Taking On Risk
The Present and Future of Risk-Based Businesses in Healthcare
I write pieces on sub-sectors and trends within health tech and S-1 breakdowns. Sign up below to get the latest in your inbox.
Health tech startups are increasingly excited about taking on risk. These value-based arrangements allow them to build new higher touch care models without having to worry about reimbursement for each individual service (fee-for-service). Over the last decade, value-based care has been partially implemented by healthcare executives and furthered by policy. But we’re seeing more and more health tech startups pitch taking on risk. Their care models offer the promise of providing more frequent care that better meets patient needs and helps them better manage their conditions.
Traditional providers offer services that can be neatly mapped to billing codes. But risk-based businesses have flexibility to take whatever actions they think will best help patients manage their health and lower system costs. In a great overview of risk-based models put out by the a16z team, Corbin Petro, the CEO of Eleanor Health (a risk-based Substance Use Disorder company), put this succinctly:
“What the evidence tells us is that creating meaning and purpose in people’s lives really helps retain them in recovery and support their mental health and wellbeing. Well, there’s no CPT [billing] code for creating meaning and purpose in people’s lives. You have to figure out another way to address that, and the clinical model and the business model go hand in hand.”
These models have always made intuitive sense. The U.S. spends 5% of healthcare dollars on more proactive primary care. Investing in more frequent touchpoints with patients before their conditions flare up and building deeper trust-based relationships should improve patients’ lives and lower system costs.
The Buzz Around Risk-Based Businesses
So why the recent surge in investor and startup interest? There are four main drivers:
1. Evidence these models work
Risk-based businesses are producing strong results. Aledade increased primary care visits 34% which reduced ER utilization, inpatient utilization and total cost of care by around 15%. In its Accountable Care Organization population, Oak Street’s smaller patient panel primary care clinics reduced overall costs by 17%. Their care model, where they spend three times the average on primary care, has driven a nearly 50% reduction in hospital admissions and ER visits while maintaining a patient Net Promoter Score (NPS) of 90. ChenMed achieved similar results lowering hospital admissions and ER visits by around 30%. These results, along with the patient satisfaction rates, suggest risk-based models really can provide patients with a better healthcare experience and improve their lives through better condition management.
2. Public policy momentum for these arrangements is growing
Government policy has always driven value-based business model adoption. Traditional value-based care policies–like the introduction of Accountable Care Organizations, a shared savings model for government-run fee-for-service Medicare–spawned a previous wave of value-based companies like Aledade that helped practices succeed in this new model. More recent policies like the 2020 Comprehensive Kidney Care Contracting model allowed nephrologists to take on risk for Chronic Kidney Disease patients. This helped buoy a new wave of risk-based kidney care companies including Strive, Monogram, Somatus and Evergreen. Additionally, these government policies are helpful for increasing private payer adoption because they establish standards around how these contracts should be structured.
The introduction of ACO Reach should continue to expand the attractiveness of these models. This new government model allows practices to take on full risk for patients who are on traditional government fee-for-service Medicare, not just Medicare Advantage. Given that 60 percent of Medicare beneficiaries are on traditional Medicare, this dramatically increases the number of patients a practice can serve while going fully at risk. It will likely compel more practices to take on risk and more than doubles the addressable market for risk-based businesses.
3. Momentum is also increasing among private plans
As a playbook emerges for setting up these contracts, private payers continue to get more comfortable with risk arrangements. The typical arrangements involve payers giving 85-90% of their patient revenue to a risk-bearing provider. This allows the payer to lock in 10-15% profit margins.
There’s also becoming a standardized way payers can sequence startups through taking on risk. Arrangements can start as fee-for-service or per member-per month fees with quality/cost savings bonuses. Then they can move into shared savings where the payer and startup split the upside. After that, they can then move to limited capitation (e.g., just for primary care spend or a particular episode of care) before going to global capitation. This allows many startups to nail their operational model and scale with the balance sheet they need to successfully take on risk.
4. Increased public market investor interest in these models
At the highest level, investors are interested in approaches that bend the currently unsustainable healthcare cost curve. Risk-based businesses seem best positioned to do this given their financial incentives and the range of ways they can lower cost and improve outcomes.
The public market debuts of Agilon and Oak Street in the last two years, which both trade >$6B, and private market valuation of VillageMD >$15B have bolstered excitement for these businesses.
These businesses look like both health insurers in the amount of revenue they recognize and providers in the way they are valued by public market investors today.
For individual patients, risk-based businesses recognize substantially more revenue since they get paid the full amount that a patient is expected to cost the health system and take on risk for spend above or below that threshold. Fee-for-service providers only get paid for the set of services they render to a patient. Risk-based businesses’ revenue model is much more similar to insurers, which traditionally trade lower than providers. But Agilon and Oak Street trade at similar multiples to their counterparts who operate more on a fee-for-service model, such as Teladoc and Lifestance.
Why? Because at scale these businesses can enjoy a more attractive margin structure than traditional insurers. For one, they don’t have to pay the high costs of claims administration. On the member acquisition side, patient interest in their specific providers and care model makes them far more differentiated than insurers who are finding acquisition increasingly competitive and expensive. Oak Street Health is already projecting scaled EBITDA of 30% for its centers which is higher than many traditional providers’ margins.
Value-based models are still relatively nascent. Despite the buzz, Oak Street, ChenMed and Iora treat <1% of Medicare patients–but the potential is huge given the massive market and early proofpoints of meaningfully improving patient lives and lowering system costs.
When Does Risk Make Sense?
With valuation momentum and opportunities for clinical impact it makes sense more companies want to take on risk. But when does this make sense?
There are three factors that make taking on risk more attractive:
1. High per-patient costs
Many value-based care models are predicated on more frequent touchpoints with patients. But for patients who already have low overall costs it can be difficult to make the economics work on any increased touchpoints. Sicker patients stand to benefit more from these higher touch care models, as well as present more opportunities for cost-savings. We can see this in later stage Chronic Kidney Disease patients who incur an average of more than $50,000 in medical expenses due to frequent hospitalizations. The same opportunity to reduce frequent hospitalizations exists in the Program of All-Inclusive Care for the Elderly (PACE) for patients who need a nursing-home level of care. Helping patients better manage their conditions can reduce these hospitalizations.
Pine Park Health Chief Commercial Officer Dan O’Neill had a great thread on the attractive economics of InnovAge’s PACE patients. Its PACE patients, each of whom have an average of six chronic conditions, cost around $95,000 per year to treat. InnovAge has an opportunity to improve care and drive savings.
This dynamic is generally why there’s been such low relative penetration of risk-based models in lower-cost commercial populations vs. Medicare Advantage and dual eligibles.
2. Long length of patient relationship
A lot of risk-based models require investments upfront to save costs down the line. Short patient timelines, including in very elderly populations where members pass away or transient populations where members switch off plans, make the economics of taking on risk difficult. This is a huge consideration for commercial and Medicaid populations. The median employee time at a job is 3.7 years. Medicaid members often go in and out of eligibility. Medicare members have historically stayed on plans the longest, making Medicare Advantage a natural place for risk-based models to start.
3. Primary relationship with patient
There’s a joke that for every dollar saved in healthcare ten different organizations now try to take the credit. As more risk-based businesses emerge, it will likely become harder to take exclusive credit for full patient cost savings. Some risk models work with partial patient costs. Carrum Health is able to just take risk for higher-cost surgeries and save money on those surgeries.
But other businesses struggle without access to full patient costs. Behavioral health companies rightly claim their interventions drive down physical health costs for their patients. But because these patients’ physical health costs are typically much higher than their behavioral health costs, it’s hard for these companies to drive substantial savings on just behavioral health costs without access to risk for physical health spend.
Situations in which the specialty care doctor becomes the quarterback for a patient’s care–largely taking over from a primary care doctor–are particularly well suited for going fully at risk. These include nephrologists, oncologists and substance use disorder providers.
Other drivers of the attractiveness of taking on risk are still up for debate. While most at-risk providers have built their businesses in denser urban settings, a new wave of startups like Homeward, Hopscotch and Main Street Health are trying to bring these models to more rural areas. And while most risk-based businesses have focused on in-person care, new entrants like Babylon, Teladoc and Patina are taking on risk with more of a virtual-first model. It will be exciting to see how these approaches fare.
What Do Risk-Based Organizations Need To Be Successful
A new stack is emerging to help health systems and independent providers take on risk and take the actions that will improve health and lower costs.
To succeed under risk, it’s essential to successfully engage the patients on a providers’ panel, prioritize patient reach outs and go through the right set of activities during patient visits. A strong care model is crucial. Risk-based models are most successful when providers can control escalations and intervene early to keep patients out of the hospital. They are also more effective when providers can help navigate patients to effective specialty care. And when flare-up hospitalizations or ER visits do happen, it’s important that providers take action quickly.
These actions can help providers succeed in aggregate and prioritize a limited set of nurses, care coordinators etc. in the places where they can have the most impact. Some of these actions are more explicitly tied to value-based contracts. These include performance measures (e.g., colonoscopy rates or flu shot rates) and the per-patient amount companies receive. The latter is directly determined by the diagnoses providers have documented for that patient via risk adjustment.
Jan-Felix Schneider provided a great overview of the players enabling these actions here, and I’ve included my own amended market map below:
The first part of the key enabling infrastructure in risk-based models is obtaining as much patient data as possible to power decisions around their care. Data from previous encounters informs what conditions they might have, what medications they might be on, etc. and can be obtained from Health Information Exchanges or directly from a payer or previous provider. Real-time data on when a patient is admitted or discharged from a hospital can help ensure that immediate action is taken at crucial points in a patient’s outcome and cost journey. As healthcare data becomes increasingly liquid more and more of these providers are emerging.
Next, this data must be normalized. Health systems and payers often get data from dozens of different Electronic Health Records (EHR). Vendors use different normalization rules. Once the data is normalized, a number of applications can be built on top with business logic to determine how to prioritize patient reach outs. Care managers and population health teams identify trends in the data. Companies in the data normalization/analytics space differ on who they serve; for example, companies like Innovaccer and Health Catalyst have heavy implementations and largely serve large health systems.
Finally, one of the biggest challenges in risk-based businesses revolves around getting busy doctors and their staff to use this information. Companies are trying a range of approaches to meet this challenge. Aledade sends staff on site and helps practices make substantial changes to their workflows. Stellar, on the other hand, offers a lighter-weight innovative payment solution that seamlessly reimburses frontline staff and providers when they take value-based actions such as scheduling a patient who is due for a follow up.
Practices certainly require other solutions beyond tech. Organizations like Agilon and VillageMD help practices contract at scale. In addition to offering practices better rates as a single scaled entity, these organizations provide consulting services to help practices succeed under these arrangements. Some organizations provide wraparound care services that help augment practice staff (e.g., having centralized nurses call patients for follow-up visits).
What’s Still Unclear?
There are still some big outstanding questions whose answers will determine the relative attractiveness of at-risk models going forward.
1. Will payers renegotiate rates?
As risk-based providers drive savings, payers may set new baselines that make large savings difficult. One dynamic that may prevent this is that organizations like Agilon and Oak Street have shown early signs of being able to drive patient volume to their payer partners. Given the competitiveness of payer markets, this may give providers the upper-hand as they can threaten to redirect their marketing efforts toward other plans.
2. To what extent are there national scale advantages to these businesses?
There are massive regional advantages for scaled providers given the ability to build deeper relationships and get better rates with local payers and systems, consumer virality and a faster path to full clinic utilization. But it’s unclear how much of an advantage Oak Street will have beyond some national contract benefits and its balance sheet when expanding into entirely new geographies. The market may consolidate around a smaller set of massive national players or have regional champions, instead.
3. How many of these models will actually drive cost savings?
Third, the space is still quite early in proving cost savings. Not every care model will work. Today there are businesses taking on massive amounts of risk with minimal proofpoints of their ability to drive savings. The extent to which some of these don’t work may cast a shadow on earlier stage funding for similar opportunities, though it will differentiate the later stage companies that have figured this out.
A final area may determine if these businesses can have even more attractive economics. As discussed at greater length in this previous overview of Babylon, there are questions around how capital-lite a care model can be and still get access to full risk and savings. Virtual-first risk-based companies may prove to be even more attractive than their physical clinic-driven predecessors.
What Else is Needed?
There are still substantial gaps for providers taking on risk where future startups could help. A few include:
1. Solutions for payments
Risk-based businesses have challenging working capital dynamics; a determination of how much money they saved is often not made until months after a year’s end. This creates massive capital needs in a company’s early years. And this lagging indicator makes it challenging to evaluate providers and staff on a day-to-day basis. Innovative payment and data solutions that help alleviate these dynamics and better allow practices to track performance in real time would greatly help risk-based businesses. Payers will also need new technology to facilitate making these payments as they don’t fit neatly into their current claims systems.
2. Tools for specialists to take on risk
While the first wave of value-based care solutions has been focused on primary care doctors, it’s natural for a set of specialists to take on risk too. While the requirements of providers such as oncologists and nephrologists will share some similarities with their primary care peers, they will need tailored analytics, wraparound care models and contracting capabilities. There is an opportunity to build the equivalent of the primary care-facing enablement businesses for specialty care.
3. More data sources
Incorporating new data sources such as social determinants of health and wearables can improve risk stratification algorithms. While this data has become more easily available, startups will need to figure out a way to surface it in a way that changes decisions but isn’t too noisy for providers. Other data from places like home healthcare providers is difficult to gather today but could also improve risk stratification.
Given the momentum in the space, we certainly should see many investments powering different risk-based models over the next few years, resulting in better care experiences and outcomes for patients. It will be exciting to get clarity on which approaches are most effective and see what kind of supporting infrastructure emerges to help providers.