Abstract
Background: The integration of AI-powered scribe technology offers potential solutions to reduce administrative burdens in clinical practice, aiming to improve both provider efficiency and patient care. The Frontenac Lennox and Addington Ontario Health Team (FLA OHT) has implemented an AI scribe pilot project to assess its effectiveness, usability, and overall impact on healthcare providers' workflow.
This evaluation investigates the experiences of healthcare providers using AI scribe technology, examining its impact on administrative efficiency, workflow integration, and provider and patient satisfaction. The study also explores enablers and barriers associated with AI scribe adoption, providing insights into how such tools can be optimized for broader use in primary care settings.
Approach: The selection of the AI scribe tool was conducted through a collaborative decision-making process led by the FLA OHT’s Primary Care Digital Health and Technologies Committee. This process included patient engagement and input from member physicians who piloted various tools to assess potential time savings and practice improvements. In year funding was allocated to purchase licenses for 80% of local primary care providers at an enterprise costing rate.
A mixed-methods and grounded theory approach was utilized, comprising quantitative usage data and qualitative feedback from healthcare providers through pre- and post-evaluation surveys, periodic check-ins, and semi-structured interviews. Surveys tracked metrics related to workload, ease of use, and impact on patient-provider interactions. Interviews were conducted to gain in-depth insights, coded thematically to capture recurrent themes and experiences across providers from various levels of scribe utilization. Key areas of focus included usability, effectiveness, safety, and change management.
Results: Preliminary findings indicate varying levels of comfort with the AI scribe among providers, with most identifying significant time savings and improved workflow efficiency. However, common challenges noted include accuracy issues in specific clinical contexts and the need for consistent vendor support. Providers reported mixed feedback regarding privacy and consent concerns and appreciated the coordinated consent support offered through the OHT. Patient perceptions of the AI scribe were positive with the majority of reports citing an appreciation of enhanced engagement as providers spent less time on documentation. Successful change management strategies, such as peer support and training sessions, were identified as critical enablers for effective implementation.
Implications: Initial results suggest that AI scribe technology can alleviate some administrative burdens for healthcare providers, improving efficiency and allowing for a greater focus on patient-centered care. However, enhancements in accuracy and tailored support for specific practice settings are essential to maximize the technology's benefits. The supported implementation of AI scribes was also crucial for engaging local primary care providers in digital technology adoption and created space for conversation around where providers would like to see future innovations to support them and their practice. Further investigation will refine these insights, contributing to a more scalable model for AI scribe deployment across healthcare systems.