We evaluated predictors of completing non-targeted HIV screening in an Urgent Care center at an academic-community partnership health center serving a racially and ethnically diverse and socioeconomically disadvantaged community. We attempted to enhance the nil HIV screening rate at baseline by establishing a protocol at the UC to offer routine screening. The overall rate of testing at the end of the study was 23.5%, within the range reported by prior studies performed in ED and UC settings [3, 7–9]. In the process, we identified a previously unrecognized, independent predictor for completing HIV screening: the health care professional who treated the patient in the UC. Of the responders to the survey, about half of providers were neutral or opposed to the UC setting being the appropriate setting for a screening test, but as illustrated by the heat map of survey responses [Figure 3b], a majority of respondents favored continuing screening. Altogether, provider opinion, whether in favor or opposition to screening, likely contributes to broad range of screening completion rates by provider. This suggests that successful implementation of non-targeted screening is unlikely to occur until there is buy-in by providers in UCs and EDs, despite strong national-level guidelines.
Providers may have offered, completed or deferred screening based on perceived time constraints or because they felt screening was of low value to the patient encounter because of lack of perceived risk, among other factors. While the latter runs counter to the principle of implementing non-targeted testing in acute-care settings, the finding is consistent with previous studies [17, 18]. Other studies have demonstrated that physicians have variable knowledge about the CDC recommendations, and when they are informed, often have doubts about guidelines, fear of mistrust from their patients, or fear of what to do with a positive result [17–19].
RN responders to the survey were largely neutral about their role as it related to the screening pilot and were relatively negative with respect to whether they would support non-targeted HIV screening moving forward. In this model of HIV screening, the particular nurse screening a patient could strongly influence the patient’s response as to whether or not he/she was amenable to screening. Since willingness to be screened was evaluated by the triage RN, the group accepting screening was likely selected based on a combination of RN opinion on screening in general, and how the RN did or did not resolve their own perceptions and patient perceptions about who is at risk for HIV.
Based on our results and a recently published study reporting that RN-initiated screening has lower uptake than provider-initiated screening , we do not recommend that future non-targeted screening programs use RN-initiated screening protocols in the ED or UC setting. However, because HIV screening inherently involves RNs for other steps in the screening process, it is important to include nursing staff in education and design of HIV screening programs. In this pilot we identified several points during the encounter where improvements could be made to increase both screening uptake and test completion. As depicted in Figure 2, in the flow through the clinic visit, there were clear points where testing opportunities were lost; while there may be several patient factors impacting uptake of screening, our pilot suggests that the person offering the screening is extremely influential in the patient’s testing decision. Not all patients were enrolled in the pilot and offered HIV screening in triage. A lower number of patients than expected were agreeable to screening, possibly reflecting initiation of screening by triage RNs.
Our study also affirmed the previously documented findings that younger age groups, Hispanics, and Blacks were more likely to complete HIV screening [11, 16].
Our study has limitations. We performed a single center study, with the intent to evaluate predictors for screening at our site only; patients may have had previous tests performed with outside providers and we relied on their self-report of these tests. Furthermore, a large proportion of visits meeting criteria for enrollment in the study were not enrolled at triage. As the pilot was performed in Massachusetts prior to April 2012, the legal obligation to use written informed consent forced an opt-in strategy. This created an additional barrier to screening that is no longer required by most states. Settings where point-of-care testing is used and where written consent is no longer required may have greater screening uptake.