of the malaria-free status can be started.
Countries engaged in the different malaria programme phases
Of the 99 countries that were considered malaria-endemic in 2012, 79 countries are in malaria
control phase; 10 countries are in pre-elimination phase; 10 countries are elimination phase.
An additional 5 countries are in the prevention of re-introduction phase. During and after the
Global Programme for Malaria Eradication, 30 countries were certified by WHO as malaria-
free in the period up to 2012. In addition, 53 countries malaria never existed or disappeared
without specific measures.
Moving from the control to the elimination phase requires changes in strategies. The priority
of the malaria control phase is the achievement of universal coverage with preventive methods
and access to treatment. Discuss the priorities of malaria elimination programmes in relation
to:
▶▶ Detection of all malaria cases
▶▶ Prevention of onward transmission
▶▶ Management of malaria foci
U13
▶▶ Management of importation of malaria parasites.
Elimination programmes require more advanced technical malaria expertise than control
programmes, particularly in malaria epidemiology, entomology and surveillance. Most
countries adopt the elimination strategy in a phased approach, targeting specific parasite
species (e.g. P. falciparum first) or geographical areas. When the number of locally acquired
cases becomes very low (e.g. below 100, 10 or even fewer nationwide), importation of malaria
parasites from abroad becomes a greater threat, and prevention of reintroduction of malaria
becomes increasingly important.
The differences in units of intervention show the significance of classifying patients on the basis
of (i) source of infection/transmission, (ii) time (recent or very recent), (iii) location (a specific
region), and (iv) mode of transmission (blood and/or mosquito). Malaria cases can be therefore
classified in 5 groups as follows:
▶▶ Induced: a case resulting from contamination with infected blood
▶▶ Imported: a case contracted outside a given place
▶▶ Relapsing: a case of P. vivax or P. ovale contracted locally some time ago or detected after
a period of unrecognized latency
▶▶ Introduced: a case contracted locally from an imported case
▶▶ Indigenous: a case contracted locally from any other category of cases, including other
indigenous cases
Exercise 13.2
a. Different malaria programme phases have different programme objectives. Specify in Figure 13.1
‘main objectives’ in each phase. Compare the results with the Figure 13.2 provided by the tutor.
b. Do malaria programme phases differ in their (i) epidemiologic objectives and (ii) transmission
objectives? Show these objectives on Figure 13.2. Compare the figure with the Figure 13.3 provided
by the tutor.
c. Based on Figures 13.3, what are the units of intervention in each phase of the programme?
d. Working in small groups, prepare a flow chart illustrating the classification of malaria cases by
origin of infection and compare it with Figure 13.4 provided by the tutor.
e. Which type of cases are difficult to classify
f. Using a copy of Figure 13.1 (to which ‘milestone for transition to next programme type’ has been
added), state the most important data sources for measuring progress towards reaching a milestone.
Compare the results with the Figure 13.5 provided by the tutor.
LEARNING UNIT 14
Surveillance system
■ ■ Define the term surveillance
■ ■ List the uses of a surveillance system
■ ■ Give examples of a surveillance system’s objectives
■ ■ Explain the connection between the objectives of a
surveillance system and its ability to detect cases
93
14.1 Definition of surveillance

Surveillance in the context of malaria control programme is defined as ongoing, systematic
collection, analysis and interpretation of disease-specific data for use in planning, implementing
and evaluating public health practice. In the context of malaria elimination programmes,
surveillance is that part of the programme designed for the identification, investigation and
elimination of continuing transmission, the prevention and cure of infections and final
substantiation of claimed elimination.1
Data collection, analysis and reporting should guide necessary interventions in the health
system. Interventions may include needs assessment, planning, malaria prevention and control
and/or health programme2 evaluation.
Surveillance is also carried out in the context of other public services, to monitor and improve
the service concerned.
14.2 Uses of surveillance

Examples of surveillance in daily life
The city’s traffic control can be considered a surveillance system. The relevant organizations
continuously and systematically provide information on the traffic status with various tools
(cameras set up in the city and/or traffic unit reports).
The important point is that the necessary interventions in traffic control are done on the basis of
the data gathered in this surveillance system. For example, individuals are guided in choosing
less crowded routes by hearing announcements on the radio about which routes have most
traffic.
In public health, surveillance may be focused on a specific disease such as tuberculosis or
malaria and/or intervention coverage such as the rate of use of bednets, compliance with the
indoor residual spaying of insecticide etc. The distribution and density of vectors (such as
different species of Anopheles mosquito) is part of malaria surveillance.
A public health surveillance system is usually designed to provide data for one or more of the
following purposes:
1. Provide baseline data: Surveillance data can be used to determine baseline information
on malaria cases and deaths, and also on risk factors such as rainfall.
2. Set the programme’s goals: Surveillance can be used to gather information on prevalence
and incidence for the design of health interventions.
3. Identify the groups at risk: Surveillance may be used identify areas or populations at
increased risk of developing a disease.
4. Evaluate the progress made by interventional programmes: Surveillance data can be
used to assess progress in reducing malaria cases and deaths.
1.
WHO 2012. Disease surveillance for malaria elimination: an operational manual. Geneva, World Health organization. http://
www.who.int/malaria/surveillance_monitoring/operationalmanuals/en/index.html
2.
Health programme: a collection of activities carried out for the purpose of solving a health-related problem and/or promotion
of an existing strategy.
SURVEILLANCE SYSTEM 95
5. Identify epidemics: Surveillance will permit the early detection of epidemics to enable
containment efforts.
The different types of surveillance systems

Passive case detection (reporting of cases)
Passive case detection is the regular or periodic collection of data from case reports or registers
from health facilities where patients seek treatment on their own initiative. The health facilities
(public and private) include facilities for outpatient or inpatient services. However, some
countries with inadequate health service coverage, or with mobile populations (e.g. immigrants,
nomadic population), often organize mobile health services (outreach sessions) to reach remote
communities on a regular basis. Malaria cases reported through such arrangements should be U14
considered as passive case detection as the patients report for treatment.
In addition, many national programmes are establishing networks of community agents
to diagnose and treat malaria cases through integrated community case management of
childhood illnesses – iCCM, especially in areas with little access to formal fixed health facilities.
In summary, data coming from public and private fixed health facilities, mobile clinics and
community agents are considered from passive surveillance.
Active case detection (looking for cases)
Active case detection is an active search for malaria cases and diagnostic testing at the com-
munity or at household level, either through regular or random visits by health-care providers.
Testing may be confined to patients with fever, or everyone may be tested (mass screening).
Active case detection can be done to fill gaps in passive case detection systems (e.g. to detect
cases in populations with limited access to services, such as migrant populations). This is some-
times known as ‘proactive’ case detection, in which a population is examined even though
there may be no evidence of confirmed cases. Active case detection may also be undertaken
in response to a confirmed case or cluster of cases, in which a defined population potentially
linked to a confirmed case is identified, and symptomatic cases are tested (possibly with a RDT
then by blood slide for confirmation) as well as asymptomatic cases (by blood slides only). This
is sometimes known as ‘reactive’ case detection.
Sentinel site surveillance

Surveillance is based on the collection of data from a sample (random or non-random) of
collecting sites as indicator data for the rest of the population. The purpose is to identify cases
of a disease early in order to obtain indicative data about trends of the disease or health event.
Standard case definitions and protocols must be used to ensure validity of comparisons across
time and sites despite lack of statistically valid sampling.
The usual sources of data collected in surveillance systems

The sources for data collection are different, but mainly public and private health facilities such
as hospitals, clinics and laboratories are used. It is important to remember that generally data
are produced by the same people who provide service. Therefore, the data collection system
should have the following characteristics:
▶▶ Simple to be carried out by workers who have not received special training
▶▶ Focused on the minimal core data to be collected
▶▶ Feasible and cost-effective
The characteristics of a surveillance system must be based on its objectives. If the objective
is to detect an epidemic, the cases need to be identified rapidly and reported accurately. For
detection of epidemics, the surveillance system should target patients and people at risk. On
the other hand, if the objective is to monitor disease trends, the expected characteristics of the
surveillance system will be different. These characteristics include timeliness of reporting such
as monthly report versus weekly report.
The surveillance system’s capacity to identify cases is assessed by 4 indicators: Sensitivity,
Specificity, Positive Predictive Value and Negative Predictive Value. Sensitivity and Positive Predictive
Value are the indicators of greater significance.
Consider the table below:
Disease (and/or risk factor)
Present Absent
Reported True positive False positive
Surveillance system
Not reported False Negative True Negative
Sensitivity will be:

True positive True cases reported by surveillance system
=
True positive + False negative All cases in the population
The significance of the surveillance system’s sensitivity lies in its ability to identify the cases in
the population. The surveillance system’s Positive Predictive Value is as follows:
True positive True cases reported by surveillance system
=
True positive + False positive All cases reported by the surveillance system
The Positive Predictive Value is important when interventional measures must be taken, e.g.
if extensive case finding is to be undertaken when a single case of measles is reported then the
Positive Predictive Value of the measles surveillance system should be high.
Consider the design of a surveillance system for SARS or plague, and the necessary predictive
value of the surveillance system. The more expensive and complex are the measures following
reporting of the cases, the higher should be the predictive value of the surveillance system.
Definition of the terms

An essential initial step in a disease surveillance system is the establishment of the case
definition. The definition must be practical and clear. Figure 14.1 shows classifications used for
definition of the cases in malaria control programmes.
SUSPECTED MALARIA CASES
TESTED NOT TESTED (presumed)
POSITIVE NEGATIVE
(confirmed malaria) (not malaria)
Figure 14.1 Flow diagram of malaria case definitions
U14
Suspected malaria cases: Patient illness is suspected by a health worker to be due to malaria.
The criteria for suspected malaria usually include fever or a history of fever, but the precise
criteria vary according to local circumstances and are established by the national malaria
control programme. All suspected cases of malaria are tested by either microscopy or an
RDT.
Tested malaria case: Suspected malaria case that received a laboratory examination for
malaria parasites by microscopy or RDT.
Not tested malaria case (presumed malaria): Suspected case that did not receive a laboratory
confirmation of malaria diagnosis but was neverthless treated for malaria.
Positive case (confirmed malaria): Suspected case that was positive after a laboratory
examination for malaria.
Negative case (not malaria): Suspected case that was negative after a laboratory examination
for malaria.
Among all suspected malaria cases, the most specific for the diagnosis of malaria is confirmed
malaria. Presumed cases include many febrile cases which do not have malaria infection.
Case definitions and the health system: Health care is usually provided by different health
workers at different levels of health care system: at community level by community health
workers and/or volunteers, at primary health care levels (clinics/health centres at periphery) by
clinicians/health workers and at referral facilities (district or provincial hospitals) by physicians.
If deemed necessary, the patient will be referred to hospital level with more specialized health
care. The definition of cases is adapted to the skills and facilities of the health care system. Case
definitions are of greater sensitivity in peripheral levels: all possibly affected cases are registered
and reported to the surveillance system and the specificity is low. Conversely, at higher levels of
service delivery such as hospitals, both sensitivity and specificity are high and the possibility of
a false positive is reduced to a minimum.1
1.
Note that a similar method is used in the diagnosis of a disease. Questions with high sensitivity are used first. In these cases if
the questions are negative then it can be known with certainty that the disease does not exist. Eventually high quality tests are
used in the differential diagnosis for approving the diagnosis which has high positive predictive value.
Performance of the health-care providers in detection of cases can affect the sensitivity of the
surveillance system at each level.
In addition, the important factors affecting the surveillance system's performance are as follows:
a. People’s seeking of medical care
b. Performance of diagnostic tests
c. Registration and reporting of cases
Two important points should be noted regarding case definitions for surveillance systems:
a. The case definition may differ at different levels of the health care systems.
b. Different countries may not use the same case definitions in their surveillance systems.
Therefore disease definitions and/or special conditions in the country’s service delivery may
not be exactly identical, as the following example illustrates.
The presence of parasitaemia has different significance for the malaria case definition in
different settings. In areas of high malaria transmission, detection of parasitaemia is not always
predictive of malaria illness. Because of acquisition of immunity, a significant proportion of the
population may have parasitaemia regardless of the main cause of the illness. On the contrary,
in many low transmission areas, parasitaemia is often associated with clinical symptoms.
Some important concepts in surveillance

The surveillance system
The purpose of surveillance is to guide public health action.
Figure 14.2 shows the formation of the surveillance system. The first step is to identify the
surveillance system’s objectives, after which the case definitions and necessary staff can be
determined.
In general, the objectives of surveillance systems for communicable diseases focus on
identification and reporting of cases, but for non-communicable diseases, notification and
distribution of risk factors are also of great importance. Surveillance programmes for non-
communicable disease risk factors such as smoking, alcohol abuse, nutrition, physical activity,
Evaluation of the
surveillance system Setting objectives
Implementation of the defined Case definitions, necessary

surveillance system staff & equipment
Figure 14.2 The surveillance system cycle

obesity, blood pressure, blood sugar and cholesterol are carried out in many countries. Risk
factors are also important for a number of communicable diseases, e.g. behavioural surveillance
in HIV/AIDS programmes.
Timeliness
For data to be useful, they should be timely. Every effort should be made to report the data
promptly on a regular basis. Data should be
Na3onal level authori3es
examined and processed without delay at each
level.
Regional level authori3es
Feedback
Surveillance feedback
For a system to operate well, those collecting the U14
data need to receive feedback. Many countries District level authori3es
have surveillance systems in place for a variety

of diseases. The following (Figure 14.3) is the Health facili3es public/private)
typical flow of information which is present in community
most surveillance systems.
Feedback is essential to motivate those actually Diagnosis
collecting the data and to provide health care
providers with data from other geographic areas.
Feedback is also useful for planning purposes, Occurrence of health event
especially in decentralized systems. The report- Figure 14.3 Flow of information in the
ing system is strengthened if suitable feedback is Figure 14.3 Flow of surveillance system
informa3on in the surveillance system
given to health workers involved in reporting on
diagnosis and treatment (private and public sector), health managers, and to the public.
14.3 Evaluation of the

surveillance system
Evaluation of the surveillance system addresses the following questions:
a. Is the surveillance system necessary?
b. If necessary, is its performance satisfactory?
The function of surveillance system includes the following:
▶▶ identify the areas or population groups most affected by malaria;
▶▶ identify trends in cases and deaths that require additional intervention, e.g. epidemics; and
▶▶ assess the impact of control measures.
The performance of the surveillance system is evaluated in 4 stages:
1. Description of the surveillance system
2. Determining the benefit (usefulness)
3. Examining its characteristics
4. Evaluation of resources
14.3.1 Description of the surveillance system

This stage requires a description of the following:
▶▶ The events under surveillance (disease, risk factor …)
▶▶ Practical case definitions
▶▶ The population covered
▶▶ The surveillance system’s goals (diagnosing epidemics, monitoring disease trends …)
▶▶ Data flow, including:
▶▶ Who completes the form?
▶▶ How is the data reported and collated?
▶▶ Where does data analysis take place?
▶▶ Is feedback provided?
▶▶ Are the surveillance reports distributed?
▶▶ At what intervals are data gathered?
14.3.2 Determining its benefit (usefulness)

The significance of the surveillance system lies in the utilization of its results. It is important to
assess:
▶▶ Who are the users of the information?
▶▶ What measures are taken on the basis of the surveillance reports?
▶▶ Are the data submitted appropriately and adequately to the decision-makers?
14.3.3 Examining its characteristics

It is important to evaluate the surveillance system in relation to the following attributes:
• Simplicity: the data entry and reporting should be easy to carry on
• Flexibility: the system should have flexibility to collect new or different information as
required
• Acceptability: the system should be well accepted by the people who collect and use the
data. People’s willingness to report and participate in the surveillance will ensure its success.
• Sensitivity and positive predictive value: the system should detect a high percentage of
cases. In addition, of the persons identified by the surveillance system as having the disease, a
high proportion actually have the disease. These are performance indicators of a surveillance
system for identification of cases.
• Timeliness: It is important to have access to data on time to contribute to decision-making.
Therefore the time spent on each stage of surveillance is important.
• Representativeness: the findings from the surveillance system should be representative
of the population covered by the system, in relation to time and geographical distribution
taking in to the source of data referring to the public sector hospital wards, community
services etc.
• Completeness: The surveillance data should be complete, in terms of data entries and
number of facility reports per expected reporting period.
• Ethical considerations: It is important to ensure that information on individuals remains
confidential.
14.3.4 Evaluation of resources

The costs incurred in carrying out the surveillance activities must be specified. These expenses
relate to:
▶▶ Human resources
▶▶ Equipment and supplies U14
Exercises
Evaluation of a surveillance system
You have been assigned to the Office of Statistics within the Ministry of Health for a 3-month
period to make an evaluation of the existing surveillance system – the government is considering
making the system more useful for decision-makers.
You decide to follow the data along the chain of reporting, beginning at clinic level, through the
District Medical Officer (DMO) and Regional Medical Officer to the Office of Statistics and
the Minister. You decide to examine the chain of reporting in a particular district where you
used to work and still know many of the health officers there.
You meet with the Chief of the Statistics Office in the Ministry of Health, who says: “The
system was established by the British, and we still use the same form we were using when I
started working here 20 years ago. In the late 1990s, the Minister decided to add 6 diseases to
the list. We now collect data on 43 different diseases by sex for 3 different age groups”.
Asked what problems he sees in the system, he replies, “Nobody seems to care much. It takes
at least 9 months to get the reports from the regions. By the time we get all the data and put our
reports together, they are out of date. I sometimes wonder if anyone ever makes use of them”.
He says he will be happy to help and gives you a copy of the most recent annual report, which
was sent out last month to all district and regional medical officers. This contains data for 2001,
although we are now in 2006. It has over 200 pages, with tables of each disease by reporting site,
sex, age, region and district; without graphs or text.
In the district you visit a hospital, a clinic and two smaller rural dispensaries. All the health
personnel interviewed tell you that they learned how to fill out the forms as part of their initial
training, but have no official guidelines on how to fill them out, and their supervisor rarely
checks their work. You do an audit of the past month and find that for one of the dispensaries
the figures in the report differ markedly from those in the register.
You ask the health workers what happens to the data after they are sent to the district level. All
say they don’t know; that is the last time they ever see the data. One of them says, “I think that
the forms just sit on the DMO’s desk and after a while he throws them away. This surveillance
system is just useless paperwork that keeps me from more important work like seeing my
patients”.
You then visit the DMO. He explains that filling out forms is a big part of his job. He fills out
8 forms a month (surveillance report, hospital reporting forms, essential drugs list, number
of vaccines administered, sanitation worker report, supplementary food distribution form,
hospital budget form, clinic budget form). Every month he plans to spend a little more time
on surveillance but never gets around to it. He is aware that some of the clinics do not report
routinely and that the numbers sometimes look suspicious, but he does not have time at the
moment to fix the problem.
When asked how he uses the surveillance data, he admits he does not use them for anything.
He shows you the 2001 report he just received from the Ministry and says, “How can you make
any sense out of this jumble of numbers?” Asked if he has ever used surveillance to detect an
epidemic in the district, he says if there is a big outbreak he usually hears about it, although
sometimes it is too late to do anything. He discovered a meningococcal meningitis epidemic
last March when 11 cases showed up in the district hospital on the same day. When he later
reviewed the surveillance records, he found that in January and February the number of cases
had been 3 times higher than they had been in January and February of the previous year.
The Regional Medical Officer repeats many of the concerns expressed by the DMO. She states
that the data might be useful for monitoring trends in priority diseases. Every time she looks
at the annual report, however, there seem to be more and more cases of everything, even those
diseases for which she knows there are good control programmes. Even if she could work out
what was going on, it is difficult to make programme decisions using information that is 4 years
old or more.
Finally, the assistant to the Minister of Health tells you that he remembers from his primary
health care courses that surveillance is a good thing; however, he and the Minister do not really
make use of surveillance data because they think that these data are old and rather unreliable.
The information is rarely used in health planning for the country; they tend to rely more on
disease prevalence data collected as part of a national health survey in 1995. He hopes that
you will work out a way to get more timely information, especially since this could be useful
in following the progress of the new primary health care programme. As part of the bilateral
funding agreement for this project, a modest amount of money will be set aside for improving
disease surveillance.
Exercise 14.1
Identify the problems you consider to be present within the system in terms of :
▶▶ Data collection
▶▶ Data analysis
▶▶ Use of information
▶▶ Relevance
▶▶ Feedback
Exercise 14.2
For three of the problems that you have identified, suggest a solution.
LEARNING UNIT 15
Indicators for monitoring and evaluation
of malaria control programmes
■ ■ Describe the importance of monitoring and evaluation
■ ■ Explain the difference between monitoring and evaluation
indicators
■ ■ Give suitable examples of input, process, output, coverage
and impact indicators
■ ■ Explain the relation of indicators with global objectives
of the malaria control programme, and also their internal
relationships
103
15.1 Monitoring and evaluation

An important managerial task is to assess whether the programme is proceeding as planned
and is reaching its goals. Hence monitoring and evaluation are key components of programme
management.
The goal of a national monitoring and evaluation system is to provide reliable information on
progress in controlling malaria in the country.
The specific objectives of a national monitoring and evaluation system can be summarized as
follows:
▶▶ To collect, process, analyse, and report on malaria-relevant information
▶▶ To verify whether activities have been implemented as planned, to identify problems and
address them in a timely manner
▶▶ To provide feedback to relevant authorities to ensure accountability and to improve future
planning
▶▶ To document periodically whether planned strategies have achieved expected outcomes
and impact
Usually, the different stages of implementation of a programme – from provision of resources
to reduction of burden of disease – are indicated by the terms input, process, output and impact
(Fig. 15.1).
Figure 15.1 Monitoring and evaluation framework

INDICATORS FOR MONITORING AND EVALUATION OF MALARIA CONTROL PROGRAMMES 105
15.2 Indicators
Exercise 15.1
a. Give an example of indicators for each stage (input, process, output and impact) of the malaria
control programme.
b. List five measurements of intensity of malaria transmission (e.g. rates, etc.) as discussed in Learning
Unit 13.
c. Are these measurements enough for assessment of the malaria programme?
Monitoring and evaluation should cover all aspects of malaria implementation to be
informative for decision-making, e.g. showing burden of malaria as well as logistics of the
programme.
A good indicator has the following characteristics:
▶▶ Validity: The indicator should be able to assess its target with precision, i.e. with high
Sensitivity and Specificity. U15
▶▶ Reliability: Upon repeated measurements of a single state the value should not vary
greatly.
▶▶ Simplicity: The data required for indicators should be easy to gather.
▶▶ Cost: The cost for gathering the information for the indicators should be locally reasonable.
▶▶ Representativeness: The indicator must represent the target population, and should not be
affected by selection bias.
▶▶ Consistency: Stability of the indicator over time allowing to study changes over time.
Finally, the set of indicators chosen should be measurable and feasible.
d. Did the measurements introduced in the Learning Unit 13 on ‘Intensity of Malaria’ meet these
criteria?
15.3 Difference between monitoring and evaluation

Indicators are used for monitoring and evaluation of a programme. It is important to know the
difference between monitoring and evaluation.
Monitoring is the routine tracking of the key elements of programme/project perfomance over
time (usually inputs and outputs). Monitoring reports provide information on progress made in
the implementation of planned activities and the constraints and/or bottlenecks that have been
faced in implementation of those activities. Monitoring helps programme or project managers
determine which areas require greater effort and identify areas which might contribute to an
improved response.
Evaluation is the periodic assessment of the change in targeted results related to the programme
or project intervention. In other words, evaluation attempts to link a particular output or
outcome directly to an intervention after a period of time has passed. Evaluations can be used
to link any two parts of the monitoring and evaluation framework (inputs, processes, outputs,
outcomes, or impact), e.g. evaluation of whether financial inputs are effectively generating the
desired training or service deliveries. Evaluation thus helps managers to determine the value of
a specific programme or project.
Usually, monitoring indicators measure inputs, process and outputs, while evaluation indicators
assess outcome and impact of interventions. The output is an immediate result of the interventions,
while outcome is a more distal result of the interventions and it can be determined at whole
target population level. This is the reason that the coverage is used interchangeably or instead
of outcome in this categorization.
Exercise 15.2
a. List some of the indicators for the stages from Input to Impact for ITN intervention.
b. Is there any difference in requirement for indicators at different levels (regional, national and
international)?
15.4 Monitoring and evaluation objectives and methods

The objectives and the methodology used in monitoring and evaluation are different. In
general, evaluations are more difficult in view of the methodological precision needed: without
this, wrong conclusions on the value of a programme or project can be drawn. Evaluations are
also more costly, especially outcome and impact evaluations which often require population-
based surveys or rigorous designs.
Evaluation must assess the programme goals. Below are the goals and targets set in the Global
Malaria Action Plan and Millennium Development Goals for malaria control at the global level
and type of indicators (Box 15.1).
Box 15.1 The goals, targets and indicators for malaria control at global level
Initiative Target Type of indicator
GMAP a
• Reduce global malaria deaths to near zero Impact
• Reduce global malaria deaths from 2000 levels by 75% in 2015 Impact
• Achieve universal access to and utilization of preventive measures
Outcome
(100% of population at risk)
• Achieve universal access to case management (100% of suspected
malaria cases are tested and those confirmed are treated with Outcome
appropriate and effective antimalarial medicines
WHAb • Reduce global malaria deaths from 2000 levels by 75% in 2015 Impact
• Reduce global malaria cases from 2000 levels by 75% in 2015 Impact
MDGs c
• Target 6: Have halted and begun to reverse the incidence of malaria
Impact
and other major diseases
a
Global Malaria Action Plan; b World Health Assembly 2005; c Millennium Development Goals
15.5 The purpose of different types of indicator

Impact (or epidemiological) indicators describe the size of the malaria burden in a country.
Coverage indicators measure the extent to which implementation of effective interventions
meets the need for programme interventions at population level.
Performance indicators measure the extent to which supply meets the need of programme
interventions, usually at the service delivery point.
Exercise 15.3
a. Table 15.1 shows some of the malaria indicators. State the purpose of each indicator as either
monitoring performance, coverage or impact evaluation.
Table 15.1 Some of the malaria indicators and their targets1

Indicator Target Purpose
Reduction in cases by
Confirmed malaria cases (microscopy or RDT), per
1 75% by 2015 compared
1000 persons per year
to 2000
Reduction in deaths by 75% by
2 Inpatient malaria death per 1000 persons per year 2015 compared to 2000
All-cause deaths in < 5 children per 1000 <5 Reduction of global malaria
3
children per year (for high-transmission areas) deaths to near zero by end 2015 U15
Proportion of individuals who slept under an ITN
4 100%
the previous night
Proportion of population at risk protected by IRS in
5 No specific target set
the last 12 months
In moderate to high transmission areas: Proportion
of women who received three doses or more of
6 100%
intermittent preventive treatment (IPT) during their
last pregnancy
Proportion of suspected malaria cases that receive
7 100%
parasitological test
Proportion of confirmed malaria cases that
8 received first-line antimalarial treatment according 100%
to national policy
Proportion of health facilities without stock-outs
9 of first-line antimalarial medicines, ITNs and 100%
diagnostics, by month
Percent of districts reporting monthly numbers
of suspected malaria cases, number of cases
10 No specific target set
receiving a diagnostic test and number of
confirmed malaria cases
b. Considering these indicators, what are the main interventions/strategies of malaria control?
c. In Table 15.2, suggest indicators for each category and corresponding intervention.
WHO (2012). World Malaria Report. Geneva, World Health Organization. http://www.who.int/malaria/publications/world_
1.
Table 15.2 WHO indicators according to the steps (Input –> Impact) of Monitoring and
Evaluation
Intervention Input/Process Output Outcome(coverage) Impact
Insecticide Residual
Spraying (IRS)
Insecticide-treated
mosquito nets
(ITNs)
Case management
Intermittent
preventive
treatment (IPT)
(only in moderate to
high-transmission
areas of sub-
Saharan Africa)
d. Why is there only a single box for impact indicators as opposed to the box per row?
15.6 Interventions and indicators by programme phase

In practice, different countries implement different strategies for malaria control based on
programme phases. That is why their indicators are also different. Table 15.3 summarizes
interventions/strategies applicable to the different programme phases.
Not all of the indicators listed in Table 15.1 should be measured in all countries. Indicators 3
and 6 are specifically used in regions with moderate to high transmission and those countries
in control programme phase. However, several of these indicators provided are applicable at a
national level and beneficial for decision-making. And, based on country-specific programme
activities, national programmes need to develop appropriate monitoring indicators for malaria
control.
Another well-known indicator is the Annual Parasite Index (API), with which is calculated
as: confirmed cases during 1 year/population under surveillance × 1000. API is the most
commonly used indicator for estimating the actual intensity of malaria transmission. Various
countries use API for comparing the risk between cities, provinces and/or countries if the
diagnostic coverage is high and kept fairly constant. When percentage of suspected malaria
cases that are tested is unknown or inconsistent (<80–90% over time) then it is difficult to
interpret trends using this indicator.
Table 15.3 Interventions and strategies by programme phase
Programme phase Intervention/strategy specific to different phases

Control Moderate Surveillance
to high Prevention
transmission ITNs/LLINs, especially for all of the at risk population
IRS where appropriate
Intermittent preventive treatment in pregnancy (IPTp) in moderate-high
transmission areas in sub-Saharan Africa
Intermittent preventive treatment in infancy (IPTi) in moderate-high
transmission areas in sub-Saharan Africa
Seasonal Malaria Chemoprevention in areas with highly seasonal
transmission in the Sahel subregion of Africa
Treatment
Prompt diagnosis and effective treatment, including at the community
level where appropriate
Low Surveillance and epidemic response
transmission Surveillance; epidemic response capacity
Prevention
IRS
ITNs/LLINs
Treatment U15
Prompt diagnosis and effective treatment of suspected cases
Pre-elimination Surveillance and epidemic response
(transition phase) Surveillance; epidemic response capacity
Focalized prevention
IRS
ITNs/LLINs
Treatment
Prompt diagnosis and effective treatment of suspected cases
Elimination and Surveillance and epidemic response
prevention of Active surveillance to inform and guide active case detection and
reintroduction epidemic response (if area is receptive to malaria transmission)
Prevention
Chemoprophylaxis for travelers going to malarious areas
Treatment
Prompt diagnosis and effective treatment in confirmed cases
Points to note in comparing the API of different regions:

▶▶ Malaria definitions should be the same, e.g. reported malaria should be defined as confirmed
through laboratory testing
▶▶ Reports are usually the result of active and passive case-finding but these should be reported
separately as dilution by either of them may affect the trend of the rates.
▶▶ If cases are compared in two regions, and the active case-finding differs, then the API of these
two regions can no longer be compared. In other words, if the risk of malaria is the same in
two regions, but case-finding is more active in one, then its API will be estimated lower.
▶▶ In order to overcome this shortcoming, another indicator is used: the Annual Blood
Examination Rate (ABER). API is useful for monitoring trends when the annual blood
examination rate (ABER) is constant and higher than 10%. ABER is the number of slides
examined/population at risk. When Region A and Region B have the same API, and the
ABER is lower in Region B, then Region B has higher malaria transmission. This is because
the same numbers of infected cases per 1000 population are found with less case detection
efforts in Region B.
Exercise 15.4
a. Do interventions have the same applicability and impact in different programme phases?
b. Considering the current programme phase and epidemiological situations in your country, identify
the key indicators required. Prepare a table of indicators similar to Table 15.1.
Annex 1 provides details about core indicators for malaria control.
LEARNING UNIT 16
Data collection methods
■ ■ Name different sources of data collection related to the main
indicators of malaria
■ ■ Specify which source is important for each group of indicators
(performance, coverage and impact)
■ ■ From a numerator and denominator of an indicator, explain
from which source it should be calculated and why that
source is chosen.
■ ■ Name the main types of surveys that can estimate coverage
indicators at a national level
111
16.1 Sources of information

Three categories of indicators were studied in Unit 15 to monitor:
1. Impact
2. Coverage
3. Performance
The sources of data collection for calculation of epidemiologic indicators can be classified in
the following groups:
▶▶ Disease surveillance
▶▶ Health Information System (routine and logistic data)
▶▶ Surveys (household, health facility)
▶▶ Vital registration
▶▶ Meteorological monitoring (see Learning Unit 19 for details on this )
Exercise 16.1
The following Table 16.1 is a modified form of Table 15.3 in Unit 15, with one additional
column.
a. Identify the measurements for the impact indicators of the corresponding interventions.
b. Define the numerator and denominator for impact indicators (morbidity and mortality).
c. What is the source of the data elements for the numerator of impact indicators?
Table 16.1 Indicators for each Monitoring and Evaluation category and corresponding
interventions
Monitoring Performance Evaluation
Input/ Outcome Impact
Output
Intervention Process (coverage) Morbidity Mortality
Insecticide Residual Spraying (IRS)
Insecticide-treated nets (ITNs)
Case management
Intermittent preventive treatment (IPT)
in moderate-high transmission areas
in sub-Saharan Africa)
Covered population
Targeted population
Population at risk
Figure 16.1 Representation of populations related to malaria control

DATA COLLECTION METHODS 113
Coverage and performance indicators have been selected with the purpose of assessing
activities and that is why they target specified populations. The targets differ according to
whether the interventions are preventive or curative. Where primary prevention is concerned,
the target is the population at risk, whereas for treatment the target is those with malaria.
There are several different concepts of “population” related to malaria control, as represented
in Figure 16.1:
1. Population in areas where both vectors and malaria parasites are currently present
2. Target populations for interventions:
▶▶ Universal: covering total population at risk living in areas where malaria is endemic
(assuming unlimited resources)
▶▶ Targeted: limited to the more vulnerable groups in the population at risk (planned based
on available resources)
▶▶ Covered: population covered by programme implementation (e.g. population actually
covered by IRS this year)
▶▶ Access: population which has access to the programme interventions
This ‘population at risk’ will be used as the denominator to calculate national malaria U16
incidence rates. In high/moderate transmission countries, often the population at risk and
the target populations are similar for calculating rates and estimating the population targeted
for treatment, ITN, and IRS. In low-transmission countries, target populations for each
intervention (ITN, treatment, IRS) may be different. The number of persons targeted for ITN
may be different from those targeted for IRS. This is because areas targeted for ITN and for
IRS may vary depending on suitability or other operational factors. A national programme
may decide to make rural populations eligible for IRS but not urban populations (because of
acceptability). In another country, the contrary may be true if the rural population is highly
dispersed or is too mobile to conduct IRS.
Exercise 16.2
a. In evaluating coverage and performance indicators, explain who is responsible for delivering
malaria interventions.
b. Explain how data related to coverage of ITN, IRS, diagnosis, treatment and IPTp (in Table 16.1)
are collected.
c. Using copy of the Table 16.1, provide the data collection sources for the indicators of the corresponding
interventions. Compare your table with Table 16.2 in the Guide for Tutors.
16.2 Surveillance and survey

Exercise 16.3
a. Describe the applications of survey and surveillance in malaria control.
b. Compare the applications of survey and surveillance in relation to the different programme phases
using Table 16.3. (Each participant should work separately with this table.)
Table 16.3 Applications of surveys and surveillance in malaria programmes
Surveillance Survey
Indicators/data elements
Question to Indicators/ Question to (e.g. use LLIN as
Programme phase Objective be answered data elements be answered intervention)
Control
Pre-elimination
Elimination
Prevention of
reintroduction
c. Which data are required in a surveillance system for patients?

Samples of surveillance forms and registers are provided in Annexes 2 to 9. Note that the data
collected in surveillance systems are in two forms: individual records and aggregated data.
d. Should data for surveillance be collected in aggregated form and or as data for each individual
(case-based)?
The rationale, objectives and applications of surveillance and survey therefore differ by
programme phase (epidemiologic conditions) as well as by transmission levels.
Exercise 16.4
Compare the characteristics of malaria surveillance in the different transmission settings and
programme phases using Table 16.4.1
1.
WHO 2012. Disease surveillance for malaria control: an operational manual. Geneva, World health Organization. http://www.
who.int/malaria/surveillance_monitoring/operationalmanuals/en/index.html
Table 16.4 Characteristics of malaria surveillance in different transmission levels and programme
phases (control and elimination)
Control phase Elimination phase

Characteristics Guide for responsesa High & moderate Low Very low
transmission transmission transmission
Incidence
Common, less common,
Cases and deaths
sporadic
Temporal variation Limited, variable
Limited, heterogeneous,
Geographical variation
focal
Fevers (proportion due to
Large, small, very small
malaria)
Health facility attendance
High, low
(proportion due to malaria)
Vectors Efficient, inefficient
Mortality reduction, case
Programme objectives reduction, eliminate
transmission
Surveillance system
U16
Low expenditure per head,
Resources resources to investigate
cases
Aggregate numbers, lists of
Data recording inpatient and deaths, lists of
all cases, case details
Inpatient cases, all cases,
Investigation
individual cases

a
second column of Table 16.4 contains possible responses which the participants should use to fill columns 3, 4 and 5.
16.3 Core analyses

There are five (or six if more than one species of malaria is present) core analyses recommended
by WHO for use by programme managers in moderate to high malaria transmission areas.1
These core control graphs are on (i) malaria incidence rates, (ii) proportional malaria incidence,
(iii) general patient attendance, (iv) diagnostic activity, (v) quality of diagnosis and reporting,
and (vi) percentage of cases due to P. falciparum. Each month, districts should update these
graphs showing trends in malaria incidence and other indicators. The core graphs can be
discussed at monthly and quarterly performance assessments, and during supervisory visits.
Health services without inpatient facilities can use graphs pertaining only to outpatient cases.
Depending on the epidemiological setting, the graphs may be disaggregated by age.
1.
WHO 2012. Disease surveillance for malaria control: an operational manual. Geneva, World health Organization.
http://www.who.int/malaria/surveillance_monitoring/operationalmanuals/en/index.html
Exercise 16.5
Working in small groups, summarize the application and interpretation of each of the following
6 core surveillance graphs over time.
1. Malaria incidence rates 4. Diagnostic effort

20 4%
Confrimed cases per 1000
Annual blood examination rate
Inpatients per 10 000
Deaths per 100 000
15 3%
10 2%
5 1%
- 0%
2008 2009 2010 2011 2008 2009 2010 2011
2. Proportional malaria incidence 5. Quality of diagnosis and reporting

80% 100%
Slide positivitiy rate
% Inpatients due to malaria
% Deaths due to malaria 80%
60%
60%
40%
40%
20%
20%
% Health facilities reporting
% Suspected cases tested
0% 0%
2008 2009 2010 2011 2008 2009 2010 2011
3. General patient attendance 6. % Cases due to P. falciparum

250 50%
Outpatients per 1000
Inpatients per 10 000 % Cases due to P. falciparum
200 Deaths per 100 000 40%
150 30%
100 20%
50 10%
- 0%
2008 2009 2010 2011 2008 2009 2010 2011
Exercise 16.6
a. What differences should be considered in designing surveillance systems when the rate of transmission
is decreasing from high to low level?
Note: consider the following issues for the comparison between high and low transmission areas regarding
surveillance systems:
▶▶ The phase of malaria programme, target for the programme, distribution of malaria foci
▶▶ Proportion of malaria diagnostic testing and treatment delivered:
a. at community level
b. in the private sector
▶▶ Number of cases
▶▶ Type of patients (inpatients and or outpatients) which have malaria
▶▶ Age pattern in cases
▶▶ Plasmodium species, proportional identification in malaria cases

▶▶ Information technology platform(s) available
b. Based on the differences identified above, what are the parameters of surveillance in low-transmission
settings?
Note: consider the following items to describe the characteristics of the surveillance system.
▶▶ Level of analysis (district/health facility catchment area)
▶▶ Level of details on individual cases (inpatients versus all cases)
▶▶ Aggregate data gathering versus individual case reporting
▶▶ Any other features
16.4 Expected impact of improved diagnosis and treatment

of malarial illness
Exercise 16.7
a. Suppose that in an area of high transmission of P. falciparum, a programme of improved diagnostic
testing, treatment and reporting of malarial illness, and of severe malaria is introduced. Indicate
in Table 16.5 the changes to be expected in the various variables listed. Present separately (i) what U16
you expect in reality and (ii) what you expect to receive in reported information. Indicate expected
changes by arrows. If you wish to indicate the change expected in other variables, add them to the list.
Once Table 16.5 is completed, compare it with the one provided by the tutor. Use the
following symbols: (no change →
); (increase →
); (decrease ). →
Table 16.5 Expected impact of improved diagnosis and treatment of malarial illness
Expected change
Variables
In reality Reported information
Incidence of infection
Prevalence of infection
Incidence of disease
Prevalence of disease
Incidence of severe malaria
Death rate (all causes)
Malaria mortality rate
Case fatality rate of malaria
Case fatality rate of severe malaria
b. What are the epidemiological and operational objectives of a programme of diagnostic testing and
treatment of malaria?
Table 16.6(a) lists numbers that could be used for the calculation of indicators. Note the
distinction between real – but unknown – numbers, and the accessible numbers listed in Table
16.6(b). The purpose of the table is not to recommend the measurement/calculation of all
numbers/indicators listed but to help the participants to discuss the relationship of indicators
to reality.
Table 16.6(a) Potentially relevant numbers

Numbers that are (or can) be known
Real numbers (unknown)
(and can be used instead of the real numbers)
a. Number of cases of malarial illness g. Number of cases treated as malariab

b. Number of cases of severe malaria h. Number of cases treated as severe malariab
c. Number of deaths due to malaria i. Number of deaths attributed to malaria
j. Estimated population at risk
d. Population at riska k. Total number of attending patients
l. Total number of hospitalized patients
m. Total number of registered deaths
e. Total number of deaths
n. Total number of deaths attributed to a cause
a
Depending on the context, either the total population or only a part of it
b
Currently WHO recommends that all suspected malaria cases are testing before treatment. For a malaria programme to be
successful, it must adopt universal diagnostic testing
Table 16.6(b) Potentially relevant indices

Real rates (unknown) Alternative (measurable) indices
Incidence of malaria illness (a)/(d) (g)/(j); (g)/(k)
Incidence of severe malaria (b*)/(d) (h)/(j); (h)/(k); (h)/(l)
Proportion of severe malaria (b)/(a) (h)/(g)
Mortality rate (all causes) (e)/(d) (m)/(j)
Malaria mortality rate (c)/(d) (i)/(j)
Proportion of deaths due to malaria (c)/(e) (i)/(n)
Malaria case fatality rate (c)/(a) (i)/(g)
Case fatality rate of severe malaria (c)/(b) (i)/(h)
Assuming “a” is inclusive of “b”
Exercise 16.8
After studying Table 16.6 (a, b), consider the following questions:
a. Is the reported incidence of malarial illness equivalent to the true incidence of malarial illness?
b. If there is a difference, can it be either greater or lesser? What are the principal explanations for
errors in reporting by excess and omission?
c. Could the difference between real incidence and reported incidence vary in:
Different places?
Different times in the same place?
Different indices, e.g. incidence of malarial illness, incidence of severe malaria and malaria mortality?
d. Is it necessary to improve the collection and reporting of information until it becomes equivalent to
the reality represented? Would this be possible?
e. What would be required to make reported information equivalent to reality? Is this necessary? Is it
possible? How could the imperfections of reported information be taken into account?
f. Are all reported cases of malarial illness sufficiently similar to each other to be counted together in
reports? If not, how do they differ from each other (e.g. distinguish true differences from differences
of assessment)? Can the main differences be taken into account in reporting? If so, how can this be
done?
g. How authentic are the malaria-associated mortality measurements?

h. If all the data are not received from the health delivery units, does this affect the results obtained?
i. What is the role of sentinel sites in gathering data from selected sources?
As discussed in Learning Unit 13, supervision of the surveillance system is one of the main
activities which can contribute to quality control of the malaria control programme.
Annex 10 provides a sample of supervisory checklist for use during visits to health facilities.
Note that the data collected in surveillance systems are in two forms: individual and aggregated
data.
16.5 Surveys
Three types of population survey specifically gather information related to malaria. These are:
▶▶ Demographic and Health Survey (DHS)
▶▶ Multiple Indicator Cluster Survey (MICS)
▶▶ Malaria Indicator Survey (MIS)
These three surveys are explained below. U16
16.5.1 Demographic and health survey (DHS)

The DHS are representative national household surveys for calculation of indicators used
in monitoring and assessment of impacts of health interventions on population, health
and nutrition. The typical DHS is undertaken once every 5 years on a large sample size of
representative population (5000–30 000 households), with the purpose of comparing changes
over time. The Interim type of DHS is done between standard studies and on a smaller
representative population (2000–3000 households), and has a shorter questionnaire. These
studies give some information on the key indicators used in monitoring performance, but they
might not have the necessary data for assessment of impact.
The domains under study in DHS include: anaemia, child health, education, family planning,
female circumcision, fertility, domestic abuse, knowledge, perspective and behaviour regarding
HIV and AIDS, prevalence of HIV, characteristics of household and person responding, infant
and child morbidity and mortality, malaria (awareness of malaria transmission, use of mosquito
nets among women and children, prevalence and treatment of fever), maternal health, maternal
mortality, nutrition, wealth and socioeconomic status and empowerment of women. (http://
www.measuredhs.com)
The advantage of measuring malaria indicators for population coverage using either the DHS
or multiple indicator cluster survey (MICS) surveys (see below) is due to the rigour of the
sampling design, and reliability over time and across countries. However, these surveys are
only carried out every 5 years. If immediate data collection is required that does not fit within
the implementation schedule of either the DHS or MICS surveys within a particular country,
it is recommended that the MIS survey be used to obtain the necessary data for measuring
malaria coverage indicators.
16.5.2 Multiple indicator cluster survey (MICS)

The MICS is a survey designed by UNICEF to help countries monitoring the status of children
and women. By using a statistical approach, the MICS offers comparable indicators of nations’
socio-economic status. This survey has been carried out since 1995 at 5-year intervals. With
each survey new indicators are added so that in addition to monitoring change, the current
priorities are also monitored. By 2012 more than 200 MICSs had been undertaken in 100
countries worldwide. Half of the Millennium Development Goal indicators and many
international commitment assessment indicators are collected in the form of MICS.
The domains under study in MICS include the following indicators: nutrition, child health
(including malaria), environmental health, fertility, education, AIDS and vulnerable children.
These programmes are mainly implemented by governments and with technical support from
UNICEF and other related organizations. (www.childinfo.org).
These surveys are only conducted periodically (every 4 to 5 years), so if a country needs to
obtain data on the core malaria indicators in the interval between large national surveys, the
Malaria Indicator Survey (MIS) is the recommended survey methodology.
16.5.3 Malaria indicator survey (MIS)

The MIS provides much-needed information about all of the internationally recognized
malaria indicators including household ownership of ITN and their use, treatment, IRS and
IPT coverage. The survey also gathers background data on the characteristics of household
members and ownership of household assets such as electricity, bicycles, radios, and indoor
plumbing. The MIS can also include measurement of parasitaemia and anaemia and for this
the MIS should be timed to coincide with the malaria transmission season. The MIS package
(http://rbm.who.int/merg.html) contains guidelines, questionnaires, and manuals to support
the conduct of the survey as well as recommended tabulations for analysing the data.
There are several basic requirements for identifying a survey to which malaria indicator
questions could be added. The survey must be a household-level survey that is designed to
collect nationally representative data (similar to the DHS, MICS, or MIS surveys). This means
that the sampling frame for the survey is based on a complete enumeration of primary sampling
units across the entire country.
The survey should include a household schedule that collects basic demographic information
on all individuals who usually reside in the household or who slept in the household the
previous night. This listing provides the necessary information to select the target individuals
for the various indicators such as children under 5 years old or women of reproductive age.
In this Learning Unit, the aim is for the participants to know which kind of questions should
be added to the surveys, and to have reviewed the type of questions that these surveys use for
obtaining malaria-related indicators.
Exercise 16.9
a. Has a survey been carried out on malaria in your country?
b. If yes, in what year? Which survey? Have you seen its results? Are the results available on the web? If
not, is it planned to publish them on the web? Do you think that the survey was necessary?
c. If the survey was necessary, what role has it played (or will play) in complementing the data on
malaria in your country?
d. Can these data replace the routine disease surveillance system?
U16
LEARNING UNIT 17
Burden of malaria
■ ■ Explain how an estimate of malaria morbidity and mortality
can be obtained from the surveillance system
■ ■ Explain the need for indicators of disease burden
■ ■ Describe the difference between inequity and inequality
123
Morbidity and mortality due to malaria are key measures of burden of malaria. This Learning
Unit has three sections. The first section presents a method to estimate malaria morbidity and
mortality on the basis of data available in the health surveillance system. The second section
discusses an indicator of disease which is obtained by combining morbidity and mortality,
providing a summary indicator of the burden of malaria in a region or country. The third section
discusses estimation of inequity and inequality in the presence of diseases, health outcomes
and access to health care, and the differences between them.
17.1 Malaria morbidity

Surveillance systems usually have some degree of under-reporting. Under-reporting was
introduced in Learning Unit 13, and is measured by measuring the completeness of reporting.
As a reminder, completeness of reporting is:
Completeness of No of HF reports received through the surveillance system

= x 100
HFa reporting No of HF reports expected through the surveillance system
a
HF = health facility
The completeness of reporting measures the rate of underreporting. For example, if the malaria
completeness of reporting is 50% in a country, and if undertreporting is random and not
associated with malaria case loads, this means that the system reports half of the cases, and so
to estimate the actual number of cases, the number reported must be multiplied by 2.
Exercise 17.1
a. If the completeness of reporting is 20% in a country, then what number should the notified
individuals be multiplied by
b. What are the limitations in using reported morbidity data generated from the malaria surveillance
system?
c. Illustrate in an algorithm the malaria cases reported in public sector through the malaria surveillance
system.
Compare it with the Figure 17.1 provided by the tutor. What are the differences between them?
Explain these differences. The algorithm allows to identify the factors affecting the malaria
surveillance system. If these factors are known, then the correction factor can be quantified, and the
true number of cases can be estimated.
d. Point out and name the measurements that are required to calculate the real number of cases in
Figure 17.1.
Compare the result with the Figure 17.2 provided by the tutor.
e. Where can the data on factors affecting surveillance system sensitivity be obtained?
– Utilization of public health facilitie (u)
– Completeness of health facility reporting (r)
BURDEN OF MALARIA 125
Since r and u are both probabilities, their values are less than 1; u = proportion of individuals
that use public services for malaria diagnosis and treatment; r = percentage of health facilities
(HF) reports that are received in the surveillance system compared to all HF reports expected.
Exercise 17.2
a. Examine Figure 17.3 (provided by the tutor) to arrive at the correct diagnosis of a malaria case.
b. Which data are needed to arrive at the correct diagnosis of a malaria case?
c. Where can the data for these variables be obtained?
d. List the variables that are required to calculate the number of malaria cases (Morbidity: M), and
show whether the relationship is direct or inverse when estimating M.
Numerical example:
In 2009, 4 500 000 persons were tested for malaria in a country, of which 150 000 tested
positive. Therefore:
150 000
SPR = = 3.3%
4 500 000
4 000 000 malaria cases were reported from public health facilities which did not have a
microscopic examination and/or RDT, and are therefore possible or unconfirmed malaria
cases. Therefore if SPR = 3.3%, then:
U17
U x SPR= estimated true malaria cases among probable cases = 4 000 000 × 3.3% = 132 000
The total number of confirmed cases reported in the public sector will be:
C + (U x SPR) = 150 000 + (132 000) = 282 000
It has also been observed that only half the public centres have sent their reports:
C + (U x SPR) 282 000
= = 564 000
r 50%
So the total estimated number of cases in the public sector should be 564 000. If the household
survey indicates that only 20% of individuals with malarial symptoms have consulted the
public sector and 80% have either not consulted or consulted the private sector then:
C + (U x SPR) 564 000
M= = = 2 820 000
rxu 20%
2 820 000 is the total number of estimated malaria cases in 2009.
e. Following the numerical example above, generate an estimate of malaria cases in your country.
If this is not possible, which data are missing and where can they be found?
17.2 Malaria mortality indicator

Exercise 17.3
a. Which sources can be used to estimate the number of malaria deaths?
In order to estimate malaria mortality indirectly, the number of P. falciparum malaria cases can be
multiplied by the Case Fatality Rate (CFR).
The CFR of P. falciparum malaria varies in different countries, and has been estimated to be 0.01–
0.3% outside Africa, whereas the number appears to be around 0.45% in African countries.
b. From the numerical example given above, how many persons have died as a result of malaria in
2009, assuming the proportion of P. falciparum and its CFR were 50% and 0.09%, respectively?
17.3 DALY: a summary indicator of disease burden combining

morbidity and mortality
Traditionally, assessments of population health status were made from a variety of data sources
including mortality rates, cause of death data, as well as information on the incidence and
prevalence of disease injury and disability. Each of these (and others) are useful sources of data
in their own right but each only addresses a limited aspect of the disease burden in a population.
There is a need for a more comprehensive measure of disease burden that simultaneously takes
into account levels of mortality, disability and ill-health.
This need arose in the context of assessment of public health impact of different diseases. The
main question is how can different diseases that have different outcomes be compared using
a single metric? What are the impact of different interventions on the health of a society? For
example, which disease is of higher priority, tuberculosis or malaria? Should more resources be
allocated to malaria or gastric cancer? In order to respond to such questions, a common metric
among diseases was necessary for measurements of disease burden and assess the results of the
health interventions.
To illustrate the application of these indicators, suppose that the mortality rate of malaria and
of a certain type of cancer is 1/1000. Which one of these creates a heavier burden? To assess
the disease burden, it is necessary to know: (i) at what age the disease occurs – death occurring
at a younger age involves greater loss of future productive years, and (ii) which disease creates
disability, and how long will this disability last? The single indicator should take all these aspects
into account.
These indicators have been created taking into account the goals of the health system.
Health system goal: to increase the quality of life and life expectancy for all.
This definition can be divided into three parts:
▶▶ Life expectancy is the average number of years of life remaining at a given age
▶▶ Quality of life refers to the general well-being of individuals
▶▶ Everyone has the right to health, a state of complete physical, mental and social well-being
(benefit from health equity)
Assessment of health therefore involves considering a) life expectancy and b) the years lost
because of poor health or disability. To know how much health has been lost, it should be
possible to calculate mortality (which reduces life expectancy) and years lost with disability
which results from non-fatal illnesses, and then combine these two together.
The Disability Adjusted Life Years (DALY) indicator was created to compare the burdens of
disease based on the principle that the most appropriate measure of the effect of chronic illness
is time. This indicator is the sum of years of life lost (mortality) and years lived with disability
(Table 17.1).
Table 17.1 Relationships between disability adjusted life years, years of life lost and years lived
with disability
DALY = YLL + YLD

Disability Years of Years
Adjusted Life Lived with
Life = Lost + Disability
Years
DALY = Mortality + Morbidity
For example, consider the next two tables Table 17.2 and 17.3, and see how DALY can be
calculated from YLL and YLD.
There is Life Expectancy (column 2) for each Age Group (column 1). This Life Expectancy
U17
means how long on average this age group is expected to live. In the ‘Global Burden of Disease’
study carried out in the early 1990’s, the life expectancy of the Japanese (the longest, 82 years
at birth) was considered as the standard. On this basis, all individuals all over the world are
considered to have the right to the highest life expectancy, and if they die before 82 years, they are
considered to have lost the number of years of life between age at death and this life expectancy.
The number of years lost is calculated by multiplying the number of deaths (column 3) of that
age group by the number of years they should have lived (Standard life expectancy-column
2), giving the results in column 4. From the sum of years of life lost in different age groups, the
amount of YLL resulting from a particular disease is obtained, as shown below.
Table 17.2 Calculation of years of life lost (YLL)
Age group (1) Standard life expectancy (2) Number of deaths (3) Years of life lost (4) YLL
….
20–24 63 2312 145 656
25–29 58 3887 225 446
30–34 53 4512 239 136
35–39 48 5321 255 408
….
Total YLL 865 646
Table 17.3 below is used for calculating YLD. For each type of disabilities related to the diseases
in question, the following parameters should be calculated: the incidence of each type of
disabilities, the level of incapacitation produced by them and their duration.
Table 17.3 Calculation of years lived with disability
Number (or incidence) Disability weight Duration of disability Years lived with
of disability (year) disability (YLD)
Outcome A 453 0.70 7.0 2219.70
Outcome B 348 0.20 0.6 41.760
Outcome C 945 0.45 0.5 212.62
Total YLD 2474.08
In calculating malaria in the Global Burden of Disease Study, three disabilities were consid-
ered: symptomatic episodes, anaemia, and neurological sequelae following cerebral malaria.
Neurological sequelae include a range of conditions such as hearing impairment, quadripare-
sis, epilepsy and visual impairment.
The disability weight must be known in order to calculate YLD. Thus a number of 0–1 has
been allocated to each disability, where 0 stands for full health, 1 for death, and disabilities
caused by disease are in the range 0–1. If the disease causes major disability (e.g. Bi-polar major
depression) the disability burden will be nearer to 1, and if it causes minor disability (e.g. a scar
that is not visible to others under ordinary circumstances) it will be nearer to 0.
Therefore, in simple terms, one DALY represents the loss of one year healthy life. Two other
prerequisites (other than disability weight) are considered in the calculation of DALYs: (1)
All ages are not of the same value in the DALY calculation; (2) In this calculation there is
discounting over time, so that the value of a year lost in the future is worth less than one lost
now.
Considering malaria, best estimates currently describe the annual global burden of malaria as:
44 million disability-adjusted life years (DALYs). In 2002, malaria was ranked the 8th highest
contributor to the total global DALYs and 2nd in Africa. According to these estimates, 3% of
the world’s DALYs are due to malaria.
It is clear from the natural history of malaria in humans that the major contribution of malaria-
associated DALYs is from P. falciparum-associated deaths, especially in children aged <5 years.
A disease that has a higher DALY is not necessarily of higher public health priority. Variables
other than burden of disease need to be considered in the prioritization of diseases, e.g.
availability of suitable interventions and international commitments, e.g. inclusion in the
Millennium Development Goals.
The validity of the DALY as a universal indicator is the subject of debate. Some criticism
concerns the method of choosing certain parameters such as disability weight, age weight, and
rate of discount. The disability weight may be different in different countries – for example the
social stigma attached to sexually transmitted diseases is not the same in all countries. Therefore
the validity of choosing a single disability weight for all countries is questioned.
17.4 Inequality and inequity

Health Inequality refers to differences of health and health care across populations. Health
Inequity is difference in presence of disease, health outcomes or access to health care, which is
not only unnecessary and avoidable, but is also considered unfair and unjust.
Consider these examples:
▶▶ Malaria is unequal in people who use and those who do not use ITN.
▶▶ IRS coverage is different in different areas.
▶▶ Access to proper malaria diagnostic testing and treatment in many countries is affected by
family income.
In all of these cases, individuals are unequal, but which of the situations are unnecessary,
avoidable and unfair? The specific of health inequalities that are judged unjust or unfair
constitute health inequities.
There are many ways of calculating inequity. Most of them provide a means of assessing the
degree of income-related differences in the distribution of a health variable. For example,
an assessment could be made whether child mortality is related to level of family income,
measuring the relation between child mortality and quintile of family income in different
regions of the same country or in different countries.
U17
LEARNING UNIT 18
PART 4
Introduction to
malaria epidemics
■ ■ Define a malaria epidemic
■ ■ Identify contributing/triggering factors
131
18.1 Definition of a malaria epidemic

Exercise 18.1
a. Have you ever been directly involved in prevention and/or control of a malaria epidemic?
b. What is your definition of a malaria epidemic, or a malaria outbreak?
c. What different types of malaria epidemics do you know? Could you assign these types to the
situations (i), (ii) and (iii) indicated in Figure 18.1?
(i)
(ii)
(iii)
Figure 18.1 Major malaria epidemic types based on the epidemic patterns
18.2 Population at risk of malaria in epidemic-prone regions

Exercise 18.2
a. Where do malaria epidemics most frequently occur, and why?
Briefly describe the main epidemiological settings in which epidemics occur.
b. In epidemic-prone regions, are only children aged <5 years and pregnant women at risk?
Explain the reasoning for your answer.
c. In your country, are there malaria-free areas, or malaria epidemic-prone areas?
If there are epidemic-prone areas, is the population at risk well identified, and regions at risk
mapped?
d. Why is assessment of risk of malaria epidemics important? Explain briefly the reasons.
e. Identify at least 3 conditions that make human populations vulnerable to malaria epidemics.
(Consider immunity, parasite and environment.
f. What are the characteristics of epidemic-prone areas? Provide some examples of environmental
settings.
INTRODUCTION TO MALARIA EPIDEMICS 133
18.3 Indicators of malaria transmission and

monitoring epidemic risk
Exercise 18.3
a. For successful malaria transmission by the vector, state at least 2 epidemiological requirements
(consider the parasite, vector and host interaction)
b. What does the basic reproduction rate (Ro) equation express? (describe the equation in words)
Ro = m a2 pn / r – loge p
A key indicator of actual malaria transmission in a given area is entomological inoculation rate
(EIR), i.e. the number of infective bites per person per time. Based on the Ro, there are direct
and indirect factors responsible for malaria transmission. Direct factors appear in the equation,
but these parameters are frequently difficult or impossible to measure.
c. In Table 18.1:
i. State the factors that have influence in positively or negatively affecting the determinants of
malaria transmission given in the first column.
ii. Which of the two categories (direct or indirect factors) can be used as indicators to predict
epidemics?
Table 18.1 Direct and indirect factors that contribute to occurrence of malaria epidemics
Determinants (direct) Influencing factors (indirect)

Vector density
Human biting
Rate of gametocyte carriers U18
Length of sporogony
Daily survival rate of vectors
The use of health statistics to differentiate endemic and epidemic prone areas
d. Are health statistics (national/local epidemiological characteristics) useful for differentiating
between endemic and epidemic-prone areas?
Explain your answer. List some of the limitations of using health data.
18.4 Precipitating factors for malaria epidemics

The level of immunity acquired by the exposed population plays a decisive role in the
occurrence and severity of malaria epidemics. Malaria immunity is not generally maintained
unless exposure and human-vector contact are frequent and regular.
Exercise 18.4
a. List the main precipitating factors for malaria epidemics in your country/other places.
b. Working in small groups, fill Table 18.2 and link/correlate these factors to the main epidemiological
and environmental consequences.
Table 18.2 Precipitating factors for malaria epidemics and their consequences
Cause Potentially leading to:

1. Human activity
Economic development activities (legal or not)
Overpopulation
War/civil disturbances
Road construction/ improvement in transport facilities
Urbanization
2. Natural disasters
Expected or unexpected meteorological events (heavy rainfall,
unusual heavy flooding, cyclones, climatic changes...)
Global and local climate changes
3. Degradation of preventive and curative health services
Deficient surveillance within the control services
Malaria prevention activities are deteriorating (such as lack/shortage
of insecticide and/or inadequate or poor spray coverage)
Reluctance of villagers concerning some malaria control activities
(especially spraying operations)
Increasing resistance of vectors to insecticide and/or parasites
to antimalarials
4. Success of prevention and control efforts
Waning population immunity
c. Is it necessary to identify precipitating / contributing factors? Explain your answer.
Discuss the following information required to explore potential contributing factors to

malaria epidemics
▶▶ In case of past epidemics in an area: Ask whether the epidemics were well documented and
go back to reports if available.
▶▶ Epidemiological data from the health information system: Study the malaria transmission
pattern in the region and classify it as a high, short- seasonal or cyclical epidemics, or other.
▶▶ Meteorological data: Look for unusual rainfall, temperature, humidity pattern correlated
with epidemics in the past or correlated with current epidemics.
▶▶ Population movements: Observe whether there had been a large group of people coming
(or passing through) from high transmission areas recently [refer to camps and refugee
situations].
▶▶ Increase in population vulnerability/susceptibility: Population affected by hunger and
malnutrition as result of unusual drought, (forced) migration (e.g. due to civil war), HIV/
AIDS prevalence, etc.
▶▶ Check whether control measures are in place.
▶▶ Examine whether these control measures are effective, e.g. case management affected by
high parasite resistance to national recommended drugs, or affected by vectors resistant to
insecticides in use, etc.
d. Indicate where to obtain the necessary information for each of the questions above.
LEARNING UNIT 19
Early warning, detection,
notification and verification of a
malaria epidemic
■ ■ Describe the usual channels for notification
■ ■ Explain the concept and rationale of an early warning and
detection system
■ ■ Describe how to identify/detect a malaria epidemic on a
timely basis
■ ■ Describe how to rapidly confirm a malaria epidemic
■ ■ Describe urgent measures to contain the epidemic
135
19.1 Systems for early detection of malaria epidemic

Exercise 19.1
a. Working in small groups, discuss and sketch a diagram showing how any unusual events/epidemics
including malaria epidemics are usually reported/notified.
b. Working in small groups, explain the rationale for setting up an early detection system.
19.2 Methods for determining epidemic thresholds

Exercise 19.2
a. What system(s) would you propose for routine early detection of malaria epidemics and what kind
of data would it be important to use?
Many epidemics occur in situations where previous data is either not available, or unreliable due to
significant contextual changes. In these circumstances, precise thresholds will be difficult to establish.
What sort of indicators should be used to monitor malaria epidemics in situations like this?
b. Median and 3rd quartile
Work in small groups. Table 19.1 provides data on malaria cases reported during a 5-year period
from Province X. Respond the questions (i) – (iv) and present the findings to the class.
Table 19.1 Malaria cases reported from Province X from 2001 to 2005
Year Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sept. Oct. Nov. Dec.
2001 1609 2235 2035 1597 4927 2442 2857 5159 9245 1490 1299 2267
2002 1214 1322 1784 1880 1863 1958 398 2815 4761 5845 2588
2003 1198 1099 2010 1411 1449 2018 1737 1902 1939 1842 2332 2321
2004 2597 2219 2988 2977 5276 3534 2822 4028 3188 3395 2269 2223
2005 2941 2449 2619 2462 2973 2200 2612 2424 8658 10158 4274 2944
i. Calculate the median and the quartiles of the cases for each month and show them on a table.
Compare your results with Table 19.1 provided by the tutor.
ii. Plot the numbers in Excel or spread sheet. If a computer and Microsoft Excel are not available,
use graph paper or simple squared paper to construct a graph of the median and the 3rd quartile.
iii. What does the 3rd quartile indicate on the graph?
iv. Given the median and quartiles obtained for the years 2001–2005 and the malaria situation
for the year 2006 with the data shown in Table 19.2 below, which month of 2006 shows an
epidemic situation? Explain why.
Table 19.2 Malaria cases reported from Province X in 2006
Year Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sept. Oct. Nov. Dec.
2006 1364 2560 2817 1656 1958 2021 2255 3169 4897 9158
EARLY WARNING, EARLY DETECTION, NOTIFICATION AND VERIFICATION OF A MALARIA EPIDEMIC 137
Exercise 19.3
a. From the malaria morbidity data given below in Table 19.3, determine the median, lower (1st) and
upper (3rd) quartiles (by bolding or shading the columns). Then compare the data for year 2005 in
the last column of the table.
b. Plot the Median, Upper and Lower Quartiles derived from the weekly data on a graph.
c. What do you conclude from the graph?
Table 19.3 Malaria morbidity data of Province Z (1996–2005)
Month/Week 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 data
Jan. 1 48 20 121 76 44 47 24 9 40 63
2 50 8 83 65 39 45 29 6 37 41
3 105 13 112 108 43 48 31 8 32 56
4 88 15 130 96 49 52 35 1 46 69
Feb. 1 72 9 88 84 44 46 38 7 41 41
2 58 14 100 76 41 48 23 4 31 45
3 52 11 74 70 35 38 27 3 28 50
4 70 14 98 97 49 68 41 8 43 39
Mar. 1 42 11 69 50 33 39 30 4 32 50
2 66 10 80 77 39 45 24 3 25 52
3 45 18 67 59 39 44 32 5 35 46
4 49 9 83 63 34 45 9 8 33 38
Apr. 1 44 18 62 47 40 44 25 5 35 42
2 45 14 76 48 36 44 23 6 25 30
3 35 22 61 53 34 35 31 8 33 23
4 57 19 60 60 42 54 27 18 42 36
May 1 53 15 106 77 43 44 24 9 35 36
2 54 23 72 55 38 42 25 9 31 45
3 69 23 95 80 34 37 27 5 33 44
4 94 37 113 95 71 76 40 24 57 76
Jun. 1 81 25 113 93 54 74 31 21 52 44
2 69 39 123 72 55 60 48 17 50 61
3 69 23 157 116 58 67 40 15 51 64
U19
4 100 41 200 113 59 94 46 18 57 79
Jul. 1 94 29 184 110 60 62 29 19 55 93
2 69 25 200 129 56 60 44 11 51 62
3 92 24 129 123 53 63 31 15 40 68
4 92 23 216 96 64 74 31 16 64 89
Aug. 1 86 23 112 89 41 77 29 9 39 81
2 95 23 202 109 73 81 26 5 29 112
3 99 28 156 154 60 90 41 13 55 132
4 159 29 271 200 104 118 41 9 62 188
Sept. 1 93 19 136 128 80 84 75 13 79 162
2 159 30 221 193 126 141 95 9 116 188
3 149 26 259 181 127 137 108 11 109 258
4 272 62 540 279 139 183 101 31 105 346
Oct. 1 165 27 302 190 100 161 28 23 85 350
2 185 36 272 220 131 174 69 28 88 322
3 159 33 181 170 100 123 68 23 83 343
4 182 46 276 221 169 179 71 25 109 413
Table 19.3 Malaria morbidity data of Province Z (1996–2005) (continued)

Month/Week 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 data
Nov. 1 122 30 230 139 103 104 42 19 85 487
2 96 13 192 128 73 86 41 11 71 378
3 79 23 153 100 58 74 32 18 51 383
4 116 18 248 147 95 114 53 12 66 265
Dec. 1 93 17 143 108 62 69 45 11 50 106
2 76 25 115 99 53 65 41 12 46 113
3 55 12 114 67 50 53 33 7 41 163
4 50 19 102 69 36 37 21 15 26 223
19.2.2 Cumulative-sum methods
Exercise 19.4
a. Use Table 19.1 to practice the cumulative-sum (C-SUM) method as follows:
To calculate the C-SUM for January, add the sum for December, January and February
for the previous 5 years, and divide the total by 15. Similarly the C-SUM for February is
calculated by adding the sum of January, February and March for the previous 5 years and
dividing by 15.
Then complete the C-SUMS for each month in Table 19.4 below.
Table 19.4 Cumulative-sum of malaria cases for each month, from Province X from 2001 to 2005
Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sept. Oct. Nov. Dec.
C-SUM
b. Plot the results on a graph. The line represents the threshold above which one should be on alert for
an epidemic.
c. Add the data for 2006 to the graph. Show both lines on the same graph and interpret them.
d. These methods use monthly data. What could be the disadvantages of using monthly data?
Ideally, data should be collected and analysed at the peripheral level on a weekly basis, and the
methods described can be used to analyse weekly figures. It is possible to use the thresholds
developed from 5 years of monthly data and apply them to current weekly data.
e. Use graph paper with the quartile threshold or C-SUM clearly marked as in exercises 19.2 (b) and
19.4 (b). As the weekly data are collected, mark the number as a column under the month in
question. The next week, add the number to that of the first week, and extend the column in
the same figure. Using a different colour for each week will make it clearer. Do the same for
the 3rd and 4th weeks. If the column is already at the threshold by the 2nd or 3rd week, it will
be possible to raise the alert of an epidemic much sooner than waiting for the whole month’s
figures. Compare the graph you have developed with the one given by the tutor.
EARLY WARNING, EARLY DETECTION, NOTIFICATION AND VERIFICATION OF A MALARIA EPIDEMIC 139
19.2.3 Incidence thresholds

Exercise 19.5
Working in small groups, discuss how the use of alert threshold and epidemic threshold could
be applied to malaria.
19.3 Early detection and overall management systems

Exercise 19.6
Working in small groups, list the measures that should be set-up in a country to detect a malaria
epidemic at an early stage, either as part of the routine information system or/and as part of the
epidemic surveillance system. Organize the measures at peripheral, district and national levels.
19.4 Verification of malaria epidemics

Exercise 19.7
a. Working in small groups, discuss and outline the steps that should be followed to verify malaria
epidemics at peripheral level.
b. Working in small groups, prepare a flow chart that covers early detection, verification and notification
of malaria epidemics. The flow chart should indicate a logical flow from the peripheral to the district
level. Present the findings to the class for discussion, and compare with the results provided by the
tutor.
c. In settings where parasitological confirmation is not available, it is necessary to monitor the
percentage of suspected cases that have confirmed malaria, by regularly sampling a given number of
suspected cases for parasitaemia in different epidemiological settings.
Working in small groups, discuss test positivity (SPR or positive RDT rate) at normal and epidemic
times in arid/semi-arid areas; and highland regions with short seasonal transmission.
Where would positivity be higher during epidemics? Why?
U19
Where would there be most need for laboratory diagnosis?
19.5 Monitoring areas of epidemic risk

▶▶ In 1980, a two-stage malaria epidemic forecasting system, based on the monitoring of mete-
orological variables and changes in the entomological inoculation rate (EIR), was proposed.1
The system is suited to the forecasting of resurgent outbreaks where comprehensive surveil-
lance systems have already been developed and operated for some time, as part of the routine
malaria control services. However, the capacity to monitor EIR as a routine component in
an epidemic forecasting system is at present beyond the capacity of most African countries.
▶▶ Experiences from intersectoral programmes (e.g. food safety) which use environmental
information systems (EIS) could also be applied to the development of monitoring and early
warning systems (MEWS).
1.
Onori E. and Grab B. (1980). Indicators for the forecasting of malaria epidemics. Bulletin of the World Health Organization
1980;58 (1):91–98.
▶▶ Kenya, Uganda and the United Republic of Tanzania, in conjunction with HIMAL have
proposed the development of a three-tiered approach for malaria epidemic forecasting, early
warning and early detection in the highlands of East Africa,1 with each tier being associated
with specific indicators and responses.
Exercise 19.8
Analyse the model in Figure 19.1 and answer the questions below. It shows forecasting, early
warning and early detection model resulting from the Salt Rock Meeting, South Africa (Anon,
1999).
a. At which level should one consider (i) long range weather forecasting, (ii) early warning based on
meteorological indicators (iii) early detection?
b. What would be the possible indicators and responses for flag 1, flag 2 and flag 3?
Figure 19.1 Forecasting, early warning and early detection model resulting from the Salt Rock Meeting,
South Africa, 1999
1.
Anon (1999). Developing new approaches for the surveillance and control of malaria epidemics in the highlands of East Africa.
Highland Malaria Project and Mapping Malaria Risk in Africa: Salt Rock, South Africa.
LEARNING UNIT 20
Prevention and early response
to confirmed malaria falciparum
epidemics
■ ■ Describe the main malaria epidemic control options
■ ■ Describe how to identify the most cost-effective malaria
epidemic prevention and control options
141
20.1 Cost-effective interventions for control of P. falciparum

malaria epidemics
Exercise 20.1
a. Working in small groups, propose a list of the most important interventions for control of malaria
epidemics.
Rank the listed interventions according to their cost-effectiveness.
Of the listed interventions, which would need to be set up at early stages of any notified malaria
epidemics?
b. What should be the guiding principle/s for a drug to have a significant effect on transmission during
epidemics?
c. Do good case management procedures have a significant effect on transmission if drugs such as
chloroquine and sulfadoxine-pyrimethamine are used?
Should artemisinin-based combinations and/or primaquine be used?
In emergency situations, should effective drugs be provided free of charge?
d. During emergencies or epidemics what is a common problem for provision of care to those sick at
home in affected remote communities? What operational solutions would you propose to reach
those affected communities?
Exercise 20.2
a. Work in small groups. For operational and biological reasons vector control options may or may not
be applicable in epidemics. Discuss under which circumstances these options should be considered in
epidemic-prone districts.
b. What stage of the vector should be targeted by vector control measures in order to have significant
impact on malaria transmission and hence on the malaria burden?
c. Under what circumstances is IRS a viable option?
d. Under what circumstances is the use of ITNs or impregnated materials considered to be a viable
option?
e. What specific vector control options/methods would you recommend in situations of complex
emergencies, e.g. in setting up refugee camps?
f. Where and when would you consider other vector control options?
Exercise 20.3
Work in small groups. Use the data in Table 20.1 to test different scenarios of early detection
and interventions for epidemics, and the implications if these measures are not taken. Data in
Table 20.1 (columns 2–5) show the number of malaria cases corresponding to time in months
(column 1). Plot them on graph paper (or by computer).
PREVENTION AND EARLY RESPONSE TO CONFIRMED MALARIA FALCIPARUM EPIDEMICS 143
Table 20.1 Number of cases recorded at different stages of detection of malaria epidemics
Time Late detection Some delay in detection Early detection

No intervention
(month) & intervention & intervention & intervention
1 12 12 12 12
2 13 13 13 13
3 14 14 14 14
4 13 13 13 13
5 11 11 11 11
6 24 24 24 24
7 49 49 49 35
8 199 169 129 39
9 190 160 120 38
10 185 155 115 15
11 150 120 90 14
12 70 40 12 13
13 60 35 14 12
14 50 25 13 13
15 26 11 12 12
16 14 14 13 13
17 13 13 12 12
a. From the data in Table 20.1, show each column (number of cases) against time on a graph. Produce
one graph for each column by comparing with column 2. Shade the graphs differently to show the
different areas under curve in each graph.
b. What are the differences between these scenarios? What does the area under the curve show in each
graph?
c. Which column (from the graph produced by the group) shows the existence of proper preparedness
and response (better malaria control programme) when compared to column 2 where no
intervention was done?
d. If climate forecast, early warning and early detection for a complete prevention of the epidemics
would work well for your programme, how would you indicate them on the same graph of column 2
U20
and what will be their time sequence? What would result from this system?
20.2 Case management and drug policy during epidemics

Managing malaria cases during epidemics
Prompt diagnosis and access to effective antimalarial medicines during a malaria epidemic is
a key intervention to minimize the malaria burden in the affected population. Most malaria
patients in epidemics are non-immune, partially immune, or otherwise vulnerable to severe
disease. The antimalarials to be used for treatment must be highly efficacious (> 95% cure),
safe and well tolerated so that adherence to treatment is high. Complete courses of treatment
should always be given in all circumstances. See module on Malaria Case Management for
further information.
Exercise 20.4
a. Currently WHO recommends parasitological confirmation prior to antimalarial treatment during
non-epidemic situations. Is this applicable to epidemic situations?
b. In a country that has epidemic-prone regions, for managing uncomplicated malaria cases during
epidemics, is it possible to use a different antimalarial medicine from the one used as first-line drug
treatment for uncomplicated malaria in endemic areas in the country? Explain your answer.
c. Epidemics can overwhelm health services and severe malaria is common if prompt effective
treatment is not made widely available. In addition to individual patient-based treatment of
uncomplicated cases through fixed or mobile clinics, is there another complementary strategy for
using antimalarials?
d. What antimalarial drugs should be recommended for use in treating severe malaria cases during
epidemics? Specify the recommendations for referral and peripheral health facilities.
20.3 Vector control options for prevention and control

of malaria epidemics
Epidemics can be either prevented or controlled by using vector control interventions like IRS
which have a direct impact on transmission in the affected areas. If the interventions are not
well implemented, epidemics will continue to progress over time and space up to their final
natural resolution and ending. If well planned and carried out at the appropriate time (early
stages of the epidemic) as a result of accurate warnings, vector control interventions can reduce
transmission by killing adult mosquitoes before they transmit the disease. The list below
summarizes the type of vector control options that may or may not be relevant to prevention
and control of malaria epidemics:
▶▶ Indoor residual spraying of insecticides (IRS)
▶▶ Insecticide-treated nets (ITNs/LLINs) and other materials
▶▶ Space spraying of insecticides (limited role)
▶▶ Larval source management (larviciding)
The available insecticides for prevention and control of epidemics are listed in Table 20.2.
Table 20.2 Insecticides for prevention and control of epidemics
For indoor residual spraying:

• Organochlorines (DDT)
• Organophosphates (Fenitrothion, Malathion; Pirimiphos-methyl)
• Carbamates (Bendiocarb, Propoxur)
• Pyrethroids (Alpha-cypermethrin; Bifenthrin, Cyfluthrin, Deltamethrin, Etofenprox, Lambda-cyhalothrin)
For larval control:

• Temephos pirimiphos-methyl, fenthion, novaluron etc.
For impregnation of bednets, curtains, blankets and clothes:

• Pyrethroids (Permethrin, Delatamethrin, Alphacypermethrin, etc.)
20.3.1 Indoor residual spraying of insecticides (IRS)

During epidemics, when political pressure is high and appropriate preparedness lacking,
vector control measures like IRS are frequently employed after the transmission peak or even
after transmission has almost ceased (see exercise 20.3 and scenarios of epidemic detection
and response).
Many epidemic-prone regions in Africa practice some form of routine vector control, mainly
indoor residual house spraying. However, in many countries, the practice may have lost the
required operational quality and effectiveness. Shortage of skilled vector control staff and lack
of adequate planning and supervision often compromise the effectiveness of IRS. Despite these
problems, IRS is the most effective measure to prevent malaria epidemics when there is a reliable
early warning system, and where the basic requirements of logistics, expertise and personnel
exist.
Exercise 20.5
a. Taking the biological factors (life-cycle of the parasite and the vector) into account, what would
be the minimum time (in weeks) needed for application of IRS to effectively prevent a malaria
epidemic?
b. Is it possible to apply IRS once an epidemic has started? If so, within what period of time?
c. Why is coverage so important for effectiveness of IRS?
d. Would IRS be more effective in areas where there is high vectorial capacity or in areas with low
vectorial capacity?
e. Working in small groups, discuss and identify the important issues (criteria) that need to be
considered when IRS operations are planned.
20.3.2 Insecticide-treated mosquito nets (ITN) and other materials

Another vector control strategy involves the use of insecticide-treated mosquito nets and other
materials. The use of ITNs is high in many parts of Africa where malaria is endemic, although
the coverage in many epidemic-prone areas is still low. If distribution of ITNs is not a practical
option given the urgency of epidemic prevention and control operations, re-treatment of nets
could provide a sufficient degree of protection in areas where a high level of coverage has been
U20
achieved. In such situations, the re-treatment service should be provided free of charge.
Exercise 20.6
a. For ITNs to be fully effective, what coverage and re-treatment rate must be attained? Why is coverage
an important criterion?
b. Why is coverage and acceptance by the community in epidemic areas/regions far less than in most
endemic areas?
c. Under which circumstances would ITN be effective in preventing epidemics?
20.3.3 Space spray application

There is little evidence of any impact of space spraying on malaria. Thus, this measure is
generally not recommended, but can be considered in exceptional circumstances (e.g. outdoor
biting and resting mosquitoes). Such activity would need to be maintained over time to have
any impact, and entomological expertise is required if this intervention is deployed.
Exercise 20.7
a. What stage of the mosquito do aerial sprays mainly target?
b. Is aerial spraying a viable option? If it is possible only in limited circumstances, specify some of them.
20.3.4 Larval source management

Larval source management (LSM) includes breeding site modification and manipulation;
larviciding and the use of biological control agents. Larviciding, in conjunction with other
vector control interventions (IRS and ITNs/LLINs), may play a role to control malaria
epidemics in areas where the breeding habitats are few, fixed and findable. This is based on the
assumption that larviciding is applied at the early stages of the malaria epidemic to potential
breeding sites.
Exercise 20.8
a. Given the limitations, where should larval control be applied in relation to prevention and control of
malaria epidemics?
b. What biological and operational factors should be taken into account when considering larval
control?
20.4 Malaria in relation to epidemic diseases

Standard case definitions need to be agreed upon and peripheral staff trained to use the
definitions and report to district authorities. Epidemiologically, the consistency of definitions
of malaria cases over time is very important and needs to be well monitored (while every effort
should be made to improve access to diagnostic testing). In known epidemic-prone districts,
malaria cases are generally reported on a weekly basis together with other epidemic diseases.
Exercise 20.9
a. In areas where there are no laboratory facilities, how would be malaria defined? And should all
patients who received antimalarial treatment be classified as malaria cases?
b. Should malaria epidemic control interventions be set up as isolated / vertical interventions or
integrated with other epidemic interventions?
c. If malaria interventions should be integrated with other epidemic interventions, which ones?
d. Of the preventive and control interventions for malaria epidemics, which one do you think is most
effective?
20.5 Measuring the impact of preventive/control measures

From historical records there is little documented information on the impact of epidemics.
In general, epidemics took the population by surprise and health authorities often lack a
preparedness plans of action. The magnitude of the malaria burden during epidemics has
been linked to the stock-out or absence of efficient drugs, absence or long delay of measures to
reduce transmission, and limited knowledge of the population and media on how the disease is
transmitted, cured, and can be avoided.
Exercise 20.10
Work in small groups. When preventive measures are undertaken in advance, or control options
implemented at an early stage, deaths and severe cases averted in a targeted population can be
calculated based on some conservative assumptions. Identify the underlining assumptions for
measuring impact of prevention and of control.
20.6 Other managerial issues for preparedness and

response during epidemics
The NMCP is responsible for advising the government on strategic technical measures to be
taken when malaria epidemics occur. It recognizes defining areas and populations at risk and
pre- defined cost-effective control options, drugs and insecticides to be used, etc. In epidemic-
prone areas, the district management teams are in charge of establishing a plan of action which
includes early warning and detection systems coordinated with field partners including NGOs.
The process to elaborate a preparedness plan of action is described in the Learning Unit 21.
Malaria epidemics can also occur in complex emergency situations and there are specific
strategies for responding to such situations.
Exercise 20.11
a. Working in small groups, discuss the best options for malaria control that would be employed in a
complex emergency situation, when there is limited knowledge of the local malaria situation.
b. Work in small groups. Consider the three graphs in Figure 20.1 showing different stages of epidemic
detection. Use the checklist in Table 20.3 to mark each intervention as correct (tick) or incorrect
(cross), if it is applicable or not applicable for the particular phase of an epidemic. You may add U20
specific comments in the corresponding box.
Fig i. Epidemic detected at Fig ii. Epidemic detected Fig iii. Epidemic detected at
its start after some progression its peak
Impact
Impact
Impact
Time Time Time
Figure 20.1 Detection of malaria epidemics at different stages

Table 20.3 Operational responses to different stages of malaria epidemics
Starting Accelerated Epidemic

No. Interventions or operational measures
epidemic epidemic peak
Ensure all clinics and health facilities are operational

1
and have sufficient drugs, equipment and trained staff
Establish treatment centres (temporary clinics or mobile clinics)

2
where access is difficult or health facility coverage is low
Ensure that the correct diagnosis and treatment is provided at all

3
health facilities and at community level
Promote pro-active clinical case detection and management/

4
referral
Reinforce the referral system and consider the introduction of
5 artesunate suppositories and intra-muscular artemether as a
temporary measure where these are not already used
Intensify/maintain effective preventive measures for pregnant
6
women as locally appropriate
Reinforce health information systems for reporting

7
and epidemic monitoring, preferably on a weekly basis
8 Conduct specific epidemic health education campaigns
Organize regular press releases/conferences/articles

9
for public information
10 IRS if area is previously sprayed
11 IRS in areas previously not sprayed
12 Space spraying
13 Insecticide-treated mosquito nets (ITN) and other materials

LEARNING UNIT 21
Post-epidemic assessment and
preparedness plan of action
■ ■ Undertake a quick assessment of the epidemic detection and
control response
■ ■ Develop a preparedness plan of action
149
21.1 The post-epidemic assessment

Exercise 21.1
a. Working in small groups, discuss a logical flow and the steps of an epidemic cycle; produce a diagram
of the cycle and present it for class discussion.
Consider what has been covered in Learning Units 18–20 and what should be included the post-
epidemic assessment exercise. All elements which are part of the classical epidemic described earlier
must be carefully analysed. Compare the diagram with Figure 21.1 provided by the tutor and
discuss any discrepancies.
b. During the post-epidemic assessment exercise, for which factors should it be possible to make a
careful analysis?
The essential elements to be assessed during the post-epidemic analysis are obtained using a
questionnaire; an example of the questionnaire is provided in Annex 11.
NMCP staff from the MoH with responsibity for developing and monitoring strategic
operations related to malaria epidemic prevention and control should lead the field exercise to
the area affected by epidemic with clear expected outcomes. They should be accompanied in
the field by the following experts:
▶▶ A national epidemiologist working in national health information system is expected to
manage the overall epidemic surveillance system
▶▶ A national meteorologist and or an expert dealing with meteorological, other relevant issues
and warning indicators
▶▶ A representative from the national disaster department
▶▶ A representative from the partners at national and/or district level
The team should preferably be from multiple sectors to provide a comprehensive overview of
problems encountered at national and district level. Occasionally, external technical assistance
can be requested by the MoH to WHO or other technical/funding agencies.
The post-epidemic assessment exercise is an essential step within the epidemic circle to identify
success and failure of all interventions (planned or unplanned), and ultimately to assess
whether the detecting systems and control options have had an impact on the malaria burden.
This important exercise is frequently neglected by implementing partners and the MoH.
Consequently, the lessons, both positive and negative, are not learnt and are not used to modify
or strengthen existing interventions for the future. Building on experience should improve the
preparedness plan of action, and encourage support provided by national and district partners.
The report should be widely distributed for partners’ information and input. The checklist
provided as Annex 11 can be used and adapted for this purpose.
21.2 Preparedness plan of action

In epidemic-prone countries, NMCP from the MoH and partners develop a national
document which defines strategic approaches to detect and control malaria epidemics, with
the ultimate aim of minimizing the malaria burden. This strategic document should be part
POST-EPIDEMIC ASSESSMENT AND PREPAREDNESS PLAN OF ACTION 151
of a comprehensive emergency plan covering all public health emergencies. The preparedness
plan of action (PPOA) should be complete, accurate and agreed among partners based on
understanding of the epidemiology, and best choice of preventive and control options for
malaria in epidemic-prone areas. The PPOA should detail all planned interventions at all levels
with all partners, and provide an estimated budget for these interventions.
Strategic elements of a preparedness plan of action should include all practical aspects which
relate to each of the interventions listed in Table 20.3. The “how” and “where” should be
explained in detail with attached additional budget if necessary. Ideally NMCP at central
level should provide epidemic-prone districts with technical guidance and should facilitate
collaboration with implementing agencies at district level, such as local authorities, media, civil
society, private companies and NGOs. All stakeholders should be consulted at the planning
stage and be involved in specific actions based on comparative advantages and mandates.
Capacity building as a key component of success has to be part of the planned activities and
should target all actors.
Exercise 21.2
Working in small groups, discuss what strategic elements need to be included in a preparedness
plan of action for prevention and control of a malaria epidemic in a logical sequence. Produce a
diagram of the steps and present it for class discussion. (Consider the discussion on indicators
for epidemics, assessment/investigation, reporting and response.)
Exercise 21.3
a. Should the PPOA for a malaria epidemic be developed in isolation or linked to other epidemics and
emergency situations?
b. Should it be developed at inter-country, national or regional/district level?
c. Working in small groups, discuss the reasons why most countries are not sufficiently prepared to cope
with epidemics.
List some key reasons and rank them according to their importance and potential for solution.
U21
153
AN1
ANNEX 1
Core surveillance indicators for malaria control
1. Confirmed malaria cases (number and rate per month or per year)
Formula 1000 * number of confirmed malaria cases
population at risk of malaria
Numerator The number of suspected malaria cases confirmed by either microscopy or RDT.
the number should include both outpatient and inpatient cases. The number should
include cases detected passively (attending health facilities or seen by community
health workers) or actively (sought in the community); it is often useful to provide
a breakdown of cases detected passively and actively. Regardless of transmission
setting, any person with a positive result in a parasite-based test (microscopy or RDT),
irrespective of clinical symptoms, should be considered to have a (confirmed) case of
malaria.
Denominator The number of people living in areas where malaria transmission occurs. Areas with
malaria transmission can be defined as those from which locally acquired, confirmed
cases of malaria are reported. Population sizes should be adjusted for population
growth according to projections from the national census or, when not available,
United Nations Population Division projections.
It is sometimes useful to use the total population of an area or country as the
denominator in order to compare overall levels of risk among geographical areas or
countries.
If rates are calculated per month, as when plotting rates on a graph, the population
size should be divided by 12 to obtain an annual incidence rate.
Breakdown High and moderate transmission: Age group (all ages, < 5 years), parasite species,
geographical area, time (year and month).
Low transmission: Sex, 5-year age groups, type of detection (passive, active,
community).
Elimination: foci, village, source of infection: imported, local (introduced, indigenous,
relapsing), induced.
Purpose To measure trends in malaria morbidity and to identify locations of ongoing
malaria transmission. this indicator is the most important measure of progress and
management in low-incidence areas.
Interpretation Trends can be affected by:
• completeness of reporting: trends can change if the percentage of health facilities
reporting in a month changes or if different sets of health facilities are included e.g.
including health posts or private hospitals.
• number of tests undertaken (slides examined, RDTs performed)
• changes in use of health facilities owing to greater availability of antimalarial drugs
or implementation of user fees
• actual changes in the incidence of malaria.
In order to conclude that changes in the value of the indicator reflect a real change in
the incidence of malaria in the community, indicators for completeness of reporting,
annual blood examination rate and total number of outpatient visits per 1000
population must be examined, to confirm that they are reasonably constant over time.
Other The indicator is also known as the annual parasite index.
2. Inpatient malaria cases (number and rate per month or per year)
Formula 10 000 * number of inpatient malaria cases
Numerator The number of inpatients with a primary diagnosis of malaria at discharge or death.
Patients who have absconded or been transferred should be excluded. All cases
should have had a parasite-based test for malaria (microscopy and/or RDT) and a
discharge diagnosis based on this test result.
Malaria inpatient numbers should include patients from both hospitals and other
facilities with beds.
malaria transmission can be defined as those in which locally acquired, confirmed
It is sometimes useful to use the total population of an area or country as the
denominator in order to compare overall levels of risk among geographical areas or
countries.
Breakdown Age group (all ages, < 5 years), geographical area, time (year and month).
Purpose To monitor the impact of programmes on severe disease. This indicator may reflect the
impact of treatment, as treatment attenuates clinical progression from uncomplicated
to severe disease.
Interpretation Inpatient cases are markers of severe disease and death and indicate failure of the
health system to either prevent or effectively treat malaria.
The numbers of inpatient cases are much larger than those of health facility deaths,
allowing trends to be more easily discerned.
This indicator is most useful in high- and moderate-transmission settings in which the
rates of severe morbidity are significant.
In some countries, inpatient cases may also include uncomplicated P. falciparum
cases (according to national guidelines) to ensure full treatment, recovery and parasite
clearance.
Trends can be affected by:
including private hospitals.
• changes in diagnostic practice e.g. introduction of more diagnostic testing may
reduce the number of inpatients in whom malaria is diagnosed.
• actual changes in the incidence of severe malaria.
In order to conclude that changes in the value of the indicator reflect a real change
in the incidence of severe malaria in the community, indicators must be examined
for completeness of reporting, changes in the percentage of cases that have had a
diagnostic test and total number of inpatient visits per 10 000, to confirm that they are
reasonably constant over time.
3. Inpatient malaria deaths (number and rate per month or per year)
Formula 100 000 * number of inpatient malaria deaths
Numerator Cases in which the underlying cause of death is malaria. All recorded malaria deaths
diagnosis based on the test result.
Data on malaria deaths from hospitals and other facilities with beds should be
included.
ANNEX 1 155
It is sometimes useful to use the total population of an area or country as a

AN1
denominator in order to compare the overall level of risk among geographical areas or
countries.
Purpose To monitor the impact of programmes on the number of malaria deaths.
Interpretation Inpatient deaths are markers of very severe disease and indicate failure of the health
system to either prevent or effectively treat malaria.
The numbers of deaths are generally small, and trends may be difficult to discern, but
clusters of deaths may occur in time and space.
This indicator is most useful in high-transmission settings in which malaria death rates
are high. Trends can be affected by:
including private hospitals.
• changes in diagnostic practice e.g. the introduction of more diagnostic testing may
reduce the number of malaria deaths diagnosed.
• actual changes in the number of malaria deaths.
malaria mortality in the community, indicators should be examined for completeness
of reporting, changes in the percentage of cases that had a diagnostic test, total
number of inpatient visits per 10 000 and number of deaths per 100 000, to confirm
that they are reasonably constant over time.
4. Malaria test positivity rate (RDT and/or blood slide)

Formula 1000 * number of confirmed malaria cases
number of patients receiving a parasitological test
Numerator Number of cases of suspected malaria confirmed by either microscopy or RDT. The
number should include both outpatient and inpatient cases. The number should
include cases detected passively (attending health facilities or seen by community
health workers) or actively (sought in the community); it is often useful to provide a
breakdown of cases detected passively and actively. Any person with a positive result
in a parasite-based test (microscopy or RDT), irrespective of clinical symptoms, should
be considered to have confirmed malaria case.
Denominator Total number of suspected malaria cases tested
Breakdown Type of diagnostic test (microscopy or RDT), parasite species, geographical area, time
(year and month), type of detection (passive, active, community)
Purpose To reflect trends in malaria morbidity and identify y areas with the most intense malaria
transmission. partially ‘corrects’ for incompleteness of reporting and RDT stock-outs
because the numerator is derived from the same source as the denominator.
Interpretation RDT and slide positivity rates can differ and should therefore be reported separately.
RDT reflects the presence of antigens and may remain positive after parasites have
been cleared by treatment.
Test positivity rates can change if parasitological diagnosis has been extended to
populations living in intense transmission areas that previously did not have access
to testing. Care should be taken, therefore, to take into account possible confounding
factors when interpreting trends.
Changes in test positivity rates do not reflect percentage changes in malaria cases or
incidence, as the number of malaria cases is part of the denominator.
5. Percentage of cases due to P. falciparum

Formula 100 * number of confirmed P. falciparum malaria cases
number of confirmed malaria cases
Numerator Number of P. falciparum cases confirmed by either microscopy or RDT. The number
should include both outpatient and inpatient cases. Mixed infections with P. falciparum
should be counted as P. falciparum.
The number can include cases detected passively (attending health facilities or seen
by community health workers) or actively (sought in the community); it is often useful
to provide a breakdown of cases detected passively and actively.
Denominator Total number of cases of malaria confirmed by either microscopy or RDT. The number
should include both outpatient and inpatient cases. The number should include cases
detected passively (attending health facilities or seen by community health workers) or
actively (sought in the community); it is often useful to provide a breakdown of cases
detected passively and actively.
Breakdown Type of diagnostic test (microscopy or RDT), geographical area, time (year and
month), type of detection (passive, active, community).
Purpose To reflect the proportion of cases due to P. falciparum and provide information on the
likelihood of observing severe cases
Interpretation Can provide information on the likelihood of observing severe cases and the extent to
which programmes should be adjusted to address P. vivax or other species.
Can provide information on the degree of malaria control, as, in areas where control
measures are scaled up, the proportion of cases due to P. falciparum may decrease;
P. vivax appears to be respond less quickly to control measures because it can tolerate
a wider range of environmental conditions and because the dormant liver stage
(hypnozoite) enables infections to persist in the absence of mosquito transmission.
P. ovale and P. malaiae may also become more frequent, but these are rare in most
settings.
As the ability to detect P. falciparum may vary by type of test (microscopy or RDT),
care should be taken to ensure that the proportion of cases due to
P. falciparum is not influenced by changes in the ratio of different types of test used,
i.e. the results of microscopy and RDT should be analyzed separately.
6. Percentage of inpatient cases with a discharge diagnosis of malaria

Formula 100 * number of inpatient cases with a discharge diagnosis of malaria
total number of inpatients
Numerator Number of inpatients with a primary diagnosis of malaria at discharge or death.
Patients who have absconded or been transferred should be excluded. All cases
discharge diagnosis based on this test result.
Malaria inpatient numbers should include patients from both hospitals and other
Denominator Total number of inpatients discharged or died (inpatients who have absconded or been
transferred should be excluded).
Total inpatient numbers should include patients from both hospitals and other facilities
with beds.
Purpose To monitor the impact of programmes on severe disease. Partially ‘corrects’ for
incompleteness of reporting because the numerator is derived from the same source
as the denominator.
Interpretation Inpatient cases are markers of severe disease and death and indicate failure of the
health system to either prevent or effectively treat malaria.
The number of inpatient cases is generally larger than the number of health facility
deaths, so that trends can be more easily discerned.
This indicator is most useful in high- and moderate-transmission settings in which the
rates of severe morbidity are significant.
In some countries, inpatient cases may also include uncomplicated P. falciparum
(according to national guidelines) in order to ensure full treatment, recovery and
parasite clearance.
Trends can be affected by:
• completeness of reporting: trends can change if different sets of health facilities
are included, e.g. private hospitals with different proportions of inpatients due to
malaria.
reduce the number of malaria inpatients diagnosed.
• changes in attendances for conditions other than malaria e.g. malaria inpatient
proportions may decrease if the number of skilled deliveries increases.
• actual changes in the incidence of severe malaria.
ANNEX 1 157
AN1
the incidence of malaria, indicators should be examined for completeness of reporting,
percentage of cases that had a diagnostic test and the numbers of inpatient cases of
malaria and other conditions, to confirm that they are reasonably constant over time.
Changes in the percentage of inpatients with a discharge diagnosis of malaria do not
reflect changes in the number of malaria inpatient cases or inpatient case incidence as
the number of malaria cases is part of the denominator.
7. Percentage of inpatient deaths due to malaria

Formula 100 * number of inpatient deaths due to malaria
total number of inpatient deaths
Numerator Number of inpatients with a primary diagnosis of malaria at death. All cases should
have had a parasite-based test for malaria (microscopy and/or RDT) and a diagnosis
based on this test result.
The numbers of malaria deaths should include patients from both hospitals and other
Denominator Total number of inpatient deaths.
The numbers of deaths should include patients from both hospitals and other facilities
with beds.
Purpose To monitor the impact of programmes on the number of malaria deaths. Partially
‘corrects’ for incompleteness of reporting because the numerator is derived from the
same source as the denominator.
Interpretation Inpatient deaths are markers of very severe disease and indicate failure of the health
system to either prevent or effectively treat malaria.
The numbers of deaths are generally small, and trends may be difficult to discern, but
clusters of deaths may occur in time and space.
This indicator is most useful in high-transmission settings, in which malaria death rates
are high. Trends can be affected by:
• completeness of reporting: trends can change if different sets of health facilities are
included, e.g. private hospitals with different proportions of deaths due to malaria.
reduce the number of cases of malaria diagnosed in inpatients.
• changes in attendance for conditions other than malaria e.g. malaria inpatient
proportions may decrease if the number of skilled deliveries increases.
• actual changes in malaria death rates.
malaria death rates, the indicators must be examined for completeness of reporting,
percentage of cases that had a diagnostic test and the numbers of deaths from
malaria and other conditions, to confirm that they are reasonably constant over time.
Changes in test positivity rates do not reflect percentage changes in the number of
malaria cases or incidence, as the number of malaria cases is part of the denominator.
8. Annual blood examination rate

Formula 100 * number of patients receiving a parasitological test
Numerator Total number of suspected malaria cases tested. this can include active and passive
case detection. Patients tested by both RDT and microscopy should be counted only
once.
Denominator Number of people living in areas where malaria transmission occurs. Areas with
cases of malaria are reported.
Population sizes should be adjusted for population growth according to projections
from the national census or, when not available, United Nations Population Division
projections.
If rates are calculated per month, as when plotting a graph, the population size should
be divided by 12 to obtain an annual rate.
Breakdown Type of diagnostic test (microscopy or RDT), geographical area, time (year and
month), type of detection (passive, active, community).
Purpose To reflect the extent of diagnostic testing in a population; aids interpretation of other
surveillance indicators.
Interpretation Higher annual blood examination rates generally reflect more complete malaria
surveillance.
Some past guidance suggests that the annual blood examination rate should be about
10% in order to provide reliable trends, but the empirical evidence for such a target is
not strong. In high- transmission settings, the rate is likely to greatly exceed 10% due
to passive case detection alone.
9. Percentage of suspected malaria cases that have had a diagnostic test

Formula 100 * number of patients receiving a parasitological test
number of suspected cases of malaria
Numerator Total number of suspected malaria cases tested. This should include those found by
passive case detection only; patients identified by active case detection should be
excluded. Patients tested by both RDT and microscopy should be counted only once.
Denominator Number of suspected malaria cases attending health facilities (i.e. passive case
detection).
Breakdown Type of diagnostic test (microscopy or RDT), geographical area, time (year and month)
Purpose WHO recommends that all suspected malaria cases should receive a diagnostic test
by microscopy or RDT, regardless of age. The indicator reflects the extent to which
malaria programmes are able to achieve this goal and where further effort may be
required.
Interpretation A value less than 100% may indicate problems in data recording, policy adoption,
financing, stock-outs of RDT at national or health facility level or clinician adherence.
10. Completeness of health facility reporting

Formula 100 * number of health facilities reports received
number of health facilities expected
Numerator Number of monthly reports received from health facilities. It may be necessary
to consider the number of health facility reports for different data elements e.g. if
outpatient cases are reported on different forms from inpatient cases.
If a health facility does not submit a report, e.g. because it is temporarily closed, a null
report showing zero cases and activities should be created and the reported marked
as received.
Denominator Number of health facility reports expected. Generally, this is the number of health
facilities expected to report multiplied by the number of months considered.
Breakdown Geographical area, time (year and month), report type (e.g. inpatient, outpatient)
Purpose Regular monitoring and follow-up can improve the completeness of reporting until all
health facilities are consistently reporting every month. Aids interpretation of other
surveillance indicators.
Interpretation The completeness of reporting of health facilities should be near 100%. Values
< 100% may indicate problems with supplies of stationary, communications, staff
availability, motivation or skills.
The indicator gives equal weight to all health facilities and therefore may not reflect
the completeness of case reporting; missing reports from district hospitals are likely to
account for a larger number of missing cases than missing reports from remote rural
health facilities.
Other If data reported from district to regional or national level are summarized by district
(rather than by health facility), the district summary form should contain two variables:
number of health facilities expected to report and number of health facilities that
reported.
If community workers report malaria information to health facilities every month, the
completeness of reporting by community workers should also be calculated. The
health facility reporting form should contain two additional data elements: number of
community workers expected to report and number that reported during the month.
159
ANNEX 2
Suggested register for community health
AN2
workers, health posts and outpatient
departments of health centres and hospitals
Residence Malaria Final
Age in Provisional New
No. Date Name (village, Sex Treatment
years diagnosis visit?
neighborhood) test result diagnosis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
(6) Age in years: Age should be recorded as < 1 or 0 for children < 1 year of age.
(7) Provisional diagnosis: may be amended in column 11 if the result of a malaria diagnostic test result is
negative.
(9) Malaria test result: The result should be recorded as +ve, –ve, or not done. If more than one species is
possible, the parasite species (P.f., P.v., P.m., P.o.) should be recorded for positive test results.
(11) Final diagnosis: Will include presumed malaria if no test was performed.
(12) Treatment: Specify if artemesinin-based combination therapy or other antimalarial treatment was given and
if patient referred. The number of suspected malaria cases can be derived from column 7.
The number of confirmed cases can be derived from column 9. The number of presumed malaria cases can
be derived by subtracting the number of confirmed malaria cases in column 9 from the number of malaria
diagnoses in column 10. Counts should apply only to new visits, which are indicated in column 8;sometimes,
columns for repeat visits are added to the right of column 11.

160
Patient attendance Total

Suspected malaria ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØOOO
OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
47
Microscopy
Patients tested by microscopy ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØOOO OOOOO
ANNEX 3
42
P. falciparum ØØØØØ ØØØØØ ØØØØØ ØOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
16
P. vivax ØØØØØ ØOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
6
P. malariae OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
P. ovale OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
Mixed OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
Positive tests (confirmed malaria) <5 ØØØØØ ØØØØØ ØØOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
12
Positive tests (confirmed malaria) ≥5 ØØØØØ ØØØØØ OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
10
health centres and hospitals
RDT testing
Patients tested with RDT ØØØØO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
4
Positive tests (confirmed malaria) <5 ØOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
1
Positive tests (confirmed malaria) ≥5 OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
Treatment
Sheet for tallying outpatient attendance at
Confirmed cases receiving antimalarial ØØØØØ ØØØØØ ØØØØØ ØØØØØ ØØØOO OOOOO OOOOO OOOOO OOOOO OOOOO
22
Presumed cases receiving antimalarial OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
(presumed cases = cases not tested) OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO OOOOO
161
ANNEX 4
Daily and weekly records of outpatient
attendance at health centres and hospitals
Month: April 2012 1 2 3 4 5 6 7 Weekly 8 9

Day: S M T W T F S total S M
Patient attendance
Suspected malaria 8 59 47 AN4

Microscopy
Patients examined 56 42
P. falciparum 18 16
P. vivax 3 6
P. malariae
P. ovale
Mixed
Positive tests (confirmed malaria) <5 8 12
Positive tests (confirmed malaria) ≥5 13 10

RDT testing
Patients tested with RDT 8 2 4
Positive tests (confirmed malaria) <5 2 0 1
Positive tests (confirmed malaria) ≥5 1 0 0

Treatment
Confirmed cases receiving antimalarial 2 0 1

Presumed cases receiving antimalarial
(presumed cases = cases not tested) 1 0 0
162
ANNEX 5
Discharge register for inpatient departments
of health centres and hospitals
Residence Length of Reason

No. Date Name (village, Sex Age YMD Diagnosis stay for
neighbourhood) (days) leaving
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

(7) YMD: units in which age is recorded: days should be used for children < 1 month, months for children < 1
year, years for others).
(8) Diagnosis: Should follow ICD classifications as far as possible; some facilities may add a column for the
ICD code.
(10) Reason for leaving: discharged, died, transferred or absconded
The total number of malaria inpatient cases should be the number discharged plus died, i.e. excluding
transferred and absconded, as a final diagnosis will not have been made.
163
ANNEX 6
Reports from health posts and community
health workers to health facilities
Patient attendance
Suspected malaria
Testing
Patients tested with RDT
Confirmed malaria <5 years
Confirmed malaria ≥5 years
Treatment
Confirmed malaria treated with antimalarial medicine

AN6
Cases not tested treated with antimalarial medicine
Cases referred

The number of variables to be reported each month should be kept to a minimum, to enhance the
completeness and quality of reporting. All health workers should understand the terms used, i.e. ‘confirmed
malaria’: suspected malaria cases with a positive test. Notes can be placed at the bottom of a form and in
standard treatment manuals as a reminder.
164
ANNEX 7
Reports from health facilities to
the district level
Areas with P. falciparum only

Outpatients
Suspected malaria
Total outpatients
Testing
Patients tested by microscopy
Confirmed malaria 5+ years
Confirmed malaria 5+ years
Inpatients
Malaria <5
Malaria 5+
Total discharges <5
Total discharges 5+
Deaths
Malaria <5
Malaria 5+
Total deaths <5
Total deaths 5+
Treatment

ANNEX 7 165
Areas with more than one species of Plasmodium

Outpatients
Suspected malaria
Total outpatients
Testing
Patients with microscopic slide examination
P. falciparum
P. vivax
P. malariae
P. ovale
Mixed
Total confirmed malaria <5 years
Total confirmed malaria 5+ years
Confirmed malaria ≥5 years
Inpatients
Malaria <5
AN7
Malaria ≥5
Total discharges <5
Total discharges ≥5
Deaths
Malaria <5
Malaria ≥5
Total deaths <5
Total deaths ≥5
Treatment

The number of variables to be reported each month should be kept to a minimum, to enhance the
completeness and quality of reporting. All health workers should understand the terms used, i.e. ‘confirmed
malaria’: suspected malaria cases with a positive test. Notes can be placed at the bottom of a form and in
standard treatment manuals as a reminder.
The tally sheet should be locally adapted. For example, if there is no P. vivax or P. ovale, those can be removed.
In settings where a multi-species RDT is used, the RDT section should be adapted to report those results.

166
ANNEX 8
Line lists of inpatient malaria cases and deaths to be reported
to district level in low-transmission seetings

(7) Type of test: RDT, microscopy or none.
(8) Species: If only P. falciparum is present, this column is not needed. If more than one species might be involved, the parasite species (P.f., P.v., P.m., P.o.) should be recorded for positive test results.
(9) ITN: insecticide-treated net.
(11) IRS: indoor residula spraying.
(16) Medicines used: Specific details to be provided to determine possibility of expired or counterfeit medicines.
Malaria prevention Antimalarial treatment Reason for
House Date Receieved leaving (dis-
Residence Type of ITN owned ITN used in 2 weeks received Date of contacted antimalarial charged/died/
Date (villlage, Pregnant? test (RDT/ by house- before admission (all IRS onset of health treatment Date Medicines absconded/
No. admitted Name surburb) Sex Age (Y/N)? micr.) Species hold (Y/N)? nights some/ none)? (Y/N)? symptoms system (Y/N) started used transferred)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17)
ANNEX 9
Line lists of all confirmed malaria cases to be reported
at district level in low-transmission settings

(6) Type of test: RDT, microscopy or none
(7) Species: If only P. falciparum is present, this column is not needed. If more than one species might be involved, the parasite species (P.f., P.v., P.m., P.o.) should be recorded for positive test results.
(8) ITN: insecticide-treated net.
(10) IRS: indoor residual spraying.
(15) ACT, artemisinin-based combined therapy; CQ, chlororquine.
Malaria prevention Antimalarial treatment
Receieved
Residence ITN owned ITN used in 2 weeks House Date of Date antimalarial Medicines
Date (villlage, Type of test (RDT/ by household before admission (all received IRS onset of contacted treatment Date (ACT, CQ,
No. admitted Name surburb) Sex Age micr.) Species (Y/N)? nights some/ none)? (Y/N)? symptoms health system (Y/N) started others)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)
167
AN9
168
ANNEX 10
Example of supervisory checklist for
malaria surveillance for countries with high
to moderate transmission
During visits to health facilities, supervisors should check that registers are kept up to date, with all fields completed;
data on report forms correspond to information in registers and tally sheets; core analysis graphs and tables are
up to date; and discussions are held about interpretation of the trends and potential action. Health facility staff
should be encouraged to investigate all malaria inpatient cases and death.
Present but not Present and up Present, up to date

Record keeping Not present
up to date to date and no mistakes
Outpatient register ✓
Discharge register ✓
Daily attendance summary book ✓

Monthly attendance summary
✓
book
Graph of suspected cases ✓

Graph of number of tests
✓
performed
Graph of number of confirmed
✓
cases
Graph of test positivity rate ✓
Reporting None 1 2 3
Number of monthly reports sent
✓
on time in last 3 months
Investigations performed in past Done &
Not done Done
3 months action taken
Malaria deaths ✓
Malaria inpatients ✓
Malaria cases ✓
Disease or progamme delivery issues that need attention
Large number of inpatient cases still from Lacienda village

Recommendation
Calculate test positivity rates as demonstrated.

Work with Laceienda village chief to encourage residents to use LLINs and attend health
centre promptly if ill with fever.
169
ANNEX 11
Example of questionnaire for malaria
post-epidemic assessment
The following questionnaire should provide the basic principless for post-epidemic assessment
and help to identify potential defects in the key components. It could also provide a framework
to assess the level of success in responding to the epidemic.
1. Have epidemic-prone areas for the country been demarcated?
If so, did the epidemic occur in a high-risk area?
2. Are early warning systems using, for example, real-time weather data made available
and shared and discussed by district management teams?
a. Did this data predict a possible epidemic in the region?
b. Was the regional malaria control station aware of the risk?
c. Was this information disseminated to all levels of malaria control?
e. Was there adequate time for planning for interventions if the predictions were confirmed?
3. Early detection system

a. Is there a well equipped surveillance system for early detection of epidemics, including
malaria, at district level in areas which are epidemic prone?
b. Were these data recorded, analysed in relation to defined thresholds at district level, providing
regular feed back to peripheral health-care facilities?
c. Were records of previous years available for comparison?
d. What method was used to analyse anomalies and define / validate thresholds (i.e. mean
+ 2 standard deviations, 3rd quartile, etc)?
e. Were these data regularly reported to a central facility?
If yes, what communication channels were used?
4. Recognition of anomalies and preliminary action taken at the periphery AN11

a. Are unexpected increase in cases detected at the periphery and action immediately taken?
b. If yes, what action was taken at the periphery first and then at district level?
c. How was the verification of the epidemic implemented?
d. How effective was the verification process? Fast enough (days)?
e. How was notification to district made and what was the lag time (days)? If more than
2 days, what caused the delay?
5. Was the response timely?

a. Was there effective communication between the local and district level?
b. What was the lag time between confirmation of the epidemic and response?
c. Were there adequate drugs and medical supplies at district level for rapid distribution?
d. Were insecticides and related supplies available (if relevant)?
e. Were there sufficient trained personnel to handle the epidemic?
f. Were there sufficient diagnostic facilities?
6. Disease and economic burden

a. How long did the epidemic last (weeks)?
b. What was the population affected?
c. How many cases were fatal?
d. What was the level of morbidity?
e. If private companies were affected, was some attempt made by the company to measure the
economic impact?
7. If the situation required mobilizing national emergency support:

a. What was the time lag for communication between district and national levels?
b. Who alerted the national level to stimulate a national response (district office, newspaper or
other media, other source)?
c. Was national support necessary? Was partners’ support necessary?
d. If other support was needed, was it effective in curbing the epidemic? [give some explanation]
8. Preparedness plan of action

a. Was there a budget allotted for malaria epidemic response?
b. Were partners involved in the development of the PPOA?
c. Were source reduction measures employed?
If yes, were they technically appropriate? Were they effective?
The questionnaire, such as the example above, should make clear what problems were faced
during the pre-epidemic and early epidemic periods when control options are expected to be
the most efficient. This knowledge would enable NMCP and partners to understand how to
strengthen or amend the existing epidemic preparedness plan.
For more information, please contact:
Global Malaria Programme

World Health Organization
20 avenue Appia ISBN 978 92 4 150601 4
1211 Geneva 27
Switzerland
E-mail: [email protected]
http//www.who.int/malaria

Epidemiological Approach For Malaria Control

Uploaded by

Copyright:

Available Formats

Epidemiological Approach For Malaria Control

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Epidemiological Approach For Malaria Control

Uploaded by

Copyright:

Available Formats

Second edition

for malaria control

GUIDE FOR PARTICIPANTS

The revision of this document was made possible through the

© World Health Organization 2013

Part 1: Introduction to basic epidemiology and statistics

Part 2: Applied malaria epidemiology

Part 3: Malaria surveillance, monitoring and evaluation

Part 4: Prevention and control of malaria epidemics

Annex 1. Core surveillance indicators for malaria control...........................................................................153

P. Ringwald, M. Robalo, A. Schapira, A. Teklehaimanot, M. Warsame, W. Were, R. Williams,

Development of the module

The intended users of this training module

Use of the Guide for Participants

Evaluation of the participants

Evaluation of the training by the participants

1.1 Types of epidemiological studies

Uses of descriptive studies

Types of descriptive studies

1.1.2 Analytic studies U1

▶▶ Cross-sectional studies (surveys)

▶▶ Case control studies

followed up or traced over a period of time.

DESCRIPTIVE ANALYTICAL / EXPERIMENTAL

Figure 1.1 Types of primary epidemiological studies

1.2.1 Secondary research

1.3 Potential errors in epidemiological studies

1.3.2 Systematic error (bias)

Information bias (also called measurement bias)

EXPOSURE (income) DISEASE (malaria)

Figure 1.2 Confounding – bednets, income and malaria

1.4 Validity and reliability

2.1 Raw data

2.5 Relative risk and risk difference

Table 2.1 Number of malaria cases, Province X, 2001–2005

YEAR CASES POPULATION

2001 30 858 492 810

a. Describe in words the trend in the number of cases.

Table 2.2 Age breakdown of malaria cases, Province X, 2006

AGE CASES % OF ALL CASES POPULATION AT RISK INCIDENCE RATE/100/year

3.1.1 Guidelines for developing tables

Rows and columns to be clearly labelled

YEAR All patients Malaria patients %

2001 136 289 16 946 12.4

Data collected from 24 district hospitals

Footnote to enhance clarity

equal Label x axis: method of classification

Figure 3.1 General graph with descriptive annotation

3.2.1 Guidelines for preparing graphs

3.2.2 Types of graphs

District A State Z District B

Geographic coordinate chart

No data 0 0 – 0.1 0.1–1 1–10 10–50 50–100 100

Figure 3.11 Malaria inpatients in Hospitals A and B for 12 months period.

▶▶ severe anaemia and acute respiratory distress

Source: Marsh et al., NEJM 332, 1399-1404, 1995

Figure 3.12 Case fatality

Population Cases in Deaths in

2001 58 520 10 241 205 175 3.5 2.0