STATISTICAL METHOD
Individual Participant Data Meta-Analysis
Statistical Consulting for Participant-Level Data
We analyze participant-level records pooled from several studies, choosing between one-stage and two-stage models, separating within-study from across-study effects, and reporting the pooled estimate with its confidence interval and every harmonization decision.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
METHOD OVERVIEW
What Is IPD Meta-Analysis?
An individual participant data (IPD) meta-analysis obtains the original participant-level records from several studies, instead of their published summary estimates, and analyzes them together while keeping every participant linked to their study. The analysis can follow two stages, estimating an effect within each study and then pooling, or one stage, fitting a single model that accounts for clustering by study. Participant-level covariates allow treatment-covariate interactions to be separated into within-study and across-study components, which summary data cannot do without ecological bias. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, provides this analysis within your project design, with the relevant checks, estimates, uncertainty measures, interpretation and publication-ready reporting.
WHEN TO ENGAGE
When Is IPD Meta-Analysis Appropriate?
This analysis fits a set of studies whose investigators can share participant-level records, and these are the situations in which researchers most often ask us to plan, run or confirm it.
Studies Ready to Share Records Investigators of several trials have agreed to share participant data, and a pooled treatment effect on one consistent outcome definition is needed.
Treatment-Covariate Interactions The question is whether the effect differs by age, severity or sex, and published subgroup results are too sparse or coarse to answer it within trials.
Outcomes Reported Inconsistently Published trials use different scales, time points or definitions, so the outcome has to be derived consistently from the raw records before pooling.
Reviewing a Completed IPD Analysis Participant data were pooled and a reviewer questioned the model. Independent Statistical Review re-examines the clustering and the interactions.
ANALYTICAL APPROACH
How We Apply IPD Meta-Analysis
Within the family of Evidence Synthesis & Meta-Research, participant data are modeled only after variables, coding and study structure have been examined.
Harmonization Before Modeling Aligning definitions across studies is usually harder than fitting the model. We preserve study-specific treatment coding, missingness and outcome timing, and log every recoding.
One-Stage and Two-Stage Models A one-stage model accounts for clustering by study, since treating all participants as one dataset is incorrect. A two-stage analysis serves as a transparent check on it.
Within-Study Interactions We center each covariate within its study and separate within-study from across-study interaction, so the estimate reflects individual-level modification, not ecological bias.
Bias and Missing Studies IPD does not remove bias in the original studies, so risk of bias is still assessed, and obtained and unobtained studies are compared on design and results to judge availability bias.
DELIVERABLES
What You Receive
As part of our Meta-Analysis & Evidence Synthesis service, the participant-level analysis is delivered with the estimates, checks and files needed to report and defend it.
Data & Variable Review Variable definitions, treatment coding, missingness and follow-up timing are reviewed study by study, and a harmonization log records every recoding made.
Method-Specific Checks Study clustering, model convergence and heterogeneity of treatment effects are examined, with a two-stage analysis and published aggregate results as comparison checks.
Statistical Results The pooled treatment effect with its 95% confidence interval, heterogeneity, within-study interaction estimates with intervals, and the study-specific estimates.
Publication-Ready Tables & Figures Forest plots of study-specific and pooled effects, interaction plots and a harmonization table, formatted to PRISMA-IPD or the journal's requirements.
Interpretation Findings are written as estimates with uncertainty, separating within-study from across-study effects, with certainty in the GRADE framework and unobtained studies noted.
Reproducible Analysis Files R scripts, Stata do-files or SAS programs are provided with the harmonization code, so each estimate can be rerun, with data kept under the sharing agreement.
ALTERNATIVE METHODS
When Another Method May Be More Appropriate
We choose the method by the question and by what data can be obtained. When records are unavailable or unnecessary, or the question involves one dataset or several treatments, these methods fit better.
01
Meta-Analysis
When published estimates are enough and only an overall effect is needed, pooling aggregate data is simpler and avoids the burden of data sharing agreements.
02
Meta-Regression
When participant records cannot be obtained, study-level moderators can be examined, although conclusions about individuals then risk ecological bias.
03
Linear Mixed Models (Multilevel Modeling)
For one dataset with participants clustered in sites, a mixed model applies directly, without the harmonization of variables across studies that IPD requires.
04
Network Meta-Analysis (NMA)
When several active treatments must be compared at once, a network is needed, and participant data from some trials can be combined with published results from the rest.
CASE EXAMPLE
How We Work in Practice
An illustration of how a participant-level analysis moves from a written plan to a reportable, defensible pooled estimate.
ILLUSTRATIVE EXAMPLE · MENTAL HEALTH EVIDENCE REVIEW · UNITED STATES
A review group at a research institute pooled participant-level data from trials of a brief psychological therapy to ask whether benefit depends on baseline depression severity.
The analysis plan, filed before the data were merged, specified a one-stage random-effects model with stratified study intercepts and a within-trial interaction between treatment and baseline severity, with a two-stage analysis as a check. The trials used different depression scales, so we derived a standardized score within each trial and recorded each recoding in a harmonization log. Baseline severity was centered within each trial, which kept differences between trials out of the interaction. Two eligible trials could not share data, so we compared their published results with the pooled estimate.
Outcome: The group received the pooled treatment effect and the within-trial interaction, each with its 95% confidence interval, the two-stage comparison, the harmonization log and scripts.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
SOFTWARE
Statistical Software for Participant-Level Data
One-stage and two-stage models are run in R, Stata, SAS and IBM SPSS, with syntax or scripts provided where applicable.
RESEARCH DISCIPLINES
Participant-Level Evidence Across Research Fields
Pooling participant data is used wherever trials can be combined, each field with its own outcome measures, data sharing norms and reporting conventions.
A Defined Process for Your Analysis
We review your dataset, research question and design before any work begins, so the scope, deliverables, timeline and fixed project fee are clear from the outset.
01
Share Your Research Materials
Send your dataset, research question, study design and any existing analysis plan or output. We will let you know if anything else is needed to assess the project.
02
Receive Your Assessment
Your statistically trained Research Coordinator reviews the project and confirms the required analysis, deliverables, timeline and fixed project price before work begins.
03
Analysis & Delivery
Once you confirm the scope, an in-house statistician completes the agreed analysis and delivers the outputs, interpretation and reporting files.
RESEARCHER VOICE
What Researchers Say
Feedback from a clinical epidemiologist who wanted the model checked and the final reporting made more defensible.
“I had results from Individual Participant Data (IPD) Meta-Analysis, but the interpretation still felt shaky because of how to preserve participants' lived experience without flattening the analysis into generic themes. They checked the model, corrected the parts that needed attention, and made the final reporting much more defensible. The final output was concise, clearly interpreted, and ready to use in our research reporting.”
RESEARCH RESPONSIBILITY
Statistical Support Without Taking Over Your Research
We provide the statistical work for your participant-level analysis, and the project, its meaning and its authorship stay entirely with you.
Agreed Statistical Support We provide the agreed methodological, analytical and statistical reporting support for your participant-level analysis, within the scope confirmed before work begins.
No Intellectual Contribution OnlineSPSS claims no intellectual contribution to your research. Our work is statistical support, and authorship remains yours.
Your Research Decisions The eligibility criteria, the data sharing agreements and the conclusions remain with you, and our statistical interpretation stays within what the participant data support.
Confidential Handling Research data and project details are handled confidentially and are not published. An NDA is available where required, and can be arranged before sensitive materials are shared.
FREQUENTLY ASKED QUESTIONS
Questions About IPD Meta-Analysis
Short answers to what researchers ask before pooling participant data, from model stages and interactions to confidentiality and the assessment process.
Should I use a one-stage or a two-stage analysis?
Both usually agree when specified well, so the choice rests on the data. A two-stage analysis estimates each study separately and then pools, which is transparent and suits large studies. A one-stage model fits all participants at once, with stratified study intercepts and a common or random treatment effect, and helps with small studies or rare events. Either way, each participant stays linked to their study. We state the choice in the plan and show the other as a check.
Why separate within-trial and across-trial interactions?
A pooled interaction mixes two things: how effects differ between participants inside each trial, and how trial averages differ between trials. The second is open to ecological bias and confounding by trial-level features. Centering the covariate within each trial isolates the within-trial component, which speaks to individual-level effect modification, and we report it with its interval.
Is IPD always better than a standard meta-analysis?
No. IPD allows consistent definitions, individual-level modeling and interaction analysis, but it still carries the bias of the original studies and needs time and data sharing agreements. If some trials do not supply data, we assess whether they differ, for instance by comparing their published estimates. When only an overall effect is needed, aggregate pooling can be enough.
Can you review an analysis that has already been completed?
Yes, we can review a completed participant-level analysis before submission or a response to reviewers. We check the harmonization decisions, the handling of study clustering, the one-stage or two-stage model, the interaction estimates and the reporting, and confirm what stands or needs to change.
How do you protect my dataset and research confidentiality?
Your datasets, output files and any protocol or analysis plan are handled confidentially, and project information is not published. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Need Support With
IPD Meta-Analysis?
Share your review question, the participant-level data dictionaries and your analysis plan. We assess how the data can be pooled and confirm the approach before any work begins.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014