#37: Scaling Specialty Care
The challenge in picking winners in digital health.
Big News
Hinge Health just filed its S-1, signaling its intent to go public later this year. When the MSK care giant does IPO, it will join only two digital health companies that have gone public in the last two years—Waystar and Tempus AI. The fact that we’ve seen just three digital health IPOs since 2023 is a sobering reminder for digital health: it’s not 2021 anymore. Halle Tecco’s blog recently covered trends in digital health IPOs, exits, and acquisitions—by her count, 37 digital health companies went public in 2021. Even after Hinge Health (and Omada Health) list on the NYSE, the total from 2022-2025 will still be under 10.
Tech-enabled Specialty Care
The Hinge Health news matters not just for market confidence but because the company exemplifies what I’m calling tech-enabled specialty care. I’m by no means the first person to use this term (and if I am, it’s probably a bad sign).
Privately, the market is saturated with these types of startups—so much so that part of my goal here is to clarify my own thinking and develop a perspective on these businesses as an early-stage investor.
Tech-enabled specialty care businesses tend to follow this framework:
Problem → The specialty faces: 1) growing demand; 2) a provider shortage; and 3) rising costs (associated with the usage of traditional services/therapeutics).
Solution → Technology (telehealth, AI, diagnostics, wearables) scales care beyond clinician capacity, supported by multidisciplinary care teams (physicians, NPs, PTs, dietitians, social workers, etc.).
Note: In some of these companies, clinicians deliver most of the care; in others, like Hinge Health, technology leads.
Business Model → Selling directly to payers (self-insured employers, private plans, Medicare and Medicaid).
Bull Case
There’s a lot to be excited about with these businesses.
Network Effects → Because technology is being used to collect data and deliver care, the quality of care gets better with each additional user/patient that engages.
Unique Data → Growing datasets become increasingly valuable to major stakeholders like pharma, who need to identify the right patients (for trials and therapy), ensure adherence, and demonstrate outcomes and cost savings to payers.
Sticky Product → Specialty care platforms are built for recurring use.
Lock in effect → Once adopted, these platforms are difficult for payers to replace.
Land and Expand → They can expand into adjacent specialties with existing customers.
Bear Case
But for everything that can go right, a lot can go wrong.
Selling to payers is hard.
It’s notoriously difficult—sometimes even harder than selling to health systems. Payers are skeptical of anything that might increase costs, so they require strong efficacy data proving that the solution improves outcomes and reduces spending. That usually means fewer ER visits, hospitalizations, and escalations of care. Getting that data typically requires a pilot with an early customer, which can be expensive but is critical to secure early traction and funding.
Automating care is risky.
Certain specialties—like MSK and dermatology—already have fewer human touch-points, so automating healthcare through technology is more feasible:
Higher-touch specialties are harder to automate—missteps can lead to worse outcomes, poor adherence, misdiagnosis, or safety issues. The specialties above lend themselves to businesses that behave more like tech companies than service providers—more appealing to venture investors, though may not actually be the best for improving outcomes at scale.
Defensibility.
Investors need to believe there is something unique about a startup that will prevent both incumbents and future startups from competing. I’ve been thinking about specialty care startups in terms of a tech-service dependency spectrum, which defines a specialty care business based on 1) How differentiated and central the tech is; and 2) How human-service intensive the care delivery is.
Tech-dominant companies build their IP around core tech like AI, with services commoditized. Examples include Piction Health and Headspace—though Headspace now includes human psychiatry. Hinge Health likely sits somewhere between this category and the next.
Balanced tech + services companies use tech to amplify human care—through triage, diagnostics, or optimized care pathways. Examples: Maven Clinic, Isaac Health.
Service-dominant companies use a combination of commoditized tech (like telehealth) and differentiated services. The “secret sauce” is in the care delivery model, clinician network, and overall member/patient experience—not in the unique technology. Examples: Cityblock Health, LifeStance Health.
This framework helps in deciding whether to look for a technology moat or a service model moat.
Path to Profitability.
Depending on where specialty care companies fall along a spectrum of tech-enabled services and “service-enabled tech”, healthy margins are important for showing that a business model has a path to profitability. Generally, the more tech-enabled and less service-oriented the business, the higher the margins will be and the clearer the path to profitability looks.
Hinge Health
So how does Hinge Health fit this (oversimplified) framework?
Solving a huge problem → 50% of all adults experience MSK pain, with the annual cost of MSK conditions totaling around $600B (which includes costs due to lost worker productivity).
Using tech to automate, personalize, and scale MSK care → The platform, which includes AI motion tracking and an FDA-cleared wearable device for electrical nerve stimulation, reduces the number of human care team hours associated with traditional physical therapy by ~95%.
Selling to Payers → Clients include the five largest national health plans based on self-insured lives (approximately 19 million lives contracted in their existing client base).
Network effects → Hinge Health’s AI model is trained on a proprietary MSK dataset and is continuously improving as each new member engages with programs.
Sticky product → 98% client retention rate for 2024.
Lock in → NDR of 117%.
Hinge has de-risked nearly every major point of failure:
Proven sales model with major payers and partnerships with PBMs.
Demonstrated improved outcomes and cost savings.
Improved gross margins from 63% (2023) to 77% (2024).
But is it defensible? That’s unclear. I’m still not sure I can articulate exactly what makes Hinge Health differentiated. Their success may simply be a combination of a first-mover advantage in taking a tech-forward approach to MSK care and executing extraordinarily well to get to scale before anyone else could. Now, they get to build technology that is truly differentiated because of the millions of patients that engage with the platform.
On the one hand, the tech-enabled specialty care market for most therapeutic areas is saturated, so investors need to believe a certain company has an unfair advantage that separates them from the rest. On the other hand, the early success of startups like Flyte Health (obesity care) and the longer tenure of Hinge Health as a leader in MSK health suggest this model can work when executed well, even if early differentiation is not all that clear.
To believe in tech-enabled specialty care is to believe in the vision shared by Hinge Health Co-founders Daniel Perez and Gabriel Mecklenburg:
“I believe that software will soon automate all non-touch aspects of healthcare such as interpreting symptoms, formulating diagnoses, crafting care plans, etc. Moreover, hardware advancements are automating and enabling self-service for select aspects of healthcare that require touch such as: monitoring (e.g. continuous glucose monitors), medication delivery (e.g. insulin pumps, connected inhalers), pain relief (e.g. Hinge Health Enso), and more.”
What I’m reading this week:
The massive opportunity in women’s longevity, Second Opinion
Animation Is Eating The World, Michael Dempsey
ChatGPT Is a Blurry JPEG of the Web, The New Yorker