Allm's AI Summary Feature Cuts Visiting Nursing Documentation Time by up to 42%

Allm, a subsidiary of DeNA (Headquartered in Shibuya-ku, Tokyo; Representative Director and CEO: Shingo Okamura), announced the launch of an AI-powered summary feature (nursing report draft generation) for its multidisciplinary collaboration system, 'Team.' This feature is capable of reducing the time required for visiting nursing documentation by up to 42%.

This new feature was developed under the Strategic Innovation Promotion Program (SIP) Phase 3, "Construction of an Integrated Healthcare System," led by the Cabinet Office and managed by the National Institute of Health Crisis Management (JIHS). Allm acted as the lead research institution for the "B-4: Nursing Support and Improvement of Medical Quality" project. The development of this task was inherited from Theme 3, "Use of Generative AI in the Construction of Integrated Healthcare Systems," under the SIP Phase 3 supplementary budget.

Background: In visiting nursing, documentation such as records, plans, and reports constitutes a significant portion of a nurse's workload. Guidelines from the Ministry of Health, Labour and Welfare (2026) emphasize the necessity of accurately documenting and storing information on patients' medical history, medication status, living environment, and usage of other services, highlighting the burden of these administrative tasks. Consequently, improving efficiency to ensure high-quality care time has become an urgent challenge.

Features of the New Function: An LLM (Large Language Model) analyzes data stored in 'Team' to automatically generate highly accurate drafts with a single click. - Seamless data linkage: Automatically retrieves reports from the previous month and plans/records from the current month. - One-click generation: No complex prompt input required; drafts are completed with a single button. - Consistency assurance: Creates coherent text based on past progress, preventing transcription errors.

Overview of Empirical Tests: Empirical tests were conducted from August to October 2025 at six visiting nursing stations in Nagaoka and Sanjo cities, Niigata Prefecture. The tests involved 15 nurses and 460 instances of functional usage, comparing report generation times with and without the AI summary feature.

Results of Empirical Tests: Significant time-saving effects were confirmed at the visiting nursing stations: - Standard cases (report generation time under 10 minutes): 42% time reduction. - High-load cases (report generation time under 20 minutes): 39% time reduction.

Ongoing use of the AI was suggested to contribute not only to administrative efficiency but also to the reduction of nurses' psychological burden.