Excessive clinical documentation is a common problem that causes physician burnout and even impacts patient care, a recent study by the American Society for Health Informatics found. It revealed doctors and nurses’ frustration with the burden of maintaining electronic health records and the time and effort required to complete the necessary documentation.
There are technologies that can aid. Many healthcare systems are using natural language processing tools, generative artificial intelligence, and ambient scripting to assist physicians with EHR documentation tasks.
But these tools are not plug-and-play. To work effectively, they must be tailored to the specific clinical needs of end users.
Dr. Dean Dalili, DeepScribe’s chief medical officer, says customization through ambient scribes is an crucial feature, especially for medical specialties, because the ability to personalize notetaking reduces the time physicians spend editing automated notes.
“Less time is spent in the physical interface” on formatting and other individual note-taking preferences, and more time is spent on patients, he said.
In this Q&A session, Dalili also discussed ambient intelligence and its benefits for enterprises. The feature compares note taking with coding standards and creates a reporting framework to check the quality of notes from clinical visits at various points of care.
Q. How does environmental clinical documentation improve the quality of care delivery and reduce burnout?
AND. To understand the impact of AI environment documentation, previous standards must be considered. Clinical documentation was kept in paper form and then transferred to the EHR [federal] policy.
This transition has had a positive impact on quality and safety, particularly in drug safety, with physicians gaining greater insight into the patient context when making decisions.
The problem, however, is that EHR advisors have compromised the quality of care delivery because physicians often engage more visually with their computers than patients do.
EHRs contribute to burnout – which is actually a phenomenon associated with high levels of mental workload. It’s part of being a clinician, but it’s made worse when you’re constantly switching gears to different types of work and translating the encounter into some visual output in a documentation format via keyboard.
AI clinical documentation allows the provider to directly contact the patient and conduct a normal conversation. This conversation becomes a source of information for which the software creates comprehensive, structured documentation. In some ways, more versatile than relying on the vendor’s memory.
Technology is always listening and sometimes picks up details that the provider forgets or doesn’t focus on. These details go into the clinical documentation, resulting in a better experience for the patient and better quality notes.
Q. Why are scribe’s customizable AI tools crucial in secondary care settings?
AND. Personalization is crucial for every doctor, but most crucial for specialists. As doctors, we fall into a rhythm.
A one-size-fits-all routing solution that generates a note with a standard structure will not be supported by most providers. There are nuances in how providers like to capture information – this could be the subjective history parts of the note, the physical examination parts, or the assessment and plan part where the provider groups parts of the treatment plan with each clinical patient. rating.
There are also note formatting elements that every doctor likes. If you are a geriatrician, you may address patients as “Mr.” or “Miss”; if you are a pediatrician, you do not want to address a teenage patient this way and may want to operate only their first name.
Having customization options helps create a note that most likely mimics the provider’s preferences, what they are used to, or what they have written in the past. This is crucial because when the output matches the supplier’s documentation preferences, it requires less editing from the supplier.
Think about the value proposition of any AI script: you don’t have to spend time talking first and then spend more time documenting the conversation. However, if you still need to edit this note to make it look the way you want, you still have a lot of extra work to do. That’s why personalization is crucial for specialties.
Specialty-specific workflows are recorded differently than a regular primary care visit. As such, there are different areas of emphasis and detail that vendors want to capture. It is crucial to set up the format not to be generic, one-size-fits-all and applicable to certain people, but rather to allow for the results and structure to be tailored so that the AI listens for elements of the visit that are unique to that specialty.
For example, in oncology, there is usually a very long data summary that outlines the patient’s diagnosis and all the data elements needed to identify their problem. Additionally, certain elements of the plan may be specific to cancer treatment – related not only to medical therapy, but also to social support, nutrition and other issues. The note for orthopedists might look very different and focus on the musculoskeletal physical examination, imaging, and so on.
Q. How is this technology different from other AI developers on the market?
AND. First, we operate a unique enormous language model that incorporates historical data from clinical visits encoded by live scribes, which helps create structured data elements. We train our LLM – unlike an LLM like ChatGPT4 which is trained online.
If you operate medical information to refine your LLM, you are more likely to obtain correct medical-related results. If you are training LLM all over the internet, there will be additional noise that may affect the content.
DeepScribe has the largest source of training data based on user notes used to create highly correct documentation, which helps build trust and adoption and minimizes time spent by suppliers on rework and edits.
The second standout feature is that the tool offers over 50 different customization elements, enabling vendors to create work closer to what they would create from scratch, across a wide range of specializations and users.
The third key differentiator is a recent category called ambient intelligence, which encompasses functionality beyond description.
This is where patient talk can be applied to any type of structured data, whether it’s a coding standard or a clinical quality standard. Based on this, the AI can determine whether the conversation met this coding standard or not.
This intelligence also allows us to create a reporting structure that allows an enterprise to instantly review physician performance across a wide range of providers. This is an opportunity to aid identify high-value clinical content at the point of care, as well as assess whether that content has been delivered.
Q. Does DeepScribe integrate with physicians’ existing workflows and technology stack?
AND. With ambient AI, the workflow is fundamentally different and requires some adjustments – the clinician must vocalize findings in a way they may not have done in the past.
Suppliers can’t just say, “That doesn’t look good.” Instead, they have to say, “Your left knee looks swollen,” so the language is quite specific for providers to adjust to. This level of detail makes AI listening more reliable.
The degree of integration depends on the EHR, and DeepScribe has integrations with Epic, athenahealth, eClinicalWorks, and over a hundred other EHRs, but if a physician wants to connect a proprietary EHR, DeepScribe can also integrate via API.