Algorithms Do Not A Company Make
May 22, 2020
By Dan Lambert, Technology Partner at SpringTide
The Health IT industry is full of companies touting useful inventions for patient care, workflow optimization and EMR upgrades. But ultimately, as every industry insider knows, there is a very high failure rate for new entrants. The high failure and premature exit rates among healthcare innovation companies can be improved if innovators commit to delivering end-to-end rather than point solutions.
We see repeating patterns of why companies can’t break in:
1) Selling to hospitals is hard. In our experience, most large purchases take over a year and require an RFP. The sales process is also tough in that new companies are going up against entrenched incentives from GE, Philips etc.
2) Reimbursements work differently. It can take 10 years+ to get a new billable CPT code and typically startups and funds don’t have that long of a runway. The device or system must also find a way to hospitals and clinics to actually be reimbursed.
3) Regulation stifles innovation. New businesses have to figure out how to avoid fee splitting, be compliant with Stark laws, maintain HIPAA compliance, avoid anti-kickback and inducement laws, and deal with State by state tele-health rules.
These factors combine to produce an industry that boasts only a small handful of IPOs, and a lot of early exits. In the face of all of these headwinds, we will highlight two emerging company types that we believe are well poised to navigate this welter of industry complexity.
PathologyWatch is a Dermatopathology AI company. The business integrates with Dermatology offices to capture volume coming out of the clinic. To the Dermatologist, PathologyWatch provides digital pathology, case integration with the EMR, and the ability for the dermatologist to bill for cases.
We invested in PathologyWatch knowing it had successfully navigated the three hurdles above. The company sells to outpatient clinics (instead of hospitals) and is reimbursed directly by insurance for the Pathology reads. It is one of the first companies to capture true clinical volume in the AI space. The company is in over 20 Dermatology clinics and poised to spread rapidly through the outpatient market.
Knowing that there will not be a computational AI reimbursement for many years, PathologyWatch specifically focused on two factors: 1) controlling the pathology sample volume and 2) controlling the AI. This is where the margin can be captured. Achieving these objectives meant that PathologyWatch had to integrate across numerous LIS and EMR systems, but the effect is end to end control of the Pathology process, to which AI can be applied.
We think that PathologyWatch will serve as a template for how outpatient medicine embraces AI over the coming decade.
Remote Monitoring of Critical Devices
Companies like 91.life, PaceMate, and Geneva Health Solutions began by understanding a major pain point for hospitals. A beleaguered hospital administrator explained – ‘We are using a lot of different technology and devices but billing for these devices is cumbersome and near impossible to track.’
This new breed of AI-powered companies recognizes two dimensions of the need: 1) data is difficult to view and organize and 2) hospitals can significantly grow top line revenue by using software to manage these devices.
These companies are working within the existing reimbursement system and selling to hospitals in a way that promises to both make the product appealing, and improve patient care. This does require a significant amount of systems integration work, and therefore a strong understanding of the clients’ existing systems.
The thesis that unites these two opportunity areas is this: Companies that just sell software in healthcare very rarely survive. The ones that succeed think deeply about root problems, and develop truly comprehensive solutions that can thrive in the complex framework of the US healthcare system.