GenAI adoption grows, but concerns over accuracy dominate
A KLAS report, its first to focus on genAI's adoption now and in the future, draws from the perspectives of 66 healthcare executives.
Photo: Andrew Brooks/Getty Images
Although the overall implementation of Generative AI in healthcare remains limited, a gradual increase in adoption rates, primarily among larger entities, is starting to take hold, according to a KLAS survey 66 of healthcare executives.
While just a quarter of respondents said they have already incorporated genAI solutions, about six in 10 (58%) expressed their intention to adopt or procure gen AI-powered applications within the next year.
The study, which was conducted across diverse healthcare settings, uncovered an inclination towards genAI tools among larger organizations, which are leveraging resources to integrate AI solutions from various providers including Epic, Google, Nuance, and OpenAI into their systems.
WHY THIS MATTERS
The results indicate the presence of sufficient resources and readily available data, coupled with dedicated data scientists, empower these organizations to more effectively develop, implement, and drive outcomes with genAI solutions.
The primary motivation for this surge in adoption is the aspiration to enhance efficiency across healthcare enterprises.
Respondents said they anticipate leveraging generative AI to streamline documentation processes, improve patient communication, and optimize workflows.
Automation of clinical notes generation, personalized patient communication, and workflow automation are expected to alleviate time-consuming tasks and reduce errors, in turn enhancing patient care and engagement.
Despite enthusiasm for the potential benefits, several uncertainties linger regarding the precise utilization of generative AI within organizational workflows.
First and foremost is the lack of a definitive strategy for AI implementation among several respondents, who also said they foresee challenges related to accuracy, cost-effectiveness, and security and privacy compliance.
Accuracy in particular was cited as a chief concern, with respondents expressing concerns over inaccuracies, bias, and AI "hallucinations" – wildly incorrect or improbably drawn conclusions – impacting patient care and decisions.
Ensuring robust security measures to protect sensitive patient data remains a paramount concern to meet stringent regulatory requirements such as HIPAA compliance.
The survey also highlighted the significant financial investment required to meaningfully implement and maintain AI infrastructure.
The initial investment might not yield immediate returns, thus necessitating careful assessment of long-term benefits and potential cost savings.
"As more organizations start leveraging generative AI, the solutions will need to drive hoped-for outcomes and prove a strong ROI to stay relevant and viable in the market long term," the report noted.
THE LARGER TREND
The healthcare industry's aversion to risk is one of the key stumbling blocks on the road to genAI adoption, even as applications for the technology can provide 99% accuracy in diagnostics, for example.
An October report from Moody's highlighted the positive impact AI could have on revenue cycles, but also cited the increased risk of deploying a technology that has still gone largely untested.
In September, Oracle Cerner integrated generative AI into its EHR systems with a product called Clinical Digital Assistant. The assistant lets providers leverage GenAI and voice commands for decreased manual workload and documentation.
CVS, Kaiser Permanente and Mayo Clinic have all invested in software startup Abridge, which has focused its efforts on leveraging Gen AI to aid in clinical documentation.
Twitter: @dropdeaded209
Email the writer: nathaneddy@gmail.com
