Why you should build in primary care: a layman’s guide to risk delegation
A profound shift in how primary care is financed is opening up opportunities for entrepreneurs to transform the landscape of American healthcare. Who will start the next Humana?
This article assumes you have a working knowledge of health insurance in the United States: for instance, Medicare, Medicare Advantage, and fee-for-service reimbursement. If you don’t, a primer is on the way shortly.
The most exciting part of our healthcare system for startups is surprisingly neither high-flying specialties nor richly-reimbursed procedures. The best opportunities are actually in humble primary care.¹
“But wait,” you cry, “how can that be? There’s no money in primary care!” Up until the past few years your skepticism would have been warranted.
But good primary care creates mounds of downstream savings and is now finally being rewarded. Both private payers and Medicare have recently expanded alternatives to fee-for-service reimbursement that pay for effective prevention, coordination, and steerage. If advanced primary care keeps patients out of the hospital, it is possible to earn over $1,000 per patient through these new models in comparison to about $85 per patient in a fee-for-service regime.
In plain English: top-quality primary care providers can earn more than ten times as much per patient in exchange for reducing unnecessary hospitalizations and slowing disease progression.
If you aren’t a policy wonk or an industry insider, you may have missed this entirely. But fear not: for those of us who have not spent time in the Humphrey Building, we are here to put risk delegation into plain English.
We believe that the strongest entrepreneurs targeting this opportunity could build something larger than today’s insurance giants — Humana, Aetna, and Centene — while improving outcomes for our seniors and saving American taxpayers billions of dollars.
How have primary care providers traditionally been paid?
Let’s start by understanding the old financing regime in order to contrast it against what is emerging.
Traditionally, providers are paid on a fee-for-service basis: for each service (e.g. a consultation, a flu shot, etc) rendered to patients, a service code is assigned. These service codes are recorded on claims and sent to insurers who pay a fixed reimbursement for each claim line.
For instance, one of the most common service codes in primary care is 99213, “Level 3 Established Office Visit” — a typical primary care consultation. A more complex visit addressing multiple chronic conditions could result in a bill for 99214. For a comprehensive annual exam, providers may bill G0438, “Annual Wellness Visit”.
Each of these codes may require 45 minutes or more of physician time and are reimbursed at the following rates by Medicare. You can look these up yourself on the Medicare Physician Fee Schedule:²
- 99213 — $77.98
- 99214 — $113.03
- G0438 — $176.92
If these codes are a practice’s bread and butter, it is hard to earn much profit. A very crude analysis suggests that a practice can earn about $85 per commercial patient per year before considering overhead and administrative costs. Medicare may yield a bit more per patient because the elderly tend to see the doctor more often than those under 65, with both higher revenue and higher associated costs:
Compare this with the gastroenterologist across the road making $350 for each 30–60 minute colonoscopy, and you start to get the picture.
Code-based reimbursement also limits how care can be delivered. If providers want to use AI to assist decisions or autonomously make diagnoses or conduct screens, they may not get paid. If providers want to ask patients to send images of a new mole or possible ear infection for remote evaluation, there may not be a code. If providers want to spend time helping a patient navigate to a good in-network specialist or find social services — same story.
In the fee-for-service world, primary care is fundamentally reimbursed for process instead of outcomes, and it is reimbursed relatively poorly.
How is primary care reimbursement changing?
At $85 to $100 per patient per year, there is not much incentive for the private sector to invest in primary care. New entrants found ways to change the dynamic in the 2010s by introducing new financing models.
One Medical added an annual membership fee of about $125 per patient per year on top of fee-for-service reimbursement. In some locations, One Medical partners with local health systems, and receives a per patient per month partnership payment from the health system in lieu of billing payers on a fee-for service basis. The resulting patient-level economics are below:
What is this partnership fee? In clinics where One Medical has a health system partnership with a group like UCSF, UCSF takes care of fee-for-service billing on One Medical’s behalf and collects reimbursement from payers. The monthly fee that UCSF provides One Medical in lieu of claim reimbursement typically exceeds what fee-for-service reimbursement would have yielded. UCSF effectively pays One Medical $200 a patient per year for preferential referrals.
Surging enrollment in Medicare Advantage plans created opportunities to innovate further on primary care financing towards the middle of the 2010s.⁶ Some names that may not be familiar like Oak Street Health, Landmark, and CareMore were simultaneously developing at-risk contracts that yield even better patient-level economics.
Risk delegation allows primary care providers to earn profit by keeping patients out of the hospital. Medicare Advantage plans hand over 85% of their premium to a primary care provider like Oak Street Health, who is then responsible for all of that patient’s costs including primary care but also specialist visits and hospitalizations.⁷ Keep the patient out of the hospital and you earn a handsome reward; fail to do so and you are on the hook for costs beyond the premium you received.⁸
The implications are that risk-bearing primary care providers can spend as much as they would like on prevention, chronic disease management, and steerage if that investment can pay off in additional savings against projected costs. Exceptional primary care is rewarded with exceptional profit, though failure to contain cost is punished with losses.
In most arrangements providers are significantly freed from the constraints of billing codes.⁹ Technology and work on social determinants of health can all be rewarded through shared savings even if a code does not exist.
What can this look like at a patient level? Much higher revenue and substantially more profit per patient:
The reason to be excited is easier to visualize when we compare the numbers above to the Fee-For-Service and One Medical patients considered earlier:
These numbers are notional but realistic. It’s sort of like if One Medical could generate twice as much cash flow for the same capital investment in a new clinic.
Oak Street Health, ChenMed, and others continue to expand to serve Medicare Advantage plans through these types of contracts. 8VC portfolio companies Cityblock and Galileo represent a second wave that uses technology in significant ways to drive quality, efficiency, and patient satisfaction. If you can shrink the “Primary Care Costs” part of the “Cost of Care” bar in any way the value is immense. Meanwhile, a good patient experience drives engagement and reduces 3rd party costs while promoting retention.
Only a very small fraction of Medicare Advantage members are managed through even all of these names combined. Medicare Advantage is adding over one million new members a year. Bearing risk on Medicare Advantage lives will continue to be a good space for some time to come.
But an even bigger opportunity is right around the corner.
Last year CMS made a policy change that allows primary care providers to switch from fee-for-service to at-risk reimbursement for the 42M regular Medicare beneficiaries outside the 22M enrolled in Medicare Advantage.¹⁰ ¹¹ This is a gargantuan opportunity for entrepreneurs.
There is plenty of nuance, but if we squint, Direct Contracting truly behaves quite similarly to risk delegation as it has been implemented in Medicare Advantage. From an official CMS deck:¹²
Medicare fee-for-service beneficiaries can be aligned with a given primary care practice if that practice is enrolled in Direct Contracting and either (A) the beneficiary explicitly consents (“voluntary alignment”), or (B) the beneficiary receives most of their primary care from that practice without explicitly signing up to be assigned (“claims-based alignment”).
If you are wondering about the finer details of the model, the acronyms in the slide and how Direct Contracting differs from Accountable Care Organizations, these are reasonable questions that we may address in a future post. For now our takeaway should be that there is a massive and under-explored opportunity for tech-enabled primary care in the Medicare fee-for-service population.
What should we build?
Every once and awhile it is possible to spot an opening in the US healthcare system where good technology and efficiency and quality are incentivized and rewarded. Primary care for our Medicare beneficiaries is one such opening and we encourage entrepreneurs to drive a truck through it.
Frankly, we were surprised to see so little discussion of the Direct Contracting model in the technology community when it was announced last April. If you are an aspiring healthcare entrepreneur, take a second and sign up for CMS’s weekly newsletter. This is one of many alternative payment models that rewards innovation.
Continuing on, a few thoughts on specific niches to explore:
We think there is some room for pure enablement — operational support and technology for existing practices — in this environment. Aledade for instance is doing admirable work helping existing independent practices transition to a value-based world though primarily focusing on ACOs. But this is a hands-on business. It is hard to teach an old dog new tricks; technology is necessarily only a small part of it.
With that said, one intriguing technology pure play would be an “iron man suit” for physicians to set up a value-based practice fresh out of residency and take advantage of these payment models. Imagine if a family practice resident could sign up online and immediately get access to an EHR, payer credentialing and billing support, population health tools, various supplier discounts and administrative support, and help acquiring patients to build a panel. The whole practice could be designed around home visits to minimize the complexity and capital expenditure of a brick and mortar office. This model would avoid the service-intensive work of shifting the behaviors of an existing practice.
But as previously mentioned, we are a fan of technology-enabled service businesses for a few reasons and feel that the biggest opportunity may be for newly-built primary care groups specializing in taking risk. What can be done beyond Oak Street and ChenMed? Or Cityblock and Galileo for that matter?
At-Risk Primary Care for Defined Populations
One angle is if there are particular patient segments defined by socioeconomic, geographic, or medical variables that are best served by a niche player. Could you make Oak Street for rural America for instance? A focused provider could offer superior results and a top-notch patient experience for specific types of individuals based on disease state or other characteristics. What about a group that is geared towards patients with cardiovascular disease, or those with dementia or Alzheimer’s disease? Ultimately markets will have multiple options and different providers who specialize in particular types of risks. If you can create value for a patient segment you can earn a reward.
At-Risk Primary Care Beyond Medicare
We may also see risk delegation for other types of insurance. Could something be built out for commercial payers or managed care organizations and our Medicaid population? Medicaid plans are perpetually cash-strapped and in the wake of COVID will have even greater pressure to contain costs while simultaneously growing coverage. A low cost solution for Medicaid that truly reduces inpatient and / or emergency room utilization could be of tremendous value to states and transform opportunities for our least well off.
Value-Based Specialty Care and Procedures
Finally we are intrigued by specialist providers that may partner with risk-bearing primary care providers and Direct Contracting Entities in the future. There may be space for e.g. value-focused orthopedic groups to emerge, similar to Arizona’s CORE Institute. As opposed to hospitals and health systems that charge on a fee-for-service basis and command high reimbursement, new entrants could take a bundled payment to manage a whole episode to reduce waste, and offering the lowest possible price would ensure volume from risk-bearing primary care.
In the Direct Contracting model, these specialists could actually join together into a Direct Contracting Entity with the primary care provider, and each earn a piece of any shared savings.
A lack of strong incentives to do better than yesterday plagues our American healthcare system. It is frustrating that exceptions are hidden away from all but industry insiders.
But if we collectively want to enjoy a higher standard of living, we need engineers to dive into the acronyms, in partnership with healthcare professionals who do not allow the preciousness of their mission to get in the way of innovation, and build the delivery system we ought to have.
Success in the new world of primary care means outsize rewards for entrepreneurs and investors, reduced costs for the American taxpayer, and a radically higher standard of care for our seniors and beyond. If politics is about tradeoffs between guns and butter, here technology and innovative service design allows us to have more of everything.
 Primary care means many things to many people. This is a pedantic argument that the author does not find interesting. Primary care is here defined as a front door that patients access before going to see specialists. It can be elastic to include e.g. chronic disease management and behavioral health, or narrower, depending on the practice.
 https://www.cms.gov/apps/physician-fee-schedule/search/search-criteria.aspx — we are showing the median rate across different MACs.
 https://www.sec.gov/Archives/edgar/data/1404123/000119312520001429/d806726ds1.htm — here we are taking per-patient revenue and cost of care from 2017 when patients were reimbursed on a fee-for-service basis instead of via partnership fees from health systems, and we exclude all membership fees from the revenue.
 This allows plans to predictably spend at their MLR limit on a given member. Note, however, that contracts can be configured in many different ways according to specific payer preferences.
 A company like Oak Street can earn profit through savings against the ceded premium, but also by effectively pushing up the premium through better diagnosis documentation, HCC capture, and Risk Adjustment. The excellent Rahul Rajkumar of Blue Cross Blue Shield of North Carolina raises the real concern that in some cases, at-risk financing will yield better HCC capture, but not any real decrease in utilization: https://www.healthaffairs.org/do/10.1377/hblog20200325.524312/full/.
 We are simplifying things here — at-risk providers will still have to generate claims as usual for various reporting requirements. Meanwhile, both regulation and payers can constrain benefits that are offered to patients; there are some strings attached to ceded premium dollars.