AI Scribes Part 1: Reducing Documentation Time and Uncovering New Ethical Challenges in Health Care

Image credit: © PRIM - stock.adobe.com

The Documentation Dilemma

Imagine a world where clinicians spend more time looking into their patients’ eyes than at computer screens. Since the first electronic medical record (EMR) emerged in 1972, clinicians have found themselves increasingly trapped behind screens. Today, clinicians dedicate a staggering 35% to 37% of their time to documentation tasks. For primary care physicians, this translates to 16 minutes of every patient visit spent on EMR duties. It’s a crisis that threatens the very essence of patient care.¹

Inside the AI Scribe-a-thon: A Clinician’s Experience

When I was invited to participate in an AI Scribe-a-thon hosted by XPC, a primary care innovation community, I jumped at the chance. The AI Scribe-a-thon was a unique, nationwide virtual event that spanned several days, inviting clinicians from across the United States to explore the cutting-edge capabilities of AI-powered medical scribe platforms. Over 3 exhilarating days, I tested 5 advanced AI scribe platforms. What we discovered was nothing short of remarkable. These weren’t your run-of-the-mill transcription services; they were intelligent systems designed to not just transcribe but also organize and write notes, setting a new frontier in medical documentation.²

As the event kicked off, a palpable energy filled the group. The criteria for evaluation were robust, aiming to push these platforms to their limits: product layout, speaker and medication dictation, and note organization.

Are AI medical scribes revolutionizing medical documentation or just another health tech nightmare?

How AI Scribes Work: Transforming Clinical Documentation

Medical AI scribes leverage advanced technologies like speech recognition, natural language processing (NLP), and machine learning to automate the documentation of clinical encounters in real-time. By accurately transcribing clinician-patient interactions into structured medical records, these systems handle complex medical jargon and medication details with ease.

This technology significantly reduces the administrative burden on clinicians, allowing them to focus more on patient care. During events like the AI Scribe-a-thon, AI scribes have demonstrated their ability to transcribe intricate patient histories, create SOAP notes, develop custom care plans, and even generate ICD-10 codes, thereby enhancing the accuracy and consistency of medical documentation.³

From Science Fiction to Medical Fact: The Power of AI Scribes

The core of these AI scribes – their speech recognition capability – is nothing short of miraculous. As I rapid-fired complex patient histories, these digital assistants captured every detail with stunning accuracy. Medical jargon, rapid-fire symptoms, intricate family histories – nothing fazed them.

And forget clunky, confusing interfaces. These AI scribes boast sleek, intuitive designs that feel more like your favorite apps.

But here’s the kicker: they don’t just transcribe; they comprehend. These AI scribes transform free-form conversations into structured clinical notes, automatically organizing information into SOAP format and extracting key clinical data points. It’s like having a brilliant student by your side, one who never tires and never misses a detail.

Enhancing Patient-Provider Interaction with AI Scribes

The sell for clinicians is this: by reducing the time spent on documentation, AI scribes allow clinicians to focus more on patient interaction. The potential to save time and reduce the administrative burden on clinicians is real. We are not talking about some far-removed return on investment that uses sleight of hand to convince you. They really have the potential to save time.

During the event, I had a mock patient tell me a history. As I was listening, I had this feeling of, “I know this is all being captured.” I could then focus on the conversation, and I really did find myself present. Man, what a feeling!

Mastering Medication Management with AI Scribes

Medication management has long been a documentation nightmare, but these AI scribes offer help there as well. They capture complex drug names, dosages, and administration routes with pinpoint accuracy. Some even flag the medications to confirm accuracy.

Customization and Flexibility: AI Scribes Tailored to Your Practice

One size rarely fits all in medicine, but that is what we deal with in the current tech landscape. The EMRs are these fixed bricks of tech that, God forbid, you need to do something else, the engineers just laugh because most times they don’t even know where to start.

These AI scribes aren’t just glorified typists – they’re poised to become indispensable clinical assistants. Imagine an AI that not only documents but also suggests differential diagnoses, alerts you to emerging treatment guidelines, and generates patient-friendly visit summaries and follow-up instructions on the fly.

Ethical Considerations and Challenges of AI Medical Scribes

Despite their amazing abilities, the reliability of AI medical scribes in capturing nuanced and complex medical information at scale remains uncertain. I did have a few times where it hallucinated false information about a patient case, which is obviously a concern. There is also an argument that clinicians could become overly dependent on AI scribes, leading to a potential loss of critical documentation skills.

Patient data security is a paramount concern with AI scribes. While current measures, such as encryption and secure data storage, aim to protect patient information, experts argue that these may not be adequate to address all potential threats.

A key ethical debate revolves around whether patients should be informed about the use of AI scribes in their care. Current standards and practices regarding AI use in healthcare emphasize the importance of transparency and obtaining patient consent. Ensuring that patients are fully aware and agreeable to AI involvement in their care is crucial for maintaining trust and ethical integrity.

Determining accountability when an AI scribe makes a mistake is complex. Should the clinician, the AI developer, or both be held responsible? Legal precedents and potential policy frameworks need to address these issues comprehensively. The ethical implications of errors made by AI scribes necessitate clear guidelines to protect patients and ensure fair accountability.

The Future of Medical Documentation: AI Scribes Leading the Way

As I reflect on my experience with these AI scribes, I'm filled with optimism for the future of clinical practice. By automating and enhancing the documentation process, these technologies promise to free up more time for what matters most – direct patient care.

The stethoscope transformed medical practice in the 19th century. In the 21st century, AI scribes may well be the tool that allows us to truly listen to our patients once again, freeing us from the documentation burden that has plagued healthcare for decades.

Even with its potential pitfalls, the future of medical documentation is here, and it's smarter, faster, and more intuitive than we ever imagined. It's not just about saving time; it's about reclaiming the heart of medicine – the sacred relationship between clinician and patient. And that's a future worth getting excited about.

Read part 2 here

Michael Rubio, PA-C, is a dermatology physician associate (PA) at Infinity Dermatology in Brooklyn, NY. He is the vice co-chair of the Society of Dermatology PAs (SDPA) Communication Committee and a contributor to the National Commission on Certification of Physician Assistants (NCCPA) development of the Certificate of Added Qualifications (CAQ) in Dermatology. He is also a co-founder of Well Revolution (www.wellrevolution.com), a same-day direct primary care platform helping to address the primary care shortage crisis in the United States.

References

1. Atherton J. Development of the electronic health record. Virtual Mentor. 2011;13(3):186-189. Published 2011 Mar 1. doi:10.1001/virtualmentor.2011.13.3.mhst1-1103
2. Mui P. X = primary care (XPC). Accessed August 12, 2024. https://www.xprimarycare.com/
3. Tierney AA, Gayre G, Hoberman B, et al. Ambient artificial intelligence scribes to alleviate the burden of clinical documentation. NEJM Catal Innov Care Deliv. 2024;5(3). doi:10.1056/CAT.23.0404