Axios Pro Exclusive Content

AI medical scribes vary widely by price and other features

Data: Axios reporting; Note: Asterisks indicate figures sourced from outside the company; Chart: Axios Visuals
Data: Axios reporting; Note: Asterisks indicate figures sourced from outside the company; Chart: Axios Visuals

High-tech medical scribes promise to revitalize the low-tech doctor's appointment, but their pricing and accuracy vary, according to an Axios analysis.

Why it matters: Ambient clinical documentation tools have become the latest hot ticket in health care, with companies drawing big checks and conference crowds.

The big picture: With several already in market, medical scribes hold the potential to reinvigorate doctor-patient interactions, which are often hampered by administrative burden and bottlenecks.

  • "My mission [starting Suki] was to make health tech invisible so physicians can focus on what they love and take care of patients," says Suki founder and CEO Punit Soni.

Follow the money: Investors are pouring venture capital into AI-powered clinical documentation companies.

  • Ambience in February raised a $70 million Series B led by Kleiner Perkins and the OpenAI Startup Fund, tipping total funds raised to $100 million, per a spokesperson.
  • Abridge in February collected a $150 million Series C led by Lightspeed Venture Partners and Redpoint Ventures, bringing total funding to $213.9 million, per PitchBook.
  • Nabla in January nabbed a $24 million Series B, bringing total capital raised to $43 million, per a spokesperson.
  • Suki in 2021 collected $55 million, bringing total funds raised to roughly $120 million, per CEO Soni.

What we did: We measured five metrics — price, accuracy, latency, specialties supported, and whether companies had 3+ EHR integrations.

  • Prices shown are per person, per month. Only Nabla and Suki provided prices. Other costs were sourced from public documents and sources familiar with the matter, since the companies declined to respond.
  • Accuracy refers to the rate at which providers accept transcriptions without edits, while latency represents the time to transcribe. The number of supported specialties was provided by the companies.
  • Spokespeople for all the companies in the chart said their tools were integrated with at least three EHRs including Epic, Cerner and athenahealth.

Zoom in: Nuance was the only company that does not currently offer medical code extraction, though a spokesperson said the company's technology supports it.

The intrigue: Charts show quantity, not quality, and we recognize that a chart can't capture every nuance.

  • For example, with regard to the number of specialties a company claims to support, the quality of that support depends on the degree to which it has fine-tuned its models to reflect those specialties' specific workflows, revenues and prior authorization requirements.

Behind the scenes: Tampa General Hospital CIO Scott Arnold told Axios at the ViVE conference that TGH physicians were excited to use Nuance's scribe tools — "and they don't normally get excited about these things," he said. "The idea of shortening their day is attractive."

Reality check: As clinical documentation tools are deployed, providers are encountering difficulties validating them, per STAT News.

  • One problem is the lack of standards to evaluate their performance.

Threat level: Clinical documentation tools may test the limits of tech's role in protecting patient privacy and reliability.

  • "This is the most personal and sensitive data you might collect. If we get that wrong in some capacity, we all understand that's going to be a huge issue," Lightspeed Venture Partners partner and Abridge investor Sebastian Duesterhoeft told Axios last month.

Editor's note: This story's chart has been corrected to reflect new information from Suki noting that its scribe's whole note accuracy is 92% (not 84%, which is the accuracy by section).

Go deeper