Committee's action has compromised continuing studies

BMJ 1995;311:1639 (16 December)

Letters

Committee's action has compromised continuing studies

EDITOR,--Douglas Carnall's report of the controversy over whetherthe third generation contraceptive pills are associated withan increased risk of thrombosis contains an error that illustrateshow bias may arise in a case-control study such as that ledby Professor Walter Spitzer.¹ Carnall states that the study"matches 980 users of the combined pill who have a cardiovascularevent with 2200 controls." In fact, the study matched 870 womenwho developed the disease of interest--that is, cardiovascularevents--with 2900 matched controls. A large percentage of thesewomen were not taking or had never taken a combined pill.

Bias could have arisen in the study if recruiting physicians,with the best of intentions, actively recruited women usingthe pill who developed the disease of interest. We believe thatthis may have happened in some instances, which would artificiallyinflate the risk. We continue to check the data for bias. Thispossibility . . . [Full text of this article]

Add to CiteULike CiteULike Add to Complore Complore Add to Connotea Connotea Add to Del.icio.us Del.icio.us Add to Digg Digg Add to Reddit Reddit Add to Technorati Technorati What's this?

Relevant Articles

Postmarketing surveillance does not catch all adverse events: Paul O'Brien
BMJ 1996 312: 577. [Extract] [Full Text]

Controversy rages over new contraceptive data: Douglas Carnall
BMJ 1995 311: 1117-1118. [Extract] [Full Text]

This article has been cited by other articles:

Prasad, R. N. V., Koh, S. C. L., Viegas, O. A. C., Shan Ratnam, S. (1999). Effects on Hemostasis After Two-Year Use of Low Dose Combined Oral Contraceptives with Gestodene or Levonorgestrel. CLIN APPL THROMB HEMOST 5: 60-70 [Abstract]
O'Brien, P. (1996). Postmarketing surveillance does not catch all adverse events. BMJ 312: 577-577 [Full text]