Related Experiment Videos
Exploring changes over time and characteristics associated with data retrieval across individual participant data
Sarah J Nevitt1, Anthony G Marson2, Becky Davie3
1Department of Biostatistics, University of Liverpool, Liverpool L69 3GL, UK sjn16@liverpool.ac.uk.
Summary
Retrieving individual participant data (IPD) for meta-analyses remains challenging, with only 25% of studies accessing all data. Factors like study design and publication venue influence data access success rates.
Area of Science:
- Medical research methodology
- Evidence synthesis
- Biostatistics
Background:
- Individual participant data (IPD) meta-analyses are the gold standard for synthesizing clinical research.
- Challenges in retrieving IPD can limit the completeness and robustness of meta-analysis findings.
- Understanding trends and factors influencing IPD retrieval is crucial for improving research synthesis.
Purpose of the Study:
- To assess if the success rate of retrieving individual participant data (IPD) for meta-analyses has increased over time.
- To identify characteristics associated with successful IPD retrieval for meta-analyses.
- To evaluate the long-term IPD retrieval experience of a specialized group (Cochrane Epilepsy Group).
Main Methods:
- Systematic review of published IPD meta-analyses (1987-2015).
- Inclusion of studies from multiple databases (Medline, CENTRAL, Scopus, Web of Science, CINAHL Plus, PsycINFO).
- Analysis of IPD retrieval rates and associated study characteristics, supplemented by a 20-year case reflection.
Main Results:
- Only 25% of 760 reviewed IPD meta-analyses retrieved 100% of eligible IPD; 43% retrieved 80% or more.
- No significant improvement in IPD retrieval rates over time was observed.
- High IPD retrieval was associated with meta-analyses including only randomized trials, having an authorship policy, fewer participants, and not being published in the Cochrane Database of Systematic Reviews.
- The Cochrane Epilepsy Group's IPD retrieval rate declined from 83% to 65% between 2005 and 2015.
Conclusions:
- Despite being the gold standard, IPD meta-analyses frequently lack access to complete datasets.
- IPD retrieval rates have not demonstrably improved, and may be declining in some contexts.
- Study characteristics, such as design and publication venue, significantly influence IPD retrieval success, highlighting areas for potential improvement in data sharing practices.