Health Care + VC: How FFS, Churn, and Clinical Efficacy Shape Markets and thus Investments
I interviewed 40 VCs, health tech entrepreneurs, health care executives, and industry experts and spent hours researching the intersections between health care, public health, and venture capital. What emerged was a multi-part report that I’m calling Preventing Prevention: Barriers to Venture Capital Investments in Upstream and Community-based Care.
This is part four in this series. Here are parts one, two, and three. You may want to read those before this one.
Funding for this work came from the Robert Wood Johnson Foundation.
Intro, Background, Etc
Venture Capitalists fund markets. They get paid to understand where markets are headed — as shaped by trends in policy, technology, and consumer behaviors — and help fund their growth.
So to understand where VCs invest, and how that may or may not facilitate improved health outcomes across populations, we must understand some of the core market forces at play within health care.
These market forces are shaped first and foremost by policy, and in a previous post called How The U.S. Health Care System Works (And Doesn’t) I dive into the financial models that are the economic pillars shaping health care markets. If you don’t have a general understanding as to how money flows through healthcare, you’ll want to read that post before this one.
This post looks specifically at 3 concepts that represent barriers to the health care system working more upstream:
- The legacy of fee for service — Why it’s still here and probably will be for long while [and to some extent, that may be ok]),
- Churn within health insurance plans — How this shapes what an insurance plan looks like and encourages cherry picking for the healthy and wealthy (leaving sicker and poorer individuals with more costly plans), and
- The availability of evidence to demonstrate clinical efficacy — Basically: How do we know whether an intervention works or not?
It’s important for me to point out that these are broad observations that I’m making based on my own understand of health care (check LinkedIn for my resume), research, and the conversations with others. Take this all with a grain of salt: There’s certainly nuance that I’m missing. The fragmentation of health care in America pretty much guarantees it.
But the important trends are, well, important. So let’s dive in.
The Legacy of Fee For Service (FFS)
Shifts away from fee for service (FFS) — which generally rewards prevention less — and towards alternative payment models (APMs) — which better reward preventive action — are not occurring strictly by mandate or force.
Rather providers are moving to APMs generally on their own accord, responding to nudges created by the Affordable Care Act and other policy-driven activity.
As you may have guessed, not all providers (doctors, hospitals, physician networks, etc) are jumping on board as quickly as others.
Which ones are making the leap and which ones aren’t? You may have guessed this one as well: It comes down to the dollar.
There’s a lot of money to be made in FFS. This is why, of course, FFS has been the dominant financing model for the bulk of U.S. health care history. If you’re a hospital on FFS, you need to see that more money can be made in an APM. Otherwise, there’s no business incentive to make the shift. (And more so: There’s a lot of incentive to lobby governments to keep FFS around.)
But also to be factored into the business equation are the transition costs associated with changing from FFS. Here we’re talking about the technological, programmatic, and human barriers that must be overcome to change how an organization operates.
Hospitals and other providers have been operating on FFS for decades. Industry executives have had to figure out how to ensure sustainability of hospitals, clinics, and insurance companies in the FFS world. Oddly, though there is so much money floating around health care, hospital systems don’t necessarily have a lot of money to play with. Ensuring business viability is no small business feat.
But these years and years of operating in a FFS world have cultivated and calcified expertise and sets of norms and understandings that rely on that FFS world for relevance.
“The leaders at the top of these organizations, in particular CFOs who handle the flow of money, got to their jobs in the FFS system.” One insurance executive told me. “They don’t have experience with other forms of accounting.”
This is the legacy barrier of the FFS system.
Changing everything, from the top down, requires an incredible business leap. Not just staffing mindsets and work flow processes, but also internal technologies must be updated. EMRs and claims processing systems were built with FFS in mind.
Coordinating Care is Hard When You’ve Never Had to Do it Before
FFS does not naturally incentivize the sharing of information nor the coordination of doctors to improve the overall care given to a patient. (This is one reason why EMRs are so terrible; they’ve never needed to be good!) Alternative Payment Models (APMs) are meant to help address this fundamental issue.
Woven into the legacy of FFS are the simple business challenges that come when historically independent organizations suddenly try to work together.
To illustrate: I spoke with one physician in a community health clinic in Northern California who worked with other community clinics to form a mega-group that would work together to share data and coordinate care. (For you nerds out there: This was a group of FQHCs that came together to form an ACO).
A primary driver in the formation of this ACO was the desire to share data on these types of patients whom were bouncing between their clinics. However, when it came down to the coordination of care as determined by this data, processes broke down.
For example, when decisions had to be made as to which provider was to take the lead on initiating engagement with a patient, they’d get caught in conversations as to who was supposed to do what delaying care and introducing incredible friction into the system.
These fundamental management problems are issues many ACOs and others jumping to APMs face. For ACOs, these problems are preventing them from living up to the narrative of care transformation long heralded by policy experts. Some providers have found the collective operations so difficult to manage that they’ve broken apart, broken back into the separate fragments of care delivery that they were prior to ACO formation.
Those Shifting to APMs Already Have Work In That Space
Those that are embracing shifts to APMs and managed care tend to be those that were already operating in some capacity outside of the FFS world.
I spoke with one executive within a delivery system of the State of New York who said “We’re fully embracing managed care in part because we were already a major safety net provider.” A sizeable portion of their patients already had Medicaid and were within their managed care program (Most of Medicaid operates off of capitation, though the extent to which this is true and details of that capitation vary by state).
To transition more fully to the new world of APMs, they’re simply building out and beefing up this arm of their operation so that it applies to all of their patients. This is of course much easier to do compared to creating a managed care operation from scratch.
This same executive pointed to a conversation he had with an executive at a different provider in the same area who said “We’re holding out on FFS until doing otherwise is literally forced on us.” The costs for them are too high. They’re quite comfortable where they are.
And healthcare VC’s know that this mindset is out there. Many of them believe that FFS will always be around. An even more skeptical take was presented by one investor who told me: “A lot of people see ACOs as just another three-letter fad. Like HMOs, ACOs will eventually pass.”
Since there’s still a lot of money to be made in FFS, investors still put a lot of money towards products and services which extract value out of FFS-financed markets. Investments into medical device companies, pharmaceuticals, and acute care interventions continue to be safer and more lucrative.
In short: We’re still operating in a world where charisma and ideology, not financial incentives, are really driving the shifts in health care financing. One insurance executive explained it this way: “Whether a provider is bought into new ways of doing business really depends on one factor: Who’s in charge.”
So even though there are strong signals from DC and general trends towards APMs and the formation of ACOs, the legacy of FFS will continue to hinder their advancement. And since tech-driven population health management underlies ACOs, the legacy of FFS hinders that advancement as well.
Churn, Churn, Churn
Churn is an oft discussed issue in health care, but for the uninitiated: “Churning” is when a person changes (or loses) insurance.
Most writings on churn focus on the impact it can have on continuity of care for people bouncing between Medicaid, CHIP, and the Exchanges. Within Medicare, the same issue occurs when individuals bounce between Medicare Advantage and traditional Medicare. And within employer-based health insurance, churn is largely driven by major life events such as getting a new job or getting married (and joining a partner’s plan).
But beyond this, an insurance plan’s churn rate affects the business decisions of that insurer, and the likelihood of an individual churning out can affect the care available to them.
Here are the Churn Rates as the Literature Presents Them
National-level data on churn rates is hard to come by, but here are general trends:
- Churn on the Exchanges is highest: About half, 1 in 2 (53%), of those buying individual coverage in 2014 changed plans the following year (source)
- Medicaid churn is also high: In one study, about 1 in 4 (25%) Medicaid respondents changed health coverage at least once in 2015 (source). With this said, Medicaid is incredibly fragmented — across states, and in many states: across counties — so there is certainly a spectrum of rates. For example, I spoke with one Medicaid Managed Care org in Northern California that showed a rate of only 1 in 10 (10%).
- Medicare Advantage (MA) churn is lower: Though there is variance across MA plans, this is the general consensus (It’s also one reason why MA plans often have VC backing). In terms of numbers, one study found about 1 in 10 MA plan enrollees (11%) in 2013 voluntarily switched to another plan in 2014. (source)
- Employer-sponsored coverage experiences the least amount of churning. There’s a spectrum here of course…
- One study found that only 1 in 17 (6%) changed employer plans from 2014 to 2015 (source)
- Small business have higher churn: 1 in 7 (14.6%) (source)
- Large self-insured employers tend to be the on lower side. For example, at CostCo, turnover is just 1 in 17 (6%) after one year’s employment (source).
Here are the Churn Rates per the VCs I Spoke To
Interestingly the rate of churn is not how churn gets presented to me by the VCs, entrepreneurs, and health plans. Rather they use a number to demonstrate average length of stay in a plan. They look like this:
- Medicaid = 1 year
- Exchanges = 1 year
- Employers = 3 to 7 years, depending on the employer and the commercial provider
- Large employers = 5 to 7 years?
- Self insured = 5 to 7 years?
- Medicare Advantage = 5 to 8, depending on the plan
So there seems to be a discrepancy in how academics publish data on churn rates and how investors think about churn rate. The investor is thinking about how long you are likely to have someone. This goes into that LTV number.
There’s certainly nuance here that I’m skipping over, but suffice it to say: This seems an interesting juxtaposition that could be explored further.
Higher-churn Insurance Markets Have Fewer Prevention and Wellness Features
In a previous post on How VC Works (And Doesn’t) In Health Care, I laid out how the Longterm Value (LTV) / Customer Acquisition Costs (CAC) ratio is a key metric that investors look at (and businesses work towards) when evaluating a company for possible investment.
Through this lens, churn impacts the numerator, the LTV: The faster an individual churns off, the lower the lifetime value of that individual. And if you’re a company focusing on prevention and health promotion services, where upfront interventions yield downstream savings, this lower LTV presents an barrier to finding a sustainable business model.
Or more pointedly, and to fully bring the fragmented complexity of the U.S. health care system front and center, there are some interventions that can yield financial ROI within certain time frames in certain markets, and others that don’t.
But: Which interventions? And in what markets?
Prevention, health promotion, and certainly the concept of “wellness” are long-terms games. How long exactly depends on the type of intervention we’re talking about, but in general: If an insurance company knows that they’re likely to only have you for only 12 months, there’s very little business incentive to even offer you a service that would occur 18 months out.
This means that in markets where churn is high, insurance plans aren’t going to even offer prevention and wellness programs as features in their plans.
This is why certain regulations exist. For example, regulations require minimum services (such as “prevention screenings”). However, beyond the minimum, it’s a business decision as to what features to include in a given health insurance plan.
An important paradigm should be noted here: Insurance companies are businesses. The product this business sells is their insurance plans. They sell to individuals as well as to small groups and employers, and the particulars of that insurance plan, the features of that plan, depends on who they’re selling to.
If You’re Loosing Customers Every Year, Then Getting New Ones Becomes the Focus
Plans operating in high-churn markets, such as the individual market, by definition lose a lot of customers every year. This means they also have to sell a lot of plans every year to make up for those lost.
This thus becomes the game: Work hard to get more revenue in from new sales every year than the revenue lost from departing customers. This game leads insurers to focus enormous energies on the revenue generating side, They work hard to determine what gets people to buy their plans. Including what types of features, what benefits, drive increases in sales.
One non-digital example of how this plays out: Most insurance plans on the individual market offer gym memberships. They offer this not because they hope to see the long-term benefits of a customer going to the gym. But they do know that individuals shopping for insurance like to see a gym membership as a feature, and they’ve figured out that it increases net revenue to offer it.
Why don’t the plans offer more customer-facing tech-based tools to promote healthy behaviors? The plans I spoke to say the numbers just aren’t there: The costs of providing those are higher than the increase in sales revenue generated.
There has been an industry trend recently on improving the customer experience of health insurance purchasing. Does this result in better health outcomes? Well… that’s not really what they’re going for. It’s good, I suppose, that the customer experience is improving, but that does not translate to better insurance packages offering more preventive and wellness-oriented services.
Insurance companies also make efforts to retain customers. It makes sense that they would. But from what I can tell, the business practice is to just assume a churn rate rather than try to change their churn rate. Someone who knows more than I do would have to explain this dynamic in a deeper way than I can.
The Data Surrounding Upstream Interventions
Determining what constitutes “prevention” is actually, annoyingly, a fairly difficult task. Our health care system is reactive: A trigger determines a (potential) problem. And an intervention then occurs to remediate this (potential) problem.
This dynamic is so true that prevention-oriented companies trying to sell into health care have to work around it. They have to pitch themselves as an intervention company, one that is explicitly entering at a particular point (the trigger) in order to prevent a concrete, identifiable (diagnosable) downstream outcome.
One CEO of a tech-based obesity-prevention company framed it this way. They’ve had to lean “more toward describing ourselves as an intervention for early metabolic disease (e.g. pre-diabetes) vs. a prevention company.” He went further. “You’ll be hard pressed to gain traction as a generalized prevention company in the U.S. health care system. “
That everything is reactive seems to suggest that “true prevention” doesn’t exist within health care. Each intervention is in response to something. We are always monitoring and searching for problems. The goal is simply to find a trigger that signals to us earlier, a trigger that is further upstream.
The questions then become: How far upstream are we able to clinically mark a problem? And do we have an intervention that we know works if we implement it at that time?
If we want the delivery system to operate more upstream, we have to move the posts within which it operates.
That is: We need more of the right data to help identify better upstream triggers, and more evidence on what interventions work at the time of those upstream triggers.
More Data Does Not Mean Better Data. And We Need Better Data.
Most of the time, a clinical trigger occurs within a doctor’s visit. For example, when you go get a routine screening, you hope for a negative result, but if it’s positive then you hope you’ve caught it while it is still in its infancy.
In addition to the routine screenings (which are not without controversy, as doing even the right thing too early can cause more harm than good), there are increasingly other data sources brought into the mix to be “screened”. This future likely includes wearables/trackers, our online behaviors, and our genes and their expressions, are being incorporated into the monitoring database. Remote monitoring, via an internet-connected health devices, will eventually get folded in.
Digital health startups generally recognize this. If you can digitize something that used to be analogue, if you can passively track something that used to be manual, you are able to gather data that informs the health system’s understanding of a patient or patient group.
But thinking upstream can also be… literal. Water quality matters. As does neighborhood air quality and ones risk of lead poisoning due to, say, the age of their home.
One structural challenge to this world of data collection for clinical purposes is the business behind data collection. Data that can be monetized is more likely to be captured and woven into the business of health care. Further, data that identifies a trigger for which there is a clear (money-making/saving) intervention is also more likely to get captured.
We’re at a wild data-gathering time. Right now we have more data than we know what to do with it. The health system is using data in ways it never has before. Many of these ways are good. But many of these ways are problematic. Indeed some researchers worry that big data is leading to big error.
As we struggle through these early days of the digital revolution, an increasingly important question will be: What’s the right data that, when paired with an algorithm and/or human, reveals the right insight at the right time?
Clinical Efficacy is Required for An Intervention to be Prescribed
Once we have that signal, the next question is: What is the right intervention?
Classic non-clinical prevention programs — such as run out of public health departments— typically lack the rigor required by much of our health care system. (This is why, perhaps, they are run out of public health departments.) These types of intervention operate off of “theories of change” or “frameworks” that are common in the public health world.
At their worst, there is little to no evidence that they will work, but were merely the idea of a (likely overworked and underpaid) public health professional. At their best they may operate off of principles of effective prevention programs. But even this “best” is insufficient, at least from the lens of traditional health care delivery.
That’s because, clinical efficacy is the standard the health system requires. And many upstream and community-based interventions lack this standard of evidence.
This isn’t to say that those upstream and community-based (public health) interventions don’t work. Rather this is a reflection of our Western, clinical-trial-driven approach to medicine, where randomized clinical trials are the gold standard.
Clinical trials (or a study of sufficient statistical power) can be expensive, and the further upstream you go, the further into the community an intervention gets, the more difficult it becomes to control for variables.
The emergence of the digital therapeutics market, where software is the drug, demonstrates the importance of this mindset for entrepreneurs who want to sell into the health care system. These interventions range from the use of digitized, internet-connected scales coupled with online coaching to address obesity to interactive digital games that leverage neural plasticity to address mental health conditions like ADHD.
Companies in the world of digital therapeutics are running their products through clinical trials, seeking FDA approval, and thus garnering the standard of evidence — clinical efficacy — needed to be worthy of a physician’s prescription.
Contrast this with a general wellness program offered by an employee or a standard wellness app on iTunes or Google Play: While one can get healthier by using them, a doctor isn’t likely to prescribe them, nor is an insurance company likely to pay for it.
There may be exceptions of course, I know I’m over-stating the point, and I’m not even arguing that clinical efficacy is what we all should be going for. My point is simply that a VC-backed startup looking to sell into health care should hold these principles in mind.
The Business Case for Prevention is Often Hard to Make
Tied to this is the reimbursement pathway and the business of health care.
Health care folks talk a lot about the social determinants of health, one framework for thinking about upstream and community-based interventions. One insurance executive relayed a short interaction she had with her superiors that :
We were having a discussion as to the importance of addressing the social determinants of health. Everyone in the room, the entire leadership team, generally understand the social importance of, say, providing housing to people with unstable living situations. However, for it to be worthwhile for us [the business that is an insurance company], we’d need to see the direct financial impact of investing in those areas.
The dots are not sufficiently connected. There is no direct line from “providing housing” to “improved bottom line”.
It’s worth noting that this particular insurer still put money towards the issue of housing the homeless, but it became a “do good” effort categorized as charity. It became, just as “do good” work is for most companies, a PR expense.
Her ask was clear: “People like me need to be better equipped to take the business arguments [of prevention] to their leadership who may be more reticent.”
Summary and What’s Next
The legacy of fee for service. Churn. Data, including proving clinical efficacy, surrounding upstream care… These were difficult to fully explain. Their impacts are both more than what I’ve stated and less, depending on the particular slice of the health care system you may be looking at.
My goals were not to be complete in my representing these things, but rather to point out that they exist, and that they exist as market and business barriers to our health care system working further upstream and towards the direction of prevention and overall wellness.
And since investors follow — indeed: fund — markets, these point also to barriers preventing VC dollars from going towards products and services that work further upstream.