[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
September 13, 1995

From Science to PracticeMeta-analyses Using Individual Patient Data Are Needed

Author Affiliations

From the Health Services Research Unit, National Institute of Public Health, Oslo, Norway (Dr Oxman); the Clinical Trial Service Unit and Imperial Cancer Research Fund Cancer Studies Unit, University of Oxford (England) (Dr Clarke); and the Medical Research Council Cancer Trials Office, Cambridge, England (Dr Stewart).

JAMA. 1995;274(10):845-846. doi:10.1001/jama.1995.03530100085040

In this issue of THE JOURNAL, Jeng et al1 report discrepancies between the results of meta-analyses using aggregate data available from the literature (MAL) and meta-analyses using individual patient data (MAP). This may be seen as further grounds for mistrusting meta-analysis in general2,3 and MAL in particular,4 but before deciding that all meta-analyses are an exercise in "mega-silliness,"5 consider the alternatives.

See also p 830.

The key characteristic of MAPs is the use of individual patient data (IPD). Meta-analyses using aggregate data also can, and frequently do, include unpublished data, so the term IPD meta-analyses is preferable to avoid confusing issues relating to using IPD and issues relating to using unpublished data.

IPD meta-analyses require international cooperation between the individuals and groups who have conducted relevant trials and a considerable amount of time, personnel, and financial resources.6 These global collaborative endeavors have yielded important information

First Page Preview View Large
First page PDF preview
First page PDF preview