Objective: Human immunodeficiency virus (HIV) continues to be a significant public health concern for adolescents and young adults. Since 2006, the Centers for Disease Control and Prevention has recommended more aggressive routine screening for HIV for patients presenting to the emergency department (ED). Our objectives were to design and validate a survey of physician barriers toward the use of rapid HIV testing in the pediatric ED and then to use this validated tool to conduct a national survey of pediatric emergency practitioners' attitudes toward rapid HIV testing in the ED.
Methods: Survey design and initial validation steps were conducted with a panel of health care practitioners familiar to HIV testing. Several variables were identified as possible barriers toward rapid HIV testing. The survey was sent via electronic software to a national sample of pediatric emergency practitioners over 2 listservs. The previously identified variables were evaluated by factor analysis for internal consistency and homogeneity, and confirmatory factor analysis was conducted via promax and varimax rotation. All factor analyses were conducted using Stata software. Once the validation was complete, the surveys were sent to groups of pediatric emergency practitioners who had previously identified as having rapid HIV testing available in their EDs. Standard descriptive statistics were used, and group differences were evaluated with t test and χ(2) test.
Results: Four factors were identified during the validation process as being the most important barriers for rapid HIV testing in the pediatric ED: self efficacy, familiarity, external barriers, and a previously unidentified factor, which we interpreted as related to barriers to the specific environment of one's own ED. A total of 80 participants returned the final, validated survey. The participants came from 9 different pediatric emergency medicine groups (5 in areas of low rates of HIV infection, 4 in areas of high rates of HIV infection). Self-reported rates of testing were not different based on HIV infection rate in the community or the respondent's level of training. High testing was more common when a guideline was reported (39%) than when it was not (13.3%; difference, 25.7%; 95% confidence interval, 2.9%-48.5%). Of the 4 factors identified, we found statistically significant differences in scores on all 4 factors between high versus low testers, with high testers disagreeing more strongly with the various barrier questions proposed. We found no difference in the factor scores between areas of high versus low HIV infection rates.
Conclusions: Our results suggest that several factors related to perceived provider barriers are associated with rates of HIV testing in the ED and that personal factors (eg, level of training) and community HIV prevalence were not associated with rates of testing. Our results confirm what has been speculated by numerous authors and provide data to inform efforts to improve compliance with national recommendations for increased testing.