To the Editor.—
This letter is to note major defects in the article by Doherty and Haven, "Medical Malpractice and Negligence: Sociodemographic Characteristics of Claimants and Nonclaimants" (238:1656, 1977). Simply stated, the authors' control group is entirely inadequate in view of preexisting knowledge about determinants of malpractice and the sociodemographic correlates of use of surgical procedures. Additionally, the statistical analysis of the data is inadequate. The defects are such that no conclusions should be drawn from the study.If one desires to assess the independent effects of sociodemographic characteristics on the risk of litigation, then it is critical to adjust for medical characteristics known to relate to the probability of a suit. Accordingly, the appropriate control group should not consist of a random sample of all hospital admissions, but rather, a sample matched, at the very least, for such broad characteristics as admitting service (orthopedics, pediatrics, medicine, and general surgery)