Case-control studies are retrospective in nature.
You have noticed something interesting, and you think you may know how this interesting situation came about. You begin to backtrack to see if you are correct.
It can be very exciting to think that we may have noticed something that no one else has noticed. This is where rules (research design) come into play to keep us honest and prudent.
Here is a simple schematic of the Case-control study:
Remember that:
- You can substitute any outcome of interest for Diseased versus No disease (e.g., chronic high blood sugar versus not chronic high blood sugar), and;
- You can substitute Treated versus Not treated for Exposed versus Not exposed.
It is important to note here what is meant by "retrospective". All this means is that you are working from outcome to suspected cause. You can start collecting data, or you can do a chart review, or you can do a combination. What makes it "retrospective" is that you are working from effect to cause, not vice-versa.
As in any good study, you need to give serious attention to:
- Stating a hypothesis (or hypotheses), or at least some well-focused research questions
- Carefully defining the population to which your study should generalize
- Selecting an appropriate research design (BEFORE you start collecting data - PLEASE get help if you are unsure)
- Ensuring the objectivity, precision, and reliability of your measures
- How you will draw justifiable and defensible conclusions (i.e., what will you need to be able to do this)
As you proceed:
- Keep it simple (don't try to answer all possible questions with one study - you can't do it.)
- Ensure that all potential cases (or a random sample) are included
- Define & select an appropriate control sample (patients identical to the case group except for the disease or other condition of interest)
- Maintain the integrity and completeness of your data collection.
Return to analytical designs.