This comparison of the US health care system and 10 other countries’ offers at least 4 clear conclusions for those curious and persistent Americans who kept asking me: Which country has the best health care?
First—and most clearly—the US health care system is not the best, 2nd best, or even 3rd best in the world. It significantly underperforms on numerous dimensions.
Once, maybe in the 1950s and 1960s, the United States had the best health care in the world. At some facilities and for some conditions, the United States currently offers fantastic care. Cancer care may be such a condition. But overall today the United States is way behind many other countries. Only 90% of the US population has coverage, while almost every other country manages to get over 99% coverage. On cost, the United States spends 27% more per person than the next most expensive country and 50% more than most countries.
For all that money, the US health care system performs poorly on many dimensions, including ones that Americans prize highly, such as patient choice of hospitals and physicians, with Canada, France, Germany, the Netherlands, Switzerland, Taiwan, and other countries exceling and outperforming the United States on this very dimension. And there is ample data that quality in the United States is—to put it charitably—inconsistent; the United States ranks poorly in many areas amenable to health care interventions, such as infant and maternal mortality. On many dimensions, US health care is not in the top 10 globally.
Second, although not the best, the United States also does not have the worst health care system in the world. As I said, the Chinese system seems to be both poorer performing for today and less innovative for the future.
The 3rd conclusion is that the United States does excel on some dimensions, particularly innovation and experimentation in payment models and care delivery (especially for chronic conditions). These innovations are a work in progress, but the United States—along with the Netherlands and perhaps the UK—is far more invested in these experiments and demonstration projects. Ironically, the system’s underperformance spurred this drive for innovation. For instance, because US hospitals are so expensive and have high rates of infections and complications, there are strenuous efforts to avoid hospitalizations and deliver what hitherto was hospital-level care in the outpatient setting and patients’ homes. Similarly, because the fee-for-service payment model incentivizes overuse of high-margin tests, treatments, and unnecessary care, the United States leads in the development of alternative payment models. This innovation is one reason I am optimistic about the US system’s long-term performance. In the next decade or 2, the United States will again become one of the best systems in the world.
Considering the US system’s underperformance and innovativeness, it is appropriate to ask what this study of other countries can teach us about what changes would improve the US system. In other words: Where should we focus that innovative energy?
As one expert wrote, “It is not desirable [or even possible] to ‘lift and shift’ health system parts from one country to another.” Nevertheless, what my audiences have asked for is possible. We can learn specific lessons from other countries that can be implemented in the United States to significantly improve the system.
This evaluation of 10 other systems reveals 6 feasible and beneficial changes to the American system: (1) ensuring universal coverage, (2) enhancing financial protection for children, (3) mandating simplification across all forms of health insurance, (4) emphasizing primary care, (5) accelerating the innovation and dissemination of best practices in chronic care coordination and mental health, and (6) regulating drug prices.
Fully 10 years after the Affordable Care Act the United States has seen great improvement in coverage, but still we can barely get to 90% coverage. Even worse, the last few years of Republican attacks on the ACA’s coverage mechanisms—such as repealing the mandate, reducing funding for advertising and navigators, and weakening the insurance exchanges—have caused the uninsured rate to rise—the opposite of what we expect with falling unemployment and a growing economy. A high uninsured rate is not a law of nature, and the United States can achieve near-universal coverage without massive, systemic change.
Most people who are uninsured today are eligible for highly subsidized, relatively affordable health insurance through Medicaid or the exchanges. Over 6 million uninsured Americans are eligible for Medicaid. Even states that have expanded the program have hundreds of thousands of residents who are eligible but have not signed up. Similarly, the health insurance exchanges are designed to help people with incomes above 100% of the poverty level but no affordable employer-based insurance—think independent contractors and Uber drivers in the new gig economy. The exchange plans cover all preventive services and annual checkups without a deductible or co-pay. With the subsidies, plans for some people have almost no premium and often cost low-income individuals less than US$100 per month. But many independent contractors or people working in small businesses that do not offer health insurance are unaware of the subsidies and/or do not have the time to shop during a limited enrollment window. As a result, millions of low-income Americans are uninsured despite the existence of programs to cover them.
What can we learn from other countries about getting to true universal coverage? Automatic enrollment with substantial subsidies has enormous power.
In Norway and the UK everyone has health coverage by virtue of being a resident. That’s all it takes. Furthermore, Norway has a simple form of auto-enrollment: all residents pay taxes to fund the National Insurance Scheme, which insures all residents. Any resident can show up, get care, and have it paid for from the system. Amazingly, the United States has a nearly identical system in Medicare Part A, which pays for hospital care for seniors.
Although auto-enrollment is a bit more complex in systems based on private health insurance, it is still achievable. In the Netherlands and Switzerland, the government automatically enrolls people who do not obtain health insurance themselves. In Switzerland every person needs to register with a canton (state) as a resident and then has 90 days to buy health insurance. After 90 days the government can randomly select a plan for new residents and require them to pay both a premium and a penalty. The Netherlands operates a similar system at the national level, and the penalty is stiff. Importantly, both the Netherlands and Switzerland provide generous subsidies to between 33% and 60% of the population to purchase the mandatory private insurance. (On subsidies, the Netherlands is much more generous than Switzerland.) Hence, despite having coverage through competing private insurance plans, the Netherlands and Switzerland have over 99% coverage through a combination of strong financial penalties, substantial subsidies for low- and middle-income families, and auto-enrollment if individuals do not initiate enrollment themselves.
Behavioral economics suggests that automatic enrollment would dramatically reduce the uninsured rate in the United States. When programs automatically enroll people, who can also opt out if they wish, participation goes way, way up. For instance, in countries that have automatic enrollment in organ donation (organ donation is assumed at the time of death unless a person has explicitly refused for religious or other reasons) more than 90% of the population agrees—or at least does not object—to organ donation. Conversely, in opt-in countries, where people need to explicitly agree to organ donation when they get their driver’s license, such as the United States, only about 15% of the population volunteers as donors.
The complex patchwork of insurance programs in the United States makes it harder to design an auto-enrollment program. Americans are eligible for different programs based on employment status, age, income, and geography. Often they need to select an insurance company and pay a premium—more burdensome choices. However, even with these complications, it is possible to have a modified version of auto-enrollment. It is not even partisan or fantastical—analysts from both the left-leaning Brookings Institution and right-leaning American Enterprise Institute have written extensively about how to leverage automatic enrollment. This could slash the uninsured rate to under 5%. Indeed, there are 6 states and the District of Columbia that have achieved over 95% coverage under the ACA, showing that small changes can significantly improve coverage.
One approach is to auto-enroll people without insurance into Medicaid. If people show up at a physician’s office, urgent care center, emergency room, or hospital without any insurance, the provider would be legally required to bill Medicaid. Medicaid would keep that person enrolled unless it determines the person is eligible for some other coverage—say, in the exchanges or through Medicare. It would then enroll them in that alternative coverage. Another possibility is to auto-enroll uninsured people into the lowest- or 2nd-lowest-cost silver plan on the insurance exchanges and add an assessment for the premium—after subsidies—to their taxes at the end of the year. Switzerland does this.
These auto-enrollment mechanisms would require changes in laws and regulations. For instance, the income eligibility determination mechanisms in Medicaid and the insurance exchanges are not aligned. For most individuals Medicaid uses a person’s previous month’s pay, while exchange subsidies are based on the previous year’s income. Using Medicaid’s income determination process in the exchanges—and eliminating the claw-back of subsidies should a person’s income be higher than expected—would greatly facilitate auto-enrollment. These legal and regulatory changes to facilitate auto-enrollment are hardly brain surgery and should not be the excuses to do nothing.
In addition to auto-enrollment, 2 key issues seem to be the premium charged and, relatedly, the level of subsidies in the insurance exchanges. Two easy ways for the United States to drive down premiums in the exchanges are to make reinsurance permanent and to pay for the cost-sharing subsidies. Reinsurance reduces the risk of super-expensive patients for insurance companies, and this, in turn, reduces premiums. Another way to reduce premiums would be to reinstate the federal government’s payment for the cost-sharing subsidies. Estimates are that this would reduce premiums about 10% to 15%. Research has shown that low-income individuals are extremely price sensitive when it comes to buying health insurance, and reducing premium sticker prices dramatically increases enrollment.
Similarly, the United States could increase subsidies for the purchase of health insurance in the exchanges. While the United States is generous up to about US$62,000 for a family of 4—the median household income—with substantial help for families buying insurance and with deductibles and co-pays, it becomes much less generous for slightly-better-off families. There has been much discussion about increasing subsidies for insurance premiums and out-of-pocket expenses and extending them to families with incomes over US$100,000. The lessons—particularly from the Netherlands—suggest this: higher subsidies than those currently in the ACA would help induce people who qualify for the exchanges to buy private insurance.
Other countries’ experiences suggest that auto-enrollment with reinsurance and more generous subsidies would be effective and important policy mechanisms to get to universal coverage—more important than public versus private financing, the role of insurance companies, a public option, and so on. The United States needs to adopt some form of auto-enrollment and improve the generosity of subsidies for middle-income families.
“Our children are our most valuable asset.”
How often do we hear politicians, business executives, and civic leaders utter this phrase? When it comes to health care in the United States, however, this seems mostly to be an empty phrase, not a living reality. This comparative study of health care reveals that in most countries children are viewed as both a treasure and an investment in the future. Consequently, society collectively pays for children’s health insurance and provides them services with no deductible or co-pay. But in the United States, the financial burden for children’s health care falls on each individual family through higher premiums, deductibles, and co-pays.
The United States has made important progress in covering children at low or no cost to their parents. In the 1990s, a bipartisan coalition led by Senator Ted Kennedy (D-MA) and Senator Orrin Hatch (R-UT) enacted the Children’s Health Insurance Program (CHIP), which covered children whose parents earned too much money to be eligible for Medicaid but who did not have private health insurance. It now provides health coverage to about 9 million children, and the uninsured rate for children has fallen about 65% since CHIP’s enactment. Not only has the United States expanded coverage for children, we also have made the benefits more comprehensive. In 2010, the ACA mandated that all insurance companies include pediatric dental and vision care as one of the 10 essential health benefits.
Although these are important improvements, they remain inadequate. About 4 million American children—or 5%—are uninsured. Outrageously, almost a third of these uninsured children live in just 2 states: Texas and Florida. In Texas alone 11% of children lack health insurance. Nearly 60% of all uninsured children are eligible for Medicaid and/or CHIP but are not enrolled, and these states create barriers to enrolling children.
Another problem is that even insured children can be a significant financial burden on their parents. On average, family premiums for employer-sponsored insurance are nearly 3 times those for individuals—US$7,188 for individuals versus US$20,576 for families in 2019. Indeed, an average worker’s contribution to an individual, employer-sponsored health insurance plan was US$1,242 in 2019 but US$6,015 for family coverage—nearly 5 times more. After paying higher premiums, families must then pay deductibles and co-pays for their children as well.
My comparative study reveals that almost all other countries liberate families from the financial burden of paying for their children’s health care. Some countries do this by directly subsidizing premiums for children’s health insurance and eliminating deductibles and co-pays for pediatric services. In the Netherlands, all children under 18 need to have health insurance, but families do not pay any additional premiums to cover children with the standard package. In Germany, all workers at a company earning similar salaries pay the same in payroll taxes for their health insurance regardless of whether they have a family or not. And workers’ family members—nonearning spouses and children—get health coverage without any additional payments from the employer or employee.
Countries without typical insurance premiums also strive to make children free to parents. In Australia, the tax (levy) for the mandatory government Medicare program does not change based on family size. In addition, Australian children are typically added to private health insurance plans at no additional cost to the family. In Norway, there is no co-pay when children under 16 go to the primary care physician to receive health services, and pediatric dental care is free. Rather than loading the premiums (or taxes) and out-of-pocket costs for a child’s health coverage on individual families, these countries see children as an investment and spread the cost across all of society.
Such policies strongly communicate family values—the valuing of families and their children. Having society pay the full cost of children’s health care also helps to fulfill the ideal of equality of opportunity by ensuring that all children will get access to the same set of essential health services regardless of their parents’ income and social status. It also serves a coldly self-interested purpose: How much future economic productivity is lost by denying children access to needed care today?
The United States—all Americans, together—should assume the financial liability for children’s health insurance. There are several policy changes that would accomplish this.
First, the United States could require employers to charge the same health insurance premiums to workers, whether they are individuals or have dependent children. This would distribute the cost of family health coverage among all workers—not just adults with families. Second, the United States could emulate almost all other countries and require private insurers to eliminate all deductibles and co-pays when children under 18 receive medical services, including visits to the pediatrician, the urgent care clinic, the dentist, or the hospital and for prescription drugs.
When it comes to public health coverage, another policy would be to ensure that 100% of children receive health coverage by enacting auto-enrollment into Medicaid and CHIP without requiring any verification of family income or other requirements. There should be no co-payments for pediatric care in Medicaid and CHIP.
Finally, because of low reimbursement rates, many physicians refuse to take Medicaid or CHIP patients or limit the number of patients with these types of coverage they are willing to treat. For 2 years, in 2013 and 2014, the ACA increased payment to physicians caring for Medicaid patients. The result was a significant improvement in both the number of physicians taking Medicaid patients and the promptness of the services the patients received. And today 19 states have voluntarily continued the policy of higher Medicaid reimbursement for physicians. To ensure that all children with Medicaid or CHIP coverage have access to pediatricians and specialists and receive timely care, the US government could permanently increase Medicaid’s reimbursement rates for pediatric services.
Enacting these policies would ensure 100% of American children receive health coverage and that this coverage is not a substantial financial burden to their families.
Every health care system in every high-income country is complex. That does appear to be a law of nature. But this study of 10 other health systems reveals that the American system is unique in that it is significantly more complex than the system of any other country. It needs simplification—urgently.
Complexity is burdensome. Patients spend too much time trying to figure out which type of coverage they are eligible for, which specific plan to sign up for, which physicians and hospitals are in network, how much they owe in deductibles and co-pays, how to dispute inaccurate bills, and so on. There is woefully little time left for patients to ensure they are getting optimal, evidence-based care. Purchasing insurance is so complex that employers hire consultants to design their employees’ health benefits. The multitude of insurance products—usually dictated by employers—and complex insurance billing processes mean that US physician practices spend 4 times more on administration compared to their Canadian counterparts.
Complexity is also financially costly, and these costs are ultimately borne by all Americans. In 2012, the National Academy of Medicine (NAM) estimated that the United States has disproportionately high administrative costs. In 2019, the Center for American Progress updated NAM’s estimate, showing that the United States spends nearly US$500 billion on billing and insurance-related costs. Nearly half of that, over US$240 billion, is classified as “excess”—read, waste—that could be eliminated with a simpler system.
American policymakers, insurers, and others did not intend to design a mind-bogglingly complex system. It is path dependence at work. Nobody considers simplicity an imperative in developing new laws, regulations, or policies. Rather than trying to simplify existing systems, lawmakers layer programs onto each other to address problems and deficiencies—and to satisfy hospitals, insurance companies, drug and device companies, and scores of other interest groups. Most importantly, they add various provisions to secure enough votes for legislation to be passed—with little attention to how the added provision will affect administration. Here is a simple example I recently witnessed. A well-intentioned congresswoman wanted to reduce patients’ burgeoning out-of-pocket costs, so she suggested eliminating deductibles and co-pays for 5 primary care visits and 3 specialist visits for established chronic conditions each year. Just imagine the complexity of keeping track of the number of visits as well as identifying which visits are to primary care doctors versus specialists treating a specific, qualifying condition.
A key lesson from studying other countries is that there is much that can be done to simplify the US health care system in 3 important areas: (1) the differing ways Americans get insurance, (2) the complexity of insurance benefits design, and (3) the relationships among insurers, hospitals, and physicians.
First, the United States needs to simplify the process of buying health insurance. The United States basically has every type of health financing ever invented—socialized medicine (the VA), single payer (traditional Medicare and Medicaid), single payer managed by private insurance (Medicare Advantage and managed Medicaid), private employer-based insurance, and individual purchase (see Table 12-2). This is preposterous.
HERE IS HOW we could simplify getting coverage: reduce the insurance options to either employer-sponsored insurance or Medicare. Within Medicare offer a choice between traditional fee-for-service or a managed care plan. This would require combining Medicare, Medicaid, and the insurance exchanges into one program that allows all people who do not receive employer-sponsored insurance to choose between traditional Medicare or a Medicare Advantage plan operated by a private insurance company. This is essentially the plan proposed by the Center for American Progress and introduced into Congress by Representatives Rosa DeLauro (D-CT) and Jan Schakowsky (D-IL).
Obviously, many details would need to be worked out—how much states would contribute when Medicaid is merged with Medicare, what businesses should contribute if they do not provide health insurance and want their employees to be in Medicare, and how much to charge individuals under 65 who enroll in Medicare. But at the core this change would make it much easier on Americans to navigate health insurance options and would simplify billing by physicians and hospitals. It may be a substantial change given the current political realities, but it is a meaningful step in the right direction that other countries have already taken.
Another simplification is to reduce the myriad health insurance benefit designs. Most countries have significant limitations on the range of services covered, co-pays, balance billing, out-of-pocket expenses, network formation, which services supplemental insurance can offer, and so on. By law, the 22 Dutch insurance companies must cover the same basic benefits and have the same minimum deductible with the same exemptions for primary care physician visits. The ability to accept higher deductibles in exchange for lower premiums is the same across companies. The only thing that really varies is the level of premiums, supplemental coverage, and the physician and hospital networks. Because important factors cannot vary between insurance companies, patient choices are focused on just a few key factors, and this also helps providers know for certain which services are covered.
Such simplification was successfully achieved in the United States within the Medigap program, in which seniors can buy private, supplemental insurance to pay for services and out-of-pocket expenses not covered by traditional Medicare. Through legislation, Congress reduced the number of benefit designs to 10 paradigmatic types. Each type had standardized benefits, limited the duration of exclusions for preexisting conditions, and required minimum medical-loss ratios. Thus, seniors’ choices were reduced to selecting among the 10 insurance designs.
The government could do the same for employer-sponsored insurance and drug benefits. Today, it is largely employers and their benefit design consultants that require variations in the design of their insurance benefit. There is woefully little evidence that these benefit designs help with improving quality, reducing costs, or improving employee health. But we are sure they are a main reason administrative costs are so high. Needing to determine what patients are covered for, their deductibles, the network, and all the other variables is administratively costly. The federal government already took a step in the direction of simplification with the passage of the ACA, which delineates 10 essential benefits that all plans must cover. The government should further reduce plan variability and define up to 10 prototypical plans with standardized benefits, co-pay and deductible levels, drug benefits, and other consumer-facing parameters. This would reduce employers’ and employees’ costs and would also allow patients to more easily compare insurance options.
Finally, simplification needs to occur at doctors’ offices, hospitals, and other facilities. There are endless paper registration forms filled out by hand, recording and rerecording addresses, past medical problems, allergies, medications, and emergency contacts. In the digital era, this is completely wasteful, and health care providers are only slowly switching to digital registration voluntarily.
In contrast, the Taiwanese system has developed a useful digital medical card that makes registration analogous to the swipe of a credit card. We should not reinvent the wheel. The federal government should mandate that within 3 years all physicians, urgent care centers, hospitals, and other providers have digital registration either by electronic card or a biomarker (fingerprint or eye scan). This would link to verifying insurance eligibility and benefits as well as past registration that could be corrected. It would also allow patients to have a reasonable upfront estimate of what their physician visits cost.
Another level of complexity relates to physician and hospital billing. Reducing the different types of health offerings to 10 should reduce the complexity of billing, but more importantly, there should be administrative simplification of bills. The United States needs a single clearinghouse for medical bills, with a standardized form that prevents each insurer from using its own mechanisms to request additional information.
Finally, prior authorization requirements, despised by doctors but required by insurers for many high-cost tests and treatments, need reform. They take time, and they are costly. Although physicians and patients hate prior authorization, the tools are effective because physicians frequently do not practice evidence-based medicine. Physicians often order unnecessary—even harmful—tests, treatments, and excessively priced medications. The only way to get rid of prior authorization is for physicians to adhere to proven care pathways.
Until that day comes, prior authorization can be simplified in 2 ways. First, it can be put in the flow of physician ordering. This will take technical improvements and upgrades in EHRs, but it is coming. Insurers could also more widely institute “gold carding,” which would allow physicians who consistently (over 95% of the time) adhere to treatment guidelines to be exempt from prior authorization. Second, rapidly shifting to alternative payment models in which physicians are held accountable for the total cost of care should, over time, reduce unnecessary and inefficient tests and treatments because such models align the physicians’ interests with adhering to evidence-based care.
Simplification has a cost, though.
It applies a handful of rules to everyone, treating all people the same. It reduces exceptions and taking account of unique circumstances. But having exceptions for each person, group, or employer—which is the current system—creates substantial and costly complexity. The US health care system has gone way too far in the direction of complexity, which imposes substantial burdens on all of us. We need to bring the pendulum back toward simplicity. And to this end, every organization and person in the health care system should evaluate what they are doing by the simplicity standard: How can I make this simpler and less time intensive?
The United States has too many adult-focused specialists, who are paid more to perform procedures and provide consultations than primary care physicians and pediatric specialists. The United States needs to prioritize primary care.
Medicare pays a primary care physician about US$160 for a 45-minute discussion of advance care planning, but cardiologists receive about US$620 for single-vessel stent placement. Does placing a stent in a heart blood vessel really require nearly 4 times more skill and brain power than a compassionate 45-minute end-of-life care discussion with a patient and their family? If so, why do most cardiologists—and other American physicians—do almost everything they can to avoid such discussions? Furthermore, specialists in the United States expend significantly more resources than primary care physicians when they manage patients with similar conditions. Like many countries, the United States has too few pediatric specialists and pays them too little. But unlike many countries, the United States systematically undervalues primary care.
My comparative study reveals that a strong primary care system enhances performance. Countries like the Netherlands, Norway, and the UK have strong primary care models. Primary care physicians manage most patients. Specialists work as consultants to the primary care physicians and only manage the most complex patients. Conversely, in Germany patients have wide choice and do not need to get a primary care referral in order to access specialists or surgeons. France has a system that is in between: primary care referrals are often necessary to access a specialist, but patients can access any specialist they want with a referral, and the specialist is not a consultant to the primary care physician.
The Dutch system has significant advantages, and it is worth shifting the US system more in that direction, even if we are unlikely to go as far as the Dutch in creating a strong gatekeeper model. First, Dutch GPs tend to assume responsibility for many more services than US primary care physicians. For example, almost all of them employ nurse practitioners who are allowed to prescribe medications. They often employ mental health nurses for behavioral health conditions, such as comorbid depression and anxiety. Primary care cooperatives provide extensive after-hours telephone triage and even in-person after-hours care so patients do not go to the emergency room in the middle of the night. In addition, patients are financially penalized for going directly to specialists, and specialists have additional administrative burdens if they see a patient without a referral. This also ensures coordination of care through the Dutch primary care physicians.
Changing Americans’ behavior regarding primary care and specialist care is not going to be easy. But one thing can be done: change out-of-pocket payments for patients. The United States could make primary care visits free to patients—no deductibles and no co-pays. This could be linked to increasing the co-pay to see specialists without a primary care physician referral.
For this to work effectively, insurers would need to change payments to primary care physicians to incentivize them to assume more responsibility for managing and coordinating patient care. This could be achieved by shifting primary care payments to capitation with financial responsibility for a patient’s total cost of care. Primary care physicians could thus realize substantial financial benefits by managing more of their patients’ care, ensuring their patients are going to physicians who follow guidelines, and avoiding unnecessary or inefficient tests and treatments.
More than 80 cents of every health care dollar in the United States goes toward patients with chronic illnesses: diabetes, asthma, hypertension, congestive heart failure, emphysema, cancer, and other conditions. To lower health care costs and improve quality, care improvement needs to focus on these patients with chronic conditions, especially patients who are at high risk of exacerbation of their illnesses.
No country has universally implemented innovative approaches to chronic care management. But I did find many bright spots and excellent models of chronic care coordination, especially in the Netherlands and in the United States. One of the lessons I learned from this study is that physician groups that perform well in managing patients with chronic illnesses all seem to adopt a similar model of care. The key ingredients are:
• Creating multidisciplinary teams around a care coordinator embedded in the physician practices to manage high-risk patients
• Identifying high-risk, high-cost patients—patients who are likely to have exacerbations—who need added attention and services
• Educating patients and their families about how their condition is treated and teaching patients self-management techniques
• Instituting frequent, proactive outreach to patients not only to ensure compliance with testing, medication adherence, and lifestyle changes but also to address stressors that might lead to exacerbations
• Implementing same-day appointments and home visits for rapid responses to preempt and address potential disease exacerbations
• Using data on patient outcomes and practice performance to iterate and improve processes of care
In the Netherlands and the United States these models are facilitated by alternative payment methods such as capitation or bundling—so that physician practices earn more money when they do a good job of coordination that results in patients’ reduced use of hospitals.
These practices work to both improve the quality of care for patients and reduce costs. For instance, the Dutch Diabeter model has a comparatively high percentage of pediatric diabetics with hemoglobin A1c under 7.5% and a very low rate of hospitalization. Many physician practices in the United States have similarly achieved significant savings with embedded, proactive chronic care coordinators as part of multidisciplinary teams. For instance, in 2018 CareMore—a Medicare Advantage plan primarily focused in the Western United States and targeting the frail elderly—had 20% fewer hospital admissions and 23% fewer bed days (a measure of admissions and length of stay) than traditional Medicare patients. Compared to traditional Medicare, it also had a lower hospital readmission rate.
Obviously Diabeter and CareMore programs could be further improved. But they are well above average. And such excellent care for patients with chronic illnesses is less a matter of invention than one of dissemination. The challenge is to scale proven models throughout the health care system. This requires changing physician payment, helping physician practices consistently implement these 6 steps in the way they care for patients, and providing physicians frequent performance assessments so they know whether and how much they are improving. In the United States insurance companies, Medicare, and Medicaid need to continue to expand alternative payment models. This scaling will require the use of performance data, learning collaboratives, and technical assistance to enable physician practices to adopt the right approach.
We have a similar need to address mental health issues. There is a high burden of mental health conditions, which raise costs substantially. About 5% of the general public have depression, and 7% have anxiety. Among cancer patients, 20% have co-occurring depression and 10% have anxiety. Among patients with coronary artery disease, 15% live with depression and 23% have anxiety.
Mental health problems translate into high costs that are not for mental health services; instead, they are for “regular” medical services because of the challenges of managing these diseases that are then complicated by mental health issues. It is hard to get depressed diabetes patients to manage their diet and take their insulin appropriately, and this drives up the costs associated with diabetes. One report showed that symptomatic depression increased costs of caring for adults with diabetes by over US$5,000 compared to diabetic patients with no depression. The same pattern is found for patients with cancer, asthma, and congestive heart failure.
Two of the biggest deficiencies in health care systems have been the lack of diagnosis of mental health conditions and the slowness of providing care once a behavioral health problem is identified. No country has systematic mental health screening of either hospitalized patients or patients with chronic illnesses. In addition, many patients with mental health issues often have difficulty getting an appointment with a psychiatrist or other mental health provider. In the United States it is not unusual for patients to wait 2 or 3 months for an initial appointment with a psychiatrist.
Although no country I looked at excelled system-wide at mental health care, both the Netherlands and the United States had some impressive models. Dutch primary care physicians are responsible for treating mild mental health conditions, and 88% employ mental health nurses to treat patients. In the United States I found a number of groups—from small practices to larger providers, such as Advocate Health Care—doing a good job. Advocate screens all hospitalized seniors for comorbid mental health problems and intervenes within 48 hours. And at some of their larger primary care practices they have embedded psychiatric social workers to care for patients with mental health conditions. These interventions have improved care and lowered costs.
Much of the success of effective programs seems to inhere in consistently implementing a few critical steps:
• Embedding mental health providers—nurses specialized in mental health care, psychologists, or psychiatric social workers—in physician practices or linking to online providers
• Routinely screening patients for depression, anxiety disorders, and substance abuse and having them be rapidly—within 24 to 48 hours—seen by the in-house or online mental health providers
• Using targeted pharmacologic and behavioral interventions with the aim of addressing the mental health problem and returning patients to functionality
• Routinely assessing patients for improvement and giving this feedback to mental health providers
These mental health interventions are not yet as firmly tested and established as the ones for the care of high-risk, high-cost chronically ill patients. They need additional evaluation and refinement. But this process needs to be accelerated. Many small changes could facilitate this process, such as having EHRs seamlessly able to record patients’ mental health screening tests and mandating reporting on mental health outcomes for practices. Payers could also assess whether incentives for physician groups that embed mental health providers in the practice lead to effectively intervening to treat mental health problems and reducing medical costs. Once successful programs have been identified, they need to be scaled.
When it comes to both drug prices and drug price regulations, the United States is an extreme outlier. With less than 4.5% of the world’s population, the United States accounts for over 40% of all drug expenditures. It is the only country that grants drug companies monopolies through patents and marketing exclusivity and then allows them to freely set prices. All of the countries I studied explicitly regulate drug prices. Consequently, there are stark differences in drug prices that are not attributable to the number of drugs Americans consume. This study revealed that it is possible to regulate drug prices without stifling innovation or reducing access for patients.
Studying other countries gives insight into an effective approach to instituting drug price regulation. First, all other countries have one price for drugs for the entire population. They may adjust co-payments, but there is one price. Thus, in the United States drug prices should apply nationally, not just to one market segment, such as Medicare or the employer-based market. Second, drug price regulation needs to be informed by independent standards—not just what the drug companies want to charge. Some countries, such as Australia, Norway, and the UK, use a rigorous cost-effectiveness analysis to establish drug prices. Other countries, such as the Netherlands, Canada, France, and Taiwan, largely rely on external reference pricing. These 2 methods place independent benchmarks for acceptable prices. Benchmarks could be used in one of 2 ways. In one approach, the highest price—whether the cost-effectiveness price or the price in other countries—could establish a maximal price for drugs, allowing companies to set their own price below that maximum. Alternatively, these pricing methods could inform negotiations with the drug companies on the final price.
Another important lesson is that the government need not conduct these negotiations. In Germany the government empowers a nongovernmental entity to negotiate the prices with drug companies under legal constraints, with arbitration if they cannot come to an agreement.
This review of practices in other countries suggests that drug price regulation did not lead to significant delays in access to newly approved drugs. The differences in time to market seemed to come more from when drug companies submit applications for drug marketing to countries and the regulatory bodies’ time for safety and efficacy assessments. In almost all countries there was a strict timeline to establish drug prices that ranged from a few months to one year. Although drug price regulation might have added a few months to the time to market, it often was because of company resistance to lower prices, and no delay was excessive. What’s more, the savings were substantial—the equivalent of US$100 billion in drug spending.
Regardless, one key improvement the United States could make to lower costs for patients and the entire system is to actively regulate drug prices.
HOW CHALLENGING IS IT to implement these 6 reforms? None of this is going to be easy. Many clearly require federal legislation. For instance, changing the way income is determined for Medicaid and the insurance exchanges, requiring that employers assess families the same insurance contributions as individuals, constraining employers to offer only one of 10 different insurance products, and regulating drug prices all require federal legislation. Given the current partisanship at the federal level, legislating on health care is hard. We cannot seem to get a law passed on surprise medical billing or drug prices, despite large public majorities supporting laws on these issues. I do think some of these proposals can be enacted, especially around shifting more toward auto-enrollment and drug price regulation. But it will take years.
Some of these suggestions could be done, at least in part, by regulation or corporate initiatives and some by greater coordination between Medicare and private insurers. They should be easier to implement. Although administrative simplification was included in the ACA, it has floundered. It could be achieved if the federal government more vigorously pursued it in collaboration with private insurance. Creating a clearinghouse for claims and a standardized claims form used by all payers would be a step in the right direction.
The federal government needs to regularly reevaluate its fee schedule—how much it pays for different physician activities. For a long time it has been urged to increase the value of primary care and lower the value paid for surgical and other procedures. Recent evidence suggests that, at least for some procedures, such as hip and knee replacements, the government overpays. Paying based on the time spent, using empirical data from EHRs, is one way to rebalance the amount spent on surgical procedures. Reevaluating the difficulty of cognitive activities versus procedures is another thing that can be initiated without legislation. But more importantly, the government could work with private insurance companies to more aggressively—and, even more importantly, rapidly—shift toward alternative payment models. For the past 5 to 10 years there have been multiple demonstration projects testing these payment mechanisms. In general, they have modestly improved quality and lowered cost. None has made quality worse or systematically increased costs. There needs to be a coordinated effort between the government and private payers to expand and make permanent these models that would convince health systems and physician practices to invest more in changing their processes of care. And at the top of the list of those changes would be giving more focus to high-risk, high-cost patients who have either chronic conditions and/or mental health issues. These changes are occurring, but they could be significantly accelerated without the need for legislation.