Health Research Methodology PDF

Health research methodology: A guide for training in research methods HEALTH RESEARCH METHODOLOGY A Guide for Training in Research Methods Second Edition WORLD HEALTH ORGANIZATION Regional Office for the Western Pacific...

Health research methodology: A guide for training in research methods HEALTH RESEARCH METHODOLOGY A Guide for Training in Research Methods Second Edition WORLD HEALTH ORGANIZATION Regional Office for the Western Pacific Manila, 2001 i  CONTENTS Foreword v Introduction vii Acknowledgements ix Chapter 1: Introduction to research 1 Chapter 2: Research strategies and design 11 Chapter 3: Descriptive epidemiological studies and clinical trials 43 Chapter 4: Experimental studies and clinical trials 55 Chapter 5: Sampling methods and sampling size 71 Chapter 6: Bias and confounding 85 Chapter 7: Basic risk measurement 97 Chapter 8: Tests of significance 107 Chapter 9: Association and causation 125 Chapter 10: Ethical aspects of health research 141 Chapter 11: Construction of a research proposal 147 Annexes: Annex 1: Questionnaire design 169 Annex 2: Descriptive statistics: Table, graphs, and charts 187 Annex 3: Organization of a workshop on research 211 methods in health sciences Index 232  Health research methodology: A guide for training in research methods WHO Library Cataloguing in Publication Data Health Research Methodology: A guide for training in research methods. Second Edition. 1. Health services research - methods. 2. Research design. I. World Health Organization. Regional Office for the Western Pacific ISBN 92 9061 157 X The World Health Organization welcomes requests for permission to reproduce or translate its publications, in part or in full. Applications and enquiries should be addressed to the Office of Publications, World Health Organization, Geneva, Switzerland or to the Regional Office for the Western Pacific, Manila, Philippines, which will be glad to provide the latest information on any changes made to the text, plans for new editions, and reprints and translations already available. © World Health Organization 2001 Publications of the World Health Organization enjoy copyright protection in accordance with the provisions of Protocol 2 of the Universal Copyright Convention. All rights reserved. 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 Secretariat 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. 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. ii  Health research methodology: A guide for training in research methods FOREWORD Scientific research plays a very important role in our efforts to maintain health and combating diseases. Research helps us create new knowledge and develop proper tools for the use of existing knowledge. Not only does it enable health care providers to diagnose and treat diseases, research also provides evidence for policies and decisions on health and development. WHO and its Member States are aware of the importance of research. However, health research has not been a priority in many developing countries in the Region. The lack of research methodology and the absence of qualified researchers hinder many developing countries to conduct health research by themselves. In many countries, the system for management and coordination of health research has not been established or is not functioning properly. WHO is committed to stimulating scientific research in developing countries. An articulate and clearly defined WHO framework and vision on research and partnership with Member States will strengthen research capacity in developing countries. The WHO Regional Office for the Western Pacific has organized more than 20 training courses on health research design and methodology in the last two decades. In 1992, the Regional Office published a training manual entitled Health Research Methodology: A Guide for Training in Research Methods. Since then, the manual, well received by readers worldwide, has been translated into Chinese, Khmer, Laotian, Mongolian and Vietnamese. To accommodate requests from readers to incorporate recent developments on research methodology and experiences of past training courses, the manual has been revised and reissued. v  Health research methodology: A guide for training in research methods We hope this revised version of the landmark manual will help scientists, researchers, health practitioners and administrators to learn and practise the concepts and principles of scientific research. The knowledge of the scientific methods will help them design and conduct research projects with precision in their own countries. The publication of the revised manual also reiterates our commitment to developing countries in the Region to help them build and strengthen the health research systems. Shigeru Omi, MD, Ph.D. Regional Director WHO Western Pacific Regional Office vi  Health research methodology: A guide for training in research methods INTRODUCTION This is a revised version of an earlier manual on Health Research Methodology and deals with the basic concepts and principles of scientific research methods with particular attention to research in the health field. The research process is the cornerstone for informed and effective decision-making, and is integral to countries’ efforts to improve the health of their populations and the effectiveness of their health systems, particularly during times of dramatic epidemiological, demographic, and economic changes that profoundly affect health systems. Research on (1) health policy and health systems, (2) epidemiology dealing with noncommunicable diseases and existing, new and emerging communicable diseases, (3) reproduction, child health and nutrition, including domestic or sexual violence, and (4) social-behaviour, including analysis of peoples’ health seeking processes and their beliefs, knowledge and practices about health and illness, conducted by multi-disciplinary teams will enhance developing countries’ efforts to fight diseases and maintain health for the public. The manual describes methods for planning and conducting scientific research: from formulation of problems to setting research objectives, to designing the study, including methods of data collection, statistical analysis as well as interpretation and dissemination of the results. The earlier manual, used as resource and guide for the conduct of workshops on health research methodology in various countries of the Western Pacific Region has been expanded to include more details on some of the commonly used statistical methods and to clarify the points raised during workshops. The discussion on biases has been expanded considerably. vii  Health research methodology: A guide for training in research methods This manual is expected to be used by the WHO Western Pacific Regional Office as a reference guide in training young scientists to conduct health sciences research. It will be used as a starting point and not as a comprehensive textbook on research methods. Many excellent textbooks are available for this purpose and are referenced in the manual. We have tried to use real life examples from the Region for illustrating the principles and methods used in the manual to make it more relevant to the regional context. The manual will be useful in planning a research project, especially in preparing a research grant application for a donor agency. In particular, the attached copy of the application form of WHO serves as a guide. The issues discussed in the manual will help the researcher to focus on issues of importance before the study is proposed and undertaken. In addition, the manual would also be useful when writing a thesis to meet academic requirements of a degree in the health field. We hope that this manual will not only provide basic information on research methods in the health field, but also stimulate the reader to inquire further into the complex area of research methodology as well as increase the productivity of the young researcher in the Region. We hope it will attract researchers to conduct further studies in the health field, be it a clinical trial or field epidemiology or study of health services. viii  Health research methodology: A guide for training in research methods ACKNOWLEDGEMENTS The World Health Organization - Regional Office for the Western Pacific acknowledges the original contributions of Professor Yung-Han Parik, Professor Ung-Ring Ko and Dr Kamini Mohan Patwary to the first edition of this manual. We also wish to acknowledge Dr Rama Nair and Dr Lye Munn Sann for their collaborative efforts in revising this manual. ix  Health research methodology: A guide for training in research methods Chapter 1 Research and Scientific Methods 1.1 Definition Research is a quest for knowledge through diligent search or investigation or experimentation aimed at the discovery and interpretation of new knowledge. Scientific method is a systematic body of procedures and techniques applied in carrying out investigation or experimentation targeted at obtaining new knowledge. In the context of this manual, research and scientific methods may be considered a course of critical inquiry leading to the discovery of fact or information which increases our understanding of human health and disease. 1.2 Categories of research 1. Empirical and theoretical research The philosophical approach to research is basically of two types: empirical and theoretical. Health research mainly follows the empirical approach, i.e. it is based upon observation and experience more than upon theory and abstraction. Epidemiological research, for example, depends upon the systematic collection of observations on the health- related phenomena of interest in defined populations. Moreover, even in abstraction with mathematical models, advances in understanding of disease occurrence and causation cannot be made without a comparison of the theoretical constructs with that which we actually observe in populations. Empirical and theoretical research complement each other in developing an understanding of the phenomena, in predicting future events, and in the prevention of events harmful to the general welfare of the population of interest. 1  Chapter 1: Research and scientific methods Empirical research in the health sciences can be qualitative or quantitative in nature. Generally, health science research deals with information of a quantitative nature, and this manual deals exclusively with this type of research. For the most part, this involves the identification of the population of interest, the characteristics (variables) of the individuals (units) in the population, and the study of the variability of these characteristics among the individuals in the population. Thus the quantification in empirical research is achieved by three related numerical procedures: (a) measurement of variables; (b) estimation of population parameters (parameters of the probability distribution that captures the variability of observations in the population); and (c) statistical testing of hypotheses, or estimating the extent to which ‘chance’ alone may account for the variation among the individuals or groups under observation. Taking chance, or probability into account is absolutely critical to biological research, and is the substance of research design. Research design, above all else, must account for and maintain the role of chance in order to ensure validity. It is statistical methods which preserve the laws of probability in our inquiry, and allow proper analysis and interpretation of results. Statistics are the tool that permits health research to be empirical rather than abstract; they allow us to confirm our findings by further observation and experiment. 2. Basic and applied Research can be functionally divided into basic (or pure) research and applied research. Basic research is usually considered to involve a search for knowledge without a defined goal of utility or specific purpose. Applied research is problem-oriented, and is directed towards the solution of an existing problem. There is continuing controversy over the relative benefits and merits to society of basic and applied research. Some claim that science, which depends greatly on society for its support, should address itself directly to the solution of the relevant problems of man, while others argue that scientific inquiry is most productive when freely undertaken, and that the greatest advances in science have resulted from pure research. It is generally recognized that there needs to be a healthy balance between the two types of research, with the more affluent and technologically advanced societies able to support a greater proportion of basic research than those with fewer resources to spare. 2  Health research methodology: A guide for training in research methods 3. Health research triangle Yet another way of classifying health research, be it empirical or theoretical, basic or applied, is to describe it under three operational interlinked categories of biomedical, health services and behavioural research, the so-called health research triangle. Biomedical research deals primarily with basic research involving processes at the cellular level; health research deals with issues in the environment surrounding man, which promote changes at the cellular level; and behavioural research deals with the interaction of man and the environment in a manner reflecting the beliefs, attitudes and practices of the individual in society. 1.3 Scientific foundations of research Several fundamental principles are used in scientific inquiry: 1. Order The scientific method differs from ‘common sense’ in arriving at conclusions by employing an organized observation of entities or events which are classified or ordered on the basis of common properties and behaviours. It is this commonality of properties and behaviours that allows predictions, which, carried to the ultimate, become laws. 2. Inference and chance Reasoning, or inference is the force of advances in research. In terms of logic, it means that a statement or conclusion ought to be accepted because one or more other statements or premises (evidence) are true. Inferential suppositions, presumptions or theories may be so developed, through careful construction, as to pose testable hypothesis. The testing of hypothesis is the basic method of advancing knowledge in science. Two distinct approaches or arguments have evolved in the development of inferences: deductive and inductive. In deduction, the conclusion necessarily follows from the premises, as in syllogism (all A is B, all B is C, therefore all A is C) or in algebraic equations. Deduction can be distinguished by the fact that it moves from the general to the specific, and does not allow for the element of chance or uncertainty. Deductive inferences, therefore, are suited to theoretical research. 3  Chapter 1: Research and scientific methods Health research, being primarily empirical, depends almost entirely upon inductive reasoning. The conclusion does not necessarily follow from the premises or evidence (facts). We can say only that the conclusion is more likely to be valid if the premises are true, i.e. there is a possibility that the premises may be true but the conclusions false. Chance must, therefore, be fully accounted for. Further, inductive reasoning is distinguished by the fact that it moves from the specific to the general – it builds. 3. Evaluation of probability The critical requirement in the design of research, the one that ensures validity, is the evaluation of probability from beginning to end. The most salient elements of design, which are meant to ensure the integrity of probability and the prevention of bias, are: representative sampling, randomization in the selection of study groups, maintenance of comparison groups as controls, blinding of experiments and subjects, and the use of probability (statistical) methods in the analysis and interpretation of outcome. Probability is a measure of the uncertainty or variability of the characteristic among individuals in the population. If the entire population is observed, the calculation of the relative frequencies of the variables provides all the information about the variability. If only a sample of individuals in the population is observed, the inference from the sample to the population (specific to the general) will involve the identification of the probabilities of the events being observed, as well as the laws of probability that allow us to measure the amount of uncertainty in our inferences. These objectives can be achieved only by the proper design of research which incorporates the laws of probability. 4. Hypothesis Hypotheses are carefully constructed statements about a phenomenon in the population. The hypotheses may have been generated by deductive reasoning, or based on inductive reasoning from prior observations. One of the most useful tools of health research is the generation of hypotheses which, when tested, will lead to the identification of the most likely causes of disease or changes in the condition being observed. Although we cannot draw definite conclusions, or claim proof using the inductive method, we can come ever closer to the truth by knocking down existing hypotheses and replacing them with ones of greater plausibility. 4  Health research methodology: A guide for training in research methods In health research, hypotheses are often constructed and tested to identify causes of disease and to explain the distribution of disease in populations. Mill’s canons of inductive reasoning are frequently utilized in the forming of hypotheses which relate association and causation. Briefly stated, these methods include: (a) method of difference – when the frequency of a disease is markedly dissimilar under two circumstances, and a factor can be identified in one circumstance and not the other, this factor, or its absence, may be the cause of the disease (for example, the difference in frequency of lung cancer in smokers and non- smokers); (b) method of agreement – if a factor, or its absence is common to a number of different circumstances that are found to be associated with the presence of a disease, that factor, or its absence may be causally associated with the disease (e.g. the occurrence of hepatitis A is associated with patient contact, crowding and poor sanitation and hygiene, each conducive to the transmission of the hepatitis virus); (c) the method of concomitant variation, or the dose response effect – the increasing expression of endemic goitre with decreasing levels of iodine in the diet, the increasing frequency of leukaemia with increasing radiation exposure, the increase in prevalence of elephantiasis in areas of increasing filarial endemicity, are each examples of this concomitant variation; (d) the method of analogy – the distribution and frequency of a disease or effect may be similar enough to that of some other disease to suggest commonality in cause (e.g. hepatitis B virus infection and cancer of the liver). 1.4 Study design The epidemiological approach is based upon statistical principles in the structuring of research design. In this approach, research can be divided into that which is basically observational in type, and that which is experimental. Observational types of studies generally employ the method of sample surveys, where a sample of the population is observed for various characteristics. This may be by actual interviews of the 5  Chapter 1: Research and scientific methods subjects, by obtaining measurements of physical characteristics, or by simply extracting information from existing sources, such as disease registries, hospital or employment records. Surveys of the cross- sectional type (where the information on cause and effect is simultaneously gathered, and the time sequence cannot be determined) are considered to be hypothesis-generating studies, whereas surveys where the observations on cause and effect differ by way of a period of time (such as case-control studies and cohort studies) are considered to be analytical in nature, and inference of associations can be made. Testing of hypotheses is best done by experiment, where all the factors other than those under consideration can be controlled. However, in human diseases, this is not often possible, due to ethical and practical considerations. Therefore, it is often replaced by so- called ‘natural’ experiments, or by carefully designed observational studies (case-control studies, cohort studies) with enough information about the ‘extraneous’ factors to be able to adjust for these factors in drawing inferences. These analytical observational studies can be retrospective (case-control) or prospective (cohort and retrospective cohort studies). These methods compare groups of individuals for differences in exposure or differences in outcome. They differ from experiments in that there is no direct intervention by the investigator, and the investigator cannot control the extraneous factors for any of the individuals under observation. In either approach, statistical reasoning using the laws of probability guides the inferential process. Some basic assumptions are made about the population, its characteristics and the probability distribution, and the likelihood of the observations supporting or contradicting the stated hypothesis, is evaluated. Based on these calculated probabilities, the hypothesis is accepted or rejected (or the state of uncertainty is left unresolved, especially when the samples observed are too small for reliability). Specific study designs are discussed later in the manual. The process of moving from hypothesis generation to hypothesis testing is illustrated below. An observation, or series of observations triggers a hypothesis; a cross-sectional survey is undertaken to generate proper hypotheses; an observational study establishes associations and supports (or rejects) the hypothesis; and an experiment is conducted to test the hypothesis. 6  Health research methodology: A guide for training in research methods Case Series > Cross-sectional Surveys > Analytic Studies > Case-control Cohort Restrospective Cohort Experiment > Randomized Trials Quasi-Experiment 1.5 Planning and management of research 1. Research programme As a complex activity, research requires careful planning, management and administration in its development and implementation. Within the constraints of the present world climate of restricted research budgets, it is becoming increasingly necessary that health research be programmed research, with clearly defined and practicably achievable objectives. Some basic steps necessary in developing a research programme include: (a) defining the intended role and scope of the unit undertaking the research; (b) determining the capabilities and resources of the research unit, to include personnel, facilities, equipment, supplies, time and budget, and accessibility of research material; (c) selecting the research topic, considering factors such as magnitude of the problem and its impact; urgency of the need for a solution; 7  Chapter 1: Research and scientific methods relevance to the aims of the funding agency; amenability of the problem to investigation; feasibility of the approach; chances of success; expected impact of a successful outcome; spin-off in terms of training of staff and other research capability strengthening elements; (d) constructing research protocols which will serve as the guiding documents for the execution, monitoring and evaluation of the research; (e) setting up a well-defined administrative structure with lines of direction, supervision, consultation and collaboration based upon task-specific job descriptions; (f) formulating a schedule of targets for consolidation of results and preparation of these results for dissemination, including publication in the scientific literature. 2. Execution of research The mechanics of conducting research follow the simple steps of formulating the problem, planning the approach (research design) and executing activities within a strategic network leading to specific objectives which will give the solution to the problem. The following provides a framework for a research proposal into which the basic elements of a research study can be incorporated (these are discussed in more detail in Chapter 11): a. Conceptualizing the problem: identifying the problem (what is the problem?); prioritizing the problem (why is this an important problem?); rationale (can the problem be solved, and what are the benefits to society if the problem is solved?); 8 Health research methodology: A guide for training in research methods b. Background: literature review (what do we already know?); c. Formulating the objectives: framing the questions according to general and specific objectives; developing a testable hypothesis to achieve the objectives; d. Research methodology: defining the population, characteristics of interest and probability distributions; type of study (observational or analytical, surveys or experiments); method of data collection, management and analysis: ◊ sample selection; ◊ measuring instruments (reliability and validity of instruments); ◊ training of interviewers; ◊ quality control of measurements; ◊ computerization, checking and validating measurements; ◊ the issue of missing observations; ◊ statistical summarization of information; ◊ testing of hypothesis; ◊ ethical considerations; e. Workplan: personnel; timetable (who will do what, and when); project administration; f. Plans for dissemination: presentation to authorities to implement the results of the research (if applicable); 9  Chapter 1: Research and scientific methods publication in scientific journals and other works (including those of the agency which funded the project) for wide distribution of the research findings. A good proposal will also contain an executive summary giving an overview of the above topics in clear and simple language understandable by lay persons, and a list of references. 1.6 The research worker Among the important qualities associated with successful research are: a spirit of adventure in seeking new facts; perseverance and patience; integrity to oneself and to the value of the scientific method; ◊ an analytical mind able to participate in critical thinking; ◊ receptivity to criticism at the professional level; ◊ openness of mind, and the ability to see the significance of the unexpected observation; ◊ objectivity. 1.7 Conclusion Scientific inquiry is one of the most challenging enterprises of mankind, and the support that it receives is a measure of the strength, vitality and vision of a society. The approach and methods of research have slowly evolved to become ever more precise and efficient. The technology is at hand to explore the unknown. The success of this however, depends as ever on the individual and collective talents of the researchers bound by the tenets of science, such as those dealing with order, inference and chance, as accounted for and encompassed by solid research design and methodology. 10  Health research methodology: A guide for training in research methods Chapter 2 Research Strategies and Design 2.1 Introduction The selection of a research strategy is the core of a research design and is probably the single most important decision the investigator has to make. Therefore, the development of a research strategy is the main focus of this manual. Essential components of the research design and the scientific basis for these will be discussed in the following chapters. The research strategy must include the definition of the population of interest, the definition of variables (characteristics of the individuals in this population), their status and relationships to one another. In testing a hypothesis, for example, an investigator may be able to assign the independence or exposure variable to a number of subjects in the study, and withhold it from others (controls) while controlling for other extraneous or confounding variables. This strategy constitutes an experiment and covers hypothesis testing through intervention. Another investigator may choose to compare people with and without exposure to a factor, and to analyse the incidence of a disease in these groups to find out if the disease is related to the exposure. This constitutes an analytical study, of which there are many varieties; this type of study also incorporates the testing of hypotheses. Still another investigator may simply describe the distribution of a phenomenon or the outcome of a programme. This constitutes a descriptive study with no intervention and no prior hypothesis. 11  Chapter 2: Research strategies and design In all of the above situations, observations are made on a group of people, and inferences are made about relationships or associations of various ‘exposures’ to ‘outcomes’. The inferences reached are always subject to uncertainty due to the variation of characteristics across the population. The accuracy of the inference depends, therefore, on the accuracy of the information collected and on the representativeness of the subjects observed to the larger group of subjects in the population, as well as on the accuracy of the statistical methods used to draw the inference. In order to develop a good research strategy, we need to understand the nature of these ‘errors’ or ‘variations’ and the methods available to measure the errors. 2.2 Errors in the inference Two common sources of error that need to be controlled result from problems with ‘reliability’ and ‘validity’. Our inference should have high reliability (if the observations are repeated under similar conditions, the inferences should be similar) and high validity (the inference should be a reflection of the true nature of the relationship). The reliability and validity of inferences depend on the reliability and validity of the measurements (are we measuring the right thing, and with accuracy?) as well as the reliability and validity of the samples chosen (have we got a true representation of the population that we are drawing inferences from?). The reliability of a sample is achieved by selecting a large sample, and the validity is achieved by ensuring the sample selection is unbiased. In statistical terms, reliability is measured using ‘random error’ and validity by ‘bias’. 2.2.1 Reliability Reliability of measurements If repeated measurements of a characteristic in the same individual under identical conditions produce similar results, we would say that the measurement is reliable. If independent, repeated observations are taken and the probability distribution is identified, the standard deviation of the observations provide a measure of reliability. If the measurement has high reliability, the standard deviation should be smaller. One way to increase the reliability is to take the average of a number of observations (the average having a smaller standard deviation – known as standard error of the mean [sem] – than the standard deviation of the individual observations). 12  Health research methodology: A guide for training in research methods Reliability of study A result is said to be reliable if the same result is obtained when the study is repeated under the same conditions. The natural variability in observations among individuals in the population is commonly known as random error. For example, if one is measuring the systolic blood pressure (SBP) of individuals, it has been observed that the measurements in large groups of people would follow a ‘normal’ distribution, so that the standard deviation of SBP is used as a measure of random error in SBP measurements. Clearly, if the standard deviation is small, repeated studies from this population are bound to come up with similar results. If the standard deviation is large, different samples from the same population will tend to differ substantially. Since we are often dealing with summary measures from samples that have standard deviations inversely proportional to the square root of the sample size, increasing the sample size increases the reliability of these measures. (More on these issues in Chapter 5.) 2.2.2 Validity A measurement is said to be valid if it measures what it is supposed to. If a measurement is not valid, we say it is ‘biased’. Bias is a systematic error (as opposed to a random error) that skews the observation to one side of the truth. Thus, if we use a scale that is not calibrated to zero, the weights we obtain using this scale will be biased. Similarly, if a sample is biased (for example, more males in the sample than the proportion of males in the population, or selecting cases from a hospital and controls from the general community in a case-control study), the results tend to be biased. Since it is often difficult to correct for biases once the data have been collected, it is always advisable to avoid bias when designing a study. (More details on biases and how to avoid them in Chapter 6.) 2.3 Experimental versus observational strategies Although an experiment is an important step in establishing causality, it is often neither feasible nor ethical to subject human beings to risk factors in etiological studies. Instead, epidemiologists make use of ‘natural experiments’, when available, or they resort (more frequently) to analytical observational studies or quasi-experiments. However, there is one area of epidemiology in which experimental 13  Chapter 2: Research strategies and design strategies are used extensively: this is the area of clinical and field trials for testing new drugs or intervention programmes. Advantages of the experimental approach include the following: The ability to manipulate or assign independent variables. This is by far the most distinct advantage of experimental strategies. It is readily illustrated by clinical trials, described in Chapter 4, in which cases of a specific disease are deliberately assigned (in random order, or by matching) to treatment and to control groups. For example, in an evaluation of the efficiency of intrauterine devices, women of a certain age and with certain other characteristics may be assigned at random or in matched pairs to physicians and nurses. A criterion for evaluation, such as the frequency of complications, is compared in the two groups. It may also be possible to manipulate the degree of exposure or the dose of the treatment. The ability to randomize subjects to experimental and control groups. Randomization makes it more likely that the distribution of some extraneous variables will be equalized between the two groups, although it is still necessary in the analysis to compare the distribution of these variables to ensure the validity of inferences drawn from the study. It is also possible in experiments (and also in some observational studies) to use matching in conjunction with randomization. In addition, randomization provides a basis for the calculation of appropriate probabilities of error in the inference. The ability to control confounding and eliminate sources of spurious association. Most of the other factors that interfere with the association under study can more easily be controlled in experiments (especially in animals) than in observational studies. The ability to ensure temporality. Determining which variables precede and which are the consequences of the intervention is more feasible in experimental studies than it is in some analytical studies, particularly those of the case-control and cohort designs. The ability to replicate findings. Experiments are often more replicable than observational studies. Replication satisfies the consistency requirement in causation. In practice, however, few clinical trials are exactly replicated. 14 Health research methodology: A guide for training in research methods All in all, the evidence for causal relationship is more compelling if it comes from a carefully executed experimental study, because selection factors that inadvertently bias observational studies can virtually be eliminated by the process of randomization. However, other sources of bias are not automatically controlled by randomization. The limitations of the experimental approach are sometimes overlooked, as the impressive advantages of experiments have led some people to reject evidence for causation if it is not based on experiments. If we were limited to the experimental approach, however, we would have to abandon most of the evidence upon which significant advances in public health have been made. Experiments also have the following limitations: Lack of reality. In most human situations, it is impossible to randomize all risk factors except those under examination. Observational methods deal with more realistic situations. Difficulties in extrapolation. Results of experiments in animal models, which are rigorously controlled, cannot readily be extrapolated to human populations. Ethical problems. In human experimentation, people are either deliberately exposed to risk factors (in etiological studies) or treatment is deliberately withheld from cases (intervention trials). It is equally unethical to test the efficiency or side-effects of new treatments without critical evaluation in a small group of human subjects. (See also Chapter 10.) Difficulties in manipulating the independent variable. It is virtually impossible, for instance, to assign smoking habits at random to the experimental and control groups. Non-representativeness of samples. Many experiments are carried out on captive populations or volunteers, who are not necessarily representative of the population at large. Experiments in hospitals (where the experimental approach is most feasible and is frequently used) suffer from several sources of selection bias. 15  Chapter 2: Research strategies and design 2.4 Descriptive studies Definition When an epidemiological study is not structured formally as an analytical or experimental study, i.e. when it is not aimed specifically to test a hypothesis, it is called a descriptive study, and belongs to the observational category of studies. The wealth of material obtained in most descriptive studies allows the generation of hypotheses, which can then be tested by analytical or experimental designs. A survey, for example a prevalence survey, could also be defined as a descriptive study, as it covers the elements of descriptive study. Conduct of descriptive studies Descriptive studies entail the collection, analysis and interpretation of data. Both qualitative and quantitative techniques may be used, including questionnaires, interviews, observations of participants, and service statistics, as well as documents describing communities, groups, situations, programmes and other individual or ecological units. The distinctive feature of this approach is that its primary concern is with description rather than with the testing of hypotheses or proving causality. The descriptive approach may, nevertheless, be integrated with or supplement methods that address these issues, and may add considerably to the information base. Kinds of descriptive studies Case series This kind of study is based on reports of a series of cases of a specific condition, or a series of treated cases, with no specifically allocated control group. These represent the numerator of disease occurrence, and should not be used to estimate risks. In an attempt to make such series more impressive, clinicians may calculate proportional distribution, which consists simply of percentages of the total number of cases that belong to a specific category of age, sex, ethnic group or other characteristic. These numbers are not rates, because the denominator still represents the cases and not the population at risk. 16 Health research methodology: A guide for training in research methods Community diagnosis or needs assessment This kind of study entails collection of data on existing health problems, programmes, achievements, constraints, social stratification, leadership patterns, focal points of resistance or high prevalence, or groups at highest risk. Its purpose is to identify existing needs and to provide baseline data for the design of further studies or action. Epidemiological description of disease occurrence This common use of the descriptive approach entails the collection of data on the occurrence and distribution of disease in populations according to specific characteristics of individuals (e.g. age, sex, education, smoking habits, religion, occupation, social class, marital status, health status, personality), place (rural/urban, local, subnational, national, international) and time (epidemic, seasonal, cyclic, secular). A description may also be given of familial characteristics such as birth order, parity, family size, maternal age, birth interval or family type. Descriptive cross-sectional studies or community (population) surveys Cross-sectional studies entail the collection of data on, as the term implies, a cross-section of the population, which may comprise the whole population or a proportion (sample) of it. Many cross-sectional studies do not aim at testing a hypothesis about an association, and are thus descriptive. They provide a prevalence rate at a particular point in time (point prevalence) or over a period of time (period prevalence). The study population at risk is the denominator for these prevalence rates. Included in this type of descriptive study are surveys in which the distribution of a disease, disability, pathological condition, immunological condition, nutritional status, fitness, or intelligence, etc., is assessed. This design may also be used in health systems research to describe ‘prevalence’ by certain characteristics – pattern of health service utilization and compliance – or in opinion surveys. A common procedure used in family planning and in other services is the KAP survey (survey of knowledge, attitudes and practice). 17  Chapter 2: Research strategies and design Ecological descriptive studies When the unit of observation is an aggregate (e.g. family, clan or school) or an ecological unit (a village, town or country) the study becomes an ecological descriptive study. As mentioned earlier, hypothesis testing is not generally an objective of the descriptive study. However, in some of the above studies (cross-sectional surveys, ecological studies) some hypothesis testing may be appropriate. Moreover, description of the data is also an integral part of the analytical study. 2.5 Analytical strategies in epidemiology Observational studies, where establishing a relationship (association) between a ‘risk factor’ (etiological agent) and an outcome (disease) is the primary goal, are termed analytical. In this type of study, hypothesis testing is the primary tool of inference. The basic approach in analytical studies is to develop a specific, testable hypothesis, and to design the study to control any extraneous variables that could potentially confound the observed relationship between the studied factor and the disease. The approach varies according to the specific strategy used. 2.5.1 Case-control studies The simplest and most commonly used analytical strategy in epidemiology involves the case-control study. It is designed primarily to establish the causes of diseases by investigating associations between exposure to a risk factor and the occurrence of disease. The design is relatively simple, except that it is backward-looking (retrospective) based on the exposure histories of cases and controls. With this type of study, one investigates an association by contrasting the exposure of a series of cases of the specified disease with the exposure pattern of carefully selected control groups free from that particular disease (Figure 2.1). Data are analysed to determine whether exposure was different for cases and for controls. The risk factor is something that happened or began in the past, presumably before disease onset, e.g. smoking, or a previous infection or medication. Information about the exposure is obtained by taking a history and/or from records. Occasionally, the suspected factor or attribute is a permanent one, such as blood group, which can be ascertained by clinical or laboratory investigation. A higher frequency of the attribute or risk factor among 18  Health research methodology: A guide for training in research methods FIGURE 2.1 DESIGN OF A CASE-CONTROL STUDY Expose (with characteristic or risk factor) Cases (those with condition) Unexposed (without characteristic or risk factor) Expose (with characteristic or risk factor) Controls (those without condition) Unexposed (without characteristic or risk factor) Example Chewers of tobacco Cases of oral cancer Non-chewers of tobacco Chewers of tobacco Those free of oral cancer Non-chewers of tobacco 19  Chapter 2: Research strategies and design cases than among controls is indicative of its association with the disease/condition – an association that may be of etiological significance. In other words, if a greater proportion of cases than controls give a history of exposure, or have records or indications of exposure in the past, the factor or attribute can be suspected of being a causative factor. Selection of cases What constitutes a case in the study should be clearly defined with regard to the histological type and other specific characteristics of the disease, such as date of diagnosis, geographical location, etc. Cases that do not fit these criteria should be excluded from the study. This design is particularly efficient for rare diseases, because all cases that fit the study criteria in a particular setting within a specific period are usually included. This allows for a reasonable number of cases to be included in the study without waiting for the occurrence of new cases of the disease, which might take a long time. For reasons of convenience and completeness of case records, the cases identified for case-control studies are often those from a hospital setting, from physicians’ private practices, or from disease registries. Newly diagnosed cases within a specific period (incident cases) are preferred to prevalent cases, since such a choice may eliminate the possibility that long-term survivors of a disease were exposed to the investigated risk factor after the onset of the disease. The selection of cases should be such that the study results are reliable and valid. For these reasons, the following guidelines should be used when selecting cases in a case-control study: a. The criteria for inclusion in the study (what constitutes a case) and criteria for exclusion from the study must be clearly specified; this will improve the validity of the results; b. The sources of cases may be: - all cases admitted to or discharged from a hospital, clinic, or private practice within a specified period; - all cases reported or diagnosed during a survey or surveillance programme within a specified period; - incident or newly diagnosed cases; - incident cases in an ongoing cohort study or in an occupational cohort (sometimes called a nested case- control study); 20 Health research methodology: A guide for training in research methods - deaths with a record of causes of death, and fulfilling other criteria for the study; - case units with a prescribed health outcome; c. If the number of cases is too large, a probability sample may be used; d. Cases selected for the study should be representative of all cases of the disease under consideration. Selection of controls It is crucial to set up one or more control groups of people who do not have the specified disease or condition in order to obtain estimates of the frequency of the attribute or risk factor for comparison with its frequency among cases. This is the most important aspect of the case-control study, as biases in the selection of controls may invalidate the study results, and bias in the selection of controls is often the greatest cause for concern when analysing data from case-control studies. a. The sources of comparison groups may be: a probability sample of a defined population, if the cases are drawn from that defined population; a sample of patients admitted to, or attending the same institution as the cases; a sample of relatives or associates of the cases (neighbourhood controls); a group of persons selected from the same source population as the cases, and matched with the cases for potentially confounding variables; on other risk factors (other than the one under consideration); b. The selection of controls may involve matching on other risk factors: Matching means that controls are selected such that cases and controls have the same (or very similar) characteristics other than the disease and the risk factor being investigated. The characteristics are those that would confound the effect of the putative risk factor, 21  Chapter 2: Research strategies and design i.e. these characteristics are known to have an association with the disease, and may be associated with the risk factor being studied. The purpose of the matching is to ensure comparability of these characteristics for the two groups, so that any observed association between the putative risk factor and the disease is not affected by differential distribution of these other characteristics. It is common to match for age, sex, race and socioeconomic status in case-control studies on diseases, as we know all of these factors affect the incidence of most of the diseases. Matching may be done on an individual basis (one-to-one matching) or on a group basis (frequency matching). Individual matching is preferable, because of the ease of analysis accounting for matching. The disadvantages of matching include a loss of precision and overmatching. Also, once a matched design is used, the matching variable is eliminated from consideration, and therefore it cannot be investigated for etiological association with the disease. For example, if we matched for marital status in a study of breast cancer, we would not know whether single or married women had different risks for breast cancer. Many epidemiologists prefer to conduct studies without matching, and use statistical methods to adjust for possible confounding during analysis, because of the increased precision and the ability to investigate any possible interaction effects. The use of unmatched controls, obtained through random sampling, allows greater flexibility in studying various interactions. What is most important is that information on potential confounding factors should be collected in the study, so that these can be adjusted in the analysis. c. The number of control groups may vary. It is sometimes desirable to have more than one control group, representing a variety of disease conditions other than that under study and/or non-hospitalized groups. Use of multiple controls confers three advantages: If the frequency of the attribute or risk factor does not differ from one control group to another, but is consistently lower than that among the cases, this increases the internal consistency of the association; 22 Health research methodology: A guide for training in research methods If a control group is taken of patients with another disease, which is independently associated with the risk factor, the difference in the frequency of the factor between cases and controls may well be masked. In such a case, the use of another control group will save the research project; Multiple controls provide a check on bias. The impact of poorly chosen controls on the conclusions of a case-control study is commonly exemplified by Pearl’s study in 1929. Pearl compared 816 malignancies identified among 7500 autopsied cases at the Johns Hopkins Hospital in Baltimore, Maryland, USA with 816 non-malignant autopsied cases matched at death for age, sex, race and date of death. Lesions of active tuberculosis were found in 6.6% of cases and in 16.3% of controls, which led to the conclusion that there was antagonism between tuberculosis and cancer. This finding could not be corroborated in animal experiments. One explanation for Pearl’s findings is that his control group inadvertently included many individuals who had died of tuberculosis, because tuberculosis patients were more frequently autopsied than were patients with other causes of death, and were thus unrepresentative of the general population of deaths. Collection of data on exposure and other factors Often data are collected through interviews, questionnaires and/ or examination of records. Occasionally, clinical and laboratory examinations are carried out, but often this is not possible, especially if the ‘cases’ include past cases which may also include some deaths. The following precautions should be taken when deciding on the data- collection strategy: observation should be objective, or, if obtained by survey methods, well standardized; the investigator or interviewer should not know whether a subject is in the case or control group (blinding); the same procedures, e.g. interview and setting, should be used for all groups. 23  Chapter 2: Research strategies and design Multifactorial case-control studies The common form of case-control study addresses one main factor or attribute at a time. It is possible, however, to investigate several exposure factors in the same study. For example, in a study in three states in the USA with a population of 13 million, all mothers of leukaemic children of 1-4 years old (diagnosed in 1959-67) were interviewed. As controls, a sample of 13 000 other women was taken. Four factors were considered, two preconceptional (preconceptional radiation and previous reproductive wastage) and two post-conceptional (in utero irradiation and viral infection during pregnancy). Analysis showed that each factor was related to leukaemia in their children (Gibson et al.,1968). Further analysis was conducted for combinations of factors, where the estimated relative risk in the absence of any of the four factors was made equal to 1.0, as shown in Table 2.1. It is apparent that the effect was the greatest among women with all four factors, and that there is synergism between the factors. Advantages of case-control studies The following are examples of the advantages of case-control studies: feasible when the disease being studied occurs only rarely, e.g. cancer of a specific organ; relatively efficient, requiring a smaller sample than a cohort study; little problem with attrition, as when follow-up requires periodic investigations and some subjects refuse to continue to cooperate; sometimes they are the earliest practical observational strategy for determining an association (e.g. use of diethylstilbesterol and clear-cell adenocarcinoma of the vagina in daughters). Enhancement of the validity of case-control studies Ways in which one can increase the validity of a study include ensuring that: the cases are representative of all cases in a particular setting; 24 Health research methodology: A guide for training in research methods the controls are similar to cases with respect to risk factors other than the study factor; multiple controls are used with consistent results; cases and controls are truly selected independently of exposure status; the sources of bias are mitigated, or at least shown not to have affected the results. (A common example is the British study of smoking and lung cancer by Doll and Hill (1952). After the cases and controls had been interviewed, it was discovered that some of the cases had been wrongly diagnosed as cancer. Reanalysis showed the persistence of the association and indicated that, in the study, the fact of being told that they had lung cancer did not bias the respondents with regard to the history they gave of smoking); repeated studies in different settings and by different investigators confirm each other (for example, the association between smoking and lung cancer has been reported by over 25 investigators from ten countries); it is possible to demonstrate a dose-response or gradient relationship (for example, several case-control studies showed that the number of cigarettes smoked per day was related to the risk of lung cancer); a hybrid design of case-control study nested in a cohort study with a defined population is used; this is a most powerful strategy. Disadvantages and biases of case-control studies The following are some of the problems associated with case- control studies: the absence of epidemiological denominators (population at risk) makes the calculation of incidence rates, and hence of attributable risks, impossible; temporality is a serious problem in many case-control studies where it is not possible to determine whether the attribute led to the disease/condition, or vice versa; 25  Chapter 2: Research strategies and design there is a great risk of bias in the selection of cases and controls. This is particularly serious when a single control group is related to the risk factor under investigation; it may be very difficult or impossible to obtain information on exposure if the recall period is long; selective survival, which operates in case-control studies, may bias the comparison; there is no way of ascertaining whether the exposure was the same for those who died and those who survived; because most case-control studies are performed in hospitals, they are liable to Berkson’s fallacy, or the effect of differing admission policies and rates; measurement bias may exist, including selective recall and misclassification (putting cases in the control group, or vice versa); there is also the possibility of the Hawthorne effect: with repeated interviews, respondents may be influenced by being under study; case-control studies are incapable of disclosing other conditions related to the risk factor: for example, in a study of the side-effects of oral contraceptives, one has to know their side-effects before a case-control design can be set up. 2.5.2 Prospective cohort studies The common strategy of cohort studies is to start with a reference population (or a representative sample thereof), some of whom have certain characteristics or attributes relevant to the study (exposed group), with others who do not have those characteristics (unexposed group). Both groups should, at the outset of the study, be free from the condition or conditions under consideration. Both groups are then observed over a specified period to find out the risk each group has of developing the condition(s) of interest. This is illustrated diagrammatically in Figure 2.2. 26  Health research methodology: A guide for training in research methods FIGURE 2.2 DESIGN OF A COHORT (PROSPECTIVE) STUDY Develop disease With characteristic Do not develop disease Reference population Sample Develop disease Without characteristic Do not develop disease Example Develop oral cancer Chew tobacco Do not develop oral cancer Population (free of oral cancer Sample Develop oral disease Do not chew tobacco Do not develop oral cancer 27  Chapter 2: Research strategies and design Design features a. Selection of cohort: a community cohort of specific age and sex; an exposure cohort, e.g. radiologists, smokers, users of oral contraceptives; a birth cohort, e.g. school entrants; an occupational cohort, e.g. miners, military personnel; a marriage cohort; a diagnosed or treated cohort, e.g. cases treated with radiotherapy, surgery, hormonal treatment. The usual procedure is to locate or identify the cohort, which may be a total population in an area or sample thereof. b. Data to be collected: data on the exposure of interest to the study hypotheses; data on the outcome of interest to the study hypotheses; characteristics of the cohort that might confound the association under study. c. Methods of data collection Several methods are used to obtain the above data, which should be on a longitudinal basis. These methods include: interview surveys with follow-up procedures; medical records monitored over time; medical examinations and laboratory testing; record linkage of sets with exposure data and sets with outcome data, e.g. work history data in underground mines with mortality data from national mortality files. In a conventional cohort study, an initial cross-sectional study is often performed to exclude persons with the outcome of interest (disease) and to identify the cohort that is free from the disease. 28 Health research methodology: A guide for training in research methods Measures of frequency Two methods are commonly used in cohort studies to measure the incidence of the disease (condition) under investigation: a. Cumulative incidence This index of disease frequency is based on the total population at risk which was, at entry to the study, free of the disease under investigation. The incidence of the disease is calculated for each stratum of exposure to the risk factor, and is the ratio of the number of new cases or events in a specified period of observation, to the total population at risk during that period. This incidence measure provides an estimate of the probability or risk of developing disease among all members of the group who were included in the study at its initiation, and were at risk of disease. Because cumulating all new cases in the total population at risk derives the measure, the term ‘cumulative incidence’ has been applied. Cumulative incidence is a proportion, not a rate, and can vary from 0 to 1, that is, no less than 0% and no more than 100% of the population at risk can acquire the disease. This measure of disease frequency is calculated as if all units or individuals had the same period of observation, but new cases are no longer at risk once they develop the disease. b. Incidence density (person-time approach) This approach is an improvement over the conventional measure of incidence, because it takes into consideration both the number observed and the duration of observation for each individual. Thus, if 30 individuals were observed as follows: 10 for two years, 5 for three years, and 15 for four years, they would contribute (10x2)+(5x3)+(15x4) = 95 person-years of observation, which would become the denominator. The numerator is the number of new cases observed in these groups over the specified period of time. This gives an incidence rate per person-year, called the incidence density. Person-years do not represent the number of persons: 400 person-years of observation could represent 400 persons each observed for one year, or 40 persons each observed for 10 years. Two drawbacks of this measure are that the exact time when the disease occurs often cannot be ascertained, and that the rate of disease development over time is not necessarily constant. 29  Chapter 2: Research strategies and design The basic measures of effect used in cohort studies are the relative risk (RR), attributable risk (AR), population attributable risk (PAR), population attributable risk percent (ARP%), and etiologic fraction (EF). These measures will be discussed in detail in Chapter 7. Advantages of cohort studies The following are some of the advantages of a cohort study compared with a case-control study: Because of the presence of a defined population at risk, cohort studies allow the possibility of measuring directly the relative risk of developing the condition for those who have the characteristic, compared to those who do not, on the basis of incidence measures calculated for each of the groups separately. In a cohort study, it is known that the characteristic precedes the development of the disease, since all the subjects are free of disease at the beginning of the study; this allows for a conclusion of cause-effect relationship (a necessary, but not sufficient, condition). Because the presence or absence of the risk factor is recorded before the disease occurs, there is no chance of bias being introduced due to awareness of being sick as in encountered in case-control studies. There is also less chance of encountering the problem of selective survival or selective recall, although selection bias can still occur because some subjects who contracted the disease will have been eliminated from consideration at the start of the study. Cohort studies are capable of identifying other diseases that may be related to the same risk factor. Unlike case-control studies, cohort studies provide the possibility of estimating attributable risks, thus indicating the absolute magnitude of disease attributable to the risk factor. If a probability sample is taken from the reference population, it is possible to generalize from the sample to the reference population with a known degree of precision. 30 Health research methodology: A guide for training in research methods Disadvantages of cohort studies The following are some of the disadvantages of cohort studies: These studies are long-term and are thus not always feasible; they are relatively inefficient for studying rare conditions. They are very costly in time, personnel, space and patient follow-up. Sample sizes required for cohort studies are extremely large, especially for infrequent conditions; it is usually difficult to find and manage samples of this size. The most serious problem is that of attrition, or loss of people from the sample or control during the course of the study as a result of migration or refusal to continue to participate in the study. Such attrition can affect the validity of the conclusion, if it renders the samples less representative, or if the people who become unavailable are different from those actually followed up. The higher the proportion lost (say beyond 10-15%) the more serious the potential bias. There may also be attrition among investigators who may lose interest, leave for another job, or become involved in another project. Over a long period, many changes may occur in the environment, among individuals or in the type of intervention, and these may confuse the issue of association and attributable risk. Over a long period, study procedures may influence the behaviour of the persons investigated in such a way that the development of the disease may be influenced accordingly (Hawthorne effect). This problem is more likely to occur in studies involving repeated contact with participants, as in studies of diet or the use of contraceptives. The participants may modify their diet or shift to another contraceptive method because of repeated probing. Behavioural changes are also a serious problem in opinion surveys, acceptability studies and psychological investigations, such as studies of the psychological sequelae of sterilization. 31  Chapter 2: Research strategies and design A serious ethical problem may arise when it becomes apparent that the exposed population is manifesting significant disease excess before the follow-up period is completed. It must be emphasized that, although the cohort study is close to the randomized trial (experiment) in terms of epidemiological power, it may still have problems of validity. Care must be taken to ensure that it satisfies other requirements of epidemiological research, particularly with regard to appropriate sampling, construction of comparison groups, handling of missing data, application of appropriate statistical methods and other prerequisites for a sound research design. 2.5.3 Historical (retrospective) cohort studies In a prospective cohort study, the investigators or their substitutes are typically present from the beginning to the end of the observation period. However, it is possible to maintain the advantages of the cohort study without the continuous presence of the investigators, or having to wait a long time to collect the necessary data, through the use of a historical or retrospective cohort study. The design of such a study is illustrated in Figure 2.3. A historical cohort study depends upon the availability of data or records that allow reconstruction of the exposure of cohorts to a suspected risk factor and follow-up of their mortality or morbidity over time. In other words, although the investigator was not present when the exposure was first identified, he reconstructs exposed and unexposed populations from records, and then proceeds as though he had been present throughout the study. Historically constructed cohorts share several advantages of the prospective cohort. If all requirements are satisfied, a historical cohort may suffer less from the disadvantages of time and expense. Historical cohort studies have, however, the following disadvantages: All of the relevant variables may not be available in the original records. It may be difficult to ascertain that the study population was free from the condition at the start of the comparison. This problem does not exist if we are concerned with deaths as indicators of disease. 32  Health research methodology: A guide for training in research methods FIGURE 2.3 DESIGN OF A HISTORICAL (RETROSPECTIVE) COHORT STUDY Disease With characteristic No disease Population at Sample risk Disease Without characteristic No disease Investigation begins here and reconstructs the history of exposure and development of disease Time > 33  Chapter 2: Research strategies and design Attrition problems may be serious due to loss of records, incomplete records, or difficulties in tracing or locating all of the original population for further study. These studies require ingenuity in identifying suitable populations and in obtaining reliable information concerning exposure and other relevant factors. Examples of such population groups include members of health insurance plans, military personnel, industrial groups (such as miners), professional groups, members of a trade union, etc. 2.5.4 Prognostic cohort studies Prognostic cohort studies are a special type of cohort study used to identify factors that might influence the prognosis after a diagnosis or treatment. These follow-up studies have the following features: The cohort consists of cases diagnosed at a fixed time, or cases treated at a fixed time by a medical or surgical treatment, rehabilitation procedure, psychological adjustment or vocational adjustment. By definition, such cases are not free of a specified disease, as in the case of a conventional cohort study (but are free of the ‘outcome of interest’). The outcome of interest is usually survival, cure, improvement, disability, vocational adjustment, or repeat episode of the illness, etc. 2.5.5 Analytical cross-sectional studies In an analytical cross-sectional study, the investigator measures exposure and disease simultaneously in a representative sample of the population. By taking a representative sample, it is possible to generalize the results obtained in the sample for the population as a whole. Cross-sectional studies measure the association between the exposure variable and existing disease (prevalence), unlike cohort studies, which measure the rate of developing disease (incidence). Rare diseases, conditions of short duration, or diseases with high case fatality are often not detected by the one-time snapshot of the cross- 34  Health research methodology: A guide for training in research methods sectional study. Therefore, cross-sectional studies are more appropriate for measuring the relationship between fairly permanent characteristics in individuals and chronic diseases or stable conditions. Design Cross-sectional studies are represented in Figure 2.4. They usually start with a reference population, from which a random sample is taken. Data are collected at the same time on the risk factor or characteristic and the condition. FIGURE 2.4 DESIGN OF A CROSS-SECTIONAL STUDY Characteristic (exposure) and disease Characteristic (exposure) and no disease Reference Sample population No characteristic (no exposure) and disease No characteristic (no exposure) and no disease 35  Chapter 2: Research strategies and design Advantages of cross-sectional studies The following are some advantages of cross-sectional studies: Cross-sectional studies have the great advantage over case-control studies of starting with a reference