Ritten AI Scribe

The age of the AI Scribe is upon us. The rapid advancement of generative AI technology has made the ever elusive dream of “fully automated progress notes” closer than ever to reality. A wave of companies pursuing this vision have emerged in the past few years, often as browser extensions or point solutions that integrate with existing EHRs.

‍

Some of these companies are focused on healthcare broadly, integrating with hospital EHR systems. Others are focused on one or more specialties, working with various vertical practice management vendors that serve those providers. This dynamic landscape can lead to a tyranny of choice for programs. Which option is best? At Ritten, we believe that the best scribe (meaning a tool that passively listens to a conversation and generates a note) will be the one offered by the EMR itself. While standalone scribe tools can be developed quickly and are driving substantial innovation, the EMR’s position as the clinical workflow owner and repository of the richest client information position it best to deliver the greatest value as these tools mature. The EMR understands the form structure, question meaning, and client data better than a point solution ever can. 

‍

Notes don’t live in a vacuum. In a well formed behavioral health note or assessment, a clinician will include details and observations not captured verbally during the session. Observations from a prior session, current treatment plan elements, and data captured outside of a single conversation have an important role to play in documentation. As the system of record for all of those other documents and data sources, the EMR is uniquely positioned to incorporate all of that other information into the scribe’s response. Choosing specific data elements to consider and implementing those integrations across multiple EMR platforms is slow, administratively tedious, and in many cases, unrealistic.

‍

Workflow over models. AI models are quickly improving and have become largely commoditized (or will be soon). Proprietary models don’t command anything close to the level of investment and expertise that large vendors like OpenAI, Google, and others can muster behind their ever-improving tools. This means that a standalone scribe tool can’t leverage a proprietary model to differentiate its performance from an organic EMR feature that sits on a comparable model. Without an advantage in its underlying AI technology, a point solution that sits on the fringes of a user’s workflow won’t be able to overcome its disjointed user experience relative to a feature within the EMR.

‍

Robust AI scribe features require more than just a button in the EMR that kicks off a recording to be truly integrated. They require, for example, the flexibility to understand complex customized forms and incorporate field-by-field level answer review and acceptance. Building a point solution to handle this level of integration across multiple EMR platforms has a high likelihood of being complicated and brittle. The EMR will always have better organic workflow integration, with faster and more flexible iteration.

‍

The Scribe is just the beginning. At Ritten, our vision for leveraging AI to speed up documentation creation goes far beyond the traditional “Scribe” concept. Consider the structure of a behavioral health treatment plan review or discharge plan. There are certainly elements of “scribing” in these documents, where clinician-patient conversations are summarized to answer specific questions. However, there are many other categories of fields as well. Some questions might ask the clinician to synthesize changes since the last time the document was created. Other sections might require a summary of care provided since a certain date, or they might be a straightforward list of medications. Ritten will use a suite of different AI tools to integrate data from across the client chart to automate these common tasks. 

‍

With these considerations in mind, we’re excited at Ritten to realize this goal of an AI assistant within a behavioral health EMR over the coming months. Our existing assessment summarization tools, recently released AI scribe, and core features like customized forms, set the stage for a broad suite of in-app AI based tools that give providers seamless support within their existing workflows. We can return much needed time to clinicians, and support engaged, data enriched client care.

‍