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.
What is the main difference between the health of 'smokers' in a cohort study and that of the 'cases' in a case-control study?
Everyone in a cohort study is well at the outset – including the smokers in our example – whereas the cases in a case-control study all have the disease 'problem' at the outset of the study.
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.