Handling Duplicate Publications in a Systematic Review

The same underlying study, particularly large or notable clinical trials, sometimes gets published more than once -- an initial primary results paper, a secondary analysis focusing on a different outcome, a conference abstract preceding the full publication -- and failing to identify these as the same underlying study risks a genuine double-counting problem in your systematic review.

Why duplicate publication happens

Researchers legitimately publish multiple papers from a single trial for different purposes: the primary results paper, secondary or subgroup analyses, long-term follow-up results, and methods papers describing the trial design in detail before results are available. This is normal, expected academic practice and not itself a form of misconduct, but it creates a genuine identification challenge for anyone conducting a systematic review of the resulting literature.

The double-counting risk this creates

If your search identifies both the primary results paper and a secondary analysis of the same trial as separate records, and you fail to recognize they describe the same underlying study population, you risk including what is effectively the same participants' data twice in your meta-analysis, artificially inflating your total sample size and giving that single trial's result disproportionate influence on your pooled estimate.

How to spot potential duplicates during screening

Matching author names, particularly the senior or corresponding author, across multiple candidate records is an initial signal worth checking. Similar or identical sample sizes, recruitment periods, and study settings across records addressing a similar question are further indicators. Explicit trial registration numbers, when reported, are the most reliable single identifier -- two publications citing the same clinical trial registry number are almost certainly describing the same underlying trial.

Confirming a suspected duplicate

Once you suspect two records describe the same trial, checking both papers' methods sections directly for matching recruitment dates, matching sample size, and matching described intervention protocol confirms or rules out the suspicion more reliably than surface-level similarity alone. Where genuine ambiguity remains, contacting the corresponding author directly to clarify is a legitimate and sometimes necessary step.

What to do once duplicates are confirmed

Include the trial's data only once in your quantitative synthesis, typically drawing outcome data from whichever specific publication reports it most completely and reliably for your particular outcome of interest -- which might mean using the primary results paper for one outcome and a longer-term follow-up paper for a different outcome from the same trial, provided you're clear this represents one underlying study contributing to potentially multiple outcome analyses, not multiple independent studies.

Reporting duplicate handling in your PRISMA flow diagram

Your flow diagram should reflect duplicate publications as a specific, named category of exclusion or consolidation, distinct from records excluded for not meeting eligibility criteria at all, giving readers a transparent view of how many candidate records were actually consolidated into a smaller number of underlying unique studies.

Multiple publications reporting different outcomes from the same trial

A specific and common scenario worth planning for explicitly: a single trial reports its primary outcome in one paper and a distinct secondary outcome, sometimes years later, in a separate publication. If your review addresses both outcomes, you may legitimately extract data from both papers, but your methods section should be explicit that both draw from the same underlying trial, and your sample size accounting in any combined synthesis should reflect this correctly rather than treating the two papers as fully independent evidence.

Software support for duplicate detection

Reference management and systematic review software increasingly include automated duplicate detection features, useful for catching straightforward duplicate database records during initial deduplication, though these automated tools are considerably less reliable at catching the more subtle case of genuinely different publications describing the same underlying trial, which typically requires the kind of manual, detail-level checking described above.

Why this deserves deliberate attention

Undetected duplicate trial data is one of the more subtle ways a systematic review's pooled estimate can be quietly distorted without any single obviously wrong step in the process -- each individual extraction might be technically accurate, while the overall analysis still overweights one trial's actual influence. Building explicit duplicate-checking into your screening and extraction workflow, rather than assuming it will be caught incidentally, is a genuinely worthwhile methodological safeguard.

Duplicate detection and your risk-of-bias assessment

Once you've identified that two records describe the same underlying trial, your risk-of-bias assessment should be conducted once for that trial as a whole, informed by whichever publication reports the relevant methodological details most completely, rather than conducting separate, potentially inconsistent risk-of-bias judgments for what are actually the same underlying study's different publications.

A note on companion papers versus true duplicates

Not every related publication from the same research group constitutes a duplicate in the problematic sense -- a genuinely distinct secondary analysis addressing a different research question, using the same trial's data but analyzing it toward a different specific outcome your review is separately interested in, is a legitimate additional source rather than a duplicate to be excluded, provided you're careful about how you account for shared sample size across your combined analyses. Keeping a clear, dedicated log of which included records were identified as companion publications from the same trial, distinct from your main study inclusion list, makes this accounting considerably easier to get right and to later verify. This modest additional record-keeping effort, maintained consistently throughout screening and extraction, pays for itself considerably when a reviewer or reader later asks a specific question about your sample size accounting. Reviewers and readers alike tend to trust this kind of careful, documented accounting considerably more than a review that simply asserts its total sample size without a clear, traceable record of how that figure was actually derived. This transparency becomes especially valuable during peer review, when a methods-focused reviewer specifically checking your sample size accounting can trace your reasoning directly rather than having to take your final total figure purely on trust. This kind of traceable accounting is precisely the standard a careful methods reviewer expects, and providing it proactively rather than reactively reflects a genuinely well-prepared, transparent submission.