Health Management, Ethics and Research: 14.  Research Strategies and Study Designs for Small-Scale Research

Study Session 14  Research Strategies and Study Designs for Small-Scale Research

Introduction

In the previous study sessions you became familiar with conducting a community survey and using the data to create a community profile for the kebele in which you will be working. You were also introduced to the general principles of health research and how they apply to any small-scale research projects that you may be able to conduct as a Health Extension Practitioner in a rural area at some time in the future. Although you may not be involved directly in research at an early stage in your career, it is important that you know about some simple research methods and how you could use them to inform your understanding of health problems in your area.

Selection of a research strategy and study design is the most important decision a researcher has to make before beginning any research activity. The main focus of this study session is to outline the essential components of research strategies and study designs, which include defining the population of interest and the study variables. These terms will be fully explained in this study session. You will also learn about the selection of an appropriate study design for your particular study population and variables, and how to collect useful data that has the potential to improve the health services you provide in your community.

Learning Outcomes for Study Session 14

When you have studied this session, you should be able to:

14.1  Define and use correctly all of the key words printed in bold. (SAQs 14.1, 14.3 and 14.4)

14.2  Identify and define appropriate study populations for a specific type of research study. (SAQ 14.1)

14.3  Identify the study variables in descriptions of small-scale research projects presented to you. (SAQs 14.1, 14.2 and 14.3)

14.4  Describe the three survey designs most commonly used in small-scale health research and comment on their relative advantages and disadvantages. (SAQ 14.4)

14.1  What is the study population?

In your work you may have identified something interesting that you would like to investigate further and which could form the basis of a possible research problem. It will probably involve particular groups, for example young people, or people with a specific health problem such as malaria. So, investigating the problem will almost certainly require a research project. These particular groups of interest are referred to as the study population. This study population is the total members of a defined class of people, objects, places or events selected because they are relevant to your research question. For example, if you want to study maternal healthcare, the study population would be all the pregnant women who are under your care. However, a study population may consist of villages, institutions, records, or events such as death due to accidents, etc.

In any research study, the study population has to be clearly defined according to particular characteristics such as age, sex, residence or geographical accessibility to health services (Figure 14.1). The way you define your study population depends on the problem you want to investigate and on the objectives of the study. For example to investigate health problems of orphans whose parents have died of HIV/AIDS, then your study population could be all children below twelve years of age whose parents have died as a result of HIV/AIDS. On the other hand, to investigate the problem of worm infestation in primary school children your study population could be all primary schools in your kebele.

Figure 14.1  Your study population could be any definable group that is relevant to your study, for example, teenage boys might be the study population for research on awareness of contraception. (Photo: Janet Haresnape)

When considering what your study population might be, there may also be practical questions to consider such as how easy is to get access to the study population you are hoping to study (Figure 14.2).

Figure 14.2  Transport in rural areas can be problematic, especially in the rainy season. Can you easily reach the study population? (Photo: Janet Haresnape)

14.2  Variables in health research

If you do have a chance to do some research in your village, then you will need to know about what sorts of variables might be suitable for study. A variable is something that can take on different values. A variable may be in the form of numbers (such as weight or age), or non-numerical characteristics (such as occupation, ethnicity, education level or gender). It can be a characteristic of a person, object, place, event or phenomenon. For instance, a water filter is an object that exists in several different types (or variables), so you could study the effectiveness of different water filters as the variable in a research study on prevention of diarrhoeal diseases. The type of house that people live in is a variable ‘place’ that affects health, and an accidental injury is an example of a variable ‘event or phenomenon’ that you could study.

Variables may take numerical (quantitative) or non-numerical (qualitative) values. Measurements such as weight expressed in kilograms or pounds, height expressed in metres or centimetres, income expressed in dollars or birr, and number of children per family are examples of numerical or quantitative variables.

Qualitative variables are attributes described in terms of the presence or absence of certain characteristics, such as diseased or non-diseased, smoker or non-smoker, married or unmarried. Qualitative variables may also be descriptive, such as occupation or political views. In the question above, items A (marital status) and E (family support) are qualitative variables.

Knowing about variables is very important, because you will not be able to undertake useful research or discuss your research findings unless you understand which variables you are dealing with. The identification of variables as non-numerical or numerical determines whether the data is qualitative or quantitative, and this in turn determines the method of analysing the data. Remember that you have already learned about qualitative and quantitative data in Study Session 12.

14.3  Types of study design

If you are interested in doing some small-scale research in your kebele, then you will certainly need to think carefully about the design of your study. Study designs can be observational (in which the researcher simply observes what is going on), or experimental (in which the researcher intervenes in some way and then observes what happens as a result).

14.3.1  Observational and experimental studies

In an observational study you just observe and analyse your study subjects or situations in their natural condition, but the researcher does not intervene in any way. However, the observations should not just be a haphazard collection of facts; observational studies should apply the same rigour as experimental studies by recording the observations in a systematic way that can lead to meaningful interpretations. You are more likely to be involved in observational studies than in experimental studies.

An experimental study is a type of research design in which some of the conditions are under the direct control of the researcher. He or she intentionally alters one or more factors (e.g. the behaviour of participants, or their circumstances and situations), and then measures the outcomes to analyse the effects of the alteration. For example, you could implement a health education campaign on the benefits of immunization (this is the intervention), and then measure any improvement in immunization coverage rates (by analysing the outcomes of your intervention).

Figure 14.3  Investigating the proper use of bed nets could be an informative subject for research. (Photo: UNICEF Ethiopia/Indrias Getachew)

One thing we have to remind you here is that it would not be ethical to introduce ITNs to some households in your community and intentionally leave other households without nets, in order to conduct an experimental study on the effect of ITNs on preventing malaria. The knowledge that ITNs protect people from the mosquitoes that transmit malaria is already well established by research. Therefore, there is no need for you to duplicate this research and deny protection to some families by leaving them out of the ITN distribution programme. This example reminds you that your actions in designing a research study must observe strict ethical guidelines (refer back to Study Sessions 7 and 8).

Figure 14.4  A research study could look at the nutritional status of all children under three years. (Photo: Janet Haresnape)

14.3.2  Types of observational studies

As a Health Extension Practitioner you will probably only be directly concerned with observational studies. These generally take the form of sample surveys (of selected individuals) or population surveys (covering everyone in the population), where the sample or the population is observed for various characteristics. This may be done by interviewing study subjects, using questionnaires or focus groups, by obtaining measurements of physical characteristics (e.g. height or age), or by simply extracting information from existing sources, such as disease registries or hospital records. The most common type of study design you will be involved in is the observational survey, in which information is systematically collected from your study population on a specific topic.

There are three types of observational survey which are most likely to use if you conduct a small-scale research project, and each has a different study design. They are:

• cross-sectional study
• case-control study
• cohort study.

These are just three of many types of research study. These three will be covered in more detail in the sections that follow.

14.3.3  Cross-sectional studies

You may remember from Study Session 10 that cross-sectional studies look at something that is going on at a particular moment in time. If you do a cross-sectional study, you would collect information about the current situation, for example the current state of a particular disease in a particular section of your community, or the current exposure of a certain group of individuals in your community to an infection. Therefore, in cross-sectional studies the information on the possible cause and the possible effect is collected at the same time. However, the problem with cross-sectional studies is that there is no way of knowing whether the possible cause is really responsible for the effect you measured – or could it be the other way around? The following example illustrates what we mean by this problem.

Cross-sectional studies aim at describing the distribution of certain variables in a study population at a certain time, for example in the last month. Examples might include studies of:

• Physical characteristics of people: e.g. surveys of the occurrence of tuberculosis, HIV, malnutrition, high blood pressure, diabetes, anaemia, etc.
• Physical characteristics of the environment: e.g. evaluation of latrine usage, access to safe water supply, distance from the nearest health centre, etc.
• Demographic or socio-economic characteristics of people: e.g. their age, education, marital status, number of children and income.
• The behaviour or practices of people and their knowledge, attitudes, beliefs or opinions, which may help to explain that behaviour: e.g. attitudes to contraception, or harmful traditional practices, healthcare-seeking behaviour if a child has measles or diarrhoea, etc.
• Adverse events that occurred in the population, e.g. the occurrence of road traffic accidents or burns, and natural disasters like flooding or drought.

Cross-sectional studies usually cover a selected sample of the population. For example, it would be impossible to study all children with measles in a country, so a sample of children with measles is chosen randomly without any bias in the selection. Data are collected at the same time from the sample of children on the risk factor or characteristic (in the example above, it is being underweight) and the disease condition (having measles). If a cross-sectional survey covers the total population of a country it is called a census (as you already learned in Study Session 10). A cross-sectional survey of a sample of the population is cheaper than doing a census because it consumes fewer resources. A cross-sectional survey may be repeated at a later date in order to measure changes over time in the characteristics that were studied. Case Study 14.1 gives a simple example of a cross-sectional study.

14.3.4  Case-control studies

If you want to find out more about health conditions in your village it would be necessary to use a slightly more complicated research design than a cross-sectional study. Usually, this will involve a case-control study. You begin by selecting two groups for comparison: one group of the population (the cases) who have a characteristic that you wish to investigate (such as mothers whose child died within one month of birth, or children with malnutrition), and another group (the controls) where that problem is absent. The aim of a case-control study is to find out what factors have contributed to the problem under investigation. Case Study 14.2 gives an example to illustrate this type of study.

It is important to ensure that cases and controls in your study come from the same study population. For example, in a case-control study of vaccination of pregnant women against tetanus, where cases of unvaccinated women are being selected from a particular health centre, controls should be selected from mothers who have attended vaccination at the same health centre. If controls are selected from another centre, they would not be from the same study population and therefore they would not be directly comparable to the cases.

14.3.5  Cohort studies

In order to study a health issue in more detail, it may be necessary to follow a group of people over a long period of time to see which members of the group (or cohort) develop the condition. As you can imagine this type of study is complicated and may be very expensive because it covers a long time. Usually this type of study is done by external agencies such as university researchers or NGOs.

The term cohort describes a specific group of people who are all in a particular situation during a certain period of time, for example, everyone who was born in a particular year, or everyone who shares a particular characteristic (e.g. all farmers, or all women of childbearing age). If these people are studied over a long period of time to see what diseases they develop, this is called a cohort study. The most common type of cohort study begins with a population who share the same general characteristics (e.g. all men aged between 40 and 50 years), but then compares those individuals in the cohort who are exposed to a disease risk factor (e.g. the tobacco smokers) with those individuals in the cohort who are not exposed to the risk factor (in this example, the non-smokers). Then you follow everyone in the cohort over a period of months or years, and compare the occurrence of the problem that you expect to be related to the risk factor (for example, lung cancer or pneumonia) in the two groups. This allows you to determine whether a greater proportion of those with the risk factor (in this example, smoking) are indeed affected by the disease.

Cohort studies may take a long time to yield meaningful results, so they are generally expensive to conduct. However, they have the great advantage of resolving the ‘cause or effect’ problem we mentioned earlier in relation to cross-sectional studies. Because everyone in the cohort is well at the beginning of the research period, it is possible to say that (for example) being a smoker is a contributory cause of lung cancer, if the smokers develop more lung cancers than the non-smokers over time.

14.4  Selecting the appropriate study type to investigate a particular problem

The choice of your study design may depend on the following factors:

• type of problem
• knowledge already available about the problem
• resources available for the study.

For example, as we said earlier, cohort studies are very expensive because they have to follow-up a large number of individuals over a long period of time until the results become clear. Some possible investigations, and the type of study design which would be most suitable in each case, are shown in Table 14.1.

Figure 14.5  Installation of a protected water source should have an impact on the number of cases of diarrhoea. (Photo: Richard Adam)

Summary of Study Session 14

In Study Session 14, you have learned that:

1. Defining the study population and study variables, and choosing the study type, are an essential starting point for every research project.
2. The way you define your study population depends on the problem you want to investigate and on the objectives of your study.
3. Variables may be numerical or non-numerical, so the data collected may be quantitative or qualitative, and this determines the appropriate method of analysing the data.
4. Study designs may be observational or experimental. Observational studies involve making observations. Experimental studies involve modifying or changing something about the study participants and looking at the effect of this intervention.
5. The three common observational study designs are cross-sectional, case-control and cohort studies, all of which involve a survey of some kind.

Self-Assessment Questions (SAQs) for Study Session 14

Now that you have completed this study session, you can assess how well you have achieved its Learning Outcomes by answering the following questions. Write your answers in your Study Diary and discuss them with your Tutor at the next Study Support Meeting. You can check your answers with the Notes on the Self-Assessment Questions at the end of this Module.

First read Case Study 14.3 and then answer the questions that follow it.

Now read Case Study 14.4 and then answer the question that follows it.