Psychiatrists' Experiences, Training Needs, and Preferences Regarding Prescription and Management of HIV Pre‐exposure Prophylaxis (PrEP) Within Psychiatric Care

Psychiatrists were recruited via a combination of methods to surmount the documented difficulties in conducting research with physicians (20). First, information about the study was shared via email with the American Medical Association Physician MasterFile, a large, national listserv of psychiatrists (n = 10,840). Emails were sent up to six times including an initial message and up to five reminders if the study instrument was not completed. Geographic filtering was applied to over‐sample the Ending the HIV Epidemic (EHE) priority jurisdictions given the increased focus on HIV prevention in these areas due to disproportionate HIV incidence (21). This was supplemented by distributing information about the study at the 2023 American Psychiatric Association annual meeting (n = 1143 U.S. attending/fellows in psychiatry), for a total of 11,983 possible respondents.

Psychiatrists were eligible for participation if they: (1) Held an allopathic (MD) or osteopathic (DO) medical degree, (2) Spent at least some percentage of their time providing direct patient care, (3) Were practicing in the U.S., and (4) Were an attending or fellow in psychiatry. Data were collected between November 2022‐October 2023. All participants received an electronic $50 Visa^® gift card as compensation.

The study instrument was developed based on review of previous studies about PrEP implementation conducted among primary care and infectious disease physicians (22, 23, 24). Items were refined with input from experts in infectious disease and psychiatry. A focus group of five psychiatrists reviewed the instrument and provided feedback on item wording, which was incorporated before distribution. The study instrument was hosted on Qualtrics® (Qualtrics Inc., Provo, UT). The full study instrument is included as Supplemental Material.

Previous PrEP Experiences: First, psychiatrists indicated whether they were aware of PrEP at the time of the study, which was defined. Next, psychiatrists indicated whether a patient requested PrEP prescription from them, and whether they previously prescribed PrEP. If a participant indicated “yes” to either of these, they were prompted to complete a follow‐up item to estimate the number of times (1‐5, 6–10, ≥11) for each. Additionally, we inquired about interest in prescribing PrEP and interest in receiving training about PrEP (yes/no/not sure).

Barriers: A series of 12 possible barriers to PrEP implementation including six system‐level barriers (eg. insurance coverage, administrative requirements, liability) and six individual‐level barriers (eg. limited knowledge, discomfort in discussing sex, out‐of‐scope) was presented. The list was developed considering previous studies with other groups of clinicians about implementing PrEP prescription 22, 25, 26. Respondents selected all barriers they thought would negatively affect PrEP implementation in their practice.

Preferences for PrEP Implementation: A series of four models for PrEP implementation was presented (eg. all psychiatrists trained in PrEP prescription, patients referred out for PrEP, psychiatrist initiates PrEP with arranged primary care follow‐up) based on previous studies (22, 25, 26). Psychiatrists indicated each of the models they felt were acceptable.

Confidence: Two domains of confidence were assessed. First, participants were presented with 10 PrEP‐related tasks (eg. ordering follow‐up lab testing, taking sexual history) adapted from the CDC guidelines, and previous research with healthcare professionals and trainees (26, 27). Second, participants were presented with a series of eight items measuring confidence managing general medical conditions (GMC; eg. hypertension, non‐opiate pain management, and asthma) (28). The GMC confidence scale was included as a baseline measure of comfort practicing outside of the traditionally defined scope‐of‐practice of psychiatry. All confidence items were phrased as: “I would feel confident…” and rated on a 7‐point scale (1 = strongly disagree, 7 = strongly agree). Items in each domain were averaged to create aggregate measures of PrEP‐related tasks (Cronbach's α = 0.91) and managing GMC (α = 0.95).

Demographics: Psychiatrists indicated their race, sexual orientation, gender identity, sex assigned at birth, age, years in practice (including residency and any fellowship[s]), and primary practice setting. Participants also indicated the state in which they practiced. If psychiatrists practiced in a state with an EHE priority jurisdiction (counties or cities), they completed a follow‐up item to indicate if they practiced in one of these jurisdictions (21).

First, we calculated frequencies and other descriptive statistics to describe the sample and outcomes. We employed a series of bivariate, binomial logistic regressions to model the associations between demographic/practice characteristics, confidence in managing GMC and PrEP‐related tasks, and likelihood of: (1) receipt of a PrEP request, (2) previous PrEP prescription, and (3) interest in prescribing PrEP, followed by three multivariable models to evaluate the combined associations between all demographic/practice characteristics, confidence domains, and each outcome. For the multivariable models, adjusted odds ratios (aOR) are presented and we assessed for multicollinearity during model construction. All descriptive outcomes were repeated, separated by primary practice setting (Supplemental Figures S1A‐2C). In addition, we performed a series of secondary analyses of the primary practice setting, barriers, and confidence comparing the group of psychiatrists who had prescribed PrEP following a patient request and the group who did not using chi‐squared tests for categorical variables and independent samples t‐tests for continuous variables.

Analyses were completed utilizing StataMP V17 (StataCorp). A p‐value <0.05 was established as statistically significant. This study was reviewed and approved by the institutional review board of the University of Chicago.

Most respondents were attending psychiatrists (n = 764, 86.8%), practiced in academic medical centers (n = 642, 73.0%), and over half reported their primary practice setting was outpatient psychiatry (n = 480, 54.6%). Geographically, the largest percentage practiced in the Northeastern U.S. (n = 304, 34.6%), and in an EHE priority jurisdiction (n = 623, 70.8%). Most completed a fellowship after residency (n = 584, 66.4%), and child/adolescent psychiatry (n = 228, 25.9%) was the most common. Most identified as heterosexual (n = 724, 82.3%), as men (n = 475, 54.0%), and as White (n = 482, 54.8%). Mean practice duration was 14.0 years (95% CI: 13.3–14.7), mean percentage of time providing patient care was 72.1% (70.4–73.8), and mean percentage of time spent in outpatient psychiatry was 70.1% (67.7–72.6). Complete demographic information is included in Table 1. We found 93.2% of psychiatrists were aware of PrEP.

Overall, 19.3% of psychiatrists had received a PrEP request from a patient (Figure 1A). Of these, most (72.4%) received a PrEP request between 1 and 5 times. Attending psychiatrists were more likely to have received a PrEP request (aOR = 2.08 [1.06–4.06], p = 0.03) compared to fellows, while those practicing primarily in consult‐liaison (CL) psychiatry were less likely to have received a request (aOR = 0.33 [0.11–0.93], p = 0.04) (Table 2). Psychiatrists practicing in one of the EHE priority jurisdictions (aOR = 1.62 [1.02–2.58], p = 0.04) were also more likely to have received a PrEP request. Compared to psychiatrists without fellowship training, those who completed an addiction medicine fellowship (aOR = 3.61 [1.78–7.29], p < 0.001) or forensic psychiatry fellowship (aOR = 2.52 [1.07–5.92], p = 0.03) were more likely to have received a PrEP request. A greater percentage of time in patient care overall (aOR = 1.01 [1.00–1.02], p = 0.01) and greater confidence managing GMC (aOR = 1.21 [1.03–1.42], p = 0.02) and PrEP‐related tasks (aOR = 1.76 [1.44–2.16], p < 0.001) were associated with greater likelihood of receipt of a PrEP request (Table 2).

We found that 17.3% of psychiatrists had prescribed PrEP to a patient (Figure 1A), and most had prescribed between 1 and 5 times (68.4%). Of psychiatrists who indicated they received a request for PrEP prescription, 65.9% prescribed PrEP, and 5.6% of psychiatrists initiated PrEP prescription without explicit request. Psychiatrists practicing primarily in inpatient psychiatry (compared to outpatient; aOR = 2.80 [1.11–7.06], p = 0.03), or in one of the EHE priority jurisdictions (aOR = 2.08 [1.23–3.54], p = 0.003), were more likely to have prescribed PrEP (Table 2). Psychiatrists who completed a CL‐psychiatry (aOR = 0.37 [0.14–0.98], p = 0.04) fellowship (compared to no fellowship) were less likely to have previously prescribed PrEP. Greater confidence in managing GMC (aOR = 1.34 [1.12–1.60], p = 0.001) and PrEP‐related tasks (aOR = 2.10 [1.67–2.65], p < 0.001) were associated with higher likelihood of PrEP prescription. Among the group who had received a PrEP request but did not prescribe (n = 58), the greatest percentage were practicing primarily in outpatient psychiatry (n = 37, 63.8%).

Of psychiatrists who indicated they had not previously prescribed PrEP, 53.9% indicated they would be interested in prescribing PrEP (Figure 1A). Psychiatrists practicing primarily in inpatient psychiatry (relative to outpatient) were less likely to indicate interest in PrEP prescription (aOR = 0.43 [0.19–0.96], p = 0.04) as were those who had completed a community psychiatry fellowship (aOR = 0.25 [0.11–0.57], p = 0.001). Greater confidence in managing GMC (aOR = 1.30 [1.13–1.50], p < 0.001) and PrEP‐related tasks (aOR = 1.53 [1.28–1.84], p < 0.001) were associated with interest in prescribing PrEP.

All participants endorsed at least one individual‐level barrier to PrEP implementation (Figure 2A). The most frequently indicated individual barrier was limited knowledge of PrEP (76.3%) and the least frequently indicated was discomfort discussing sexual activity with patients (4.2%). High percentages also endorsed the need to manage follow‐up care for patients on PrEP (59.5%) and the view that PrEP was outside of the scope‐of‐practice for psychiatrists (45.6%) were additional barriers.

All participants also endorsed at least one system‐level barrier to PrEP implementation (Figure 2B). The most frequently indicated system‐level barrier was administrative requirements from insurers for PrEP (48.0%) and the least frequently endorsed was liability of PrEP prescription (20.5%). Limited time during visits to discuss both PrEP and the patient's psychiatric needs (43.0%) and lacking a referral process to Infectious Diseases for patients identified as HIV seropositive (31.4%) were additional barriers. Analysis of anticipated barriers separated by primary practice setting followed this overall trend (Supplemental Figures S1 and S2).

Overall, the percentages of psychiatrists endorsing each barrier were similar in the groups of psychiatrists who had prescribed PrEP following a patient request (n = 112) and who had not (n = 58) and the overall sample. Compared to those who had prescribed following a request, those who did not were more likely to endorse that PrEP was outside of their scope of practice (χ²[n = 170] = 6.93, 25.0% vs. 44.8%, p = 0.008) and that management of the follow‐up care for PrEP (χ²[n = 170] = 5.59, 46.4% vs. 65.5%, p = 0.02).

The greatest percentage of psychiatrists (63.1%) indicated that a model in which a psychiatrist would prescribe an initial prescription of PrEP to a patient with prompt follow‐up with primary care or infectious diseases was their preference (Figure 1B). Smaller but still sizeable percentages indicated that having all psychiatrists in the practice trained in PrEP prescription and management (55.5%), referral outside the practice for PrEP prescription (41.1%), and having a single psychiatrist trained as the PrEP specialist for the practice (28.3%) were acceptable implementation methods (Figure 1B). Overall, 67.3% of psychiatrists were interested in receiving training on PrEP (Figure 1B). Online training about PrEP prescription was the preferred method overall (88.7%).

Overall confidence for PrEP‐related tasks was 3.95 (3.88–4.03) which was greater than mean confidence managing GMC (3.64 [3.54–3.73], p < 0.001) (Figure 3). Psychiatrists self‐reported the greatest confidence in referring patients to infectious diseases specialists for newly identified HIV diagnoses (M = 5.41, [5.29–5.52]) and lowest in ordering follow‐up labs and testing for patients taking PrEP (M = 3.06, [2.92–3.20]). Mean confidence managing GMC (M = 4.72 [4.44–5.00] vs. M = 3.49, [3.17–3.82], t[168] = 5.36, p < 0.001) and PrEP‐related tasks (M = 5.03 [4.82–5.25] vs. M = 4.09, [3.82–4.37], t[168] = 5.16, p < 0.001) were greater among the group of psychiatrists who had prescribed following a patient request compared to the group who did not.

The science and efficacy of PrEP for HIV prevention is well‐established, but clinical adoption has not reached the necessary scale. Implementation strategies and adjunctive interventions are needed to bridge this gap to expand access to PrEP for all patients who are vulnerable to HIV (31). Many prior implementation strategies to increase PrEP use have been reported in the literature (31). Adaptation of existing interventions for implementation in psychiatric care may be an efficient and rapidly scalable method for increasing PrEP use among PLWMI, reducing HIV incidence in this underserved population, and addressing the system barriers to PrEP implementation reported by psychiatrists (32).

It is notable that the greatest percentage of psychiatrists who previously prescribed PrEP practiced primarily in inpatient psychiatry (30.8%). Implementation of a protocol to initiate PrEP prescription prior to discharge from the inpatient psychiatry with linkage to PrEP navigation services upon discharge may be a promising future direction to reach patients who face other difficulties in accessing care. Such a model would align with psychiatrists' indicated preference for initial PrEP prescription with subsequent follow‐up. A similar structure has been applied in other acute care settings like emergency departments (ED) in which patients with HIV risk‐factors are referred for timely PrEP initiation as an outpatient (33). Existing ED PrEP navigation models have not included PrEP prescription while in the ED (33). But, PrEP prescription may be possible at discharge from inpatient psychiatry with appropriate follow‐up appointments and HIV testing arranged given that psychiatric hospitalization is longer and gives more time for care coordination as opposed to an ED visit. PrEP initiation for medically hospitalized patients with HIV risk‐factors was feasible and acceptable in a recent pilot study suggesting this may be possible in other types of inpatient care (34).

This sort of a model may also be feasible in settings that already have some degree of psychiatric and primary care service co‐location, such as in Federally Qualified Health Centers (FQHC) or look‐alike facilities. However, this model is not in place in all mental health care settings. Implementing a model in which psychiatrists begin PrEP with prompt referral would require identification and organization of referral streams that are unique to the existing workflows of a care setting. This sort of practice model may also limit prescription as not all patients seeking psychiatric care have established primary care follow‐up to continue PrEP prescription. Importantly, we found that most psychiatrists who'd previously prescribed PrEP had done so between 1 and 5 times suggesting prescription is not sustained. This has been identified in other studies as well reinforcing the need for interventions aimed at PrEP persistence especially with ongoing HIV risk (35).

One such practice model may be informed by previous collaborative practice initiatives to scale the use of buprenorphine for opioid use disorder. The Physician Clinical Support System‐Buprenorphine (PCSS‐B) is a previously described model in which primary care clinicians acted as the primary prescribers of buprenorphine with a national support system of addiction medicine specialists and psychiatrists acting as a resource (36). This arrangement may be adaptable to PrEP prescription by psychiatrists with a select group of infectious disease or primary care clinicians serving as a resource for psychiatrists who are identifying HIV risk‐factors in patients and prescribing PrEP. Such a model would allow psychiatrists to manage the day‐to‐day aspects of PrEP care, with expert availability when needed. This model may also address the identified barriers of needing an urgent referral process should a patient be identified as newly diagnosed with HIV and in treating STIs diagnosed upon screening. Similarly, the University of California San Francisco operates the National Clinician Consultation Center for HIV care and prevention‐related clinical questions as well as providing online resources about PrEP (https://nccc.ucsf.edu/). An additional possibility is collaborative care models with pharmacists within psychiatric practice given increasing engagement of pharmacists in HIV prevention and mental healthcare (37, 38).

Additionally, psychiatrists and psychiatric care facilities have longstanding experience with LAI antipsychotic and anti‐craving medications, including clinical services for regular administration of both intramuscular and subcutaneous formulations. LAI‐cabotegravir represents the first LAI‐PrEP option with possible approval of lenacapavir, which is in clinical trials to evaluate effectiveness as PrEP administered subcutaneously at 6‐month intervals. Implementation of co‐located service models in which LAI‐PrEP and LAI‐psychotropic or anti‐craving medications are administered at synchronized intervals may also represent an efficient way of improving PrEP uptake. Leveraging the existing clinical workflows and physical/staffing infrastructure for medication storage and administration within psychiatric care settings may also provide valuable lessons for integration of LAI‐PrEP in other clinical contexts with less experience with LAI medication formulations (39).

The findings of the present study should be interpreted considering several limitations. First, most of the participants in the study were practicing in academic medical centers which does not represent the full spectrum of psychiatrists and practice settings. Many patients receive mental health care in community settings or through collaborative practice arrangements which are an important target for future study to support PrEP prescription for PLWMI. Second, we intentionally oversampled psychiatrists practicing in the EHE priority jurisdictions given the significant need for additional HIV prevention focus and disproportionate incidence. However, these jurisdictions are primarily large, urban centers and thus we did not capture a significant number of responses from rural areas of the U.S. Future work should specifically focus on recruitment from rural regions which experience unique barriers to HIV elimination, especially limited prescriber availability and access. Scope of future work should also be expanded to include additional practitioner groups, especially non‐physician and non‐psychiatrist providers (eg. physician assistants, nurse practitioners, general practitioners) who comprise much of the workforce in rural settings (40). Finally, we acknowledge the low response rate of this survey which may limit generalizability of these findings to describe PrEP prescription within psychiatric care. Data collection from physicians, while essential, is difficult given time constraints in clinical practice. We attempted to overcome this with in‐person recruitment at a large meeting of psychiatrists however still achieved a low response rate. However, our response rate is similar to prior studies of physicians related to HIV PrEP (20).

PrEP for HIV prevention is an effective and safe intervention however uptake among psychiatric patients has not met the public health need of this vulnerable population. As the primary clinicians caring for people living with mental illnesses, psychiatrists have an important role in expanding access to PrEP to prevent HIV, but data to support these efforts are sparse. The results of this study suggest that psychiatrists are interested in prescribing PrEP but need training to support broader integration. Many aspects of the clinical management of PrEP prescription are protocol‐based and can be easily integrated into psychiatric care, leveraging existing workflows. Training about PrEP for HIV prevention that is specifically focused on the needs of psychiatrists should be developed to support PrEP prescription by psychiatrists to increase access for patients vulnerable to HIV.