Let’s hope that Orwell’s memory hole remains in good repair. As 1984 fans will recall, that appliance incinerated reminders of things more conveniently forgotten.

Today we need it to cleanse memories of managed care’s profit-driven abuses, so we can proceed, unimpeded by history, with accountable care organizations (ACOs) and bundled payment—the linchpin of reforms recommended by the Society of General Internal Medicine (SGIM)’s National Commission on Physician Payment Reform (and endorsed, in this issue of JGIM, by Drs. Ho and Sandy1).

We support The Commission’s calls for rebalancing compensation for cognitive vs. procedure-related work, and reforming the Relative Value Scale Update Committee (RUC) and Medicare’s sustainable growth rate (SGR). But its main prescription echoes the 1971 “Health Maintenance Strategy” proposal that ushered in the managed care fiasco.

Back then, Ellwood and colleagues proffered health maintenance organizations (HMOs) as the market-based alternative to national health insurance.2 They argued that fee-for-service “works against the consumer’s interest … the greater the number of contacts and days used, the greater the reward to the provider.” Their solution: HMOs paid “a fixed annual fee… The economic incentives of both the provider and the consumer are aligned…[with] A performance reporting system of proven reliability…[providing] accurate information on the comparative performance of alternate sources of health care” and “surveillance of the characteristics of populations served and services provided” to guard against cherry-picking and care denials.

That same year president Nixon made HMOs the centerpiece of his healthcare agenda, because (as captured on tape) “this [HMO strategy] is a private enterprise one… the incentives are toward less medical care, because the less care they give, the more money they make.” Employers and insurers soon followed Nixon’s lead, and by the mid-1980s, many providers rushed to create their own HMOs.

But egregious abuses followed. Headlines blared; patients sued over vital services denied; and HMO whistle-blowers told horrifying tales of office celebrations triggered when reviewers discovered loopholes allowing the denial of transplants.

Physicians were pressured to withhold care, and to hide that pressure from patients; bonuses of up to $150,000 annually were offered to doctors who minimized specialty referrals, inpatient care, etc.3 Our protest of those incentives, and a contract provision forbidding their disclosure (a “gag clause”) led to “delisting”. Award-winning physicians—who often attract unprofitably sick patients—were also delisted. An academic leader admonished physicians: “[We can] no longer tolerate having complex and expensive-to-treat patients encouraged to transfer to our group.”4

In the end, Americans concluded that fee-for-NON-service was even worse than fee-for-service.

HMOs lived on, but retreated from shifting risk to providers, relying instead on mother-may-I-style cost containment, like pre-authorization.

Now SGIM’s Commission has joined the growing policy bandwagon to reanimate the HMO strategy. There are semantic changes—ACO has replaced HMO, and when insurers drop expensive doctors (e.g. the 1,000-member Yale Medical Group5), it’s called “network optimization” not “delisting”. In a new twist on gag clauses, today’s ACO patients (e.g. seniors in Medicare’s Pioneer ACOs) aren’t told they’re enrolled. But the diagnosis and prescription are unchanged.

As in 1971, fee-for-service is the culprit. A shift to “bundled payment, capitation, and increased financial risk sharing.” is the solution, with “risk adjustment… to avoid physicians and other providers cherry-picking the healthiest patients”; and “quality measures… to assure that evidence-based care is not denied as a cost-saving mechanism.”

Twentieth century risk adjustment and quality monitoring were overmatched by HMOs’ gaming and deception. Despite additional decades of work to devise bullet-proof risk adjustment, gaming remains so powerful and pervasive that cost and quality rankings are often distorted, or even inverted. No solution is on the horizon.

Gaming to Win

Doctors, hospitals and health plans supply diagnoses, the raw material for risk-adjustment algorithms. But diagnoses reflect the aggressiveness of workups and coding, not just the patient’s clinical condition. Under risk-adjusted capitation and bundled payment, making patients look sicker on paper yields higher payments.

Overdiagnosis appears widespread. In high-cost regions, aggressive testing is common and labels patients with more seemingly serious diagnoses.6 But this apparently greater severity-of-illness is artifactual.6 How could aggressive testing lead to overdiagnosis? Obtaining echocardiograms on asymptomatic octogenarians would turn up many stiff ventricles, allowing diagnoses of “diastolic congestive heart failure (CHF)”. While few such patients would have the grave prognosis and high costs of symptomatic CHF, the diagnosis would boost their risk scores, and hence ACOs’ capitation payments (and risk-adjusted outcomes).

“Upcoding” is ubiquitous among hospitals paid diagnosis-related groups (DRG)s (bundled payment per admission) and capitated Medicare Advantage (MA) plans, which use it to extract overpayments of $30 billion annually.7 Paradoxically, Centers for Medicare & Medicaid Services’ (CMS) efforts to refine risk adjustment probably increased gaming-induced overpayments7; boosting capitation payment for comorbidities made asymptomatic CHF very profitable. Outright cheating may also be common; an Inspector General’s audit found unsupported comorbidities in 45 % of cases.

We’ve seen first-hand how coding practices impact quality scores. A public hospital where we worked scored poorly on risk-adjusted mortality—42 % above expected. In response, administrators hired consultants to comb charts for ill-described comorbidities and coach interns on wording choices that would boost risk scores; e.g. “hypo-magnesemia” rather than “Mg = 1.6” ups the risk score (and DRG payment). Within 6 months, risk-adjusted mortality fell to 14 % below expected, and Medicare reimbursement climbed $3 million.

Sadly, the upcoding/overdiagnosis arms race makes such practices mandatory. Our hospital looked bad because it was judged against hospitals that had already adopted coding coaches—we’d encountered them for years at the prosperous medical mecca nearby. Similarly, the 95 % of MA plans that cheated on a quality measure pushed down the rankings of the 5 % that told the truth.8 In quality measurement, honest guys finish last, as do primary care doctors caring for the poor, the mentally ill and non-English speakers.9

Less skillful gaming is one reason that mission-driven providers lose under bundled payment. But even if good-guy ACOs play the seemingly victimless upcoding game, they face daunting odds. Less scrupulous rivals willing to overdiagnose, cherry-pick and skimp on care can outcompete them on price and “quality”.

Moreover, giant organizations enjoy a decisive advantage. When one patient can cost millions, deep pockets are essential to assume risk. Even more important, insurers and hospital systems that dominate local markets can extract deals, and hence profits, not available to smaller competitors. The profit-advantage of bigger, more ruthless players makes them more attractive to investors and bondholders, vital sources of the capital needed to expand and modernize.

During the HMO era, many local providers entered the capitated market, but few survived. Undercapitalized, saddled with unprofitable patients and lacking clout to get the best price, some folded; others were swallowed by large national insurers.

Currently, insurers and hospital systems are bulking up. In 45 states one or two insurers now control more than half the market.10 UnitedHealth bought a Medicare ACO with 2,300 physicians; Wellpoint a chain of clinics; and Humana an in-home care manager with 1,500 providers and an urgent care/occupational health clinic firm. The proportion of office-based physicians employed or closely affiliated with hospitals grew from 41 % in 2000 to 72 % in 2010.10 In the past 5 years alone, 835 hospitals have merged; today “the typical hospital market . . . has one dominant system [and] two to three smaller systems.”.11 Even the largest cities will soon be left with only a handful (or less) of mega-ACOs.

The SGIM Commission rests its hope for cost control on ACO-type payment. Yet a system dominated by profit-maximizing oligopolies is a perilous route to savings. Moreover, the studies most often cited as evidence that ACO-like contracts bend the cost curve provide scant reassurance; the claimed savings from utilization reductions evaporate after factoring in bonuses providers earned for “shared savings” and “quality”.12 , 13 In Medicare’s demonstration program, upcoding created an illusion of lower costs, but (according to the Congressional Budget Office) virtually no real savings.

Purchasing Value

The Commission’s recommendation to pay based on quality seems a no-brainer. But there’s reason for skepticism about pay for performance (P4P), even if we could overcome the challenges of upcoding and accurate quality measurement. Doctors’ poor performance seldom stems from lack of motivation, and monetary incentives often worsen performance for cognitively complex tasks, especially when preexisting (intrinsic) motivation is high.14 Rewarding a narrow set of behaviors may distort, rather than improve global quality—the medical equivalent of teaching to the test. Do we really need to make our reimbursement system more complex?


Like the SGIM Commission, we rue the toxic incentives of the current fee-for-service system. But in the profit-maximizing milieu of American medicine, capitation risks making things even worse. “Risk-sharing” too often means that physicians earn bonuses for denying care—a danger perceived by patients, who take a dim view of capitation.15 Risk-sharing is not simply the inverse of fee-for-service, but of fee splitting, the illegal practice of kickbacks for referrals.

There are many bad ways to pay doctors, and no particularly good ones. Other nations have achieved better outcomes, lower costs and fairer compensation of physicians using a variety of methods: fee-for-service, capitation, and salary; none is clearly best. The common theme isn’t mode of payment, but a universal system with regulations that restrain costs and minimize the opportunities for profit and the risk of loss.

Payment reform should focus not on manipulating greed, but on dampening it. Then the real motivations for good doctoring—altruism, social duty, and the glow we feel when we help our patients—can flourish.