Assessments of Girls' Genital Findings and the Likelihood of Sexual Abuse

OBJECTIVES: To measure agreement about genital examination findings among physicians who rate themselves as skilled in evaluating children for suspected sexual abuse, to compare these physicians' descriptions and interpretations with consensus standards developed by an expert panel, and to investigate the effects of physician and case characteristics on agreement. STUDY DESIGN: Questionnaires including 7 simulated cases, each consisting of a brief history and 1 photograph of a girl's genitalia, were mailed to random samples of 2 groups: the members of 4 physician organizations concerned with child abuse or pediatric gynecology, and pediatricians at large. Among the surveyed physicians who rated their own skill in evaluating cases of suspected sexual abuse as higher than average, we measured agreement, both overall and between those with the most and with less clinical experience, and assessed their conformity with consensus standard descriptions and interpretations. RESULTS: We received responses from 548 (50.9%) of 1076 physicians; 414 responses (75.5%) were analyzable. Two hundred six physicians (50%) rated themselves as skilled in assessing children for sexual abuse. On average, 45% of these physicians' descriptions and 72.6% of their interpretations conformed with the consensus standards. In 4 cases, between 5% and 20.7% of these physicians described genital findings that the expert panel had considered absent from the photographs. Conformity with standard interpretations tended to be higher in cases with photographs concordant with the accompanying, unambiguous histories (P=.06). The most experienced physicians resembled the expert panel more closely than did the less experienced self-rated skilled physicians in interpreting 3 simulated cases (P< or =.001). CONCLUSIONS: Assessments of girls' genital findings by physicians who rate themselves as skilled in examining children for suspected sexual abuse often differ. In some cases, among physicians who all rate themselves as skilled, assessments made by very experienced physicians may conform more closely to consensus standards than do assessments made by less experienced physicians.