This article first appeared in JAMANetwork.com. It is co-authored with Rahul Rajkumar, MD.
To some it might appear as though the US health care system has made little progress over the past 10 years. Ten years ago, many perceived that health care was unaffordable for many people in the US, and that the quality of care that many people receive was unnecessarily variable. These statements are still largely true.
However, a closer look reveals progress being made on these persistent challenges. Stimulated by payment innovation through the Center for Medicare and Medicaid Innovation and federal policy more broadly, Medicare per capita spending increased at rate of 1.8% per year between 2010 and 2018, the lowest rate of growth in history.1 Innovation in health care delivery has largely followed the opportunities created by payment innovation. Two of the most important areas of innovation have been the Medicare Advantage program, in which health plans receive risk-adjusted capitated payments from the federal government to manage a population of Medicare beneficiaries, and the Medicare Shared Savings Program, in which health care organizations and clinicians can be rewarded in exchange for assuming financial risk for managing the quality and total cost of care of a population.
According to the Health Care Payment Learning and Action Network, in 2018, only 35.8% of payments to health care clinicians flowed through payment models that included some accountability for cost and quality of care. Furthermore, only 14.5% of the payments flowed through models that included both upside and downside risk.2
Much of the last decade of health policy making was spent designing alternative payment models (APMs) and then “advanced” alternative payment models (A-APMs) and encouraging primary care physicians to adopt them. However, current federal policy offers primary care physicians and specialty physicians a choice between the Merit-Based Incentive Payment System (MIPS) and participating in true APMs, such as accountable care organizations (ACOs). In 2018, 17% of physicians choose ACO models with 2-sided risk.3 MIPS is less of a risk-based value-based payment model and more of a pay-for-performance program. It is designed to be a budget-neutral fee-for-service program with quality metrics, information exchange, cost, and practice improvement domains as opposed to a payment model designed to save money. Moreover, MIPS as a scheme combines modest but easily attained bonus payments with a relatively low probability of financial penalty. A-APMs, on the other hand, include higher levels of risk and therefore reward clinicians with more money for more cost savings and ensure that the Medicare program saves money by transferring risk to physicians. As a results, A-APMs create stronger incentives than APMs or MIPS for clinicians to redesign care to be lower cost.
A major challenge to building a collaborative value-based delivery system arises when downstream specialist physicians still operate on a fee-for-service basis. As a result, specialists remain focused on maximizing relative value units, delivering diagnostic tests, offering procedure-intensive approaches to care, and prescribing expensive medications. This undermines the efforts to move primary care clinicians into A-APMs, as the care model favored by specialists is the exact opposite approach favored by primary care clinicians.
This has led to actions designed to help primary care clinicians reduce the number of specialty referrals they generate. Entrepreneurs have created businesses offering virtual specialty consultations to help primary care clinicians retain control of diagnostic workups and avoid facility fees. It has led to larger risk bearing for primary care physician groups who hire their own specialists. It has also caused risk-bearing clinicians and health care organizations to form “compacts”4 with specialists whereby primary care physicians bundle their referrals to specialist physicians who agree to parsimonious approaches to care.
A better approach would be to broaden the value-based payment portfolio by creating APMs for specialists, and to ensure that participating in A-APMs is more attractive economically for specialists than participating in MIPS.
A major shortcoming of the current system is that investments in health have very short payback periods. In all APMs, savings for primary care physicians are measured in 12-month time frames. Insurance companies have slightly longer time frames: commercially insured members may stay with a health plan for 2 to 4 years; Medicare Advantage members may stay with a health plan for 5 to 7 years. These time frames make it challenging to invest in interventions that have longer-term benefits, even when these investments may produce substantial improvements in health.
For example, bariatric surgery or carbohydrate restriction can reverse type 2 diabetes in some patients and effectively eliminate the need for costly diabetes medications.5 The challenge for primary care physicians in APMs and payers is that the cost of these interventions is not recouped entirely in year 1. Therefore, these interventions are more likely to be adopted by employers, unions, and pensions that are responsible for health care costs for populations over long time periods.6
Several policy levers can help address this problem. One approach could be assigning a value to specific preventive care services and prepaying the savings to clinicians and health care entities for delivering preventive care. Another approach may be allowing for multiyear enrollment in health plans so that health plans have an incentive to make long-term preventive care investments in their member populations. More boldly, a new funding mechanism for breakthrough preventive care, such as health-related bonds, which are held by investors, and which increase in value along with a person’s long-term health, could create a strong incentive to address social determinants of health and healthier lifestyle interventions.
As the Centers for Medicare & Medicaid Services continue to drive adoption of APMs and, hopefully, specialty care APMs, a major barrier to system change is that commercial insurance companies largely pay clinicians and health care organizations using a fee-for-service approach. Furthermore, hospitals make most of their margin by caring for commercially insured patients. This makes redesigning care models to support Medicare’s APMs misaligned with economic incentives.
What is needed is a revenue-neutral transition to non–fee-for-service payment models for commercially insured patients. One model could be moving to a capitated payment model for hospital services that is initially equal to the current annualized spending for hospital services, so hospitals receive monthly payments for the availability of their services rather than be paid on a fee-for-service basis. To drive adoption, the federal government could make this hospital payment model a requirement for the Federal Employees Health Benefits program and Affordable Care Act exchange plans, and a feature of Employee Retirement Income Security Act plans. Removing hospital profits from fee-for-service activities for commercially insured patients could have a profoundly positive influence on redesigning care to be less hospital and procedure oriented.
Measuring quality of care and risk adjustment are critical components of any framework for health care delivery innovation. However, current methods for performing both functions are costly and cumbersome. Risk adjustment alone adds an estimated $16 billion annually to health care costs.7 Today, health plans manually review charts for approximately 80% of eligible patients to identify risk adjustment codes and identify gaps in care, and risk-bearing clinicians and health care organizations spend considerable time documenting the presence of diseases rather than treating them. Importantly, commercial, Medicare Advantage, and Medicaid payers should adopt the same patient attribution, quality metrics, and risk adjustment (when appropriate) methods as Medicare.
An even better approach would be a new risk adjustment model derived from objective data that do not rely on inputs from clinicians or payers. Examples of such data include laboratory values, medication prescribing, and patient surveys. Using these data that already exist in the electronic medical record and surveys already collected by Medicare could obviate the need to collect and review records and spend clinician time on clinical documentation.
By implementing these 4 strategies, it may be possible to make another bend in the curve of health care cost growth, and align incentives for the specialty care and hospital systems with primary care services for the first time in support of value-based care. These changes in incentives could also spur another wave of innovation that benefits patients and physicians and reduces the cost of health care.
Corresponding Author: Bob Kocher, MD, 3340 Hillview Ave, Palo Alto, CA 94301 (firstname.lastname@example.org).
Published Online: August 16, 2021. doi:10.1001/jama.2021.13153
Conflict of Interest Disclosures: Dr Kocher reported being a partner in the venture capital firm Venrock, in which he invests in health care IT and services companies. He is a board member of Devoted Health, Lyra Health, Aledade, Sitka Health, Virta Health, Newco Health, and Premera Blue Cross. Dr Rajkumar reported being an employee for Optum Care Solutions, receiving personal fees for advising from Google and PicassoMD, receiving personal fees for board membership from Health Precision, and having equity in Advantia Health and OM1.
- 1. 2020 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Fund. April 22, 2020. Accessed July 9, 2021. https://www.cms.gov/files/document/2020-medicare-trustees-report.pdf
- 2. Mitre Corporation Healthcare Payment Learning and Action Network. APM Measurement Effort. Infographic. Published 2019. Accessed July 9, 2021. https://hcp-lan.org/workproducts/apm-infographic-2019.pdf
- 3. Gonzales-Smith J, Bleser WK, Muhlestein D, Richard R, McClellan MB, Saunders RS. Medicare ACO results for 2018: more downside risk adoption, more savings, and all ACO types now averaging savings. Health Affairs blog. October 25, 2019. Accessed July 9, 2021. https://www.healthaffairs.org/do/10.1377/hblog20191024.65681/full/
- 4. Gorenstein D. Playing hardball: doctors start negotiating with doctors. Minnesota Public Radio Marketplace. October 26, 2016. Accessed July 9, 2021. https://www.marketplace.org/2016/10/26/playing-hardball-doctors-start-negotiating-doctors/
- 5. Hallberg SJ, Gershuni VM, Hazbun TL, Athinarayanan SJ. Reversing type 2 diabetes: a narrative review of the evidence. Nutrients. 2019;11(4):E766. doi:10.3390/nu11040766PubMedGoogle Scholar
- 6. Winkler R. Digital-health startups are booming. Their customers are overwhelmed. Wall Street Journal. May 3, 2021. Accessed July 9, 2021. https://www.wsj.com/articles/digital-health-startups-are-booming-their-customers-are-overwhelmed-11620039601
- 7. Grand View Research. US medical coding market size, share and trends analysis report by classification system, by component, and segment forecasts, 2021-2028. February 2021. Accessed July 9, 2021. https://www.grandviewresearch.com/industry-analysis/us-medical-coding-market