The medical diagnosis of child sexual abuse is based on a combination of factors, including disclosures by the child, specific physical examination findings and sexually transmitted infections. Genital infections with Neisseria gonorrhoeae (GC) are regarded as definitive evidence of sexual contact in children, but current microbiologic culture methods have low sensitivity for detecting vulvovaginal GC infections in prepubertal girls.
To compare the three strategies of culture alone, nonculture alone (Gen-Probe PACE 2) and sequential testing for genital GC (nonculture followed by culture of nonculture positives), we conducted a decision analysis to identify optimal testing strategies using published test sensitivities, specificities and pretest probabilities of genital GC. We surveyed community child welfare professionals to determine values for each potential health outcome.
Sequential testing yielded a higher overall health outcome value (0.855) than culture (0.834) or nonculture methods (0.838). These results were insensitive to changes in GC prevalence, test characteristics and individual health outcome values over clinically important ranges. With a baseline prevalence of 10%, more children (4 per 100) with GC infection were missed by culture than were inaccurately labeled as having GC infection (0.5 per 100) by nonculture methods. Sequential testing yielded fewer false negative results (0.8 per 100) than culture and had no false positives.
With the use of clinical decision analysis, sequential testing using nonculture methods followed by culture was preferred to culture or nonculture methods alone in the evaluation of GC in prepubertal girls with genital discharge. Current recommendations for GC testing should be reassessed with clinical decision analysis as new test methods or strategies become available.
From DeVos Children’s Hospital Child Protection Team, Michigan State University, Grand Rapids (VJP); and Department of Epidemiology, Michigan State University, East Lansing (MJR), MI.
Accepted for publication March 14, 2003.
Address for reprints: Vincent J. Palusci, M.D., M.S., DeVos Children’s Hospital, 100 Michigan Street, MC-178, Grand Rapids, MI 49503. Fax 616-391-3206; E-mail Vincent.Palusci@Spectrum-Health.org.