Letter

Unit of analysis errors should be clarified in meta-analyses

BMJ 2003; 326 doi: 10.1136/bmj.326.7385.397 (Published 15 February 2003)
Cite this as: BMJ 2003;326:397.1

Access to the full text of this article requires a subscription or payment. Please log in or subscribe below.

  1. Ruth E Thomas (r.e.thomas@abdn.ac.uk), research fellow,
  2. Craig R Ramsay, senior statistician,
  3. Laura McAuley, review group coordinator,
  4. Jeremy M Grimshaw, director, clinical epidemiology programme
  1. Health Services Research Unit, University of Aberdeen, Aberdeen AB25 2ZD
  2. Cochrane Effective Practice and Organisation of Care Group, Cochrane Effective Practice and Organisation of Care Group, Institute of Population Health, University of Ottawa, 1 Stewart Street, Ottawa, ON, Canada K1N 6N5
  3. Institute of Population Health, University of Ottawa, Ottawa Health Research Institute, 1053 Carling Avenue, Ottawa, ON, Canada K1Y 4E9

    EDITOR—Weingarten et al present a comprehensive study in what is a complex area of research.1 We were, however, unclear whether any of the included primary studies had unit of analysis errors and how the authors dealt with such studies in their meta-analysis.

    Unit of analysis errors occur in cluster randomised trials when individual patients' data …

    Access to the full text of this article requires a subscription or payment

    Subscribe

    Article access

    Article access for 1 day

    Purchase this article for £20 $30 €32*

    The PDF version can be downloaded as your personal record

    * Prices do not include VAT

    Rapid responses

    THIS WEEK'S POLL

    BMJ most popular