categories

HOT TOPICS

Scaling to $500M in Revenue: ModMed CEO Daniel Cane (Part 5)

Posted on Tuesday, Nov 19th 2024

Sramana Mitra: Double click down for me on the AI aspect. I know you were doing AI right up front, but of course, AI has evolved. The market has evolved. Market’s receptivity to AI has evolved. AI has come of age, so to speak in the last fifteen years. Where are you with AI? What are you doing with AI? What are you seeing in the adoption of AI?

Daniel Cane: I’m a technologist, so this has been an incredibly exciting opportunity for us. We’ve been using data to create algorithms for fifteen years. None of this is particularly new to us. What is new is the foundational models like large language models and computer vision models that are state of the art.

We’ve already delivered solutions to our customers that leverage large language models. One simple example involves the still widespread use of faxes in healthcare—yes, it’s ridiculous, but it’s a reality.

In many practices, it’s someone’s job to open faxes, read them, determine what kind of document it is, and figure out which patient’s chart it should be attached to. This process is time-consuming and prone to errors.

AI excels at reading documents, identifying their type, extracting key information, and matching it to the correct patient with a high degree of confidence. It can then prompt a human to confirm the match.

We’ve developed and deployed a system like this, and the feedback from users has been phenomenal. Practices that previously had staff dedicated to combing through fax queues have seen dramatic time savings.

Sramana Mitra: What is even more infuriating is that some practices require a fax referral. Otherwise, they’re not going to accept a new patient.

Daniel Cane: Let’s not get down the fax rabbit hole. It’s sad, but that’s the state of health care today until HIPAA catches up with more modern means of faxing because it’s analog is HIPAA compliant, which is why all these doctors fax it.

Sramana Mitra: Let’s come back to AI.

Daniel Cane: The far more exciting potential for AI lies in addressing significant cost centers and alleviating them. There are people in healthcare investing substantial resources trying to ‘out-doctor’ a doctor—but that’s not ModMed’s focus, nor is it something our users struggle with.

I’ve yet to meet an orthopedic surgeon who says, ‘I can’t interpret an MRI or CT scan, or a dermatologist who says, I can’t recognize skin cancer anymore—it’s just too hard. No one is telling us that. What we are hearing is, I’m spending a fortune on staff to scribe for me, read information, or scan faxes. We’re spending hours every day on phone calls to patients. Can’t AI handle things like answering calls, letting patients pay their bills, or rescheduling appointments?

And the answer is: absolutely, AI can do those things. That’s why we’ve heavily invested in our AI for patient engagement, known as the Clara solution. Clara was an acquisition we made about three years ago, and it has since facilitated over a billion messages. These often start as simple text messages—like the appointment reminders you’ve likely received from specialists in recent years.

Clara ensures patients are reminded to check in and even enables conversations with AI. If the AI can handle the query, it resolves it; if not, it categorizes the conversation and routes it to the appropriate queue for the practice—whether that’s scheduling, billing, or clinical questions.

By eliminating the need for staff to sift through a giant inbox of questions, Clara streamlines workflows and saves significant time. The more AI can handle routine questions, the better it is for both the patient and the practice.

Sramana Mitra: It’s more practice management and workflow AI.

Daniel Cane: That’s patient engagement, but we’re also heavily focused on practice management with our vaccine solution. We focus clinically, on the AI scribe. The AI scribe is going to save a tremendous amount of time and effort for practices.

With the new generation iPad featuring the M4 chip—it’ll be funny listening to this ten years from now—we’ll be among the first to fully leverage the processing power of that iPad chip for AI-driven transcription. This isn’t just about standard transcription; it’s about specialty-specific language models. Our AI will effectively ‘speak’ dermatology, ophthalmology, orthopedics, podiatry, and more.

The transcription will happen directly on the device in record time, eliminating the need to send recordings to the cloud and deal with the delays of the past—like the old Siri days. This system will be lightning-fast.

We’ve been training our models using hundreds of thousands of examples to go beyond creating summaries for the EHR. This AI will generate actionable, structured data. It will understand diagnoses, morphologies, prescriptions, treatment plans, lesion sizes—you name it.

It’s going to be incredible.

This segment is part 5 in the series : Scaling to $500M in Revenue: ModMed CEO Daniel Cane
1 2 3 4 5 6 7

Hacker News
() Comments

Featured Videos