Guidance for individual participant data meta-analysis

WWARN’s study groups combine multiple study data through pooled analyses to investigate specific scientific questions. By combining studies, they are able to draw conclusions that would not be possible with any of the smaller individual studies. Nevertheless, there are many challenges with pooled analyses and how they are reported. These individual studies can include: both randomised and non-randomised information and data collected using different methods, and may use different study designs.

Systematic reviews and meta-analyses of individual participant data (IPD) have long been recognised as a gold standard approach. They offer many advantages over analyses that use aggregated data extracted from publications. These include: increased opportunity to identify and include unpublished studies and obtain full outcome data (thereby reducing the potential for bias arising from the absence of unpublished studies and unreported outcomes); checking and transforming data to common scores or measures; standardising analyses across studies; and undertaking more flexible and powerful statistical analyses including exploration of potential effect modifiers.

At present, only a small percentage of all systematic reviews and meta-analyses follow the IPD approach. However, the increasing interest in data sharing and ‘big data’ is likely to lead to more frequent use of IPD meta-analyses to review therapeutic techniques. Therefore, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement, published in 2009, has been adapted to provide specific guidance relating to individual participant data. The ‘Preferred Reporting Items for a Systematic Review and Meta-Analysis of Individual Participant Data: The PRIMSA-IPD Statement’, recently published in The Journal of the American Medical Association, provides a reliable and consistent set of gold standard guidelines.

“The reporting of WWARN’s research has been following the principles of the PRISMA guidelines, and we welcome the publication of the PRISMA-IPD Statement,” says Prof Philippe Guérin, Director of WWARN. “Following this specific statement will provide clear guidance and consistency for future meta-analyses. It also clearly illustrates the challenges involved in optimally conducting IPD meta-analyses, challenges that we are willing to meet to maintain our high standard of research output.”

The original PRISMA Statement includes a 27-item checklist and flow diagram, predominantly used for systematic reviews and meta-analyses of randomised trials using aggregate data. The PRISMA-IPD Statement extends these guidelines by including three new criteria relating to individual participant data that help researchers to check the integrity of the data, report any issues, and explore effect, for example whether certain groups of individuals benefit more from the intervention than others. It also covers issues relating to study design, such as methods of obtaining the individual participant data, and handling trials for which individual data were unavailable.

“The IPD approach has an established history in reviews of interventions in cancer and cardiovascular disease, and is being used increasingly across a broadening range of healthcare areas,” says Prof Lesley Stewart, Chair of the PRISMA Steering Group and Director of the Centre for Reviews and Dissemination at the University of York, UK. “The aim of these guidelines is to help researchers undertaking systematic reviews and IPD meta-analyses to report their findings in a full and transparent manner and in a way that can be understood by those who need to know about and act on their findings. We encourage all authors and peer reviewers of such research to use the checklist, and are delighted that the guidelines are being championed by WWARN and used to support meta-analyses in malaria research.”

Dr Kasia Stepniewska, Head of Statistics, at WWARN added: Our Lumefantrine Pharmacokinetic/Pharmacodynamic and ACT Africa Baseline Study Groups, which are currently under peer review, have already completed their reporting using these guidelines. Going forward, all WWARN studies will adhere to these thorough and transparent reporting criteria. These guidelines enable us to ensure that our reports are standardised to the same level as other pooled analyses of patient data, helping to ensure the reproducibility of our work and careful interpretation of the findings.”

Related publications:

Lesley A. Stewart et al. Preferred Reporting Items for a Systematic Review and Meta-analysis of Individual Participant Data: The PRISMA-IPD Statement. JAMA. 2015;313(16):1657-1665. doi: 10.1001/jama.2015.3656.