How to Write a Medical Thesis Discussion Chapter

Discussion is interpretation, not a second results section

The discussion chapter should explain what the main findings mean, how they fit existing evidence, and what they cannot establish. Begin with the most important study objective and avoid opening with a long restatement of all baseline tables.

A usable paragraph structure

1. Principal finding: State the result in words and, if useful, cite the precise estimate.
2. Comparison with prior work: Identify a genuinely comparable study and summarize its actual design, population and result with a verified citation.
3. Possible reasons for agreement or difference: Consider selection criteria, case mix, measurement, treatment, follow-up or analytic adjustments.
4. Implications and limitations: Explain clinical relevance cautiously and avoid unsupported causal language.

Example of cautious wording

“Our study observed a higher frequency of the outcome among exposed participants. The comparison should be interpreted considering the observational design and potential residual confounding.” This sample sentence deliberately avoids invented percentages and unverified literature claims.

Separate limitations from recommendations

Discuss selection bias, information bias, missing data, sample size, precision, measurement limitations and confounding as they relate to your actual methods. State what the study contributes and which further research could resolve uncertainty. A non-significant p-value alone does not prove equivalence between groups.

Practical checks before submission

• Compare like-for-like populations, follow-up periods and outcome definitions.
• Use original papers, not second-hand summaries, to verify comparative data.
• Keep conclusions proportionate to the effect estimates and study design.
• Do not invent the names, years, results or references of comparator studies.

Use ResearchTry Literature Search to identify related studies, then confirm each paper before using its findings in the discussion chapter.

Frequently asked questions

Should I compare every result with previous papers?

Prioritize prespecified primary outcomes and the most important secondary findings; avoid superficial comparisons of unrelated samples.

What belongs in limitations?

Discuss selection, measurement, confounding, precision, missing data, applicability and design-specific sources of uncertainty.

Further reading and primary guidance

• ICMJE manuscript preparation: statistics, tables and references
• EQUATOR Network: STROBE observational study reporting guideline