An assessment of population-based screening guidelines versus clinical prediction rules for chlamydia and gonorrhea case finding

Abstract

IntroductionMuch remains to be learned regarding the epistemology and utility of guidelines and clinical prediction rules (CPR), as well as the extent to which knowledge about risk at a population level might be pertinent to any given patient in terms of case finding accuracy. In the current paper, we offer an empirical examination that juxtaposes population-based guidelines and CPR for sexual health decision-making. Materials and methodsWe analyzed electronic medical records from asymptomatic patient visits involving tests for chlamydia or gonorrhea between 2000 and 2012 at nine publicly funded STI clinics in British Columbia to compare the case-finding accuracy for infection risk under two scenarios: (1) if the population had been screened using the Public Health Agency of Canada (PHAC) screening guidelines for chlamydia and gonorrhea; or (2) if the population has been screened using a CPR. Performance metrics evaluated included the area under the ROC curve (AUC). ResultsIn total, 35,818 individuals met the study inclusion criteria. The overall infection rate was 3.0%. Using the PHAC guidelines, the discriminatory performance of using any versus no risk factors and counts of risk factors were: AUC=0.55, 95% CI: 0.54–0.56 and AUC=0.64, 95% CI: 0.63–0.66, respectively. The model used to derive the CPR demonstrated good discrimination (AUC=0.73, 95% CI: 0.71–0.74). ConclusionsThe current paper provides empirical evidence that demonstrates that population-based guidelines may not necessarily be a perfect fit for application at the individual level. Thus, we recommend risk estimation algorithms for use in sexual health services and programs.