Academic Research for MDs – Developing a Study Question
I’m the Clinical Research Fellow here at the Clinical and Translational Research Support Center. I wanted to take this opportunity to talk about my experience here at the center and share what I’ve been learning that might help current residents and fellows in academic training programs with their research requirements.
Right off the bat, let me say medical school does not prepare physicians to do research. We get one, maybe two lectures on study design and statistics, and another one on research ethics. There were two or three questions on a block exam second year, one or two on Step 1 and one on Step 2. I remember my Step 2 question because it was the easiest question on the test and it let my brain briefly congeal for the 15 seconds it took me to answer.
So most MDs graduate medical school and enter residency with a casual relationship to research. You need an article for journal club or to impress an attending. Let’s see what was in the New England Journal of Medicine last month. Skim the abstract, maybe the conclusions. I’m sure everything else makes sense. P value less than 0.05, it must be true, great, research is easy.
The above is not academic research. We were taught re-search in medical school, how to find an article that pertains to our subject of interest and summarize the findings of said article. In academic research you are expected to create the article future medical students will bring to their attendings and journal clubs. We read articles not just to have something to talk about in rounds, but to ask the next question in your field and improve patient care. Here on the blog I’ll be discussing the research process and provide an outline that has helped us here at the center successfully take studies from an idea all the way to publication. The first step in academic research is to have a question.
The Study Question
Here in the center developing a study question is considered the easy part. Questions flow naturally from clinical practice. But is your question a good question. To find out we recommend finding a mentor in your institution whose work you are interested in, and work with them. This should be someone who has time to meet with you regularly. The Department Chair whose appointment book is full till next year is probably not the best choice. Talk to your mentor and see what ideas they have about your question. They know what is going on in the field and can guide you as to where the field is going and what the important questions are.
Your question should include the population of interest, the predictor variable (independent variable), the comparison group (control), the outcome of interest (dependent variable), and the time frame of the study. For example: ” Did one eyed left handed Eskimos (population of interest) catch more fish(outcome of interest) than one eyed right handed Eskimos(comparison group) during the summer of 1921(time frame)?” This is a well constructed question but it is a good study question.
To answer that, we recommend that your study question also be able to answer the following five questions;
1. Is it feasible? Do you have access to the time, money and expertise necessary to answer the question? If you answer no, you just saved yourself two of the three.
2. Is it interesting? Are you interested in the subject and more importantly is anyone else? If your mom is the only one that will want to read this thing at the end, you’re not getting published.
3. Is it novel? Are there 30 articles out there that ask the exact same question you want to ask and all of them came up with the same answer? Then your question is not worth asking. It’s already been answered.
4. Is it ethical? Does your question treat patients fairly and is it reasonably safe for all involved? If your question can’t past this test hopefully it won’t get past you IRB, but save yourself some time.
5. Is it relevant? Does your question have a good answer when someone asks “so what, why should I care”? This is especially true in clinical research. You could have a P value of 0.00001 but if your question has no clinical relevance,who cares.
I know some of these seem harsh, but it is important to take a dispassionate look at your question before you go to the time and effort to develop a study. Here at the center we like to think of studies like babies. Everybody thinks their baby is beautiful because it’s theirs. Ask someone objective, someone who would tell you if you had an ugly baby, to answer the questions above about your study question. That will be the test of a good study question.
In future posts I’ll be addressing the rest of the research process, from study design, protocol creation, data collection and analysis to manuscript creation and publication. If you need assistance or advice on any part of the research process in the mean time, don’t hesitate to contact us here at the center. My contact information is below. We’d be glad to help.
Daniel Curran MD
Clinical Research Fellow
Clinical and Translational Research Support Center
University of Louisville Division of Infectious Diseases