Abstract Body

Background

In people with HIV (PWH), multimorbidity often leads to polypharmacy, which increases the risk of adverse drug reactions, drug–drug interactions (DDIs), poor adherence, and functional decline, underscoring the need for safer prescribing strategies. We describe the implementation of a multidisciplinary, AI-supported quality improvement (QI) intervention at the Modena HIV Metabolic Clinic (MHMC) to optimize medication management in aging PWH.

Methods

We conducted a cross-sectional, single-center QI project involving PWH aged ≥50 years. The intervention included four components: (1) AI-driven medication reconciliation through patient-submitted photographs of drugs; (2) Clinical decision support via NavFarma®, integrating comorbidities and identifying risks including potentially inappropriate medications (PIMs), DDIs, anticholinergic burden, nephrotoxicity, and QTc prolongation; (3) pharmacist review of AI-generated reports and therapeutic recommendations; (4) multidisciplinary decision-making to incorporate recommendations into care plans.

Results

Between November 2024 and July 2025, 181 PWH submitted medication photographs. Median age was 63 years; 72% male. Median CD4 count was 747 cells/μl, and HIV duration 31 years. Agreement between the electronic patient chart (EPC) and patient-reported medications was complete in 61.3%, while 38.7% showed discrepancies (28.2% undocumented EPC prescriptions; 34.3% undocumented patient medications). Among new prescriptions, 13 (61.9%) were high intensity statins, while discontinued drugs were low intensity statins (11, 57.9%) and proton pump inhibitors (3, 15.8%). Daily pill burden was 6, with 10 active agents. Among 98 PWH evaluated by NavFarma® and pharmacists, 135 recommendations were generated, 48.2% clinically relevant; only 1.7% led to discontinuation. PIM included major DDIs (70.4%, often involving CNS depressants, antiretrovirals, and cardiovascular drugs), QTc prolongation risk (44.9%), anticholinergic burden (26.9%), and nephrotoxicity risk (38.3%).

Conclusions

This multidisciplinary, AI-supported QI intervention was feasible and effective in identifying medication discrepancies and safety risks in aging PWH. The program captured agreement rates along with anticholinergic burden, QTc prolongation, nephrotoxicity, and DDIs, highlighting key areas for intervention. These findings support the integration of AI-assisted pharmacist reviews into HIV care to optimize prescribing and enhance medication safety in complex patients.