In this urban, essential hospital located in a city with high HIV burden among PWID, an average of 6.4% of patients seen by the ACS with plausible injection drug use were prescribed PrEP or PEP at time of discharge. This proportion increased over time, from 4.2% in 2020 to 7.5% in 2022. At lowest, it was on par with the higher end of national averages, and at highest, exceeded published rates [1, 10, 11]. Secondarily, the ACS was involved in the care of the vast majority (83.9%) of admitted patients who received PrEP or PEP at discharge.

To our knowledge, this is the second study describing PrEP and PEP prescribing to patients seen by an ACS [13], and demonstrates promise for leveraging the inpatient ACS to increase uptake of pharmacologic HIV prophylaxis in an at-risk population with high rates of overdose, homelessness, and polysubstance use. This complements the ways in which the ACS supports other evidence-based forms of HIV prevention, including harm reduction counseling [27], provision of new injection equipment [28], and initiation of gold-standard MOUD, such as methadone and buprenorphine [29]. A high proportion (92.9%) of patients in our study who were seen by the ACS and received PEP or PrEP also received MOUD at time of hospital discharge. Additional data on ACS harm reduction services are needed to fully characterize contributions to HIV prevention.

Secondarily, we demonstrated that the ACS was involved in most hospitalizations that included a PrEP or PEP discharge prescription. While the ACS does not write orders for PrEP and PEP directly, our EMR note template for the ACS explicitly prompts providers to ask patients about eligibility for PrEP and PEP and provide recommendations accordingly. Furthermore, our addiction medicine fellows, who work on the ACS, receive formal teaching on PrEP and PEP for PWID. This is important in light of data demonstrating that many medical providers hesitate to provide PrEP or PEP to PWID due to lack of comfort or knowledge [19,20,21]; anecdotally, there is also ambiguity about which specialty should “own” PrEP or PEP, particularly in the inpatient setting. Though we cannot demonstrate causality between ACS involvement and PrEP/PEP prescribing, our data lends support to the value of the ACS in stepping up to fill a potential void, and the inclusion of PEP and PrEP within addiction specialists’ scope of practice. As such, other institutions aiming to increase PrEP and PEP prescribing could consider similar, systematic interventions to support the role of an ACS in this space (e.g., developing note templates, implementing formal education on PrEP and PEP in PWID).

Inpatient hospitalization has been previously described as a “reachable moment” to provide care to patients who may have difficulty accessing traditional outpatient medical settings, including as a means to link patients to MOUD after nonfatal overdoses and other complications of OUD [12, 30]. Incorporating pharmacologic HIV prevention into the scope of practice of an ACS capitalizes on this opportunity, particularly when patients are already hospitalized with infectious complications of SUD, as in our sample, where over 50% were admitted for an injection-related infection. While PrEP has been traditionally considered an outpatient medication, for high-risk hospitalized patients with SUD, there are benefits to PrEP initiation prior to the post-acute care transition [18].

This study has several limitations. First, the goal of our study was primarily descriptive and we therefore do not include a comparison time period when the ACS did not exist or a comparison group of PWID without ACS involvement. While a control group would have been ideal, each of these would have been represented very different sets of patients from which appropriate comparisons could not be drawn. The ACS at our institution has been in existence since 2015, and substance use patterns and HIV prevalence among PWID have changed substantially since that time. For instance, since 2015, fentanyl has become ubiquitous in the opioid drug supply and concurrent stimulant use has become common [31, 32], both of which drive an increased number of injection events per day. Additionally, a 2015–2018 HIV outbreak among PWID in Lawrence and Lowell, MA, cities close to Boston, prompted a Massachusetts Department of Public Health and CDC EpiAid investigation and raised local awareness about the need for increased PEP and PrEP delivery to PWID [33]. Thus a “pre-ACS” control group would have represented a very different population than that included in the present study. In terms of comparing patients seen versus not seen by the ACS, the patients seen by the ACS typically represent those with more unstable SUD and higher-risk IDU. Even with propensity matching, it would be difficult to build a control cohort with a similar degree of medical complexity, severity of addiction, and burden of psychosocial barriers. In future studies seeking to estimate causal effects of ACS involvement on PrEP and PEP utilization, the use of an appropriate control group should be considered.

Second, in our study, demonstration of causality is further limited since the ACS provides recommendations to primary teams but does not directly prescribe medications. However, documentation from ACS notes supports recommendations and indications for PrEP and PEP for the majority of these patients. Third, the patients who received PrEP or PEP and were seen by the ACS were primarily White and English-speaking. We lacked race/ethnicity data on the larger patient population seen by the ACS overall and were thus unable to compare patients who received PrEP or PEP to all patients seen by the ACS. These disparities potentially reflect well-documented inequities in SUD care delivery [34]. Additionally, the racial and language breakdown of our study population is consistent with the population of PWID accessing care at our institution, indicating a broader opportunity to improve SUD care delivery—including PrEP and PEP—to diverse populations who inject drugs.

Fourth, patients were categorized as using a substance only if their chart had documentation to support a use disorder by Diagnostic and Statistical Manual V criteria, which may have underestimated prevalence of SUD. Fifth, when identifying the indication for PrEP and PEP, sexual risk may have been less thoroughly evaluated and documented by clinicians; anecdotally, we observed more consistent evaluation of injection-related risk. This theory is supported by a recent study of PrEP and PEP eligibility in an outpatient SUD bridge clinic setting, showing that while 85.7% of patients were assessed for injection-related risk, only 23% were assessed for sexual risk [35]. Sixth, we abstracted data for patients prescribed TDF/FTC, the form of PrEP most commonly used for PWID at our institution, but did not evaluate patients prescribed TAF/FTC or cabotegravir off-label, and thus may have underestimated PrEP delivery. Seventh, we excluded patients with Hepatitis B, since initiation of PrEP in these patients requires a more complex risk/benefit discussion and at our institution, is typically deferred to a specialty team; we chose to focus on PrEP initiations clearly within the scope of the ACS. Finally, while we document considerable month-to-month variation in the receipt of PrEP or PEP for likely eligible patients, we lacked data on clinician and patient factors that may contribute to such variability, as well as patient refusals, which may be significant [36]. Understanding these factors is important in potentially increasing the receipt of PrEP and PEP during hospitalization.