Assessing the Effectiveness of Large Language Model LLM-Enabled Nurse Treatment Planning in 2 Indian Districts A Pilot Study

What this Trial Is About

Researchers are evaluating whether AI-enabled, nurse-led treatment planning can match or improve the quality of clinical reasoning and management compared to standard physician-led care in adults aged 18 years and older. The study focuses on patients in rural and semi-urban India presenting with hypertension, diabetes mellitus, fever, breathlessness, or musculoskeletal pain. It aims to determine if nurse consultations supported by a large language model LLM achieve clinical quality scores that are not worse than those of physician consultations and to assess patient acceptance and satisfaction with AI-assisted nurse care. Participants receive two consultations during the same visit one led by a nurse using an AI-based clinical decision support tool and one by a physician providing standard care. The nurse-led consultation involves routine history taking and clinical assessment, with interaction through a digital interface to the LLM for assistance in diagnosis, reasoning, and treatment planning. The physician consultation follows usual clinical practice without AI support. The study compares these two approaches in a randomized order within each participant. Throughout the study visit, both consultations are audio recorded for blinded clinical quality evaluation. After the nurse LLM consultation, participants complete an exit survey measuring communication, trust, and satisfaction. Researchers also gather nurse-reported feedback on acceptability and feasibility through interviews after nurses complete at least 10 AI-assisted consultations. The main outcome measured is the clinical quality of the consultations immediately after both visits, with additional assessments of patient experience and nurse perspectives over up to nine months.

Assessing the Effectiveness of Large Language Model (LLM)-Enabled Nurse Treatment Planning in 2 Indian Districts