An online decision support tool for selecting HIV treatment regimens may help primary care doctors and others to care for people living with HIV (PLWH), the tool's developers reported in AIDS.
The computerized tool, called HIV-ASSIST, uses both patient- and virus-specific attributes, such as comorbidities, treatment history, or mutations, to help clinicians select appropriate antiretrovirals for a specific patient.
The algorithm's recommendations were validated against those of 17 experienced HIV providers in 10 hypothetical scenarios, four of which involved treatment-naive patients, three of which involved virally suppressed treatment-experienced patients, and three of which involved treatment-experienced patients with viremia. For that latter group, providers selected 20 different antiretrovirals due to the heterogeneity of prescribing practices in such cases, study authors noted.
HIV-ASSIST and the experienced human prescriber agreed in 99% of treatment-naive cases and in 88% and 84% of experienced cases, respectively. However, the tool is not intended to be a complete substitute for clinical judgement, study authors cautioned: Among other limitations, the tool does not include newer medications, such as doravirine, nor does it factor in potential financial costs to the patient (e.g., copayments).
Going forward, the program will include newer medications, may be incorporated into electronic medical records, and may available on a mobile platform, its developers promised.