Individual participant data meta-analysis is a research method that combines information from individual participants across multiple studies to better understand how treatments work and whether outcomes differ between groups of people.
In Post-Traumatic Stress Disorder (PTSD) research, this approach allows researchers to examine treatment outcomes in greater detail than is usually possible when combining published study averages.
Rather than relying only on summary statistics, researchers obtain anonymized participant-level information from the original studies. They can then analyze factors such as symptom severity, age, treatment response, and dropout rates across multiple trials.
This method has been used to examine Eye Movement Desensitization and Reprocessing (EMDR) alongside other psychological treatments for PTSD.
The main advantage is that researchers can investigate not only how treatments compare overall, but also whether certain participant characteristics are associated with different treatment outcomes.
However, individual participant data meta-analysis depends on access to reliable, sufficiently complete information from the original studies. Its findings still need to be interpreted in light of those studies' quality and limitations.
What Does Individual Participant Data Meta-Analysis Mean?
A meta-analysis is a research method that combines results from multiple studies investigating a similar question.
For example, several randomized controlled trials may compare EMDR with another psychological treatment for PTSD.
Each trial might report how much participants' PTSD symptoms changed after treatment.
A conventional meta-analysis combines those reported results to estimate the overall difference between treatments.
An individual participant data meta-analysis (IPD meta-analysis) takes a different approach.
Instead of collecting only published averages and treatment-effect estimates, researchers seek the original participant-level datasets from the studies.
These datasets may include:
Participants' ages and other background characteristics.
PTSD symptom scores before and after treatment.
The treatment each participant received.
Whether participants completed treatment.
Whether participants met the study's criteria for treatment response or remission.
Other clinical characteristics recorded during the original trials.
The information is typically anonymized before being shared, meaning researchers work with participant records rather than personally identifying information.
Combining these records allows researchers to conduct more detailed analyses across studies.
How Is Individual Participant Data Meta-Analysis Different From a Traditional Meta-Analysis?
The main difference is the level of information researchers analyze.
A traditional meta-analysis generally uses results already summarized in published research papers.
An IPD meta-analysis uses the underlying participant-level data when those records can be obtained.
| Research feature | Traditional meta-analysis | Individual participant data meta-analysis |
|---|---|---|
| Main information used | Published study-level results | Original participant-level datasets |
| Comparison of treatments | Combines estimated treatment effects | Reanalyzes participant outcomes across studies |
| Participant characteristics | Often limited to published summaries | Can examine characteristics recorded for individual participants |
| Outcome measurements | Relies largely on reported measures | Can sometimes standardize and recalculate outcomes |
| Missing information | Limited by published reporting | May allow more detailed analysis of missing participant values |
| Data access | Usually possible from publications | Requires access to original datasets |
| Research questions | Overall treatment effects and study-level differences | Overall effects and more detailed participant-level relationships |
Both approaches can provide valuable evidence.
An IPD meta-analysis is particularly useful when researchers want to understand whether treatment outcomes differ according to individual characteristics rather than only comparing average results.
However, access to participant-level data does not automatically make a study more reliable. The quality, completeness, and representativeness of the included evidence remain important.
How Do Researchers Conduct an Individual Participant Data Meta-Analysis?
An IPD meta-analysis usually involves several stages, beginning with a systematic review of the existing research.
1. Identifying relevant studies
Researchers search scientific databases for studies addressing a specific treatment question.
For example, they may identify randomized controlled trials comparing EMDR with other psychological treatments for PTSD.
A randomized controlled trial assigns participants to different treatment conditions using a random process.
Researchers establish eligibility criteria before selecting studies for inclusion.
2. Requesting participant-level datasets
After identifying eligible studies, researchers contact the original investigators to request anonymized participant data.
These requests may involve data-sharing agreements and procedures for protecting participant privacy.
Not every research team can provide its original dataset.
Some records may no longer be available, while others may be subject to restrictions on data sharing.
3. Checking and organizing the information
Researchers examine the datasets for missing values, inconsistencies, and differences in how information was recorded.
For example, one PTSD trial may use a symptom questionnaire with a different scoring range from another trial.
Researchers may need to place the scores on a common measurement scale before combining them.
4. Combining the datasets
The information is organized for statistical analysis.
Researchers retain information about which original study each participant belonged to.
This matters because participants from the same study may share characteristics related to recruitment, treatment delivery, or assessment methods.
5. Analyzing treatment outcomes
Researchers compare outcomes across treatment groups while accounting for differences between the original studies.
They can also investigate whether specific participant characteristics are associated with treatment results.
6. Reporting the findings and limitations
A reliable analysis explains which studies were included, how much data were obtained, what statistical methods were used, and whether important information was unavailable.
These steps help readers understand how the overall conclusions were reached.
Why Is This Method Useful in PTSD Treatment Research?
PTSD research includes studies involving different populations, experiences, and treatment approaches.
Some trials focus on military veterans, while others involve survivors of accidents, interpersonal violence, or other traumatic events.
Participants may also differ in age, symptom severity, treatment history, and other characteristics.
Individual studies are not always large enough to investigate how these differences relate to treatment outcomes.
An IPD meta-analysis can combine suitable datasets to explore these questions more systematically.
Examining individual differences
Researchers can investigate whether treatment outcomes vary according to characteristics measured in the original studies.
Improving the consistency of analyses
Participant-level data may allow researchers to calculate treatment outcomes using more consistent definitions across studies.
Investigating treatment completion
Researchers can examine whether certain participant characteristics are associated with completing or discontinuing treatment.
Exploring predictors of improvement
Participant-level information may help identify characteristics associated with greater or smaller symptom changes.
These advantages make IPD meta-analysis useful for understanding treatment outcomes beyond a single overall average.
What Has Individual Participant Data Meta-Analysis Found About EMDR?
A directly relevant study was published by Wright and colleagues in Psychological Medicine in 2024.
The researchers conducted a systematic review and individual participant data meta-analysis comparing EMDR with other psychological treatments for adults with PTSD.
The study investigated four main outcomes:
PTSD symptom severity: How strongly participants experienced PTSD symptoms following treatment.
Treatment response: Whether participants showed a substantial reduction in symptoms.
Remission: Whether participants no longer met the study's criteria for PTSD at the post-treatment assessment.
Treatment dropout: Whether participants discontinued treatment before the final assessment.
The researchers also examined whether available participant characteristics influenced treatment outcomes.
How many studies and participants were included?
The systematic search identified 15 eligible randomized controlled trials.
However, the researchers obtained usable individual participant data from only eight trials.
The resulting IPD meta-analysis included 346 participants.
The comparison treatments included trauma-focused cognitive behavioral therapies, relaxation therapy, Emotional Freedom Techniques, and other psychological approaches.
The researchers analyzed outcomes using statistical methods that accounted for participants being grouped within their original trials.
What did the study find?
The main analysis found no statistically significant differences between EMDR and the other psychological treatments for:
Reducing PTSD symptom severity.
Achieving treatment response.
Reaching PTSD remission.
Treatment dropout.
Both EMDR and the comparison treatments were associated with improvements in PTSD symptoms.
The absence of a statistically significant difference means the analysis did not establish that one treatment consistently outperformed the others on these outcomes.
However, it does not prove that the treatments are identical or equally effective for every person.
The findings apply to the available studies, participants, treatment comparisons, and statistical methods used in the analysis.
What Did the Study Find About Individual Differences in Treatment Outcomes?
One reason for using participant-level data was to investigate whether specific characteristics were associated with differences in EMDR outcomes.
These characteristics are sometimes called moderators.
A moderator is a factor that may influence the size or direction of a treatment effect.
For example, researchers may investigate whether the difference between two treatments varies according to a participant's age or initial symptom severity.
The 2024 study explored several characteristics, including age, gender, employment status, relationship status, and available clinical information.
Employment status
The researchers found that employment status was associated with some treatment outcomes.
Among participants receiving EMDR, those who were unemployed had higher PTSD symptom severity after treatment than those who were employed in the relevant analysis.
Unemployed participants receiving EMDR were also less likely to meet the study's treatment-response criterion.
However, the symptom-severity association was identified in the analysis of participants with complete outcome data and was not consistently demonstrated across all analyses.
The findings do not establish that employment itself causes a difference in treatment response.
Other circumstances associated with employment status may also be relevant.
Gender and treatment dropout
The researchers found that male participants receiving EMDR were more likely to discontinue treatment than female participants in the analyzed sample.
This was an exploratory finding based on the available trial data.
It does not mean that men generally respond less favorably to EMDR or that gender determines whether someone will complete treatment.
The study's limited sample size means further research is needed to establish how consistently this association appears.
Other participant characteristics
The researchers investigated additional characteristics but did not identify consistent statistically significant moderating effects for the other variables examined.
Not every original trial provided every characteristic, which limited the analyses researchers could perform.
These findings demonstrate the potential value of participant-level research while also showing why subgroup conclusions require caution.
What Is the Difference Between a Predictor and a Moderator?
Researchers often examine predictors and moderators when studying PTSD treatments.
Although the terms are related, they describe different research questions.
Predictor
A predictor is a characteristic associated with an outcome.
For example, researchers may find that people who begin treatment with higher PTSD symptom scores tend to have higher scores after treatment.
This relationship could occur regardless of which treatment they receive.
Moderator
A moderator is a characteristic associated with a difference in how treatments compare.
For example, researchers might investigate whether EMDR produces a different relative effect compared with another therapy among participants with different baseline symptom levels.
Finding that one subgroup has better outcomes does not automatically establish moderation.
Researchers need to test whether the difference between treatments changes across the participant characteristic.
This distinction is important because IPD meta-analysis can support more detailed examinations of treatment moderators than are often possible using published study averages.
How Can Individual Participant Data Help Researchers Understand PTSD Symptom Severity?
Participants in PTSD treatment trials may begin treatment with different levels of symptoms.
Some report relatively moderate difficulties, while others report greater symptom severity.
Researchers can use participant-level data to examine whether initial symptom severity is associated with later outcomes.
In the 2024 EMDR meta-analysis, higher PTSD symptom severity before treatment was associated with higher symptom severity afterward.
This was a predictive relationship.
It does not mean that participants with more pronounced initial symptoms cannot improve.
Instead, it indicates that baseline symptom levels were statistically related to the symptom levels observed after treatment.
IPD meta-analysis can help researchers examine these patterns across multiple studies while considering other available characteristics.
Can Individual Participant Data Meta-Analysis Show Which PTSD Treatment Is Best for Each Person?
Not necessarily.
An IPD meta-analysis can help identify patterns that may eventually support more personalized treatment decisions.
However, identifying a statistical association is different from creating a reliable method for predicting an individual's outcome.
For example, a study may find that a particular characteristic is associated with treatment response.
That finding does not automatically establish that everyone sharing the characteristic will respond similarly.
Individual treatment outcomes may be influenced by several factors that were not measured or included in the original trials.
Before a research finding can reliably guide personalized treatment selection, it generally needs to be confirmed in additional studies and evaluated for its predictive accuracy.
The goal of participant-level research is to improve understanding of treatment differences, not to reduce individual experiences to a single statistical category.
What Are One-Stage and Two-Stage IPD Meta-Analyses?
Researchers can analyze individual participant data using two common statistical approaches.
One-stage analysis
In a one-stage approach, researchers analyze participant-level data from all available studies together within one statistical model.
The model accounts for the fact that participants come from different studies.
This approach can examine overall treatment effects and relationships involving participant characteristics.
Two-stage analysis
In a two-stage approach, researchers first analyze the participant-level data within each study separately.
They then combine the resulting study-level estimates using meta-analysis methods.
Both approaches can produce reliable results when appropriately designed.
The choice depends on the research question, available data, statistical assumptions, and characteristics of the included studies.
Why does the distinction matter?
In the 2024 EMDR study, the main one-stage analysis did not identify a statistically significant difference in PTSD remission between EMDR and the other treatments.
A two-stage analysis did find a statistically significant remission difference favoring EMDR.
The researchers considered the one-stage result more reliable in this setting because of the small number and size of the included trials.
This illustrates why readers benefit from examining the analysis methods rather than relying on a single reported result.
What Are the Limitations of Individual Participant Data Meta-Analysis?
Although the method provides detailed information, it also has important limitations.
1. Original datasets may be unavailable
Researchers depend on access to participant-level information from the original study teams.
In the 2024 EMDR analysis, usable data were obtained from eight of 15 eligible trials.
The seven unavailable datasets represented a substantial portion of the identified evidence.
This creates a possibility of data availability bias, meaning the studies that provide individual data may differ from those that do not.
2. Combining data does not remove problems in the original studies
An IPD meta-analysis cannot automatically correct weaknesses in the trials it includes.
For example, differences in treatment delivery, outcome measurement, participant recruitment, or trial design may still affect the findings.
Researchers therefore assess the quality and potential bias of the original studies.
3. Some participant characteristics may be missing
Not every trial collects the same background or clinical information.
In the 2024 study, insufficient reporting prevented researchers from examining some potentially relevant variables, including symptom duration and number of treatment sessions attended.
4. Subgroup findings may be uncertain
Even when several trials are combined, particular participant groups may remain small.
This can limit the reliability of analyses examining treatment differences between subgroups.
5. Statistical findings depend on the analysis methods
Researchers may use different approaches to manage missing outcomes, standardize measurements, or estimate treatment effects.
Clear reporting helps readers understand how these decisions influence the conclusions.
6. The findings may not apply to every person with PTSD
The 2024 EMDR analysis included predominantly participants with chronic PTSD.
Its results may not apply in the same way to people with more recent traumatic experiences or to populations not well represented in the original studies.
These limitations explain why the method is valuable but does not automatically produce definitive answers.
How Do Researchers Assess the Quality of an IPD Meta-Analysis?
Researchers use established reporting guidance to improve the transparency and consistency of IPD meta-analyses.
One important framework is the Preferred Reporting Items for Systematic Reviews and Meta-Analyses of Individual Participant Data (PRISMA-IPD).
Published in 2015 by Stewart and colleagues in the Journal of the American Medical Association, this guidance addresses the reporting requirements specific to participant-level research.
It emphasizes clear descriptions of how studies were identified, how participant data were obtained, and how the information was analyzed.
| Quality consideration | What readers can look for |
|---|---|
| Study selection | Clear criteria for identifying eligible trials |
| Data availability | Number of studies providing and not providing participant data |
| Data checking | Procedures for identifying errors and inconsistencies |
| Outcome definitions | Clear explanations of symptom improvement, response, and remission |
| Statistical methods | Description of how participant data were combined and analyzed |
| Moderator analyses | Explanation of which participant characteristics were examined and whether analyses were planned |
| Missing information | How missing participant values and unavailable datasets were handled |
| Limitations | Discussion of uncertainty, bias, and generalizability |
The guidance also recommends that researchers explain whether subgroup analyses were specified in advance or developed after examining the data.
This distinction helps readers understand whether a result reflects a planned research question or an exploratory finding that needs further investigation.
How Does IPD Meta-Analysis Contribute to the Broader Evidence for EMDR?
EMDR has substantial research support as a treatment for PTSD.
It is also commonly used for anxiety, depression, grief, addictions, and other distressing experiences.
Most treatment research examines whether interventions reduce symptoms or improve functioning within a particular group.
Conventional systematic reviews and meta-analyses help bring those findings together.
Individual participant data meta-analysis adds another level of detail.
It can investigate whether differences in participant characteristics influence treatment outcomes and whether findings remain consistent after applying common statistical methods across studies.
The 2024 study by Wright and colleagues provides an example of this contribution.
It found no statistically significant difference between EMDR and other psychological treatments in its main comparisons while identifying exploratory relationships involving employment status and gender.
These findings complement the broader evidence on PTSD treatment.
They do not replace clinical trials, establish that one therapy is preferable for every individual, or determine the effectiveness of treatment formats that were not included in the original studies.
What Questions Could Future IPD Meta-Analyses Explore?
Future research could combine larger participant-level datasets from PTSD treatment trials.
Greater access to original study information would allow researchers to explore whether treatment outcomes differ across a wider range of participant characteristics.
Researchers could also investigate how baseline symptoms, trauma history, treatment completion, and other measured factors relate to longer-term outcomes.
Another important area involves consistency in outcome measurement.
When studies use comparable measures and share participant-level information, researchers may be better able to investigate changes in specific PTSD symptoms.
Further research could also examine differences between treatment formats, including therapist-guided online treatment and in-person treatment, when suitable comparative trial data are available.
The usefulness of these investigations will depend on study quality, data availability, and whether the relevant characteristics were measured consistently.
What Is the Main Takeaway?
Individual participant data meta-analysis combines information from individual participants across multiple studies rather than relying only on published study averages.
In PTSD treatment research, this allows researchers to examine overall treatment outcomes and investigate whether individual characteristics are associated with differences in improvement, remission, or treatment completion.
A 2024 meta-analysis comparing EMDR with other psychological treatments identified 15 eligible randomized controlled trials and obtained usable participant-level data from eight, involving 346 participants.
The main analysis found no statistically significant differences between EMDR and the comparison treatments in PTSD symptom severity, treatment response, remission, or dropout.
The researchers also identified exploratory associations involving employment status and gender, although the available data limited the certainty of these findings.
The main value of IPD meta-analysis is that it can help researchers look beyond average treatment effects and investigate why outcomes may differ between participants.
However, it does not guarantee more accurate conclusions when important datasets are missing or the original studies have limitations.
By combining participant-level information with careful research methods, IPD meta-analysis can contribute to a more detailed understanding of EMDR and other PTSD treatments.
References
Wright, S. L., Karyotaki, E., Cuijpers, P., et al. (2024). EMDR v. other psychological therapies for PTSD: A systematic review and individual participant data meta-analysis. Psychological Medicine, 54(8), 1580–1588.
https://doi.org/10.1017/S0033291723003446
Wright, S. L., Karyotaki, E., Bisson, J. I., et al. (2022). Protocol for individual participant data meta-analysis of interventions for post-traumatic stress. BMJ Open, 12(2), e054830.
https://doi.org/10.1136/bmjopen-2021-054830
Stewart, L. A., Clarke, M., Rovers, M., et al. (2015). Preferred Reporting Items for a Systematic Review and Meta-analysis of Individual Participant Data: The PRISMA-IPD Statement. JAMA, 313(16), 1657–1665.
https://doi.org/10.1001/jama.2015.3656
PRISMA Statement. Individual Participant Data reporting guidance.
