01 · The Question
What If Related Papers Report Different Parts of the Same Study?
You identify several publications from one study. The main paper reports the primary clinical outcome. A companion article focuses on quality of life. Another reports harms, while a later paper presents long-term outcomes.
Which outcomes belong to the study? Potentially all of them.
A study can measure many outcomes and distribute them across several reports. Cochrane notes that multiple journal articles from one study may each focus on different aspects of the investigation. The methodological challenge is therefore not to force every paper to report the same outcomes, but to reconstruct the complete outcome record while preserving the fact that all of those results originated from the same underlying study.
03 · What You Need to Know
Different Outcomes Can Be Complementary, Selective, or Apparently Conflicting
One Study Can Measure Far More Than One Paper Reports
Most substantial studies collect multiple variables. A randomized trial might measure symptoms, functioning, quality of life, treatment adherence, adverse events, and resource use. A cohort study may similarly support analyses of several outcomes over multiple follow-up periods.
Journal articles rarely reproduce every analysis generated by a project. Different reports can therefore provide complementary windows onto one study. Cochrane explicitly recognizes that a study may have multiple journal articles focusing on its design, main results, or other results.
The existence of outcome-specific papers does not turn those papers into independent studies. If necessary, first establish whether the publications actually report the same underlying study .
Build an Outcome Map Across All Reports
Rather than reading each paper in isolation, map the study's outcomes across its linked reports. At minimum, record the outcome domain, specific measure, time point, analytic population, and source report.
Report
Outcome Domain
Measure
Time Point
Analytic Population
Main report
Symptoms
Scale A
3 months
Full analysis population
Companion report
Quality of life
Scale B
3 months
Participants with available measure
Harms report
Adverse events
Event counts
Treatment period
Safety population
Follow-up report
Symptoms
Scale A
12 months
Participants retained at follow-up
This immediately shows whether the publications are providing complementary outcomes, repeated measurements of the same outcome, or potentially competing results.
Different Outcome Domains Are Usually Not a Conflict
If one paper reports anxiety and another reports quality of life, the reports do not disagree merely because their outcomes differ. Each may contribute to a different outcome specified in your review.
The important constraint is that both reports remain attached to the same study. If the review contains separate syntheses for anxiety and quality of life, the same study may legitimately contribute once to each synthesis.
This is not participant double-counting in the same sense as entering two correlated estimates into one conventional meta-analysis as though they were independent studies. Each synthesis addresses a different outcome.
The Same Outcome at Different Times Is Also Not Automatically a Conflict
A paper reporting an outcome at three months and another reporting it at 12 months may both be correct. The difference reflects follow-up time rather than contradictory findings.
Your protocol should define how time points are handled. You might have short-term and long-term outcome windows, for example. Cochrane recommends specifying outcome domains, measurement tools, time points, and other result characteristics sufficiently clearly to guide data collection and avoid post hoc selection.
Different Measures of the Same Construct Create Multiplicity
A more difficult situation occurs when the same participants are assessed using several instruments intended to measure the same outcome domain. One paper might report depression using Scale A, while another uses Scale B at the same follow-up.
If both measures satisfy your outcome definition, you now have multiple eligible results from the same study. Cochrane notes that studies often report several results for the same outcome because of different measurement scales, subgroups, or follow-up times, and recommends prespecifying how choices among eligible results will be made.
A protocol might establish a hierarchy of preferred validated instruments, specify one time point per outcome window, or use another defensible selection strategy. Depending on the review, statistical methods capable of handling multiple correlated effects may also be appropriate.
Do Not Select the Most Favorable Outcome
Suppose Scale A produces a large statistically significant effect while Scale B produces a small nonsignificant effect. Selecting Scale A because it makes the intervention appear more effective would allow the observed result to influence the selection process.
Outcome and result selection should instead follow rules established independently of the observed direction, magnitude, or statistical significance whenever possible. This is one reason systematic-review protocols need sufficient specificity about outcome definitions and measurement choices.
Compare Publications With the Protocol or Registration
When available, study protocols and prospective registrations can help reconstruct which outcomes were planned and how they were originally defined. This can be particularly informative when different publications emphasize different subsets of outcomes.
The purpose is not to assume that every difference from a protocol is misconduct. Study plans can change for legitimate reasons. Rather, the protocol provides another source against which the publication history can be understood.
Sometimes the Papers Really Do Conflict
Different outcome reporting should be distinguished from inconsistent reporting of the same result. Suppose two publications both claim to report the same outcome, population, and time point, yet provide different event counts or effect estimates.
First look for methodological explanations. Are the analysis populations different? Was one analysis adjusted and the other unadjusted? Did one publication use complete cases while another used imputation? Is one report a correction? Are the outcome definitions actually identical?
Cochrane notes that information from multiple reports can be inconsistent and recommends that review authors plan how discrepancies will be resolved. If inconsistencies cannot be resolved from the reports, contacting investigators may be appropriate.
Different outcomes
Reports describe distinct constructs or endpoints, such as symptoms and quality of life. This may be complementary reporting.
Different results for the same outcome
Reports appear to describe the same construct, population, and time point but provide incompatible values. This requires reconciliation.
Do Not Assume the Primary Report Always Overrides a Companion Paper
A primary report is useful for organizing the study, but it is not necessarily the best source for every outcome. A dedicated harms paper may provide far more complete adverse-event data. A long-term follow-up paper may be the only source for the time point your review needs.
Cochrane recommends retaining secondary reports because they may contain additional outcomes and valuable study information.
Outcome-Specific Publications Still Share a Study Population
If several papers use the same participants, their results remain statistically related even when the publications focus on different outcomes. This becomes particularly important when multiple results are being combined within a single synthesis.
Before pooling several estimates, establish whether the reports use the same participants . If they do, use the companion reports without treating shared participant contributions as independent .
Watch Out
Do not interpret the absence of an outcome from one paper as proof that the study never measured it. Check companion publications, protocols, registrations, supplementary material, and other available study sources before making that conclusion.
04 · A Practical Example
Reconstructing Outcomes Spread Across Four Publications
Hypothetical Example
One Trial With Outcomes Published Separately
Imagine a trial of an educational intervention involving 800 students. Your review considers academic performance, anxiety, and student well-being.
Paper A Reports academic performance at the end of the semester for 790 students.
Paper B Reports anxiety at the same time point for 742 students who completed the psychological assessment.
Paper C Reports student well-being for 735 participants.
Paper D Reports academic performance one year later for 610 students retained in follow-up.
These are not four independent studies. They are four reports contributing different information about one trial.
If your review has separate syntheses for academic performance, anxiety, and well-being, the appropriate result from each report may contribute to its corresponding outcome. Paper D may contribute to a separately defined long-term academic-performance analysis.
Your study record should make the relationships explicit: one trial, several reports, several analytic samples, several outcomes, and several time points. That structure is more accurate than either discarding three papers or pretending you have four independent trials.
06 · What This Means for You
Create a Study-Level Outcome Map Before Extracting Final Results
For studies with multiple reports, organize extraction around the underlying study while retaining the source of each result. Cochrane permits either extracting each report separately and reconciling the forms afterward or extracting information from all reports directly into one study-level form.
A simple decision framework
If reports provide different outcome domains that are eligible for your review
Extract each relevant outcome and assign it to the appropriate synthesis while keeping the reports linked to one study.
If reports provide the same outcome at different eligible time points
Apply your prespecified time-point or outcome-window definitions.
If several measures of the same outcome are eligible
Apply a prespecified measurement-selection hierarchy or an analytical method that accommodates correlated results.
If two reports provide incompatible values for apparently the same result
Check populations, definitions, analyses, corrections, and other sources before applying your documented discrepancy rule.
If an expected outcome appears to be missing
Check protocols, registrations, companion papers, supplementary materials, and other available study sources before concluding that it was not measured or reported.
The objective is to reconstruct what the study actually measured and reported, then apply your review's eligibility and synthesis rules consistently. The paper in which a result happens to appear should not determine its eligibility.
07 · A Quick Checklist
What to Check When Related Papers Report Different Outcomes
Before selecting outcomes from multiple reports, check:
That all reports have been linked to the correct underlying study.
Every outcome domain relevant to your review that appears across the linked reports.
The specific measurement instrument or outcome definition used in each report.
The follow-up time associated with each result.
The analytic population and sample size behind each outcome.
Whether several eligible results represent alternative measurements of the same outcome rather than independent evidence.
Whether the protocol or registration clarifies planned outcomes, measures, and time points.
Whether apparently conflicting values can be explained by differences in population, analysis, measurement, or follow-up.
Whether your selection rules were specified independently of the observed magnitude or significance of the results.
11 · Cite this Guide
How to Cite This Guide
This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.
Recommended (Field Guide)
APA
MLA
Chicago
Copy Citation
