Listen to this article · 7 min listen

The field of healthcare AI is filled with stories of companies that crashed and burned, but the most instructive failures aren’t the obvious flameouts. They’re the ones that did everything right on paper, securing a clear regulatory path and publishing strong clinical results, and still failed to find a market. For investors trying to make sense of this space, figuring out why these promising ventures die gives you a far more durable framework than just trying to copy the winners. The paper trail is permanent. You can pull the FDA filings and read the clinical studies long after a company’s gone, which makes for a perfect post-mortem.

FDA Clearance Gets You to the Starting Line

Pear Therapeutics is the textbook case. The company managed to get multiple FDA clearances for its prescription digital therapeutics (PDTs), which was a huge deal in a field that was just taking shape. That validation from the FDA’s Center for Devices and Radiological Health (CDRH) confirmed its products were safe and effective for their intended use. Investors often see that kind of clearance letter and think a major risk is off the table. After all, the FDA has authorized over 1,600 AI/ML medical devices as of September 2026, with 1,614 device decisions logged through June 29, 2026. But here’s the detail that matters: a whopping 96.2% of those clearances came through the 510(k) pathway, which just shows the device is “substantially equivalent” to something else already on the market, according to the FDA official AI device list. That’s a world away from proving standalone clinical value, and it’s a distinction that smart money is learning to spot.

An FDA clearance, even a more rigorous one through the De Novo pathway, fundamentally just covers the technical and clinical safety. It says nothing about whether patients will actually use the thing or if anyone will pay for it. When a company like Pear goes under, the autopsy doesn’t begin by looking at the technology or the regulatory paperwork. It starts with a much simpler question: did people keep using the product?

Published Studies Don’t Guarantee Engagement

On top of its regulatory wins, Pear Therapeutics also published clinical results, which should have sealed its credibility. In evidence-based medicine, peer-reviewed studies are what separate real medical products from wellness-app hype. You can see what a solid evidence base looks like by checking out a company like Hello Heart. They’ve built an incredible body of proof with over six peer-reviewed publications that cover more than 48,000 people, and their data demonstrates a sustained reduction in blood pressure over three years, as published in a JAMA Network Open Hello Heart study. That kind of validation is exceptionally rare in a field where most AI claims are backed by little more than a press release.

But here’s the hard truth. You can have FDA clearance and a stack of glowing studies, but a product’s real-world value depends entirely on patient engagement. If patients don’t use the digital therapeutic consistently, the clinical benefits you proved in a trial simply don’t materialize. It’s that simple. The public record on Pear Therapeutics shows that for all its success on paper, the real-world engagement numbers just weren’t there. This reveals the most critical lesson: efficacy in a controlled study, with motivated patients and researchers watching, almost never translates directly to sustained adherence in the messy reality of everyday life. Patient engagement is the first signal to watch, long before you even get to commercial questions.

Looking at the Competition

To really understand what happened with Pear, it’s useful to put its story next to other digital health companies like Omada Health and Virta Health. They’re all playing in the same sandbox, trying to get patients to make significant, long-term behavioral changes to improve their health.

  • Omada Health: Famous for its digital diabetes prevention program, Omada built its entire model around combining technology with human coaching and behavioral science. They recognized from the start that an app alone isn’t enough to change deep-seated habits.
  • Virta Health: Virta is focused on reversing type 2 diabetes by using a low-carb diet paired with continuous remote care from their team. Their model is incredibly dependent on getting patients to stick to a strict nutritional plan and interact constantly with their digital tools and care providers.

Both Omada and Virta have wrestled with the patient engagement dragon and survived, though not without their own challenges. It shows that while the tech is the platform, the human interaction part is what makes or breaks these models. The key difference from Pear seems to be that while all three were trying to drive behavioral change, the specific ways Omada and Virta built and maintained user engagement were what in the end determined their ability to stay in business.

The Post-Mortem: Start with Engagement

For any investor trying to do diligence in healthcare AI, the Pear flameout offers one simple, powerful insight: your analysis has to start with engagement. Don’t get distracted by the technology or even the regulatory status. The FDA CDRH’s own data shows how far a product can get down a cleared path before it smacks into the wall of commercial reality, do people actually use it, and will they keep using it? A product can be perfectly engineered and have data from a successful clinical trial, but its market viability is zero if it doesn’t find a way to fit into a patient’s life for the long haul.

This isn’t just a one-off story. It fits a broader pattern in digital health. A systematic review published in the Lancet Digital Health AI clinician study that covered 65 randomized controlled trials found that AI-assisted clinicians beat unassisted clinicians 71% of the time. Meanwhile, AI working alone only beat clinicians in 25% of cases. The conclusion is obvious: human-in-the-loop designs, where AI helps a person do their job better, are what actually works. For digital therapeutics, that means the patient’s active engagement, driven by good coaching, smart feedback, or just making the tool easy to use with their doctor, is the most important part of the system.

So the lesson is pretty blunt. If you’re building or funding a healthcare AI company in the therapeutic space, you need demonstrable, sustained patient adherence. That data is more important than the novelty of your technology or a framed copy of your FDA clearance letter. Without it, you’re just building the next cautionary tale.

Frequently Asked Questions

Does FDA clearance guarantee market success for AI health ventures?

No, FDA clearance primarily addresses technological and clinical safety, signaling that products meet safety and efficacy standards for their intended medical use. However, it does not inherently guarantee patient engagement or commercial success, as evidenced by cases like Pear Therapeutics.

What is the primary reason AI health ventures with FDA clearance and published clinical results might still fail in the market?

The primary reason is a lack of sustained patient engagement. Even with regulatory clearance and proven efficacy in studies, if patients do not consistently use the digital therapeutic as prescribed, the demonstrated clinical benefits cannot be realized, making market viability tenuous.

What critical factor should investors prioritize when evaluating AI health ventures, beyond regulatory status and clinical validation?

Investors should prioritize patient engagement. A product can be technically sound and clinically proven, but if it fails to integrate into patients’ daily lives in a meaningful and lasting way, its market viability becomes tenuous. The post-mortem analysis of failures should begin with engagement.