Health Tech Startups, Simplified

Demystifying the Flatirons, Oscars, and Clovers of Healthcare

Hoyt Gong
The Startup
6 min readJun 26, 2019

--

As the healthcare Unicorns on the block, you’ve probably heard the names before: Clover Health, Flatiron Health, Collective Health… but what exactly are they?

Backed by plenty of big names in VC and raising multiple rounds with sustained progress, these healthcare startups are bringing the digital age into medicine, projected to reach nearly $400 billion by 2024. From analytics platforms to care management systems, the technology that modern healthcare startups are pushing have gained traction with providers, investors, and the common public alike.

Yet while it’s easy to generalize them as ‘startups focusing on healthcare using some kind of data and technology,’ we can do better than that. What kind of analytics does the company perform? What does care integration actually entail? What is a health management platform anyway?

Healthcare terminology AND ML/AI buzzwords? That’s a lot to digest there.

Beyond showing Silicon Valley that healthcare is a profitable industry, these digital health startups are actually expanding health access and impacting patients. I’ve broken down some potential jargon in order to explain the ‘how’ and ‘why’ behind what makes these companies groundbreaking and worthwhile.

Flatiron Health

TL;DR: A cancer-focused software platform, for hospitals and life sciences companies, that aggregates various types of data from multiple healthcare stakeholders into actionable real-world evidence.

Named after the Flatiron district in the heart of Manhattan, this NY-based health software company spans multiple users from scientists to pharmaceutical companies to physicians in order to improve oncology research using a shared technology platform.

Their platform collects unstructured data from sources like sequencing databases, insurance claims, and academic literature repositories and structured data from hospital EHR systems into an analytics engine. This engine then provides a more comprehensive picture of healthcare through the resulting real-world evidence (RWE) generated.

The platform is specifically tailored for cancer research across these various healthcare stakeholders and adds value in a variety of ways:

  • Providers can track metrics about their cancer care delivery and continuously adapt from patient patterns into targeted treatment plans
  • Hospital administration can manage company oncology programs and optimize reimbursement billing questions
  • Academic research institutes can make better data-driven decisions on sourcing patients for clinical trial studies

Clover Health

TL;DR: A Medicare Advantage insurance plan partner for seniors, backed up by health monitoring, in-home concierge services, and personalized collaboration with your respective doctor.

We turn towards customer segments (e.g. patients) instead of healthcare stakeholders, looking at how we are changing insurance for the elderly. For starters, Clover is a health insurance (Medicare Advantage) plan at its core. As such, Clover makes a majority of its revenue from the government.

While Medicare is available for all Americans 65+ yrs, its traditional coverage gaps pose barriers for elderly populations (e.g. Medicare doesn’t cover dental, vision, and sometimes prescription drugs). An often lesser known alternative are the Part C Medicare Advantage plans that offer greater coverage at an equal or lesser premium price.

Clover takes this concept of a Medicare Advantage plan and focuses on delivering greater services and health support to their users with — you guessed it — technology.

The Clover technology platform collects and analyzes health and behavioral data to improve outcomes and lower patient spending costs. How?

As a use case, the platform is able to detect at-risk conditions based on monitored patient behavior data, thereby allowing seniors to take preventative measures in avoiding hospital admission. By also partnering with the caregivers of the seniors, Clover can further accelerate care coordination from the provider’s perspective and provide the doctor with useful analytics that may lead to a certain course of treatment.

23&Me

TL;DR: A human genome research company that enables its users to study their ancestry through genotyping gene variants, along with a Therapeutics Division for drug discovery and research.

Probably the most direct-to-consumer facing company of the ones in this article, 23&Me provides insight to your ancestry.

Long story short, you spit into a tube, mail it off, and 6–8 weeks later you get an electronic report with your specific genetic information like an ethnicity breakdown by percentage.

But here’s where we can ask questions —how accurate are the results? What do they do with the data afterwards?

How the Science Works

As of February 2018, 23&Me has genotyped over 3,000,000 individuals globally. For clarification, the company does not sequence, but rather genotypes, or tests for gene variants, of your DNA. By using a glass microarray chip and testing for hybridization to existing variants, the laboratory compiles a report of how your DNA genetic variants matches up to variants of people with a history of being X ethnicity or having Y disease. This means results are only accurate to the extent that patterns match up to the millions of people’s DNA in databases before you took the test. Note that this becomes a sample size situation: 1) Unless you have an ultra-rare variant, the results are fairly accurate, but 2) underrepresented variants in databases (i.e. from less common ethnicities) also lead to greater risk for false positives.

On Data Selling

Does 23&Me sell your data? To an extent, with informed consent.

We do use and share aggregate information with third parties in order to perform business development, initiate research, send you marketing emails and improve our services.

It’s important to understand that aggregate information is not individual information. 23&Me is not sending emails of Jane Doe’s AGCTCGAA data for example, but anonymizing user data to larger statistics — like a median or 75th-quartile metric. Aggregation largely prevents privacy leaks through a concept called differential privacy that you can read more about.

Oscar Health

TL;DR: A health insurance company targeting both common consumers and small businesses that aims to simplify logistics in getting healthcare.

I won’t dive into details of health insurance (copays, coinsurance, deductibles), but neither does Oscar when it comes to their customers who may not be getting employer health insurance. Basically, if scheduling appointments and dealing with healthcare bills give you logistical headaches, Oscar’s aim is to make that as painless as possible.

It does this by giving customers the following benefits:

  • Discounted doctor visits + 24/7 phone call availability (telemedicine)
  • Direct specialist visits (e.g. no having to go through referrals)
  • Offers fitness tracking devices and rewards to users for data tracking
  • Tailors doctor/drug recommendations based on data collected from patient visits; free coverage to an extent

Note that Oscar’s customers must still be wary of the specific policies under their plan. While Oscar does help simplify an understanding of insurance plans, the user is responsible for ultimately knowing what can and cannot be covered under their Oscar plan.

But Wait, There’s More

I’ve stopped at these four companies above since the list could continue endlessly (Collective Health, Bright Health, Verge Genomics, etc) and left out some additional details for the companies above for simplicity. It’s important to note though that the article here didn’t include any projects that large technology corporations are pursuing in healthcare. Nor does it include any of the recent innovations that pharmaceutical companies and healthcare intermediary actors have been developing.

At large, technology and analytics are changing the game in healthcare but will require a bit of due diligence to really understand how such products work. Before hopping on the AI/ML bandwagon in healthcare, it is critical to see just how analytics are being applied, what value it actually adds, and where data then gets handled. In seeing through the hype and buzzwords can we begin to formalize just how much our healthcare landscape is changing.

--

--

Hoyt Gong
The Startup

Big Data Analytics Teaching Assistant @Upenn | Wharton x Computational Biology | http://hoytgong.com