[Skip to Content]
[Skip to Content Landing]
December 12, 1986

Vaginal Spermicides and Congenital Disorders: The Validity of a Study-Reply

Author Affiliations

Boston Collaborative Drug Surveillance Program Waltham, Mass
University of Massachusetts Medical School Worcester

JAMA. 1986;256(22):3095-3096. doi:10.1001/jama.1986.03380220061019

In Reply.—  Dr Watkins has selectively reclassified the exposure status of exposed cases without engaging in a similar effort for exposed noncases. The technique is one used in advocacy, but it is not good science.The ease with which clinical notes can be misconstrued even by a disinterested observer led us originally to prefer records of acquisition of spermicides as an impartial, though admittedly imperfect, proxy for exposure. Our classification of women's exposure to spermicides was based on a consideration of the typical time between purchases of spermicides at Group Health, which we took to be a reasonable indicator of the duration of potential use. Since pharmacy records do not specify dates of use or nonuse, there will inevitably be nondifferential misclassification of exposure in a study based on pharmacy data. The same is true of studies based on clinic records. In either case, the effect of the error will