The U.S. healthcare sector faces two dilemmas. The first concerns high levels of expenditures. The United States spends a larger fraction of its

national income on healthcare than any other advanced economy. High spending might be acceptable if it produced outstanding outcomes. But health results in the U.S. do not compare well with those of other wealthy countries.

The primary reason for high healthcare expenditures in the U.S. is the ability of dominant players to mark up their prices far above marginal costs. This exercise of market power makes some firms and individuals richer, but makes the nation poorer overall. Addressing the dilemma of high expenditures thus requires reducing market concentration, increasing competition and improving how healthcare markets work.

The second dilemma concerns the growth of healthcare spending. Since 1975, it has on average risen 2 percent faster per year than GDP. If these relative growth rates persist for another 75 years – the long-range planning horizon for government entitlement programs like Medicare – the health sector will absorb 75 percent of GDP! Such trends seem to point toward a dystopian future where health spending crowds out the consumption of other necessary goods and services.

Conventional wisdom regards these two dilemmas as manifestations of the same problem: overspending on health. However, they result from different economic forces and call for distinct, though sometimes overlapping, responses. This chapter explains why health expenditures grow faster than GDP and what to do about it. Our answer emphasizes ways to align the incentives for innovation more closely with creating economic value and reducing costs.

Growth in Health Expenditures

The phenomenon of healthcare costs growing faster than GDP is not unique to the United States – the share of GDP going to healthcare has increased across all wealthy nations. There are two leading explanations for this: expensive new medical technologies and rising national incomes.

The technology for delivering health services is radically different now than in 1960. However, this flourishing of new technology does not satisfactorily explain why costs rise faster than GDP. New technologies need not find a market simply because they are available. This is particularly so if there is good reason to believe that many new technologies do not offer much added value for patients. We call the new medical technologies explanation of the growth in health spending the Field of Dreams theory because its logic echoes the great line from that 1989 film, “If you build it, they will come.”

Rising national incomes is the alternative and, in our judgment, the better explanation. Estimates across countries and regions within countries find that a 1 percent increase in GDP correlates with a 1.6 percent increase in health expenditures. If the percentage increase in health spending exceeds the percentage increase in national income, it follows that health spending will constitute a growing share of national income.

But why should more prosperous nations spend an ever-larger proportion of their growing income on health? For most things we buy, more is better, but the additional satisfaction falls with each additional increment consumed. The third donut may taste good, but probably not as good as the first.

The psychological tendency for satiation acts as a natural brake on consumption. People may continue to spend more on donuts as they grow more prosperous, but they will also direct their rising income to other goods and services whose marginal value may be higher. No particular good or service should take up a growing share of GDP as societies grow more affluent.

To understand why expenditures on healthcare differ from other expenditures, consider the following thought experiment. Suppose a society was determined to spend exactly the amount on healthcare that would make each of its citizens as well off and satisfied as possible. The optimal healthcare expenditures would be determined by the consequences of taking a percent of GDP away from other consumption and spending it on healthcare. This reallocation generates a loss (individuals consume less) and a gain (individuals live longer and healthier lives). If the loss is less than the gain, society will make its members happier by increasing healthcare expenditures as a percent of GDP.

Society should thus continue to allocate more to health so long as the gains from doing so exceed the losses in forgone non-health consumption. Individuals in a poor society may gain more from spending on roads, food, and housing and be better off with a smaller share of GDP going to health. As society grows more affluent, consumers value each additional unit of consumption less. As consumers become increasingly satiated from non-health consumption, the society in our thought experiment will allocate less to non-health goods and services and more to health expenditures that increase the length and quality of life. Based on these tradeoffs and guesses about the growth of U.S. GDP, one estimate is that the optimal solution to this thought experiment is for the United States to spend more than 30 percent of GDP on healthcare by the middle of the century.

This result suggests that healthcare costs will constitute a rising share of GDP for decades to come. Indeed, it may be that 30 percent is an underestimate because improvements in healthcare technology can increase the desirability of spending even more on healthcare.

Consider that since 1970 there has been a dramatic reduction in mortality from heart disease – particularly in men over age 50. As a result, more people survive heart attacks, increasing the value of other medical services. People who do not die of cardiac disease in their 50s need knee replacements in their 60s and treatment for Alzheimer’s in their 70s and 80s. They also place greater value on improvements in the treatment of these conditions. These mutually reinforcing effects of healthcare innovation are not small. One study estimates that reductions in mortality since 1970 have raised the value of further health progress by nearly 20 percent.

Spending a growing share of national income on health need not be a source of concern. Indeed, under the right circumstances, it increases our prosperity. To explain, let’s extend our thought experiment to consider two healthcare systems: one is the current system – let’s call that System A. The other an alternative, hypothetical system, System B, that is identical to System A in all respects except one. System A has implemented technologies and management practices such that a 10 percent increase in health spending improves life expectancy by 1.25 percent. System B, in contrast, requires only a 5 percent increase in health spending to achieve the same increase in life expectancy.

In both systems, demand for healthcare increases as society grows wealthier. However, System B will spend a higher fraction of GDP on its health sector than System A. That’s because System B gets more for each additional dollar spent on healthcare, making society more willing to trade reduced consumption growth in other areas for more longevity and a higher quality of life.

Healthcare expenditures as a portion of GDP will grow faster in System B than in System A. But the faster growth rate will not impoverish the nation because System B is better at converting foregone consumption into improved health.

Technologies, practice patterns and business models that convert resources into health outcomes result from many years of prior discovery and innovation. When the patent system fails to support the development of vaccines or antibiotics while encouraging life-cycle management of existing drugs, when cost-reducing innovations regularly “go missing,” the result is a set of technologies and processes that are less efficient at converting national income into desirable health outcomes. These distortions make the U.S. less well off than it could be, a state of affairs that is neither desirable nor necessary.

Opportunities

The conventional policy stance calls for suppressing the rise of healthcare costs, but this is a recipe for failure. Restricting healthcare spending by, for example, tying the growth of Medicare to the growth of GDP will not turn off the economic, technical and psychological forces that cause demand for health-care to grow faster than national income. A better strategy would be to strengthen incentives for innovations that create economic value and reduce costs.

Our recommendations fall into two categories: (1) better aligning the incentives of the patent system with the creation of economic value, and (2) making it easier for innovators and implementers to make money by reducing the cost of healthcare.

Incentives for Creating Economic Value

For vaccines, cancer drugs, antibiotics and the life-cycle management of drugs, the incentives inherent in the patent system are not well-aligned with creating economic value. The heart of the problem is that patents offer innovators time-limited market exclusivity as the primary incentive for innovation and discovery.

This incentive is problematic in several ways. It leads to monopoly pricing that inhibits the use of valuable therapies, time limitations that skew the choice of disease targets for drug development and reward life-cycle management practices, and reliance on “willingness to pay” as the measure of value, which biases innovation toward the health issues of wealthy individuals and countries – and which cannot account for benefits not reflected in the market price.

In this section, we discuss ideas to better align patent incentives with the creation of economic value in healthcare.

Subscription Models

Consider the incentives for the inventors of new antibiotics. For the most part, new antibiotics are simply more expensive substitutes for older antibiotics – until the microbes develop resistance to the old repertoire. At that point, new antibiotics become very valuable. Thus, the economic value of new antibiotics results from the option they offer in responding to a future contingency.

Resistance to current antibiotics may not emerge during the period of market exclusivity created by patents, and this uncertainty reduces the value of the patent protection. But introducing a new antibiotic before resistance to current treatments arises gives the bacteria a head start in developing resistance to the new drug. Wise social stewardship thus ought to push the introduction of new antibiotics as far as possible into the future.

A subscription model would remedy many of the deficiencies of the current system. One recent proposal calls for a fixed recurring payment linked to clinical performance metrics. Such a system pays for the availability of new antibiotics rather than their use, encouraging both innovation and wise stewardship.

In this model, an expert committee would set criteria for determining participation in the program – presumably focusing on high-need areas for drug development. Next, a proposed antibiotic would be evaluated to determine if it meets the qualification criteria. Then, the manufacturer would work with the federal Center for Medicare and Medicaid Services to agree on the evidentiary basis for demonstrating clinical value, and the Food and Drug Administration would evaluate the drug’s efficacy. Finally, CMS would negotiate subscription payments with the drugmaker in exchange for guarantees that the drug-maker will make the antibiotic available when needed.

The subscription revenues must be set high enough to induce drugmakers to invest in antibiotic development, but not so high as to exceed the expected value created by the new drug. Hospitals could gain access to the new drug for a low price. But in exchange, they would be required to implement CDC-recommended antimicrobial stewardship protocols, adhere to appropriate use practices and provide data to allow CMS to monitor performance.

The Frankfurt Stock Exchange welcomes a new pharmaceutical company.

Louisiana used a variation of a subscription model to ensure access to antiviral drugs for hepatitis C for its Medicaid and incarcerated populations. Unlike the subscription proposal for antibiotics, Louisiana’s model was not intended to increase incentives for innovation but to increase access by underserved populations to a valuable innovation that would otherwise be priced out of reach. However, both models use a subscription – paying for access, not for use – to address pricing problems arising from the patent system.

Hepatitis C is a deadly infectious disease that can now be cured by patented antiviral medications. These one-treatment medications are typically sold at a high price: $80,000. Treating just the Louisianans with hepatitis C enrolled in Medicaid would thus have cost an estimated $760 million – more than the state spends on K-12 education, veterans affairs or corrections. So Louisiana negotiated an agreement with pharmaceutical companies in which it committed to spending a fixed amount annually on hepatitis C medication in exchange for universal access to the drug for its Medicaid and incarcerated population. The drugmaker still charges high monopoly prices to those willing and able to pay those prices, and the state “subscribes” on behalf of defined populations who cannot pay the monopoly price.

Advanced Market Commitments

Patents rely on “willingness to pay” measures of value. For most goods and services, consumer willingness to pay is a good enough indicator of social value, even though it creates some inequality. It is acceptable, for example, that Tesla entered the car market with vehicles priced and designed for the needs of high-income buyers.

However, high-income individuals – and societies – are willing to spend more than low-income individuals and societies on better health. They are also eager to spend a considerably higher percentage of their already higher income. Thus in the health sector, a market test based on willingness-to-pay skews the incentives for innovation sharply toward improving the health and longevity of high-income individuals and high-income countries. And this results in underinvestment in preventives and treatments for diseases affecting poor countries (e.g., malaria) or marginalized populations in the United States (e.g., sickle cell disease or addiction).

An advance market commitment could offset some of these distortions. Under an AMC, sponsors (governments, philanthropies, the World Health Organization) estimate in advance the social value of a treatment and commit to paying that value to developers who can meet their specifications. Payments only occur after regulators determine that the treatment is safe and effective. AMCs do not specify any particular technology – any treatment that meets requirements is eligible.

The benefit of AMCs is illustrated most clearly in the vaccine market for diseases endemic to poor countries. R&D directed toward health problems in developing countries receives a small portion of the total investment in health R&D globally. A 2005 Center for Global Development report estimated that $6 billion is spent annually on diseases such as malaria, schistosomiasis and Chagas disease compared to global R&D expenditures of $100 billion. The same report estimated that the total size of the vaccine market in developing countries is $500 million, too small to attract much R&D investment.

In 2007, five countries and the Bill & Melinda Gates Foundation pledged $1.5 billion toward a pilot AMC targeting pneumococcal conjugate vaccine. At the start of the project, the World Health Organization estimated that pneumococcus annually killed more than 700,000 children under age five in developing countries.

The AMC launched in 2009 to provide 200 million treatments annually. Firms committing to the agreement would receive a share of the $1.5 billion AMC fund proportionate to the number of doses they supplied, paid out as a per-dose subsidy. In 2010 two drugmakers, GSK and Pfizer, each committed to supplying 30 million doses annually. By 2016, over 50 million children per year were being immunized for PVC. In 2019, a third vaccine developed by the Serum Institute of India qualified for the AMC program.

The experience with the AMC vaccine suggests that it may have saved 700,000 lives at a very favorable cost. AMCs, in other words, may be a cost-effective way to undo some of the distortions resulting from the willingness- to-pay market test built into the patent system.

The Covid-19 pandemic provides another striking example of the impact of AMCs on vaccine development. Under Operation Warp Speed, the United States offered a variety of contracts to manufacturers to spur the development of a Covid-19 vaccine. For example, Pfizer’s agreement offered a $2 billion payment contingent on licensure or emergency-use authorization for 100 million doses. The result of the program was the development of highly effective vaccines for a novel pathogen, using a new mRNA technology, in less than 12 months.

Two-Part Pricing

A drugmaker that uses its patent monopoly to set prices above marginal cost captures a larger share of a smaller economic pie. If a patent holder sells their invention at a price equal to marginal cost, they maximize the value created by their treatment. But all the value flows to buyers – there is no money left over to reward the innovator.

A well-known solution to this problem is for the patent holder to sell the drug at a price equal to its marginal production cost and then charge a separate licensing fee for access to the treatment. Because the licensing fee is separate from the price of the drug, the fee does not discourage patients from purchasing the medication. Economists call this strategy two-part pricing because of its two distinct elements: a low price close to marginal cost, plus a licensing fee.

Actually, the institutional structure of the market for patented drugs available to insured patients in the United States offers a rough approximation to two-part pricing. At the center of this market sit entities called pharmacy benefit managers. PBMs operate formularies that make designated drugs available to consumers at low co-pays that are close to the drugs’ marginal costs.

Insurance premiums constitute the fixed fee portion of two-part pricing. Insurance companies use these premiums to pay PBMs a price per drug that significantly exceeds the co-pay. PBMs, in turn, use the revenues from the insurers to pay the price above marginal cost demanded by drugmakers.

A downside to this complicated structure is that the considerable value created by the two-part pricing strategy does not flow to consumers or to drugmakers when the market for PBM services is not very competitive. PBMs can use their position as intermediaries to keep the additional revenue created by the low co-pays. Consequently, the incentives for innovation are much lower than they could be.

One way to address this inefficiency would be to levy a tax on PBMs and use the revenues collected to invest in R&D. A well-designed tax on PBMs would not raise co-pays or distort the flow of funds between insurers, PBMs and drugmakers. Instead, the tax would extract some of the monopolistic surplus from PBMs that could then be used to enhance innovation.

Buyouts and Auctions

Are alternatives to the patent system’s time-limited market exclusivity conceivable? Yes, according to a proposal inspired by the now-ancient example of the patent for daguerreo-type photography. The French government purchased this patent in 1839 and placed the technique in the public domain. As a result, daguerreotype photography was rapidly adopted worldwide and benefited from many subsequent improvements.

If the government were to buy out the patent rights to innovations at their estimated private value and then make the innovations freely available, it would eliminate many distorted incentives in today’s system. By releasing the discoveries to the public without requiring a licensing fee, this system might also stimulate follow-on innovations built on the patented invention.

A central challenge for any system of patent buyouts is determining the right price. A possible solution to this problem is an auction. Potential investors would bid on the value of the new technology, and the winning bids would reveal investor willingness to pay.

In most cases the winner would not gain ownership of the patent to the technology – the government would just use the winning bid to determine compensation for the inventor. However, to create incentives for bidders to reveal their actual assessment of the innovation, a random sample of the proposed projects would actually be sold to the winning bidders.

A second major challenge is the risk that a government buyer might misuse its monopsonistic purchasing power and force the sale of patents at confiscatory prices that discourage future innovation. To prevent this, innovators should have the option of selling their patents to the government via auction or using the conventional patent process. In other words, patent buyouts would supplement rather than replace the existing patent system.

Incentives to Make Cost-Reducing Innovation Profitable

It’s hard to make money by reducing the cost of healthcare, and this fact shapes the direction of innovation. The essence of the problem is that providers and patients make decisions about care while the benefits of lower costs accrue primarily to the insurer/ payer. A simple solution follows logically from the problem: payers should share cost savings with providers and patients. However, this straightforward solution would require many other changes.

Jump-Start Shared Savings

As discussed earlier in the book, providers act as common agents for many payers. And common agency complicates the adoption of shared-savings arrangements.

The issue is “spillovers” – incentives provided by one payer also benefit other payers who contract with the provider. Spillovers can lead to outcomes where no shared-savings incentives are offered at all. We refer to this outcome as a “sticking point equilibrium” because all the players are trapped in an undesirable situation that is hard to exit without some external forcing mechanism.

However, sticking points are not economic black holes from which escape is impossible. With bold action, a single large payer can jump-start an escape. Suppose a large enough payer (a government program or a big private insurer) adopts aggressive shared-savings contracts. In that case, providers (like hospitals) may be willing to incur the high start-up costs involved in making their care delivery processes more cost-efficient. These improvements will then be available to all the provider’s patients, including those whose bills are the responsibility of other payers.

Having made the costly transition to new processes, providers are better positioned to respond to cost-savings contracts offered by these other payers, which supports the spread of more shared-savings contracts throughout the healthcare economy.

One can see an example in a recent study of the trial Comprehensive Care for Joint Replacement program. In the trial conducted on enrollees in traditional Medicare, the government insurer altered its customary fee-for- service payment model with an arrangement in which Medicare paid for a hip or knee replacement with a single bundled payment that covered all costs during an episode of care. Providers who keep the costs of the replacements below the bundled payment get to keep the savings.

The study was possible because Medicare’s CJR trial rolled out bundled payments in some (randomly selected) communities and not others. Comparing communities with and without CJR, bundled payments made care more cost-effective. Specifically, patients in traditional Medicare were around 10 percent less likely to be discharged from a hospital to expensive post-acute care facilities. Surprisingly, however, the study also found that introducing the CJR program had almost the same effect on hospital discharges for Medicare Advantage enrollees who were not included in the CJR trial.

Similar results were found with comparisons between hospitals rather than communities. Hospitals that were most responsive to CJR – those who made the most significant change in the discharge destination for their traditional Medicare patients – also changed the discharge destination for their Medicare Advantage patients.

Before the CJR, Medicare Advantage plans wanted hospitals to reduce discharges to post-acute care facilities but could not motivate them to alter their practices. However, when traditional Medicare – the nation’s largest purchaser of healthcare services – shifted to bundled payments, it suddenly became worthwhile to make the necessary changes. This pattern is just what you would expect to see when jump-starting out of a sticking point equilibrium.

Common agency leads to a new view of the federal government’s role as jump-starter of last resort in shaping the health sector. Medicare and Medicaid are such large purchasers of services that, with sufficiently bold initiatives, they can stimulate new cost-effective practices throughout the private sector.

Economic models of common agency suggest that jump-starting becomes progressively easier the more a provider’s patients are concentrated with a particular payer. Indeed, common agency problems disappear altogether if providers work with only a single payer – as is the case, for example, with Kaiser Permanente and the Department of Veterans Affairs.

Americans in Mexico for less expensive dental care.

Appreciate the Limitations of Financial Incentives

Even with a jump-start or a single payer, shared-savings contracts are only semi-effective ways to create incentives for cost reduction. Information problems, contract design difficulties and implementation obstacles limit the effectiveness of such contracts. The most challenging information problems concern quality of care and risk adjustment.

Consider, for example, the quality issues that emerged in a recent analysis of nursing homes. The nursing home sector has a higher proportion of for-profit providers than other parts of the healthcare system (about 70 percent). The biggest payers are Medicaid and Medicare, and each pays a prospectively set amount per day of care. This prospective payment system offers providers powerful shared-savings incentives – the facility gets to keep every dollar of savings below the per diem payments.

Since 2009, private equity firms have become increasingly active in purchasing nursing homes. Private equity firms buy existing companies by borrowing money for the acquisition and placing this debt on the purchased firm’s balance sheet. Combining powerful managerial incentives with high levels of new debt supercharge economic pressures to improve operational efficiency. And in some settings, the intensive pressures brought by private equity stimulate improved management practices.

In the case of nursing homes, however, private equity does not have such a salutary effect. One study pooled data from about 7.4 million Medicare patients in over 18,000 nursing homes between 2000 and 2017. PE ownership increased the probability of death during the stay (or the following 90 days) by 1.7 percentage points.

This effect amounts to a 10 percent increase over the baseline 17 percent mortality rate, and implies more than 20,000 additional lives lost over the 12 years of the study.

An obvious conclusion from these results is that private equity acquisition of nursing homes is not a good idea. The sheer magnitude of lives lost, however, raises a deeper question. How could such degradation of quality go unnoticed for so long?

Detecting the effect of private equity ownership required millions of observations and advanced econometric methods. Since big data and mathematical sophistication are not typical of the capabilities one finds in these facilities, managers are probably not cynically trading off deaths and dollars. A more plausible explanation for the degradation of quality is managerial inattention facilitated by limited information about the quality of care.

This explanation should not let private equity off the hook. More conscientious managers could have tracked the more common and less severe precursors to increased mortality. They might have noticed, for example, an increase in the use of antipsychotic drugs or the incidence of bedsores and used this information to ensure that quality was not degrading. But in an environment with high-powered cost-cutting incentives, such initiatives are not the priority. The case of private equity in the nursing home sector is a cautionary tale about the risks of powerful shared-savings incentives in healthcare.

Risk adjustment poses another informational challenge to shared-savings contracts. To illustrate the challenges of risk adjustment, compare compensation in traditional Medicare with Medicare Advantage plans. Physicians in traditional Medicare record diagnoses and the services provided to the patient. They receive a payment from Medicare based on services provided and must be meticulous about recording services with the correct procedure codes.

In Medicare Advantage Plans, the government pays private insurers a per-patient amount based on the prospective costs of the enrollee. Patients with costly chronic diseases tend to be expensive to insure, so prior diagnoses for chronic conditions are used to estimate an enrollee’s “risk score.” For example, insurers who enroll more people with diabetes get more per enrollee.

This approach amounts to a shared-savings contract. If the insurer can find ways to deliver care at less than the expected costs, they can keep the savings. But in addition to encouraging cost-effective treatments, the shared-savings approach also creates powerful incentives to ensure that physicians code diagnoses to maximize reimbursements.

Manipulating diagnosis codes to increase revenues is known as upcoding. A recent study found evidence that physicians in Medicare Advantage plans do upcode. Estimates indicate that upcoding increases the average risk score in Medicare Advantage enrollees by 6-8 percent relative to traditional Medicare – equivalent to 39 percent of enrollees suddenly acquiring a diagnosis of diabetes!

Risk adjustment can also amplify socially unacceptable health disparities. For example, the Medicaid program in Texas transitioned from a fee-for-service system to a managed care system in which insurers received a capitated [per patient] payment for each enrollee based on historical costs in the locality. And in the Medicaid program, the cost of treating Black newborns was more than 80 percent higher than for Hispanic newborns – a difference of more than $4,200 per baby.

Because the capitated payments were not “risk-adjusted” to reflect these differences, Medicaid Managed Care plans in Texas could benefit by discouraging Black mothers from enrolling and encouraging Hispanic mothers to enroll. Consistent with this incentive, the rollout of Medicaid managed care programs in Texas appeared to significantly improve the birth outcomes of Hispanic babies, but not Black babies. In California, where Medicaid Managed Care operated with less powerful shared-savings incentives, there is no evidence that it led to increased disparities between Black and Hispanic infants.

Comparing Texas and California Medicaid offers indirect evidence of unintended consequences from the design of shared-savings contracts. More direct evidence of the difficulties of contract design comes from a study of shared-savings contracts in long-term care hospitals. These were initially created to solve a potential side effect of the perspective payment system for acute care hospitals with unusually long lengths of stay.

Before 2002, LTCH were paid their estimated daily cost. However, that year, the long-term care hospitals were transitioned to prospective payments under which they were paid a lump sum based on the patient’s diagnosis. Designers of the PPS were concerned that these hospitals might take the lump sum payment and discharge the patient too soon. The program established short-stay thresholds to discourage premature discharge of patients. The hospitals received a fixed daily amount if stays were shorter than the relevant threshold. After reaching the threshold, payment shifted from a per-day basis to a per-stay basis in which the hospital received a large lump sum to cover the remaining hospitalization costs.

This lump sum averaged $13,500 per patient, paid one day after the patient crossed the threshold. Not surprisingly, this induced a significant spike in discharges on precisely the day of the lump-sum payment.

This example illustrates that shared-savings contracts can be gamed and that efforts to limit such gaming typically create other unintended distortions. Designing efficient shared-savings contracts, in other words, is hard.

The contract-design problem is greatly compounded when contracts have to share savings that accrue to payers in the future – but this is precisely the contracting that is most useful for encouraging cost-reducing innovations for managing chronic diseases. For these diseases, early actions generate savings that manifest in future years. But neither economists nor insurers nor regulators have workable ideas for managing shared-savings contracts geared toward such future costs.

An additional limitation of shared-savings contracts concerns implementation within organizations. Many shared-savings contracts target entire organizations. For example, Accountable Care Organizations aim to promote process innovations that reduce the cost of care and improve care coordination between providers within the ACO using organization-level shared-savings incentives. However, organizational incentives do not directly translate into incentives for those who work in those organizations. The “free riding” problem in healthcare organizations can dilute gains from shared savings so dramatically that the incentives will not pay for themselves.

High-powered financial incentives can spur innovation. However, such incentives can also lead to unintended and undesirable consequences. Meaningful non-financial motivators arising from organizational culture or professional norms must therefore accompany financial incentives.

Improve the Data Infrastructure

Understanding the impact of interventions on the cost and quality of care requires data from large numbers of patients. But in the heavily fragmented U.S. healthcare sector, few organizations operate at a scale sufficient to support such data collection.

One such entity that does is Medicare, and the widespread availability of Medicare claims data has allowed researchers to learn a great deal. However, traditional Medicare claims data are limited. And private insurance claims data, including from the Medicare Advantage program, are generally unavailable for analysis.

The data infrastructure of U.S. healthcare is an archipelago of data islands. Only a few organizations, such as Kaiser Permanente and the federal Department of Veterans Affairs, have built bridges across these islands by developing data sets that span insurance categories and geographic jurisdictions, integrating medical records and claims data. Investment in a robust national data and analytic infrastructure – including incentives to encourage greater exchange of information between electronic medical records – could facilitate cost-reducing innovation if combined with a suitable catalyst.

What might such a catalyst look like? The UK’s National Institute for Health and Care Excellence (NICE) is illustrative. NICE was established in 1999 to support decision-making on new technologies and was asked to consider both costs and benefits. Because it was established as part of Britain’s National Health Service, its mandate included a commitment to provide the best value for taxpayers’ money and the most effective, fair and sustainable use of finite resources.

Where possible, NICE calculates a treatment’s incremental cost-effectiveness ratio – often expressed as cost per QALY (quality-adjusted year of good health) gained. In the UK context, interventions costing less than £20,000 per QALY (about $25,000) are considered cost-effective. Where the evidence is uncertain, conflicting, insufficient or not robust, NICE liaises with the research community to investigate the issues.

NICE has significant limitations. For instance, it is concerned with building legitimacy with relevant stakeholders, which can cause it to proceed slowly. Nonetheless, it provides essential information for clinicians and policymakers seeking to adopt cost-reducing innovations.

But the NICE model is not directly applicable to the United States. Our closest approximation to the UK’s National Health Service, Medicare, is legislatively prohibited from developing or deploying a dollars-per- QALY measure to determine coverage or reimbursement. Removing these restrictions on Medicare would substantially boost prospects for cost reduction. A private nonprofit organization, the Institute for Clinical and Economic Review, has attempted to partially fill the vacuum left by the restrictions on Medicare. The ICER was founded in 2006 as an organization within Massachusetts General Hospital, but became an independent nonprofit in 2013. It conducts evidence-based reviews of healthcare interventions, such as drugs, devices and diagnostics.

The example of ICER illustrates the influence of culture and social norms on the trajectory of innovation. The relevant norm, in this case, is the distrust of centralized power that underlies so much of the U.S. political tradition. The same analyses we accept from a private sector organization (ICER) are forbidden when produced by a government agency (Medicare). But such research is a public good. Like all public goods – think education or environmental protection – it is undersupplied by private sector actors alone. Consequently, our social norms result in chronic underinvestment in the information essential for producing and deploying cost-reducing innovations.

Increase Market Competition

There are many ways that society benefits from competitive markets. Consider the effect of horizontal mergers on incentives to innovate. Imagine two hospitals offering services that compete for the same customers. Hospital A has a dominant market position, but Hospital B is developing a new cost-reducing process that could allow it to eventually win business that otherwise would have gone to Hospital A. This possibility is a powerful incentive to pursue the innovation – but only because B needn’t consider the costs its success will impose on A.

After a horizontal merger between A and B, however, the merged entity will consider the costs of A’s lost sales. Suppose A has a robust market position pre-merger. In that case, the combined entity may find it most profitable to slow down or stop the development of B’s new process to avoid losses to A’s existing business.

Innovation can also be suppressed by exclusionary contracts that restrict market entry. Consider the case of Johnson & Johnson’s biologic drug, Remicade. J&J offered a “loyalty rebate” program for Remicade. The rebate would only be paid if a hospital bought all related biologics for treating the same conditions from J&J.

Foreseeing this, a potential biosimilar manufacturer might choose not to develop a competitor. And the failure to enforce prohibitions against such agreements may partly explain why the development of biosimilars in the United States lags behind the EU.

A recent policy review suggests three ways to make healthcare markets more competitive:

1. Reduce or eliminate policies that encourage consolidation or that impede entry and competition.
2. Strengthen antitrust enforcement.
3. Create an agency responsible for monitoring and intervening in healthcare markets.

Discouraging consolidation may require reform of payment policies used by Medicare and commercial insurers that reimburse physicians more for a procedure performed in a hospital rather than an outpatient setting. Reform to state laws and regulations could also help. Consider, for example, state certificate of need (CON) laws. These laws require a state health planning agency to approve significant health investments such as construction of a new facility. CON laws aim to inhibit unnecessary or duplicative capital expenditures, but incumbent healthcare providers can also use them to put potential competitors at a disadvantage.

Any willing-provider laws, on the books in many states, require health insurers to include any provider in their network who wishes to participate and to pay them at in-network rates. These laws aim to protect consumer choice, but they may also undermine competition among providers for inclusion in a network, thereby reducing the incentive to innovate.

Strengthening antitrust enforcement requires improving the capacities of the Federal Trade Commission and the Antitrust Division of the Department of Justice. In addition to strictly scrutinizing mergers, antitrust regulators should focus on potential or future threats to competition. Absent such scrutiny, dominant firms may prevent future competition through exclusionary conduct.

Congress could strengthen the hand of regulators in this regard by including consideration of harms to future competition in antitrust enforcement mandates. Such changes could be made more impactful by creating a specialized court to hear cases brought under federal antitrust laws, thereby ensuring that judges with the relevant expertise hear these complex cases. In addition, executive orders and laws could require the government to consider the potential effects of new legislation and regulation on competition, and generally promote new firms’ entry into markets.

Mobilize Professional and Social Norms

Imagine a circumstance in which the financial incentives for cost reduction were strong, the data, analytic and reporting capabilities were state of the art, and anticompetitive behavior was kept to a minimum by vigorous monitoring and enforcement. Would the skew in innovation away from cost reduction correct itself?

We believe the answer is no – not without a change in the professional and social norms at play in the health sector.

Professional norms in healthcare are shaped heavily by the culture of the medical profession. This culture is absorbed early in medical training and is imbued with positive values about service, healing, discipline and accountability. However, the culture of medicine also has a dark side. It is factional, hierarchical, exquisitely sensitive to threats to professional prerogatives, autonomy and status, and slow to change. At its best, it spurs its adherents to heroic service on behalf of patients, as we have seen during the Covid-19 pandemic. At its worst, it leads physicians to burnout, cynicism and despair.

One bright spot is Kaiser Permanente’s well-developed model of prepaid group practice, and its long-term relationships with its members create potent incentives for adopting value-enhancing and cost-reducing innovations. In addition, the extensive patient data and the research capabilities of its Division of Research provide precisely the kind of information infrastructure needed to make evidence-based decisions about the cost and quality of care.

However, even in this environment, physicians resisted innovations such as the concentration of surgical volume in a handful of centers of excellence to keep quality up and costs low. The reporting of quality metrics and the development of treatment guidelines also engendered resistance and were characterized as denigrating the “art” of medicine and opening a doorway to second-rate “cook-book” medicine. Implementing these changes took skillful physician leadership and years of careful persuasion.

The example of Kaiser Permanente illustrates that even when innovation is well defined and reinforced with appropriate incentives, people still need clarity about what to do in the face of change. Such clarity generally takes the form of shared understandings and expectations that go far beyond the content of formal job descriptions or standard operating procedures.

This informal clarity is the secret sauce behind successful organizations and explains why propagating even well-defined innovations can be challenging. Toyota, for example, makes no effort to hide its Toyota Production System methodology, also known as Lean, from competitors because it believes they lack the informal clarity – the organizational culture – needed to make the system work. Professional norms can conceivably be the source of such clarity in healthcare organizations. Often, however, they are not.

Earlier in the book we argued that norms could stimulate cost-reducing innovation to the extent that innovations align with ethical commitments and professional obligations. But physician culture does not make clear that stewardship of scarce social resources is a moral duty for the medical profession. Without clear ethical obligations, innovations often encounter the dark side of physician culture, where concerns over professional turf, status and autonomy slow innovation.

Changing this reality must begin in medical training. Time for pursuing scholarly or scientific research is incorporated into medical training, but such time is rarely used to expose trainees to the tools and methods of designing and implementing innovation. Physicians are taught the ethics of caring for individual patients, but not the ethics of appropriate stewardship.

Bright spots are emerging to address these gaps. At Stanford University, for example, the Byers Center for Biodesign teaches interdisciplinary teams of physicians, engineers and scientists to use design methods to create new medical devices. Texas A&M University’s ENMED program trains “physicianeers” in a joint medicine and engineering degree program whose graduates seek to design efficient new medical technologies. Programs like these need to become the norm at medical schools and universities across the country.

The blind spots in physician culture that make cost-reducing innovation difficult mirror a more profound confusion about healthcare in our civic culture. Although healthcare is bought and sold similarly to other goods and services, it has a distinct ethical foundation. This foundation is illuminated by a simple question posed by the economist Uwe Reinhardt in the Journal of the American Medical Association in 1997: “Should a child of a poor American family have the same chance of receiving adequate prevention and treatment as the child of a rich family?”

In other words, is healthcare a social good that should be available to all on equal terms, or is it a private consumption good whose financing is an individual responsibility? If the former, physicians, hospital administrators, nurses and related professionals must adopt an ethical commitment to the wise use of the resources needed to meet these social commitments.