Finding the Right Recipe for Documentation

Documentation Recipes

To determine how an organization’s documentation recipe is likely to impact EHR satisfaction and efficiency, each measured EHR deployment in the Arch Collaborative was sorted into one of eight groups depending on the organization’s average adoption level (as self-reported by physicians) of speech recognition, scribes, and dictation (adoption is considered high if it is above the median). While satisfaction and efficiency vary across these groups, not all variation is statistically significant. Overall, documentation method is a poor predictor of EHR experience and provider burnout (mathematically has only a limited impact). Statistically significant correlations are called out in future pages.

Shifting the Paradigm

Surprised by the lack of impact that documentation method has on EHR satisfaction and clinician burnout? KLAS encourages organizations to consider that how these technologies and services are implemented can play a significant role in how well they move the needle.

As noted above, the success of scribes is often dependent on the quality of the training and the expectations set for providers. Similarly, use of speech recognition technology today is not always accompanied by sufficient user education and workflow redesign. Though the technology is likely to change drastically in coming years (thanks to the development of ambient intelligence), KLAS asserts that proper training and user assistance will still be important success factors.

Outcomes by Recipe

Overall, the most effective organizations focus on only a couple of documentation methods (including personal entry and excluding dictation), likely because specializing in a couple of documentation methods allows for better support and training. Organizations with dictation or with many means of documentation see lower success.

Don’t Just Check the Box for Speech

Over 2,000 of the surveyed physicians commented on their speech recognition technology, with 45% reporting frustrations. The high number of positive comments indicates that the technology can be utilized well when accompanied by excellent EHR proficiency and speech recognition education. The quotes below are representative of the most common sentiments about speech recognition:

“I LOVE the new [version of our speech recognition solution]. I feel like after just a few days, it knows me better than the other version did after years of me yelling at it.”

“My biggest complaint is with [our speech recognition solution]. It’s not reliable. It shuts down and crashes Epic on a daily basis. I expected this to be resolved with the latest Epic upgrade, but it was not. I usually stop using [the speech recognition solution] for the day if it crashes. The cycle repeats itself the next day. When it works smoothly, it makes the workflow more efficient.”

“Please provide the opportunity for ongoing, hands-on, at-the-computer training for both [our speech recognition solution] and Epic. I believe giving employed physicians the opportunity to observe Epic-savvy staff use and optimize the system would be a great financial investment for the corporation. Additionally, having an optimization-minded advocate from our organization looking at user problems and resolving them with Epic would really reduce job frustration and improve user efficiency.”

The data shows that the use of speech recognition technology does not guarantee documentation success—an organization can achieve very strong or very weak results depending on how the technology is implemented and how well users are trained. With how speech recognition technology is commonly implemented today, providers often report better efficiency, similar or lower EHR satisfaction, and sometimes lower confidence in their own documentation. In general, physicians report better outcomes when organizations make an all-in bet with speech recognition (i.e., don’t use scribes or dictation).

Scribes

Among those using scribes, the differences are significant depending on how the data is analyzed.

The physicians who are given scribes are often the most frustrated, so it is not surprising that as a group, they report lower EHR satisfaction and worse chart-closure rates than their peers who use other documentation methods. However, the longitudinal data indicates that these metrics can improve with use of a scribe. Next year, KLAS will be able to show trends in how scribes impact physician burnout rates.

In Summary

It is increasingly clear that organizations looking for a quick fix to their EHR satisfaction challenges, either through scribes or speech recognition, are likely to be disappointed. Those that adopt these technologies and processes in tandem with strong change management and strong EHR proficiency will likely see efficiency gains.

It should be noted that the results achieved today could change as ambient intelligence supplements speech recognition technology, as organizations adopt better implementation and training programs for their technology environment, and as organizations improve how they implement scribe programs. As these technologies and services evolve, the Arch Collaborative will continue to monitor their impact and provide insights into how they can best be implemented and what results organizations can expect.