In this evaluation, we found no statistically significant differences in EHR-related and financial metrics between DAX users and the control group. However, exploratory results suggested that modest reductions in note time could result from using DAX at a high utilization level or deploying DAX to select clinician subgroups. Taken as a whole, these findings suggest that AI-enabled documentation’s efficiencies may translate to decreased markers of burnout for a subset of clinicians, and perhaps more broadly when the implementation of DAX achieves a higher adoption level. However, widespread implementation of DAX in its current form is unlikely to generate appreciable gains for health care systems looking to increase productivity.

To our knowledge, this is the first study to investigate whether this ambient-listening AI tool improves both efficiency and financial metrics from a health care system standpoint. These findings have important and timely implications as health care systems weigh the cost of adopting new technology. Indeed, the hype and novelty of ambient-listening AI tools have outpaced the evidence to support or refute claims that these tools are transformational in terms of time savings and efficiency. Consequently, health care systems, which already operate on small margins in hypercompetitive environments, run the risk of overpaying and not realizing the expected benefits. This study also contributes to the evidence-based incorporation of EHR use metrics into the evaluation of emerging technology and creates a foundational standard for comparison of outcomes across future studies.

There are several possible explanations for why we did not see large improvements in either EHR time or financial metrics. For example, while overall note time decreased, clinicians may have simply repurposed that time for other EHR activity they might otherwise not have done. In terms of financial metrics, clinicians with more free time may choose to use that time for other clinical activities or to leave work sooner. While our primary outcome found modest overall time saved (as viewed through a health system lens); our analysis identified subgroups among DAX users that saved substantial time. For example, 18% of participants saw a reduction of more than 1 hour a day in the EHR. While exploratory and unadjusted, these findings raise additional questions for future research to explore which user subgroups might achieve disproportionate benefits from using DAX. Additionally, a subset of clinicians noted that the documentation tool did indeed save them time; but rather than seeing more patients, these time savings allowed them to sleep more, spend less time working at home, and make existing encounters more focused and personal.⁵ Future research will need to evaluate whether modest time savings can add patient capacity or improve quality of care for existing patients, to support the expense of adding new technology. Additionally, in our organization, ambulatory clinicians are responsible for determining the level of service for billing encounters. Clinicians indicated that DAX captures more details from the clinician–patient interaction and gives clinicians documentary support for all the condition management they performed, often resulting in more comprehensive notes.⁵ With expansion to a larger cohort of users and over a longer time frame, this greater comprehensiveness could lead to more complete billing for services rendered, but additional study will be required to track its financial impact.

Research that evaluates how technology impacts clinicians’ daily work, in a variety of clinic specialties, has mixed findings.^3,4,7,11-14 Some studies have found that using AI tools helps create chart notes quickly and easily, thereby reducing EHR-related burnout and increasing clinician satisfaction.^3,4,11,12 However, many of these studies also show that these chart notes may need validation by clinicians due to possible inconsistencies and misinformation.^12-14 In a recent AMA survey, clinicians from different specialties expressed high levels of enthusiasm for using AI-powered tools to reduce documentation burden but also shared concern that it may impact the patient–physician relationship.¹⁵ With the evolution of the technology, DAX has addressed many implementation challenges and quality concerns. For example, DAX has improved its capability to accurately distinguish and transcribe when multiple people are talking, even when linguistic variations are present — a major barrier in other early-stage digital scribe tools.¹⁶ DAX can also perform near real-time note creation, allowing clinicians to retrieve transcribed notes within 30 seconds of completing the visit; this rapidity addresses clinical workflow challenges identified with older solutions.^7,14,17

Our study aligns with results from another study using a different ambient AI tool that was found to reduce time in notes for primary care physicians.¹³ Although we did not observe statistically significant improvements in total time in EHR or work time outside of work, in our separate qualitative study clinicians said DAX eased their cognitive burden, saved time on documentation, and allowed them to have more personal time.⁵ Even though “time saved” may not be captured with objective EHR use metrics, the clinicians’ reflections on improved work–life balance are noteworthy and demonstrate the potential for such tools to mitigate clinician burnout.

A few limitations should be noted. First, this is not a randomized controlled trial, and while we adjusted for a variety of factors, there will inherently be unmeasured confounding. In particular, the willingness to use DAX is a potential confounder. However, prior studies suggest that provider characteristics such as age, years of practice, specialty, and provider type are possible determinants of clinical adoption of AI^18,19; hence, the willingness to use DAX might be controlled to some extent by adjusting for the existing covariates in the analyses. In the fast-paced setting of emerging technology, health systems frequently must deploy the most rigorous evaluation possible based on the real-world situation. In this case, randomization was not possible, but the phased roll-out enabled the addition of a control group, which, while imperfect, adds a contemporaneous comparison.

Second, we were not able to obtain comprehensive editing time in DAX from all participating clinicians because the vendor only had data starting midway through the study (July). Hence, our data may underestimate the actual time in EHR and notes, if significant time was spent editing in the DAX editing window. However, our observations suggest that most clinicians transferred the drafted note into Epic prior to review and editing, because approximately 85% of recordings did not have edit times provided by DAX. Of the approximately 15% of recordings with an edit time, the average time was 177 seconds (about 3 minutes).

A third limitation that may have biased EHR time toward the null hypothesis is a learning curve among DAX users that may have resulted in inefficiencies. However, we conducted sensitivity analyses to include a 2-week washout period and the results remained the same.

Fourth, during the study period, Epic reported a widespread issue resulting in misclassification or underestimation of user audit log data, impacting EHR-Time₈, WOW₈, and Note-Time₈ metrics. Any potential undermeasurement would not be different between DAX users and the control group, so we do not expect this issue to influence our findings, and it highlights the value of a contemporaneous control group.

AI-powered ambient clinical documentation software presents a promising strategy to alleviate the documentation burden faced by outpatient clinicians. Our findings suggest that the tool did not make them more efficient in different aspects of EHR or financial metrics. However, the potential for substantial time savings and preventing clinician burnout emerged in our exploratory analysis and qualitative substudy.⁵ The pricing of this and other emerging technology will need to reflect demonstrable real-world benefits to make widespread implementation feasible and sustainable in a health care environment facing cost-control pressures. Further research may clarify which subgroups will gain efficiency from using the ambient documentation, and reveal opportunities to enhance the efficiency and seamless integration of the technology into workflows.

A data sharing statement provided by the authors is available with the full text of this article.

Disclosure forms provided by the authors are available with the full text of this article.

Cayla Mansel from Operations Performance Improvement, Medical Group, and Brittani M. Porter from Finance, Medical Group, helped pull operational and financial metrics from Atrium Health. Todd M. Banks from Ambulatory Services Department helped pull operational and financial metrics from Atrium Health Wake Forest Baptist Health. David Harold Russ from Human Resources assisted with the provision of clinicians’ demographics for analyses. Teammates mentioned above did not receive compensation beyond their usual salary for their contribution.