Using HIV Risk Prediction Tools to Identify Candidates for Pre-Exposure Prophylaxis: Perspectives from Patients and Primary Care Providers

STI/HIV risk prediction model development-A novel use of public data to forecast STIs/HIV risk for men who have sex with men

A novel automatic framework is proposed for global sexually transmissible infections (STIs) and HIV risk prediction. Four machine learning methods, namely, Gradient Boosting Machine (GBM), Random Forest (RF), XG Boost, and Ensemble learning GBM-RF-XG Boost are applied and evaluated on the Demographic and Health Surveys Program (DHSP), with thirteen features ultimately selected as the most predictive features. Classification and generalization experiments are conducted to test the accuracy, F1-score, precision, and area under the curve (AUC) performance of these four algorithms. Two imbalanced data solutions are also applied to reduce bias for classification performance improvement. The experimental results of these models demonstrate that the Random Forest algorithm yields the best results on HIV prediction, whereby the highest accuracy, and AUC are 0.99 and 0.99, respectively. The performance of the STI prediction achieves the best when the Synthetic Minority Oversampling Technique (SMOTE) is applied (Accuracy = 0.99, AUC = 0.99), which outperforms the state-of-the-art baselines. Two possible factors that may affect the classification and generalization performance are further analyzed. This automatic classification model helps to improve convenience and reduce the cost of HIV testing.

Systematic review to understand users perspectives on AI-enabled decision aids to inform shared decision making

Artificial intelligence (AI)-enabled decision aids can contribute to the shared decision-making process between patients and clinicians through personalised recommendations. This systematic review aims to understand users' perceptions on using AI-enabled decision aids to inform shared decision-making. Four databases were searched. The population, intervention, comparison, outcomes and study design tool was used to formulate eligibility criteria. Titles, abstracts and full texts were independently screened and PRISMA guidelines followed. A narrative synthesis was conducted. Twenty-six articles were included, with AI-enabled decision aids used for screening and prevention, prognosis, and treatment. Patients found the AI-enabled decision aids easy to understand and user-friendly, fostering a sense of ownership and promoting better adherence to recommended treatment. Clinicians expressed concerns about how up-to-date the information was and the potential for over- or under-treatment. Despite users' positive perceptions, they also acknowledged certain challenges relating to the usage and risk of bias that would need to be addressed.Registration: PROSPERO database: (CRD42020220320).

Using Machine Learning to Identify Patients at Risk of Acquiring HIV in an Urban Health System

BACKGROUND

Effective measures exist to prevent the spread of HIV. However, the identification of patients who are candidates for these measures can be a challenge. A machine learning model to predict risk for HIV may enhance patient selection for proactive outreach.

SETTING

Using data from the electronic health record at Parkland Health, 1 of the largest public healthcare systems in the country, a machine learning model is created to predict incident HIV cases. The study cohort includes any patient aged 16 or older from 2015 to 2019 (n = 458,893).

METHODS

Implementing a 70:30 ratio random split of the data into training and validation sets with an incident rate <0.08% and stratified by incidence of HIV, the model is evaluated using a k-fold cross-validated (k = 5) area under the receiver operating characteristic curve leveraging Light Gradient Boosting Machine Algorithm, an ensemble classifier.

RESULTS

The light gradient boosting machine produces the strongest predictive power to identify good candidates for HIV PrEP. A gradient boosting classifier produced the best result with an AUC of 0.88 (95% confidence interval: 0.86 to 0.89) on the training set and 0.85 (95% confidence interval: 0.81 to 0.89) on the validation set for a sensitivity of 77.8% and specificity of 75.1%.

CONCLUSIONS

A gradient boosting model using electronic health record data can be used to identify patients at risk of acquiring HIV and implemented in the clinical setting to build outreach for preventative interventions.

Assessing Patient Acceptance of an Automated Algorithm to Identify Ciswomen for HIV Pre-Exposure Prophylaxis

The use of HIV pre-exposure prophylaxis (PrEP) in cisgender women (ciswomen) lags far behind their need. Data elements from the electronic medical record (EMR), including diagnosis of a sexually transmitted infection (STI), can be incorporated into automated algorithms for identifying clients who are most vulnerable to HIV and would benefit from PrEP. However, it is unknown how women feel about the use of such technology. In this study, we assessed women's attitudes and opinions about an automated EMR-based HIV risk algorithm and determined if their perspectives varied by level of HIV risk. Respondents were identified using best practice alerts or referral to a clinic for STI symptoms from January to December 2021 in Chicago, IL. Participants were asked about HIV risk factors, their self-perceived HIV risk, and their thoughts regarding an algorithm to identify ciswomen who could benefit from PrEP. Most of the 112 women who completed the survey (85%) thought they were at low risk for HIV, despite high rates of STI diagnoses. The majority were comfortable with the use of this algorithm, but their comfort level dropped when asked about the algorithm identifying them specifically. Ciswomen had mixed feelings about the use of an automated HIV risk algorithm, citing it as a potentially helpful and empowering tool for women, yet raising concerns about invasion of privacy and potential racial bias. Clinics must balance the benefits of using an EMR-based algorithm for ciswomen with their concerns about privacy and bias to improve PrEP uptake among particularly vulnerable women.

Addressing Unmet PrEP Needs in Women: Impact of a Laboratory-Driven Protocol at an Urban, Essential Hospital

Background

HIV pre-exposure prophylaxis (PrEP) uptake in women remains low. We developed a laboratory result-driven protocol to link women with a positive bacterial sexually transmitted infection (STI) to HIV PrEP at an urban safety-net hospital.

Methods

Electronic health records of women with positive chlamydia, gonorrhea, and/or syphilis tests were reviewed, and those eligible for PrEP were referred for direct or primary care provider-driven outreach. We assessed the proportion of women with STIs who received PrEP offers, acceptance, and prescriptions before (July 1, 2018-December 31, 2018) and after (January 1, 2019-June 30, 2020) implementation to evaluate changes in the delivery of key elements of the PrEP care cascade (ie, PrEP offers, acceptance, and prescribing) for women with STIs after protocol implementation.

Results

The proportion of women who received PrEP offers increased from 7.6% to 17.6% (P < .001). After multivariable adjustment, only the postintervention period was associated with PrEP offers (odds ratio [OR], 2.49; 95% CI, 1.68-3.68). In subgroup analyses, PrEP offers increased significantly among non-Hispanic Black (OR, 2.75; 95% CI, 1.65-4.58) and Hispanic (OR, 5.34; 95% CI, 1.77-16.11) women but not among non-Hispanic White women (OR, 1.49; 95% CI, 0.54-4.05). Significant changes in PrEP acceptance and prescriptions were not observed in the sample overall.

Conclusions

A laboratory result-driven protocol was associated with a significant increase in PrEP offers to Black and Hispanic women with STI. These results provide concrete suggestions for health systems seeking to increase PrEP access and equity among women.

Current ARTs, Virologic Failure, and Implications for AIDS Management: A Systematic Review

Antiretroviral therapies (ARTs) have revolutionized the management of human immunodeficiency virus (HIV) infection, significantly improved patient outcomes, and reduced the mortality rate and incidence of acquired immunodeficiency syndrome (AIDS). However, despite the remarkable efficacy of ART, virologic failure remains a challenge in the long-term management of HIV-infected individuals. Virologic failure refers to the persistent detectable viral load in patients receiving ART, indicating an incomplete suppression of HIV replication. It can occur due to various factors, including poor medication adherence, drug resistance, suboptimal drug concentrations, drug interactions, and viral factors such as the emergence of drug-resistant strains. In recent years, extensive efforts have been made to understand and address virologic failure in order to optimize treatment outcomes. Strategies to prevent and manage virologic failure include improving treatment adherence through patient education, counselling, and supportive interventions. In addition, the regular monitoring of viral load and resistance testing enables the early detection of treatment failure and facilitates timely adjustments in ART regimens. Thus, the development of novel antiretroviral agents with improved potency, tolerability, and resistance profiles offers new options for patients experiencing virologic failure. However, new treatment options would also face virologic failure if not managed appropriately. A solution to virologic failure requires a comprehensive approach that combines individualized patient care, robust monitoring, and access to a range of antiretroviral drugs.

Performance of HIV Infection Prediction Models in Men Who Have Sex with Men: A Systematic Review and Meta-Analysis

Effective ways to identify and predict men who have sex with men (MSM) at substantial risk for HIV is a global priority. HIV risk assessment tools can improve individual risk awareness and subsequent health-seeking actions. We sought to identify and characterize the performance of HIV infection risk prediction models in MSM through systematic review and meta-analysis. PubMed, Embase, and The Cochrane Library were searched. Eighteen HIV infection risk assessment models with a total of 151,422 participants and 3643 HIV cases were identified, eight of which have been externally validated by at least one study (HIRI-MSM, Menza Score, SDET Score, Li Model, DHRS, Amsterdam Score, SexPro model, and UMRSS). The number of predictor variables in each model ranged from three to 12, age, the number of male sexual partners, unprotected receptive anal intercourse, recreational drug usage (amphetamines, poppers), and sexually transmitted infections were critical scoring variables. All eight externally validated models performed well in terms of discrimination, with the pooled area under the receiver operating characteristic curve (AUC) ranging from 0.62 (95%CI: 0.51 to 0.73, SDET Score) to 0.83 (95%CI: 0.48 to 0.99, Amsterdam Score). Calibration performance was only reported in 10 studies (35.7%, 10/28). The HIV infection risk prediction models showed moderate-to-good discrimination performance. Validation of prediction models across different geographic and ethnic environments is needed to ensure their real-world application.

Development of a Human Immunodeficiency Virus Risk Prediction Model Using Electronic Health Record Data From an Academic Health System in the Southern United States

BACKGROUND

Human immunodeficiency virus (HIV) pre-exposure prophylaxis (PrEP) is underutilized in the southern United States. Rapid identification of individuals vulnerable to diagnosis of HIV using electronic health record (EHR)-based tools may augment PrEP uptake in the region.

METHODS

Using machine learning, we developed EHR-based models to predict incident HIV diagnosis as a surrogate for PrEP candidacy. We included patients from a southern medical system with encounters between October 2014 and August 2016, training the model to predict incident HIV diagnosis between September 2016 and August 2018. We obtained 74 EHR variables as potential predictors. We compared Extreme Gradient Boosting (XGBoost) versus least absolute shrinkage selection operator (LASSO) logistic regression models, and assessed performance, overall and among women, using area under the receiver operating characteristic curve (AUROC) and area under precision recall curve (AUPRC).

RESULTS

Of 998 787 eligible patients, 162 had an incident HIV diagnosis, of whom 49 were women. The XGBoost model outperformed the LASSO model for the total cohort, achieving an AUROC of 0.89 and AUPRC of 0.01. The female-only cohort XGBoost model resulted in an AUROC of 0.78 and AUPRC of 0.00025. The most predictive variables for the overall cohort were race, sex, and male partner. The strongest positive predictors for the female-only cohort were history of pelvic inflammatory disease, drug use, and tobacco use.

CONCLUSIONS

Our machine-learning models were able to effectively predict incident HIV diagnoses including among women. This study establishes feasibility of using these models to identify persons most suitable for PrEP in the South.

Health Care Providers' Views on Clinic Infrastructure and Practice Models That May Facilitate HIV Preexposure Prophylaxis (PrEP) Prescribing: A Qualitative Meta-Synthesis

HIV (human immunodeficiency virus) preexposure prophylaxis (PrEP) is an effective biomedical HIV prevention tool. Increasing PrEP use among populations disproportionately affected by HIV is one of the key efforts in the United States' Ending the HIV Epidemic (EHE) initiative and the HIV National Strategic Plan for the United States. Given that PrEP is available only through prescription, it is important to explore structural, organizational, or environmental factors that could facilitate or impede health care provider's PrEP prescribing behavior. The purpose of this systematic review (PROSPERO [CRD: 42019138889]) is to identify qualitative studies that addressed this topic and conduct meta-synthesis using the thematic synthesis method to identify major themes on the characteristics of clinic infrastructure or clinic models that providers consider as facilitators of PrEP prescribing in the United States. Eighteen citations representing 15 studies were included in this review. Five overarching themes were identified: (1) routinized HIV risk assessment; (2) interdisciplinary/coordinated PrEP teams or services; (3) clinic capacity to provide essential PrEP-related services; (4) low out-of-pocket patient costs; and (5) access to the priority populations. Some of these themes are consistent with the recommendations of CDC's PrEP clinical guidelines and the EHE initiative. More recent studies that include perspectives of diverse providers, timely analysis of these studies, and implementation research to assess strategies to address the current practice gaps are needed to further promote PrEP prescribing among providers in the United States.

Understanding Pre-Exposure Prophylaxis Adherence in Young Women in Kenya

OBJECTIVE

To present detailed analyses of long-term pre-exposure prophylaxis (PrEP) use and associated behaviors and perceptions among young Kenyan women.

DESIGN

Prospective, observational cohort.

METHODS

The Monitoring PrEP among Young Adult women Study involved 18 to 24-year-old women at high HIV risk initiating PrEP in Kisumu and Thika, Kenya. Visits for PrEP counseling and dispensing, HIV testing, and socio-behavioral data collection occurred at Month 1 and quarterly for 2 years. PrEP adherence was measured with pharmacy refill and real-time electronic monitoring, plus tenofovir diphosphate levels in 15% of participants. HIV risk behavior and perception were assessed by self-report in weekly short message service surveys from Months 6-24. Predictors of adherence were assessed with multivariable logistic regression analysis.

RESULTS

Three hundred forty-eight women (median age 21, VOICE risk score 7) were followed for 617 person-years. Pharmacy refills steadily declined from 100% (Month 0-1) to 54% (Months 22-24). Average electronically monitored adherence similarly declined from 65% (Month 0-1) to 15% (Months 22-24). Electronically monitored adherence had moderately high concordance with tenofovir diphosphate levels (67%). High average adherence (5+ doses/week) was seen at 385/1898 (20%) participant-visits and associated with low baseline VOICE risk score, >1 current sexual partner, ≤1-hour travel time to clinic, and the Kisumu site. short message service-reported behavior and risk perception were not associated with adherence. Four women acquired HIV (incidence 0.7/100 person-years).

CONCLUSIONS

PrEP adherence was modest and declined over time. HIV risk was inconsistently associated with adherence; clinic access and site-level factors were also relevant. Relatively low HIV incidence suggests participants may have achieved protection through multiple strategies.

A Primary Care Intervention to Increase HIV Pre-Exposure Prophylaxis (PrEP) Uptake in Patients with Syphilis

Identifying candidates for HIV pre-exposure prophylaxis (PrEP) is a barrier to improving PrEP uptake in priority populations. Syphilis infection is an indication for PrEP in all individuals and can be easily assessed by primary care providers (PCP) and health systems. This retrospective study evaluated the impact of a multidisciplinary provider outreach intervention on PrEP uptake in patients with a positive syphilis test result in a safety-net hospital-based primary care practice. The PCPs of PrEP-eligible patients with a positive syphilis result were notified via the electronic medical record (EMR) about potential PrEP eligibility and institutional HIV PrEP resources. Rates of PrEP offers and prescriptions were compared in the pre (8/1/2018-12/31/2018, n = 60) and post (1/1/2019-5/31/2019, n = 86) intervention periods. Secondary analyzes evaluated receipt of appropriate syphilis treatment and contemporaneous screening for HIV, gonorrhea, and chlamydia. No significant differences in the overall proportion of patients offered (15% vs 19%) and prescribed (7% vs 5%) PrEP were observed between the pre- and post-periods. Overall, 7% of positive tests represented infectious syphilis. The rate of appropriate syphilis treatment was equivalent (57% vs 56%) and contemporaneous screening for other sexually transmitted infections was suboptimal across the entire study period. Although any positive syphilis test may be an easily abstracted metric from the EMR, this approach was inclusive of many patients without current HIV risk and did not increase PrEP uptake significantly. Future research into population health approaches to increase HIV prevention should focus on patients with infectious syphilis and other current risk factors for incident HIV infection.

Primary Care Providers' Perspectives on Using Automated HIV Risk Prediction Models to Identify Potential Candidates for Pre-exposure Prophylaxis

Identifying patients at increased risk for HIV acquisition can be challenging. Primary care providers (PCPs) may benefit from tools that help them identify appropriate candidates for HIV pre-exposure prophylaxis (PrEP). We and others have previously developed and validated HIV risk prediction models to identify PrEP candidates using electronic health records data. In the current study, we convened focus groups with PCPs to elicit their perspectives on using prediction models to identify PrEP candidates in clinical practice. PCPs were receptive to using prediction models to identify PrEP candidates. PCPs believed that models could facilitate patient-provider communication about HIV risk, destigmatize and standardize HIV risk assessments, help patients accurately perceive their risk, and identify PrEP candidates who might otherwise be missed. However, PCPs had concerns about patients' reactions to having their medical records searched, harms from potential breaches in confidentiality, and the accuracy of model predictions. Interest in clinical decision-support for PrEP was greatest among PrEP-inexperienced providers. Successful implementation of prediction models will require tailoring them to providers' preferences and addressing concerns about their use.

HIV Information Acquisition and Use Among Young Black Men Who Have Sex With Men Who Use the Internet: Mixed Methods Study

Background

HIV disproportionately affects young Black men who have sex with men (YBMSM) in the United States. eHealth holds potential for supporting linkage and engagement in HIV prevention and care and the delivery of HIV information to YBMSM.

Objective

This study aims to investigate HIV information acquisition and use among YBMSM who use the internet.

Methods

A web-based self-administered survey and semistructured interviews were conducted. The survey findings informed the development of the interview guide. Descriptive statistics were used to characterize the survey sample, and interview data were analyzed thematically using modified grounded theory methodologies.

Results

Among the internet sample (N=83), the average age was 29.2 (SD 3.5) years, 41% (n=34) of participants self-reported living with HIV, 43% (n=36) were HIV-negative, and 15% (n=13) were unsure of their HIV status. Most participants (n=79, 95%) acquired HIV information through the internet while using a mobile phone. Web-based HIV information was intentionally sought from consumer health information websites (n=31, 37%), government health information websites (n=25, 30%), and social media (n=14, 17%). Most men incidentally acquired HIV information via advertisements on social media sites and geospatial dating apps (n=54, 65%), posts on social media sites from their web-based social ties (n=44, 53%), and advertisements while browsing the internet (n=40, 48%). Although the internet is the top source of HIV information, health care providers were the most preferred (n=42, 50%) and trusted (n=80, 96%) source of HIV information. HIV information was used to facilitate the use of HIV prevention and care services. The qualitative sample included YBMSM across a range of ages and at different points of engagement in HIV prevention and care. Qualitative findings included the importance of the internet as a primary source of HIV information. The internet was used because of its ease of accessibility, because of its ability to maintain anonymity while searching for sensitive information, and to mitigate intersecting stigmas in health care settings. Participants used HIV information to assess their risk for HIV and AIDS, support their skill building for HIV prevention, inform patient–doctor communication, and learn about HIV prevention and treatment options. Men expressed concerns about their diminishing access to online spaces for HIV information exchange among YBMSM because of censorship policies on social media sites and the stigmatizing framing and tone of mass media HIV-prevention advertisements encountered while using the internet.

Conclusions

YBMSM in this sample had high utilization of eHealth for HIV information acquisition and use but diminished access to their preferred and most trusted source of HIV information: health care providers. Future eHealth-based HIV interventions culturally tailored for YBMSM should aim to reduce intersectional stigma at the point of care and support patient–provider communication. The findings demonstrate the need for community-informed, culturally tailored HIV messaging and online spaces for informational support exchange among YBMSM.

Age-Specific Risk Scores Do Not Improve HIV-1 Prediction Among Women in South Africa

Supplemental Digital Content is Available in the Text.

HIV-1 risk scoring tools could help target provision of prevention modalities such as pre-exposure prophylaxis. Recent research suggests that risk scores for women aged 18–45 may not predict risk well among young women aged 18–24. We evaluated the predictive performance of age-specific risk scores compared with the existing non–age-specific VOICE risk score, developed for women aged 18–45.

Patient-Led Decision-Making for HIV Preexposure Prophylaxis

PURPOSE OF REVIEW

Shared decision-making is a process that involves bidirectional exchange of information between patients and providers to support patients in making individualized, evidence-based decisions about their healthcare. We review the evidence on patient-led decision-making, a form of shared decision-making that maximizes patient autonomy, as a framework for decisions about HIV preexposure prophylaxis (PrEP). We also assess the likelihood that patient-led decision-making occurs for PrEP and describe interventions to facilitate this process.

RECENT FINDINGS

Patient-led decision-making is likely to be uncommon for PrEP, in part because healthcare providers lack knowledge and training about PrEP. Few evidence-based interventions exist to facilitate patient-led decision-making for PrEP. There is a need for rigorously developed interventions to increase knowledge of PrEP among patients and healthcare providers and support patient-led decision-making for PrEP, which will be increasingly important as the range of available PrEP modalities expands.

"Mini Dial-A-Nurses" and "Good Brands": What Are the Desirable Features of Online HIV/STI Risk Calculators?

A wide variety of risk calculators estimate individuals' risk for HIV/sexually transmitted infections (STI) online. These tools can help target HIV/STI screening and optimize clinical decision-making. Yet, little evidence exists on suitable features for these tools to be acceptable to end-users. We investigated the desirable characteristics of risk calculators among STI clinic clients and testing service providers. Participants interacted with online HIV/STI risk calculators featuring varied target audiences, completion lengths, and message outputs. Thematic analysis of focus groups identified six qualities that would make risk calculators more appealing for online client use: providing personalized risk assessments based on users' specific sexual behaviors and HIV/STI-related concerns; incorporating nuanced risk assessment and tailored educational information; supplying quantifiable risk estimates; using non-stigmatizing and inclusive framing; including explanations and next steps; and developing effective and appropriate branding. Incorporating these features in the design of online HIV/STI risk calculators may improve their acceptability among end-users.

Reducing the Discussion Divide by Digital Questionnaires in Health Care Settings: Disruptive Innovation for HIV Testing and PrEP Screening

BACKGROUND

Health care provider assessment of patient sexual behavior and substance use is essential for determining appropriate prevention interventions-including HIV pre-exposure prophylaxis (PrEP)-for sexual minority men (SMM). We sought to explore acceptability and utility of using electronic surveys to conduct health behavior assessments in clinical settings among SMM.

METHODS

Among a US nationwide sample of SMM (n = 4187; mean age = 38.3 years; 60% white; 82% HIV-negative), we examined associations of demographics, recruitment venue, sexual behavior characteristics, and recent substance use with participants' comfort communicating verbally and through electronic survey with a health care provider about sexual and substance use behavior.

RESULTS

On average, SMM had greater comfort communicating through electronic survey vs. verbally. In our fully adjusted analysis, preference favoring electronic surveys more strongly than verbal communication differed by age (β = -0.07, P ≤ 0.001). SMM with a bachelor's degree or more (β = 0.04, P < 0.05), those recruited from nonclinical settings (β = 0.06, P ≤ 0.001), and those without primary care providers (β = 0.04, P < 0.05) favored electronic surveys more strongly in the fully adjusted multivariable model. SMM who reported any recent casual sex partners (β = 0.05, P < 0.01), those never tested for HIV (β = 0.03, P < 0.05), and HIV-negative/unknown men not on PrEP (compared with PrEP users; β = 0.09, P ≤ 0.001) also favored electronic surveys in the fully adjusted model.

CONCLUSIONS

Reducing communication barriers by incorporating electronic surveys into patient assessments could help identify HIV testing and PrEP needs for SMM most susceptible to HIV acquisition. Nonetheless, no one screening strategy is likely to work for most SMM, and multiple approaches are needed.

Delivering PrEP to adults with "low" or "no" HIV risk and youth: experiences and perspectives of PrEP providers

Health care providers have an important role to play in expanding PrEP uptake to populations disproportionately burdened by HIV. In this study, in-depth, semi-structured qualitative interviews were conducted with 20 PrEP providers in Los Angeles, California to explore their experiences and perspectives prescribing PrEP. Overall, the findings indicate that providers generally adhered to US CDC clinical guidelines in determining PrEP eligibility. However, they also identified special considerations with respect to adults with "low" or "no" HIV risk. Three themes were developed from the data related to the delivery of PrEP to such individuals: 1) patients may not disclose sexual behaviours; 2) patient autonomy and agency; and 3) the importance of PrEP in gay male monogamous relationships. Two additional themes were identified related to prescribing PrEP to young people under the age of consent: 4) the need for more PrEP and sexual health education with youth and 5) challenges in providing PrEP to youth without parental consent. Findings highlight the importance of providing PrEP to patients who could potentially benefit from adoption, whether or not they present with clear behavioural indicators for PrEP. Providers should also consider the potential barriers to delivering PrEP to youth to ensure successful adoption among members of this population.

Artificial Intelligence and Machine Learning for HIV Prevention: Emerging Approaches to Ending the Epidemic

PURPOSE OF REVIEW

We review applications of artificial intelligence (AI), including machine learning (ML), in the field of HIV prevention.

RECENT FINDINGS

ML approaches have been used to identify potential candidates for preexposure prophylaxis (PrEP) in healthcare settings in the USA and Denmark and in a population-based research setting in Eastern Africa. Although still in the proof-of-concept stage, other applications include ML with smartphone-collected and social media data to promote real-time HIV risk reduction, virtual reality tools to facilitate HIV serodisclosure, and chatbots for HIV education. ML has also been used for causal inference in HIV prevention studies. ML has strong potential to improve delivery of PrEP, with this approach moving from development to implementation. Development and evaluation of AI and ML strategies for HIV prevention may benefit from an implementation science approach, including qualitative assessments with end users, and should be developed and evaluated with attention to equity.

Perceptions of Patient HIV Risk by Primary Care Providers in High-HIV Prevalence Areas in the Southern United States

The southern United States accounted for 52% of new HIV diagnoses in 2015. Visits to primary care providers (PCPs) offer opportunities for routine HIV screening. However, of at-risk persons in the United States who visited a health care provider within the previous year, >75% were not offered a test for HIV. Perceptions of patient population risk by PCPs could offer insight into these missed opportunities, and inform development of HIV testing interventions for PCPs to increase routine screening. During April-October 2017, we conducted online surveys regarding PCP's perceptions of patient HIV risk in six areas of the South with high-HIV prevalence. Surveys queried HIV-related knowledge, beliefs, attitudes, and practices. Free-text responses to the question "Are there any unique or special risk factors relating to HIV infection in your patient population?" were analyzed using NVivo for applied thematic analysis. Of 526 respondents, the mean age was 47 years with 65% white, 13% Asian/other, 13% black, 6% Hispanic/Latino; 71% female; 93% straight/heterosexual; and 35% offered HIV screening correctly based on standard of care. Main themes revealed were as follows: (1) provider perceptions of patient risk factors (e.g., "injection drug use is rampant"), (2) provider perceptions of patient barriers to access and care (e.g., "concern for parental notification and cost for treatment"), and (3) provider misconceptions of HIV risk and patient stigmatization (e.g., "I have a low-risk population"). Our findings suggest that PCPs in the South may warrant education regarding local HIV prevalence and routine HIV screening and prevention practices.

Factors Associated with HIV Seroconversion Among Women Attending an Urban Health Clinic in the South: A Matched Case-Control Study

Compared to knowledge about HIV risk factors among men in the south, less is known about risk factors for women. We conducted an individually matched case-control study to identify factors associated with HIV seroconversion among women. Cases had a clinician-assisted visit (CAV) between 2011 and 2016 at an Atlanta-based public health clinic before HIV diagnosis. Controls were women who visited the clinic but remained HIV negative. Controls were matched to cases in a 2:1 ratio on race, age at first CAV, and date of first CAV. Conditional logistic regression was used to develop a best-fitting model for characterizing HIV risk. Of 18,281 women who were HIV negative at their first visit, 110 (0.6%) seroconverted before 2019. Of these, 80 (73%) had a CAV before HIV diagnosis. Having multiple gonorrhea episodes, a syphilis episode, a greater number of sex partners in the past 2 months, anal sex, history of drug use, history of exchanging drugs or money for sex, and heterosexual sex with >1 sex partner in the last month were individually associated with HIV seroconversion. In multivariate analyses, having a syphilis episode [odds ratio (OR) = 4.7, 95% confidence interval (CI): 1.3-16.3], anal sex (OR = 2.8, 95% CI: 1.0-8.1), and injection drug or crack cocaine use (OR = 33.5, 95% CI: 3.6-313.3) remained associated with HIV. Women having all three risk factors were six times more likely to seroconvert compared to women without these factors. Our results offer insights into which women in a southern HIV "hotspot" may be at greatest risk for HIV.

Acceptability of electronic healthcare predictive analytics for HIV prevention: a qualitative study with men who have sex with men in New York City

BACKGROUND

Large data sets, also known as "big data", shared in health information exchanges (HIEs), can be used in novel ways to advance health, including among communities at risk for HIV infection. We examined values and opinions about the acceptability of using electronic healthcare predictive analytics (eHPA) to promote HIV prevention in men who have sex with men (MSM). Our aims were twofold: (I) to evaluate the perspectives of MSM with diverse race/ethnicity and age on the acceptability of predictive analytics to determine individual HIV risk and (II) to determine acceptability of having targeted prevention messaging based upon those risk estimates sent directly to the consumer.Method: Two of the authors facilitated 12 focus groups (n=57) with adult MSM without HIV, living in NYC. Groups were divided by ethnicity (Black, Latino, and White) and age (under 35 and 35 and over). Participants were recruited through HIV prevention sites, community-based organizations, social media, and Internet sites that serve these communities. Grounded theory methods were used to analyze the data with Dedoose.

RESULTS

We identified six main themes related to acceptability: (I) reach, relevance, and potential uptake of using predictive analytics to establish HIV risk and deliver targeted prevention messaging; (II) patient-provider communication; (III) public health and individual rights; (IV) perceptions of intervention effectiveness; (V) electronic health data security; and (VI) stigma. Within each thematic domain, MSM discussed concerns, benefits, and provided recommendations for implementation.

CONCLUSIONS

MSM in this study were supportive of the use of "big data" and technology to reach marginalized populations and improve public health, yet expressed concerns about the relevance, effectiveness, and security eHPA. Efforts to advance eHPA for HIV prevention should address these concerns, especially among the most-at-risk communities of color. Development of eHPA for HIV prevention should involve targeted messaging that addresses specific concerns regarding eHPA security, accuracy, and relevance.