BRS - Type of Study

BRS - What Type of Study?

Choosing a research design for your study is a balance between doing the best possible study to answer the research question and the resource, time, monetary and ethical constraints involved in doing the study.

First, get your research question clear:

1. What sort of Patients or Populations are you interested in? (think gender, pre-existing disease, age, etc.)

2. What exposures or interventions are you interested in? (a drug therapy, a risk factor, an exposure to a possible harm, etc.)

3. Are there comparisons that are relevant (no intervention, absence of a risk factor, etc.)

4. What are the outcomes you are interested in measuring? (disease presence, mortality, side effects, proxy measures of disease (like lab tests), etc.)

Be sure to focus. It's usually best to bite off a small piece of a larger question than try to study the whole topic at once.

Once you have the question nailed down, you should determine which study is best.

1. Is good quality data already available (both exposure and outcome)? If so, you should consider a retrospective study. If not, you must do a prospective study (or possibly a cross-sectional study, though these are limited in their ability to make inferences).

2. For retrospective studies:

   1. Can you sort the groups by exposure status, then figure out if they had the outcome? If so, then you'd want a RETROSPECTIVE COHORT study.

   2. If you know the outcome, and want to look back for exposure, then you want a CASE-CONTROL study.

3. For prospective studies:

   1. Is there an intervention (a drug, an educational program, a diet, etc.) you're testing? If so, the ideal is a RANDOMIZED CONTROLLED TRIAL, but you may have to settle for a PROSPECTIVE COHORT study.

   2. If you're assessing the effect of an exposure over which you (the researcher) don't have any control, then you want a PROSPECTIVE COHORT study.

Review the lectures on study design if you are having trouble with this part.

Study design quiz questions - ONE, TWO, THREE, FOUR, (BMJ)