How to Write the Discussion Chapter of an MD/MS Thesis

The Discussion is the chapter where examiners find out whether you understand your own results. Tables and p-values show what you found; the Discussion explains what it means, how it fits with earlier studies, and how much weight it can bear. Many MD, MS and DNB theses lose marks here, not because the data are weak, but because the chapter simply repeats the Results or quotes long lists of studies without comparison.

This guide gives you a clear structure, sentence templates you can adapt, a safe way to compare your findings with previous studies, how to write about non-significant results, and the criticisms examiners raise most often.

1A Structure That Works

Follow your university’s guidelines if they prescribe a format. Otherwise, this order keeps the chapter logical and easy to examine:

1. Restate the key findings – one short paragraph summarising the main answer to each objective, without repeating every number.
2. Compare with previous studies – objective by objective, or variable by variable, in the same order as your Results chapter.
3. Explain agreement and disagreement – why your results match some studies and differ from others.
4. Possible mechanisms – the biological or clinical explanation for what you observed, with references.
5. Strengths of the study – specific design features, stated without exaggeration.
6. Limitations – honest, specific, with their likely effect on the results.
7. Implications – for clinical practice in your setting, and for future research.
8. Link to the Conclusion – a closing paragraph that leads naturally into your Conclusion chapter.

2Opening: Restating the Key Findings

Begin with a brief reminder of what the study did and what it found. This is not a repeat of the Results; it is a summary of the answers. Keep it to one paragraph.

• “The present study was conducted to [aim] among [number] [population] at [setting].”
• “The main findings were that [finding 1], [finding 2], and that [finding 3] was not significantly associated with [outcome].”

Readers and examiners who jump straight to the Discussion should understand your study from this paragraph alone.

3Comparing Your Results with Previous Studies

This is the main body of the chapter. For each finding, compare it with the studies you described in your Review of Literature, and add any relevant studies published since. Your Review of Literature evidence matrix is the best starting point here, because it already lists each study’s design, setting, sample and results.

A paragraph pattern for each finding

1. State your finding with the key figure.
2. Name studies that agree, with their figures and settings.
3. Name studies that disagree, with their figures and settings.
4. Explain the likely reasons for the difference.

Sentence templates

• Agreement: “In the present study, [finding, with figure]. This is comparable to the findings of [Author] et al., who reported [figure] among [population] in [setting] [ref].”
• Partial agreement: “Our finding of [result] is broadly in line with [Author] et al. [ref], although they observed a [higher/lower] [measure] of [figure].”
• Disagreement: “In contrast, [Author] et al. [ref] reported [figure]. This difference may be explained by [reason].”
• Synthesis: “Taken together, most studies, including ours, suggest that [conclusion], although the size of the effect varies with [factor].”

Reasons that explain differences

When your results disagree with another study, look for a concrete reason rather than writing “this may be due to various factors”. Common ones are:

• Differences in study population (age, severity, comorbidities, rural vs urban, referral centre vs community).
• Different definitions, cut-offs, diagnostic criteria or measuring tools.
• Different study design (cross-sectional vs cohort, single-centre vs multicentre).
• Sample size and the resulting precision of the estimate.
• Different time periods, treatment protocols or follow-up duration.

Comparing without overstating

Choose your verbs carefully. A single-centre study of modest size does not “prove” anything; it “suggests”, “indicates” or “is consistent with”. An association in a cross-sectional study is not evidence of cause. Write “was associated with”, not “caused” or “led to”.

4Explaining Mechanisms

After comparison, explain why the finding makes sense biologically or clinically. For example, if you found a link between a biomarker and disease severity, briefly describe the known pathophysiology that could connect them, and cite the source. Keep this focused: one or two paragraphs per major finding, not a repeat of the basic-science section from your Review of Literature.

When your finding is unexpected, say so plainly and offer possible explanations, marked clearly as possibilities: “One possible explanation is…”, “This may reflect…”.

5Handling Non-Significant Results

A non-significant result is still a result, and it deserves honest discussion rather than a single line or silence. The key idea is that a p-value above 0.05 does not show that there is no difference or no association; it shows that your study did not find enough evidence of one. Absence of evidence is not evidence of absence.

• Report the effect size and confidence interval where your analysis produced them. A wide interval that includes both meaningful benefit and harm tells the reader the study was inconclusive, not negative.
• Consider sample size and power. If your sample was calculated for the primary outcome, secondary comparisons may be underpowered. Say this directly.
• Compare with other studies. Did other studies with larger samples find an association? Did similar-sized studies also find none?
• Avoid spin. Do not describe a non-significant difference as “a trend towards significance” to make it sound positive, and do not drop the result from the Discussion.

Template: “No statistically significant association was found between [X] and [Y] (p = [value]). This differs from [Author] et al. [ref], who reported [finding] in a larger sample of [number]. The smaller sample in the present study may have limited the power to detect a difference of this size, and the finding should be interpreted with caution.”

For more on interpretation, read understanding p-values in medical research. If you want to show how sample size affects power, the sample size calculator can help you explain it to your guide.

6Strengths and Limitations

Strengths

Mention only real, specific strengths, such as a prospective design, standardised measurement by a single trained observer, blinding, a validated tool, or a population not studied before in your region. Avoid generic claims like “this is a very important study”.

Limitations

Examiners respect honest limitations; they become suspicious when there are none. Typical ones in Indian PG theses:

• Single-centre, hospital-based sample, which may limit generalisability to the community.
• Sample size limited by the time available for the thesis.
• Cross-sectional design, which cannot establish temporal sequence.
• Short follow-up period.
• Possible selection, recall or observer bias, depending on design.

For each limitation, add a sentence on its likely effect and how you tried to reduce it: “Being a hospital-based study, the sample may over-represent more severe cases; however, consecutive sampling was used to reduce selection bias.”

7Implications and the Link to the Conclusion

Close the Discussion with what your findings mean in practice and what should be studied next. Keep implications proportionate to your evidence and relevant to your setting: “In resource-limited settings similar to ours, [finding] may help clinicians [action]. Larger multicentre studies with longer follow-up are needed to confirm [finding].”

The final paragraph should lead into the Conclusion chapter by restating the overall answer in one or two sentences. The Conclusion itself should then answer each objective directly, without new references or discussion.

8Common Examiner Criticisms

Also make sure the studies you cite are formatted consistently; our Vancouver referencing guide covers the common errors. If your statistical tests or their interpretation are still unclear, revisit statistical analysis for a medical thesis before you write.

9Getting Help with Your Discussion

A strong Discussion is the difference between a thesis that passes and one that impresses. If you would like feedback on your draft, help matching your results to the right comparison studies, or a review of how you have handled non-significant findings, the PubMedico team works with MD, MS and DNB residents on this chapter regularly. Reach us on WhatsApp at +91 78787 33004.