Most case-control studies are retrospective studies and are conducted on a selection of available cases. For example, a medical care facility wants to identify the cause of a particular disease among a group of people (case). The past patient medical history or history of exposure to a suspected risk for the group is compared with the past history of another group called the controls. The controls are similar to the case group, but don't have the disease. A case-control retrospective study is based on outcomes measured prior to exposure. Case-control studies are a relatively quick and inexpensive method of establishing evidence in a retrospective study. Samples required are small and the concept is easily understood. The main difficulty with this type of study is that it usually does not provide convincing evidence. Data collection has already been completed and may be subject to bias like flawed information, improper selection controls and selective recording of data.
In a cohort study, there are two sub-types: the retrospective cohort study and the prospective cohort study. Cohort studies are an investigation into a specific trait of a particular group that was observed or will be observed over a certain period of time. Examples of cohort studies include people who have been taking a certain medication, people who have a specific medical condition, animals that live near a certain type of pollutant or a study of poverty from the sociological point of view. Cohort studies probe deep into the problem and study groups can be sub-divided. For example, smokers might be a broad category in the study; smokers who are overweight could be a sub-group. This type of study is expensive and requires large sampling groups. It takes time to complete and is prone to attrition bias, which is a skew in the data caused by people dropping out of the study.
Cross sectional studies are conducted using a representative subset of the population. It is a type of observational study used to describe both absolute and relative risks. They can be used to describe a specific feature of the population: for example, the prevalence of a specific illness. A cross-sectional study cannot differentiate between a new illness and long-established conditions of an illness. It can only measure the frequency of conditions relating to the illness. The study can identify the existence of the health problem, but not a cause-and-effect relationship.
In a matched case-control study, the subjects of the control group are selected to match the case group in relation to specific characteristics like age or severity of the disease. The aim of the study is to make a comparison between the case group and control group with similar characteristics to increase precision of the comparison. For example, a study of pneumonia can be done by choosing the case and study groups based on factors like similarity of diagnosis at the time of admission and age. The case group receives non-invasive ventilation treatment and the control group receives invasive ventilation treatment. The results of the study show whether the invasive or non-invasive treatment is more effective.