AI Model Uncovers Hidden HIV Risk Factors for Black Women

Social Determinants Outshine Behavior in New Atlanta Study

New research highlights that societal and healthcare access issues, rather than just personal behaviors, are more potent indicators of HIV risk for Black women in the Southern U.S. The findings stem from an extensive analysis of over 300,000 women’s medical records.

Unlocking Deeper Risk Insights

Machine learning was employed to construct a predictive model, analyzing electronic health records from women treated in Atlanta between 2012 and 2022. Researchers, led by Dr. Meredith Lora of Emory University, aimed to improve HIV risk assessment for Black women, who face a disproportionately higher risk compared to White women. Their model incorporated structural social determinants and healthcare utilization patterns, alongside traditional factors.

Key predictors identified included younger age, Black race, residence in areas with high HIV incidence, and frequent changes in contact information prior to diagnosis. Women who underwent more frequent HIV screenings were also more likely to receive a diagnosis, suggesting potential underlying risk factors. Furthermore, individuals with HIV were more inclined to seek care in emergency departments rather than primary or women’s health settings.

Interestingly, the study noted that “seeking sexual health was more important to the model than STI positivity.” This suggests that a proactive approach to sexual health awareness might be a stronger indicator than actual STI diagnosis alone.

Bridging the Prevention Gap

Dr. Monica Gandhi, Director of the UCSF Bay Area Center for AIDS Research, commented on the study’s significance, noting the historical difficulty women face in accurately assessing their personal HIV risk. She pointed to a stark reality: while over 2.2 million individuals in the U.S. need pre-exposure prophylaxis (PrEP), only about 336,000 have prescriptions, according to CDC data (CDC, 2023).

“Machine learning using electronic medical records [EMR] can make risk prediction more accurate by reporting both health and social factors,” Dr. Gandhi explained. She expressed surprise at sexual health seeking behavior being a stronger predictor than STI positivity, suggesting women’s HIV risk awareness might be higher than previously understood in smaller studies.

“I was surprised by the fact that seeking sexual health was a stronger predictor of HIV than STD positivity, which indicates that women actually are more aware of their risk for HIV in the US than suggested in previous studies that usually involved smaller sample sizes.”

—Dr. Monica Gandhi, Director, UCSF Bay Area Center for AIDS Research

Dr. Gandhi also found the increased risk associated with frequent address or phone number changes noteworthy, highlighting mobility as a potential risk factor, similar to patterns observed in sub-Saharan Africa. The preference for emergency department care among those with HIV also points to critical intervention opportunities within these settings.

Future Directions in HIV Prevention

The study’s authors emphasized that while improved models can identify more candidates for PrEP, their real-world impact hinges on effective implementation and support for uptake and behavior change. Dr. Gandhi advocates for targeted PrEP services in historically Black neighborhoods and counseling for mobile individuals on HIV prevention, especially PrEP.

Future qualitative research should explore why women choose emergency departments for sexual health services, aiming to improve re-engagement with care. Understanding women’s knowledge of PrEP and whether it was previously offered could also illuminate further prevention strategies.