Knowledge, sex, and region associated with primary care providers prescribing adolescents HIV pre-exposure prophylaxis

Although HIV pre-exposure prophylaxis (PrEP) effectively and safely prevents HIV among adolescents, uptake of PrEP is low. Adolescents must have primary care providers (PCPs) prescribe them PrEP, making PCPs critical actors in PrEP delivery. However, research has primarily investigated determinants of PCPs’ intention to prescribe adolescents PrEP rather than the determinants of performing the behavior itself. We examined the demographic, clinical practice, and implementation determinants of PCPs previously prescribing PrEP to adolescents. PCPs were recruited from a national Qualtrics panel of licensed medical providers in the United States from July 15-August 19, 2022. The Theoretical Domains Framework informed the implementation determinants measured. A multivariable logistic regression was used. PCPs who were more knowledgeable of the CDC guidelines (aOR 2.97, 95% CI 2.16–4.10), who were assigned male at birth (aOR 1.64, 95% CI 1.03–2.59), and who practiced in the Western region (aOR 1.85, 95% CI 1.04–3.30) had greater odds of prior prescribing adolescents PrEP. Provider-based educational interventions should be designed, implemented, and tested to encourage PCPs to prescribe PrEP to eligible adolescents.


Demographic characteristics
Participants were asked about their age, assigned sex at birth, gender identity (male, female, genderqueer, nonbinary, transgender, agender, another option), sexual orientation (gay or lesbian, bisexual, queer, unsure or questioning, straight or heterosexual, pansexual, asexual, another option), ethnicity (yes or no as Hispanic), race (American Indian or Alaskan Native, Asian or Asian American, Black or African American, Hispanic or Latin American, Middle Eastern or Arab American, Native Hawaiian or Pacific Islander, White, Another option), and political ideology (1 = very conservative, 4 = middle of the road, and 7 = very liberal).We recoded demographic variables as 0 = female and 1 = male; 0 = gender minority and 1 = heterosexual; 0 = sexual minority and 1 = heterosexual; 0 = person of color and 1 = Non-Hispanic White; political ideology was kept as a continuous variable.We examined these demographic variables because prior research has shown demographic differences in PrEP prescription willingness or prior PrEP prescription, such as providers' sex and race/ethnicity 24 .

Clinical practice characteristics
Participants were asked about their provider type (physician MD or DO, nurse practitioner, physician assistant), their zip code and state where their practice or clinic was located, if they had ever taken a sexual history of an adolescent aged 13-18 years old in their current clinic or practice setting (0 = no and 1 = yes), ordered or recommended an HIV test for an adolescent in their current setting (0 = no and 1 = yes), ordered or recommended an STI or STD test to an adolescent in their current setting (0 = no and 1 = yes), and if they had ever prescribed an adolescent aged 13-18 PrEP (0 = no and 1 = yes).We recoded states in the four US Census Regions 25 .We recoded zip code into rural/urban status using the Index of Relative Rurality 26 .The Index of Relative Rurality is a continuous rural-urban classification scale, where 0.00 is the most urban and 1.00 is the most rural.

Theoretical domains framework determinants
Participants were asked how much they agree or disagree with the Theoretical Domains Framework determinants (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, 5 = strongly agree).All the determinants consisted of a single item to reduce participant burden, participant time, and survey costs.Items were adapted from the Determinants of Implementation Behavior Questionnaire 22 .The 11 items were knowledge (I am aware of how to prescribe sexually active adolescents HIV PrEP following the CDC guidelines), skills (I have the skills to prescribe sexually active adolescents HIV PrEP), professional role (Prescribing sexually active adolescents HIV PrEP is consistent with my professional role), belief capacity (I am confident that I can prescribe sexually active adolescents HIV PrEP), optimism (I am optimistic when prescribing sexually active adolescents HIV PrEP), belief consequence (If I prescribe sexually active adolescents HIV PrEP, it will benefit public health), attention (When I need to concentrate to prescribe sexually active adolescents HIV PrEP, I have no trouble focusing my attention), environmental resource (In my clinic, all necessary resources are available to prescribe sexually active adolescents HIV PrEP), social influence (Most professional in my organization think that I should prescribe sexually active adolescents HIV PrEP), emotion (I would enjoy prescribing sexually active adolescents HIV PrEP), and intention (I intend to prescribe sexually active adolescents HIV PrEP).
www.nature.com/scientificreports/Analysis Data were analyzed using Stata 18. First, descriptive statistics were ran on all variables: number (N), percentage (%), mean (M), and standard deviation (SD).Second, a bivariate analysis was done to find correlations between all variables in relation to prior HIV PrEP prescription (0 = no, 1 = yes).Point-biserial correlations (r pb ) were used for continuous variables and chi-square tests (χ 2 ) for categorical variables.Finally, a multivariable logistic regression was ran using the statistically significant variables from the bivariate analysis, with the outcome variable being the prior prescription of PrEP to an adolescent.We report the adjusted odds ratio (aOR) and 95% confidence intervals (95% CI).For all analyses, a p-value (p) ≤ 0.05 was statistically significant.

Demographic characteristics
Of the 528 respondents, 55.68% were assigned male at birth and 44.32% were assigned female at birth.Most PCPs identified as cisgender (97.92%) and heterosexual (92.61%).About three-quarters were Non-Hispanic White (70.45%), while one-quarter were a person of color (29.55%).The average age of PCPs was 52.47 years old (SD = 10.84).PCPs were nearly equally distributed among political ideologies (M = 4.07, SD = 1.68): 37.13% lean overall conservative, 22.92% lean middle of the road, and 39.96% lean overall liberal.See Table 1 for more information about the frequency of demographic characteristics.

Clinical practice characteristics
Table 1 also depicts the clinical practice characteristics of the sample.Most providers self-reported as physicians (96.59%), with 11 nurse practitioners and seven physician assistants.Providers were equally split among pediatrics (51.14%) and family medicine specialists (48.86%).Providers practiced throughout the US, with 36.36% in the South, 26.52% in the Midwest, 20.08% in the Northeast, and 17.05% in the West.The mean Index of Relative Rurality score was 0.29 (SD = 0.12), indicating most of the sample practiced in an urban county.Nearly all providers took a sexual history of an adolescent patient aged 13-18 within their current clinic or practice setting (96.02%).Similarly, most providers ordered or recommended an HIV test (92.42%) and an STI or STD test (95.64%) to adolescents aged 13-18 within their current clinic or practice setting.Most providers heard of PrEP before participating in the study (89.02%), and 29.92% had prescribed PrEP to an adolescent aged 13-18.

Theoretical domains framework characteristics
Table 2 shows the frequencies of the 11 Theoretical Domains Framework determinants.All of the Theoretical Domains Framework determinants, but one, had a neutral mean.Belief consequence-also known as instrumental attitude or outcome expectation-had a mean of 4.23 (SD = 0.78), with 88.07% of participants overall agreeing that prescribing PrEP to adolescents will benefit public health.Approximately three-quarters of participants strongly agreed or agreed that prescribing PrEP to adolescents is consistent with their professional role (77.46%,M = 3.93, SD = 0.98), they have the skills to prescribe PrEP to adolescents (73.48%,M = 3.74, SD = 1.08), they are confident or have self-efficacy they can prescribe PrEP to adolescents (70.46%,M = 3.73, SD = 1.03), and they intend to prescribe adolescents PrEP in the future (70.45%,M = 3.81, SD = 0.86).Fifty-seven percent of participants overall agreed that they have clinical resources to prescribe adolescents PrEP (57.76%, M = 3.45, SD = 1.12), and they have no trouble focusing their attention when prescribing adolescents PrEP (57.76%, M = 3.95, SD = 0.97).About half of the participants strongly agreed or agreed that they are aware of how to prescribe adolescents PrEP based on the CDC guidelines (54.92%,M = 3.28, SD = 1.19), and they are optimistic when prescribing adolescents PrEP (50.00%, M = 3.44, SD = 0.91).Nearly one-third of participants overall agreed that they would enjoy prescribing PrEP to adolescents (39.77%,M = 3.28, SD = 0.94), and most professionals in their organization think that they should prescribe adolescents PrEP (37.12%, M = 3.27, SD = 0.96).
The only demographic variable that was correlated with prior prescription of PrEP was sex assigned at birth, with prior prescription being positively associated with assigned male at birth (χ 2 = 13.24,p < 0.001).Regarding clinical practice variables, prior prescription of PrEP to adolescents was positively correlated with prior ordering or recommending of an HIV test to an adolescent (χ 2 = 10.38,p < 0.001), prior ordering or recommending of an STD/STI test to an adolescent (χ 2 = 5.17, p = 0.05), prior hearing about PrEP (χ 2 = 27.82,p < 0.001), if the provider was located in the Western region of the US (χ 2 = 6.47, p < 0.05), and if the provider specialized in pediatrics (χ 2 = 10.70,p < 0.01).

Discussion
This is the first national study to analyze the demographic, clinical practice, and implementation determinants of PCPs prior PrEP prescription behaviors to adolescents after the FDA approved PrEP for minors.Approximately 30% of our sample reported they ever prescribed PrEP to an adolescent patient, and this is similar other studies 12,19 that reported about one-third of providers in their sample ever prescribed PrEP to an adolescent.Our  www.nature.com/scientificreports/study was informed by the Theoretical Domains Framework 20 , a common implementation framework used to examine the implementation factors of providers delivering an evidence-based practice 21 .One variable per the Theoretical Domains Framework (knowledge), demographic characteristics (sex assigned at birth), and clinical practice characteristics (region) were significant.Our findings have important implications for behavioral-based education, theory, and practice on prescribing PrEP to adolescents.Knowledge about the CDC's guidelines has been shown to be a significant determinant in PCPs' intention to prescribe adolescents PrEP 12,16,19 , and our finding suggests that knowledge extends into actually prescribing adolescents PrEP.This illustrates the need for PCPs to be educated about the CDC's guidelines, as knowledge is associated with PCPs prescribing PrEP to adult patients [27][28][29][30][31][32][33] .Multiple on-site trainings 34 might be more effective in increasing PrEP prescription behavior than single session trainings 35 .Moreover, medical schools could incorporate the CDC's guidelines into their medical education and medical residency programs, as medical students report PrEP is not included in their training [36][37][38][39] .Future research should examine what facilitators and barriers medical school, residency, and clinic leaders have in providing PrEP training.
While our study found that the only statistically significant Theoretical Domains Framework determinant was knowledge, prior research has highlighted several other determinants, including PCPs' attitudes, self-efficacy and skills, professional roles, perceived norms, and clinical environment resources [11][12][13][14][15][16][17][18][19] .This might not be surprising, as providers who have more positive attitudes about an evidence-based practice and more clinical resources are more likely to deliver that evidence-based practice.A possible explanation is that knowledge might be a moderator or mediator for other determinants.Indeed, a study 33 utilizing the Information-Motivation-Behavioral Skill Model (IMB) demonstrated that PrEP-related information directly affected the PrEP-related attitudes and skills among PCPs.While the Theoretical Domains Framework explained 42% of the variance of PCPs' intention to prescribe PrEP to their patients 19 , the IMB explained 50% of the variance 33 .Future studies might use the IMB as their theoretical framework in investigating the determinants of PCPs prescribing PrEP to teens.
We found that PCPs who are males at birth prescribed PrEP to adolescents more than PCPs females at birth.Nearly all of our sample were cisgender.While Owens et al. 19 found sex was not a determinant of PCP's willingness to prescribe PrEP to adolescents, Leech et al. 40 found females were less likely than males to have prescribed PrEP to adult male patients.We hypothesize three reasons for why male PCPs prescribed adolescents PrEP more than female PCPs.First, adolescent 41 and adult patients 42 prefer to seek PCPs who share in the patient's sex.Because HIV affects males at birth more than females at birth in the US, male PCPs may have more interactions than female PCPs with PrEP-eligible and PrEP-seeking patients.Second, systematic reviews have noted that male providers adopt new prescriptions earlier than female providers 43,44 .Third, more family medicine/general practice physicians are males (57.7%) than females (42.3%) 45 .
Our study showed that PCPs in the Western region had 1.85 greater odds of prescribing adolescents PrEP in the past.This result corroborates with the literature that consistently shows PrEP use is higher in the Northeastern and Western regions than in the Southern and Midwestern regions [5][6][7]9 . Itis likely that PrEP prescriptions are higher in the Western region due to the availability of PrEP clinics and PrEP providers 46,47 as well as policies such as Medicaid expansion and PrEP drug assistance programs 5,6,48 .For example, the concerted efforts by public health entities in California expanded access to PrEP through community pharmacies, Medicaid expansion, and PrEP assistance programs 49,50 .Public health and HIV entities could advocate for their state legislatures to adopt these PrEP access policies.www.nature.com/scientificreports/ Moreover, Wetern and Northeastern states tend to allow minors to access HIV prevention services, including PrEP, without parental/guardian permission 51,52 .Thirty-four states and D.C. allow minors to consent to HIV prevention services; however, only 18 of these address confidentiality protections.Public health, HIV, and adolescent health entities might want to educate providers about their state's minor consent/confidentiality laws, as providers tend to be unfamiliar with their state's law, given the wide variability of minor consent laws 53,54 .Public health, HIV, and adolescent health entities might want to advocate for their state legislatures to increase the legal capacity of minors to consent to HIV prevention services.Research is necessary to understand the facilitators, barriers, and strategies that effectively change state laws that expand HIV prevention services to adolescents.

Limitations
This study has limitations.First, our sample was primarily composed of physicians, with nurse practitioners and physician assistants making up 3% of the sample.More research is necessary to examine the implementation determinants of nurse practitioners in states where they are given full prescriptive authority.Second, we utilized cross-sectional data, meaning we can claim correlations but not causations.Third, our patient vignette did not include the adolescent's demographic characteristics.Similarly, participants provided their agreement and disagreement to questions about PrEP delivery to adolescents aged 13-18 years old rather than a specific age (e.g., 16 years old) or an age range (15-17).However, some states define minors those who are 18-21 51,52 .Medical providers have noted in prior research that patient characteristics and providers' perceptions of patients play a part in their intention to prescribe and in their actual prescribing of PrEP to adolescents [12][13][14][15]17,18 . Fourt, our study only analyzed prior prescription of PrEP and not other USPSTF recommendations regarding PrEP identification, discussion, and care, such as taking an HIV risk assessment to identify adolescents at risk of contracting HIV, educating adolescent patients and their parents about PrEP, or discussing the risks and benefits of the different PrEP medications and modalities (e.g., Truvada® v. Descovy® v. Apretude®; daily oral v. long-acting injectable) 4 .

Conclusion
This study, utilizing data collected from a national sample of PCPs, demonstrates that PCPs who are more knowledgeable about the CDC's guidelines had greater the odds of prescribing adolescents PrEP.It might be beneficial to implement provider-targeted PrEP training and education programs to medical students, medical residents, and medical providers.Additionally, findings from our study suggest that scientists who study provider PrEP-related behavior might want to use theoretical frameworks that posit that information affects other constructs.We might have seen more PCPs prescribing PrEP to adolescents in the Western region because of state policies.Public health, HIV, and adolescent health entities might want to advocate state legislatures to adopt laws that allow adolescents to consent to HIV prevention services.Such education and legal interventions might increase the rates in which PCPs prescribe PrEP to eligible adolescents, therefore decreasing HIV rates among adolescents and young adults.