How to Choose a Research Question for an MD Thesis

Your thesis question shapes the next two to three years of work. A good question is not the most ambitious one — it is the one you can answer well with the patients, time and resources you actually have.

Start from a genuine clinical uncertainty

Look at your daily work. Which decisions do consultants in your department make differently? Which outcomes are rarely audited? Questions that come from real practice are easier to justify and more likely to interest examiners and journals.

Use the FINER criteria

• Feasible — adequate number of eligible patients within your data collection window, and the tests or tools are available.
• Interesting — to you, your guide and your specialty.
• Novel — confirms, extends or refutes earlier findings, or studies them in a different population.
• Ethical — acceptable to your Institutional Ethics Committee.
• Relevant — the answer could influence practice or future research.

Structure it with PICO

Write the question in terms of Population, Intervention or exposure, Comparison and Outcome. For example: “In adults with community-acquired pneumonia (P), is a CURB-65 score ≥ 3 at admission (I) compared with a score < 3 (C) associated with 30-day mortality (O)?” A PICO-structured question makes the design, sample size and analysis much easier to decide.

Check the numbers early

Before committing, check how many eligible patients your unit sees and do a rough sample size estimate. Many theses run into trouble because the required sample cannot be recruited in 12–18 months.

Search the literature before you finalise

Spend a few focused hours on PubMed. If your exact question has been answered repeatedly in similar populations, consider a different angle — a different outcome, subgroup or setting.

Agree it with your guide

Present two or three options with a one-paragraph justification for each. Your guide’s experience with feasibility and departmental priorities is invaluable.