You’ve probably done research in undergrad and already know the basics – mean, median, p-values, and all the usual stuff from biostats class. So instead of repeating that, this post is a quick guide to everything else you might need to actually apply that knowledge in med school.
1. How clinical research fits into the bigger picture
1.1 If you’re looking for a project or research mentor
It’s usually not difficult to get involved in research once you’re in med school – you just have to know where to look. If you already know your area of interest, start by checking the research section of that department on your school’s website, or browse the faculty profiles to see what specific attendings are working on.
Also, I like going through my school’s match list to see which students matched into the specialty I’m interested in. Then I look at who they worked with and reach out to those PIs to ask if they have any available projects.
Once I identify a few potential PIs I’d like to work with, I usually:
- Check their research profile
- Look for recent publications (if a PI has a few publications within the past few years, there is a high chance that they may have an ongoing project)
- See if any of their papers include med student co-authors – an excellent sign (I always look for PIs that enjoy mentoring med students)
The final step is reaching out by email.
“Hi Dr. ____, I’m an M1 interested in learning more about [topic]. I was wondering if you had any ongoing projects that I could potentially assist with.
I hope to begin [this semester/summer], and I’m available to continue remotely during the academic year. I’d love to meet briefly to learn more about any potential opportunities to contribute to your work.
For your reference, I have attached my CV. Thank you so much for your time and I look forward to hearing from you.”
If you don’t hear back in 1-2 weeks, it’s okay to follow up once politely.
1.2 If you’re on wards trying to figure out what the next best step for a patient is…
- First, check Uptodate or Amboss
- If the answer isn’t readily available and your attending asks you to search literature or you offer to search literature…
Start by framing your question into a PICO question (you will learn this in med school, but I didn’t realize how PICO was integrated into clinical practice until M3 year when I picked up this tip from an attending). This helps focus your research and makes searching literature way easier.
PICO = Population (Who?) + Intervention (What are you testing?) + Comparison (Compared to what?) + Outcome (What are you measuring?)
Example:
In adults with type 2 diabetes (P), does GLP-1 agonists (I) lead to lesser MACE (O) than the current standard of care alone (C)?
2. How to Search literature
I like to use PubMed, but there are other databases such as TRIP or Cochrane Library. Start broad, then filter down.
- Click “Advanced” to build multi-layered searches
- Combine terms with:
- AND = both terms must be present
- OR = either term
- NOT = exclude results
Then apply filters:
- For physicians in practice, the best evidence usually comes from meta-analyses, systematic reviews and randomized controlled trials (RCTs).
- Clinical trials may include both RCTs and non-randomized trials (less reliable, but check this if you want to cast a wide net)
- However, if none are available, you can also look for cohort or case-control studies. `
3. Interpreting research results
After reading the “entire” paper, here’s a quick guide to analyze and understand the results section:
- P-value < 0.05: Common threshold for statistical significance.
- Confidence Interval (CI) shows the range of likely values.
- For mean differences, if CI includes 0 → not significant
- For odds ratios or hazard ratios, if CI includes 1 → not significant
Example:
- HR = 1.02 (95% CI: 0.89–1.17) → Not significant because 1 is inside the CI.
- HR = 0.87 (95% CI: 0.78-0.97) → Significant because 1 is not inside the Cl.
Note that there’s a difference between statistical significance and clinical significance.
- Statistical significance: P < 0.05, Cl not crossing the null (1 for HR/OR, 0 for mean differences)
- Clinical significance: does this matter clinically? Look at the HR/OR and see how much risk is actually reduced
Example:
HR = 0.99 (95% CI: 0.98–1.00), p = 0.03 → statistically significant, but barely reduces risk.
HR = 0.50 (95% CI: 0.40–0.60), p < 0.001 → both statistically and clinically significant.