UPMC, KLAS Research study examines AI adoption trends, governance barriers

A recent study conducted by UPMC's Center for Connected Medicine and KLAS Research delves into the trends of AI adoption within the healthcare industry, as well as the barriers hindering effective governance. The report reveals that despite the widespread implementation of AI solutions, many organizations are still struggling to establish the necessary infrastructure and governance frameworks.

Seventy-six percent of respondents reported utilizing third-party AI solutions, while only 44% had a dedicated data platform or environment for testing these tools. The findings highlight that 92% of respondents test third-party tools before deployment, emphasizing the importance of thorough evaluation.

According to Rob Bart, M.D., UPMC's chief medical officer, "What's emerging from this research is a clear recognition that implementation is only the first step. Health systems are now focused on building the governance structures, testing capabilities and organizational strategies necessary to ensure AI delivers meaningful and measurable value."

The study surveyed 27 healthcare leaders, including representatives from health systems and ambulatory care organizations. Clinical documentation tools emerged as the most commonly adopted AI solution, with 52% of respondents reporting their use, followed by revenue cycle, coding, and billing applications (36%).

Data analysis primarily occurs within electronic health records (EHRs) or vendor analytics tools, with some organizations also relying on cloud data warehouses/lakehouses and multiple data marts or warehouses. Five respondents admitted to being uncertain about their primary data storage methods.

The report identifies data quality issues, such as manual workarounds, spreadsheets, and inconsistent definitions across teams, as the top pain points for health systems. Missing or incomplete data, unstructured data, and timeliness problems further complicate the process of transforming clinical and operational data into usable inputs.

Analysts agree that the challenges of AI implementation are not solely technical but also operational and governance-related. Internal barriers to executing AI strategies include resource, budget, time, or talent constraints; governance, security, and compliance concerns; change management, adoption, and education; use-case and vendor selection; return on investment (ROI) or financial cases; and workflow redesign or operational engagement.

The findings suggest that the next phase of AI strategy will hinge less on enthusiasm and more on execution, encompassing strong governance, clear problem statements, realistic resource planning, robust change management, and careful consideration of AI integration into daily work processes.