Effect of AI Assisted Documentation in Primary Health Care on Time Saving, Patient Satisfaction and Health Care Provider Satisfaction

A Randomized Clinical Trial on the Effectiveness of AI-Assisted Documentation

Background There is currently no research evidence from randomized trial settings on the effectiveness of AI-assisted documentation. The aim of this study is to provide evidence regarding cost-effectiveness, professional experience, and patient experience.

Design Physician consultations participating in the study are randomized into AI-assisted and traditional documentation groups in a 1:1 repeated crossover design. The goal is to include approximately 1,000 consultations per group.

Methods Professionals will be asked to provide their own assessment of potential time savings, and the time spent on documentation will be measured using technical data from the documentation tools (cost-effectiveness). Additionally, professionals will complete baseline and follow-up surveys (professional experience), and patients will be sent a survey following the consultation (patient experience).

The study will include all acute and non-acute visits for the enrolled clinician. The study excludes sector-specific work and specialized consultations, such as maternity clinics, memory clinics, intellectual disability consultations, as well as non-Finnish-language consultations and interpreter-assisted consultations. The system has been primarily designed to monitor discussions between two individuals, and therefore, consultations where a guardian or companion accompanies the patient are also excluded from the study.

Inclusion Criteria:

• Clinician practices at a site in Western Uusimaa wellbeing services county where AI-assisted documentation tool is available

Exclusion Criteria:

• none

The study will include all acute and non-acute visits for the enrolled clinician. The study excludes sector-specific work and specialized consultations, such as maternity clinics, memory clinics, intellectual disability consultations, as well as non-Finnish-language consultations and interpreter-assisted consultations. The system has been primarily designed to monitor discussions between two individuals, and therefore, consultations where a guardian or companion accompanies the patient are also excluded from the study.