A multidisciplinary, AI-supported quality improvement intervention to manage polypharmacy in aging people with HIV

IF 3.2 3区医学 Q2 INFECTIOUS DISEASES

HIV Medicine Pub Date : 2026-07-07 Epub Date: 2026-04-11 DOI:10.1111/hiv.70234

Jovana Milic, Antonia Pugliese, Michela Belli, Gian Luca Lonardi, Caterina Ruffilli, Tommaso Albano, Marco Visicaro, Martina Ricciardetto, Pierluigi De Cosmo, Chiara Mussi, Francesca Gandolfi, Cristina Mussini, Costantino Grana, Giovanni Guaraldi

{"title":"A multidisciplinary, AI-supported quality improvement intervention to manage polypharmacy in aging people with HIV","authors":"Jovana Milic, Antonia Pugliese, Michela Belli, Gian Luca Lonardi, Caterina Ruffilli, Tommaso Albano, Marco Visicaro, Martina Ricciardetto, Pierluigi De Cosmo, Chiara Mussi, Francesca Gandolfi, Cristina Mussini, Costantino Grana, Giovanni Guaraldi","doi":"10.1111/hiv.70234","DOIUrl":null,"url":null,"abstract":"<div>\n \n \n <section>\n \n <h3> Objectives</h3>\n \n <p>Aging people with HIV are increasingly affected by multimorbidity and polypharmacy, which heighten the risk of drug–drug interactions (DDIs) and potentially inappropriate medications (PIMs). This study evaluated a multidisciplinary, AI-supported quality improvement intervention designed to optimize polypharmacy management in older people with HIV.</p>\n </section>\n \n <section>\n \n <h3> Methods</h3>\n \n <p>People with HIV aged ≥50 years attending the Modena HIV Metabolic Clinic (MHMC) were invited to submit photos of their medications via WhatsApp. Images were processed by AI for optical character recognition and automatically reconciled with the electronic patient chart (EPC). AI recognition accuracy was 94% when validated against manual review. Pharmacists reviewed AI-generated reports from the NavFarma® decision support system, generated alerts for PIM, defined according to Beers and the STOPP/START criteria, DDIs, anticholinergic burden (ACB), and risks of QTc prolongation and nephrotoxicity. Primary outcome was agreement between patient-reported and EPC-recorded medications. Secondary outcomes included pill burden, total prescribed drugs and actionable alerts.</p>\n </section>\n \n <section>\n \n <h3> Results</h3>\n \n <p>Of 181 participants (median age 63 years; 72% male), 111 (61.3%) showed complete agreement between EPC and patient lists, while 70 (38.7%) had discrepancies. Pharmacist evaluation identified major DDIs in 70.4% of cases, ACB in 26.5%, QTc-prolonging drugs in 81.6% and nephrotoxic agents in 95.9%. Participants with ≥10 total prescribed drugs had higher frailty, pill burden and PIM.</p>\n </section>\n \n <section>\n \n <h3> Conclusions</h3>\n \n <p>AI-assisted medication reconciliation combined with pharmacist review improved the identification of PIM and medication-related risks, supporting safer prescribing in people with HIV. This model aligns with international calls to improve prescribing safety and offers a scalable framework for integrating digital tools into multidisciplinary HIV care.</p>\n </section>\n </div>","PeriodicalId":13176,"journal":{"name":"HIV Medicine","volume":"27 7","pages":"1109-1119"},"PeriodicalIF":3.2000,"publicationDate":"2026-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/hiv.70234","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"HIV Medicine","FirstCategoryId":"3","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1111/hiv.70234","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2026/4/11 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"INFECTIOUS DISEASES","Score":null,"Total":0}

引用次数: 0

Abstract

Objectives

Aging people with HIV are increasingly affected by multimorbidity and polypharmacy, which heighten the risk of drug–drug interactions (DDIs) and potentially inappropriate medications (PIMs). This study evaluated a multidisciplinary, AI-supported quality improvement intervention designed to optimize polypharmacy management in older people with HIV.

Methods

People with HIV aged ≥50 years attending the Modena HIV Metabolic Clinic (MHMC) were invited to submit photos of their medications via WhatsApp. Images were processed by AI for optical character recognition and automatically reconciled with the electronic patient chart (EPC). AI recognition accuracy was 94% when validated against manual review. Pharmacists reviewed AI-generated reports from the NavFarma® decision support system, generated alerts for PIM, defined according to Beers and the STOPP/START criteria, DDIs, anticholinergic burden (ACB), and risks of QTc prolongation and nephrotoxicity. Primary outcome was agreement between patient-reported and EPC-recorded medications. Secondary outcomes included pill burden, total prescribed drugs and actionable alerts.

Results

Of 181 participants (median age 63 years; 72% male), 111 (61.3%) showed complete agreement between EPC and patient lists, while 70 (38.7%) had discrepancies. Pharmacist evaluation identified major DDIs in 70.4% of cases, ACB in 26.5%, QTc-prolonging drugs in 81.6% and nephrotoxic agents in 95.9%. Participants with ≥10 total prescribed drugs had higher frailty, pill burden and PIM.

Conclusions

AI-assisted medication reconciliation combined with pharmacist review improved the identification of PIM and medication-related risks, supporting safer prescribing in people with HIV. This model aligns with international calls to improve prescribing safety and offers a scalable framework for integrating digital tools into multidisciplinary HIV care.

Abstract Image

查看原文本刊更多论文

一项多学科、人工智能支持的质量改进干预，以管理老年艾滋病毒感染者的多重用药。

目的：老年艾滋病毒感染者越来越多地受到多种疾病和多种药物的影响，这增加了药物-药物相互作用（ddi）和潜在不适当药物（PIMs）的风险。本研究评估了一项多学科、人工智能支持的质量改善干预措施，旨在优化老年艾滋病毒感染者的综合用药管理。方法：Modena HIV Metabolic Clinic （MHMC）邀请年龄≥50岁的HIV感染者通过WhatsApp提交他们的药物照片。采用人工智能对图像进行光学字符识别，并与电子病历自动核对。人工智能识别的准确率为94%。药剂师审查了NavFarma®决策支持系统中人工智能生成的报告，根据Beers和STOPP/START标准、ddi、抗胆碱能负担（ACB）、QTc延长风险和肾毒性生成PIM警报。主要结局是患者报告的药物和epc记录的药物之间的一致性。次要结局包括药片负担、总处方药物和可采取行动的警报。结果：181名参与者（中位年龄63岁，72%为男性），111名（61.3%）显示EPC与患者名单完全一致，而70名（38.7%）存在差异。药师评价确定主要ddi的病例占70.4%，ACB占26.5%，qtc延长药物占81.6%，肾毒性药物占95.9%。总处方药物≥10种的受试者虚弱程度、服药负担和PIM较高。结论：人工智能辅助的药物调节结合药师审查改善了PIM和药物相关风险的识别，支持艾滋病毒感染者更安全的处方。这一模式符合国际上提高处方安全性的呼吁，并为将数字工具整合到多学科艾滋病毒护理中提供了可扩展的框架。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

HIV Medicine 医学-传染病学

CiteScore

5.10

自引率

10.00%

发文量

167

审稿时长

6-12 weeks

期刊介绍： HIV Medicine aims to provide an alternative outlet for publication of international research papers in the field of HIV Medicine, embracing clinical, pharmocological, epidemiological, ethical, preclinical and in vitro studies. In addition, the journal will commission reviews and other feature articles. It will focus on evidence-based medicine as the mainstay of successful management of HIV and AIDS. The journal is specifically aimed at researchers and clinicians with responsibility for treating HIV seropositive patients.