Note: This post is adapted from Assessing the Value of Ambient AI: Here’s What’s Working for Healthcare Organizations by Matthew Ko, DeepScribe CEO and co-founder, originally published on LinkedIn. You can read the full piece here.
Ambient AI has become one of the most regular topics of conversation in healthcare, with its capacity to help clinicians deeply engage patients and reclaim hours spent on administrative work. But healthcare organizations have a question: How well does ambient AI fulfill that promise?
Recent studies from institutions like the University of Pennsylvania and Stanford University have set out to quantify the impact of ambient AI in healthcare settings. While studies like these generally indicate that ambient documentation saves clinicians time and helps reduce burnout rates, the quantifiable ROI results have been somewhat underwhelming. The Stanford study reported a time savings of just over half a minute per encounter; the UPenn research noted two minutes saved per visit.
These modest time savings highlight a critical issue in assessing the value of ambient AI: the lack of standardized metrics. Healthcare organizations often define the value of ambient AI by time saved and through qualitative provider feedback. But this approach overlooks a crucial factor in realizing value: clinician adoption of ambient AI.
Adoption: A Key Indicator of Success
Adoption rates serve as a cornerstone for gauging an organization's capacity for success with ambient AI. The higher the adoption, the more likely clinicians are doing well integrating the technology into their daily workflows.
At DeepScribe, we've consistently achieved adoption rates of 75-80% in large specialty settings. For these organizations, high adoption rates reflect a fundamental shift in how clinicians interact with technology and integrate it within patient care. They also represent engaged teams who trust their AI product and may be more willing to rely on functionality beyond documentation, such as features related to revenue cycle management or quality reporting.
Factors Contributing to Strong Adoption
One significant factor contributing to adoption rates is the customization of ambient AI solutions for specific medical specialties. The UPenn study highlighted this need, assessing that clinicians showing lower optimism after using the tool may come from specialties whose workflow was less conducive to a general ambient scribing approach.
This underscores the importance of tailoring AI models to understand more than just language – they must comprehend visit types and create context-aware notes that avoid unnecessary repetition and accurately reflect current conversations and care.
Another critical element in driving adoption is the quality and accuracy of AI-generated notes. Ambient AI companies focus on highlighting customer stories about saved time and reduced burnout, but there is a concerning lack of rigor and accountability surrounding the quality of AI-produced clinical notes. Healthcare organizations should prioritize asking AI vendors about their quality metrics, how they're determined, and how the results may translate for their particular medical specialty.
Additionally, collaboration between healthcare organizations and AI vendors is required for successful implementation and adoption. This partnership should focus on customizing AI solutions to fit the specific needs of different medical disciplines and, most important, ensuring that clinicians have ongoing input in the process. By working together, benefits can be realized at every level – from the exam room to the billing department.
Ultimately, the true value of ambient AI in healthcare should extend far beyond time savings. By focusing on clinician adoption rates as the key success metric – the greater the adoption, the greater the outcomes – healthcare organizations can gain a more comprehensive understanding of how well the technology is integrated into their workflows and its potential for driving more meaningful improvements in patient care and operational efficiency.
text