In Chapter 8, we distinguished between the “Ps” and “Qs” of health care costs: prices and quantities. Cost equals price multiplied by quantity
Strategies to control costs on the payment side can primarily target either prices or quantities (see Table 9-1).
Under California’s fee-for-service Medicaid program, Dr. Vincent Lo’s reimbursement for a routine office visit remained below $25 for 8 years.
The Medicare program reduced Dr. Ernesto Ojo’s fee for cataract surgery from $900 to $804.
Instead of paying all hospitals in the area the going rate for magnetic resonance imaging (MRI) brain scans ($1,400), Apple a Day HMO contracts only with those hospitals that agree to perform scans for $1,000, and will not allow its patients to receive MRIs at any other hospital.
Metropolitan Hospital wants a contract with Apple a Day HMO at a per diem rate of $2,000. Because Apple a Day can hospitalize its patients at Crosstown Hospital for $1,700 a day, Metropolitan has no choice but to reduce its per diem rate to Apple a Day to $1,700 in order to get the contract. In turn, to make up the $300 per day shortfall, Metropolitan increases its charges to several other private insurers.
Choice Ref PPO asks each hospital in its market region to submit a bid for the total charge for a knee replacement. Most bids are at least $40,000, but High Value Hospital, which has a reputation for good quality care, submits a bid of $33,000. Choice Ref informs the patients enrolled in the plan that they can choose any hospital for a knee replacement, but that Choice Ref will only pay the hospital $35,000 and the patient will be responsible for paying providers any charges above that amount. After this policy is put into effect, 75% of Choice Ref patients needing knee replacement get their operations at High Value Hospital.
In Canada and most European nations, a public or quasipublic agency regulates a uniform fee schedule for physician and hospital payments. Often, negotiations occur between the payers (payer is a general term that includes both purchasers and insurers—see Chapter 16) and professional organizations in establishing these fee schedules. In the United States, as discussed in Chapter 4, Medicare, Medicaid, and many private insurance plans have replaced “usual, customary, and reasonable” physician payment with predetermined prices for particular services. Competitive approaches to controlling prices have also been attempted in the United States. In the 1980s, California initiated competitive bidding among hospitals for Medicaid contracts, with contracts awarded to hospitals offering lower per diem charges. Private insurance plans have also used competitive bidding to bargain for reductions in physician and hospital fees. A recent variation on competitive payment strategies is “reference pricing,” such as the approach used by Choice Ref PPO. Public disclosure of prices—price transparency—is gaining ground in the United States; such disclosure may deter hospitals and physicians from setting excessively high prices (Bai & Anderson, 2015).
Controlling prices has produced some limited success at restraining the growth of overall health care expenditures. Lower prices explain much of the difference in health expenditures between the United States and other nations. For example, in 2012 the regulated fee for coronary artery bypass surgery in European nations ranged from $14,000 to $23,000; the average fee paid by private insurance plans in the United States was $73,420 (Klein, 2013). A study of reference pricing implemented by one large purchaser in California for joint replacements found that it reduced payments by about 30%, although the purchaser saved only 0.26% in total costs because joint replacements are a relatively small contributor to total costs (Lechner at al., 2013). Two major problems limit the potency of price controls for containing overall costs, particularly when prices are regulated at the fee-for-service level.
The first problem occurs when price controls are implemented in a piecemeal fashion by different payers. Providers, like Metropolitan Hospital, often respond to price controls imposed by one payer by increasing charges to other payers with less restrictive policies on fees—a phenomenon known as cost shifting. The cost-shifting problem may be avoided when a uniform fee schedule is used by all payers (as in Germany) or by a single payer (as in Canada).
The quantity of services provided often surges when prices are strictly controlled, leading analysts to conclude that providers respond to fee controls by inducing higher use of services in order to maintain earnings (Bodenheimer, 2005).
Price controls have the appeal of being a relatively painless form of cost control insofar as they do not limit the quantity of services provided. However, variations in fee schedules may compromise access to care for certain populations; Medicaid fee-for-service rates to physicians are far below private insurance rates in most states, making it difficult for Medicaid patients to find private physicians who will accept Medicaid payment. In nations with uniform fee schedules, concerns have been voiced that ratcheting down of fees may result in “patient churning” (high volumes of brief visits), with a deterioration in quality of care and patient satisfaction.
Utilization (Quantity) Controls
Because the effectiveness of price controls may be limited by increases in quantity, payers need to consider methods for containing the actual use of services. As indicated in Table 9-1, there are a variety of methods for attempting to control use. We begin by examining one strategy, changing the unit of payment, that we introduced in Chapter 4. We then describe additional mechanisms that attempt to restrain the quantity of services.
Changing the Unit of Payment
Dr. John Wiley is upset when the PPO reduces his fee from $75 to $60 per visit. In order to maintain his income, Dr. Wiley lengthens his day by half an hour so he can schedule more patient visits.
Dr. Jane Stuckey is angry when the HMO reduces her capitation payment from $20 to $15 per patient per month. She is unable to maintain her income by providing more visits because more patient visits do not bring her more money. She hopes that more HMO patients will enroll in her practice so that she can receive more capitation payments.
One simple way to get a handle on the quantity factor is by redefining the unit of payment. In Chapter 4, we discussed how services may be bundled into more aggregate units of payment, such as capitated physician payment, diagnosis-related group (DRG) episode-of-care hospital payment, and bundled payments that combine both physician and hospital payment for an episode of care.
The more aggregated the unit of payment, the more predictable the quantity tends to be. For example, in the case of Dr. Wiley receiving fee-for-service payment, there is a great potential for costs to rise due to increases in the number of physician visits, surgical procedures, and diagnostic tests. When the unit of payment is capitation, as in the case of Dr. Stuckey, the quantity factor is not the number of visits but rather the number of individuals enrolled in a practice or plan. From a health plan’s perspective, the C = P × Q formula still applies when paying physicians by capitation, but now the P is the capitation fee and the Q is the number of individuals covered. Other than by raising birth rates, physicians have little discretion in inducing a higher volume of “quantities” at the capitation level for the health care system as a whole. Similarly, under global budgeting of hospitals, P represents the average global budget per hospital and Q is the number of hospitals.
Shifting payment to a more aggregated unit has obvious appeal as a way for payers to counter cost inflation due to the quantity factor. Life is never so simple, however. In Chapter 4, we discussed how more aggregate units of payment shift financial risk to providers of care. Another way of describing this shifting of risk is that one person’s solution to the quantity problem becomes another person’s new quantity problem. A hospital paid by global budget instead of by fee-for-service now must monitor its own internal quantities of service lest these quantities drive hospital operating costs over budget. To the extent that providers are unsuccessful in managing resources under more aggregated forms of payment, pressures mount to raise the prices paid at these more aggregated payment units.
Changes in policies for units of payment rarely occur independent of other reforms in cost-control strategies, making it difficult to isolate the specific effects of changing the unit of payment. For example, physician capitation usually occurs in the context of other organizational and cost-control features within a managed care plan. Group- and staff-model HMOs receiving capitation payments from employers and paying physicians by salary have been shown to reduce costs by reducing the quantity of services provided, in particular by reducing rates of hospitalization (Hellinger, 1996; Bodenheimer, 2005).
For hospitals, changing Medicare payments from a fee-for-service to an episode-of-care unit under the DRG-based system in 1983 resulted in a modest slowing of the rate of increase in Medicare Part A expenditures. However, hospitals were able to shift costs to private payers to make up for lower DRG revenues, and national health expenditures as a whole were not affected by Medicare’s new payment mechanism (Rice, 1996). Medicare’s use of bundled episode-based payments for physicians and hospitals is a more recent development; some evidence shows that this method can reduce costs (Catalyst for Payment Reform, 2014). Global hospital budgeting in Canada has been a key element of that nation’s relative success at containing hospital costs (Commonwealth Fund, 2015).
The health care system in Germany and in some Canadian provinces has countered the open-ended dynamic of fee-for-service payment by introducing global budgeting, called expenditure caps, for physician payment (Bodenheimer, 2005). Under Canadian expenditure caps, a budget is established for all physician services in a province. Although individual physicians continue to bill the provincial health plan on a fee-for-service basis, if increases in the use of services cause overall physician costs to exceed the budget, fees are reduced (or fee increases for the following year are sacrificed) to stay within the expenditure cap. Evidence from Canada suggests that implementation of expenditure caps was associated with stabilization of physician costs in the mid-1990s (Barer et al., 1996). In the United States, Medicare adopted a less-stringent version of an expenditure cap for physician fees, known as the “sustainable growth rate,” that was abandoned in 2015 after many years of unsuccessful enforcement (Aaron, 2015). Expenditure caps for physician payments allow the payer to focus on the aggregate C part of the equation—in this case, the total physician budget. ACO shared savings programs, discussed in Chapter 6, are a related strategy attempting to provide a global expenditure feedback loop to modulate fee-for-service payments.
Randy Payton has an insurance policy with a $2,000 deductible and 20% copayment for all services; if he incurs medical expenses of $6,000, he pays the first $2,000 plus 20% of $4,000, for a total of $2,800.
Joseph Mednick’s health plan requires that he pay $20 each time he fills a prescription for a medication, with the health plan paying the cost above $20; because he suffers from diabetes, hypertension, and coronary artery disease, the copayments for his multiple medications cost him $1,200 per year.
Cost sharing refers to making patients pay directly out of pocket for some portion of their health care. In managed competition, cost sharing occurs as part of the financing transaction at the point of purchasing a health insurance plan. In this section, we discuss the more traditional notion of cost sharing—using deductibles, copayments, and uncovered services as part of the payment transaction to make patients pay a share of costs at the point of receiving health care services.
The primary intent of cost sharing at the point of service is to discourage patient demand for services. (Cost sharing also shifts some of the overall bill for health care from third party payers to individuals in the form of greater out-of-pocket expenses.) As discussed in Chapter 3, when individuals have insurance coverage, they are more likely to use services than when they have no insurance. While protection against individual financial risk is one of the essential benefits of insurance, insurance coverage removes the market restraint on costs that occurs in a system of out-of-pocket payment.
Cost sharing at the point of service has been one of the few cost-containment devices subjected to the rigorous evaluation of a randomized controlled experiment. In the Rand Health Insurance Experiment, individuals were randomly assigned to health insurance plans with varying degrees of cost sharing. Individuals with cost-sharing plans made about one-third fewer visits and were hospitalized one-third less often than individuals randomized to the plan with no cost sharing (Newhouse et al., 1981).
Although the randomized controlled trial provides an excellent laboratory for scrutinizing the effect of a single cost-containment mechanism, some observers have cautioned that analyses based on controlled research designs may produce results that cannot be generalized to the real world of health policy. For example, the United States has a greater level of cost sharing than many industrialized nations, but also the highest overall costs. Studies have found that when cost sharing begins to produce lower use of services for a large population of patients rather than for a small number of patients in an experiment, physicians may increase the volume of services provided to patients with better insurance coverage. Moreover, 70% of health care expenditures are incurred by 10% of the population—people who are extremely ill and generate huge costs through lengthy ICU stays and other major expenses. Cost sharing has little influence over this component of care. Compared to the micro-world of one not–very-sick patient deciding whether to spend some money on a physician visit, patient cost sharing in the macro-world may remove only a thin slice from a large, expanding pie (Bodenheimer, 2005).
The Rand experiment also evaluated the influence of cost sharing on appropriateness of care and health outcomes. Cost sharing did not reduce medically inappropriate use of services selectively, but equally discouraged use of appropriate and inappropriate services. Study patients (especially those with low incomes) with cost sharing received less preventive services and had poorer hypertension control than those without cost sharing (Brook et al., 1983). Patients are less likely to purchase needed medications under cost-sharing policies, for example Medicare Part D’s “donut hole” (see Chapter 2), leading to worse control of chronic illnesses and more emergency hospitalizations (Hsu et al., 2006; Goldman et al., 2007; Schneeweiss et al., 2009). These studies suggest that cost sharing is not a painless form of cost control.
Cost sharing for emergency department care may reduce inappropriate use of emergency services without adversely affecting appropriate use or patient health outcomes (Goodell et al., 2009). Cost sharing may be a painless form of cost control when used in modest amounts, not applied to low-income patients, and designed to encourage patients to use lower-cost alternative sources of care (e.g., clinics instead of emergency departments) rather than to discourage use of services altogether.
Thelma Graves suffers from a severe hyperthyroid condition; she and her physician agree that she will undergo thyroid surgery. Before scheduling the surgery, the physician has to call Ms. Graves’ insurance company to obtain preauthorization, without which the insurer will not pay for the surgery.
Fred Brady is hospitalized for an acute myocardial infarction. The hospital contacts the utilization management firm for Mr. Brady’s insurer, which authorizes 5 hospital days. On the fourth day, Mr. Brady develops a heart rate of 36 beats/min, requiring the insertion of a temporary pacemaker and prolonging the hospital stay for 10 extra days. After the fifth hospital day, Mr. Brady’s physician has to call the utilization management (UM) firm every 2 days to justify why the insurer should continue to pay for the hospitalization.
Derek Jordan has juvenile-onset diabetes and at age 42 becomes eligible for Medicare due to his permanent disability from complications of his diabetes. He is admitted to the hospital for treatment of a gangrenous toe. Under Medicare’s DRG method of payment, the hospital receives the same payment for Derek’s hospitalization regardless of whether it lasts 2 days or 12 days. Therefore, the hospital wants Derek’s physician to discharge Derek as soon as possible. Each day, a hospital UM nurse reviews Derek’s chart and suggests to the physician that Derek no longer requires acute hospitalization.
Utilization management involves the surveillance of and intervention in the clinical activities of physicians for the purpose of controlling costs (Grumbach & Bodenheimer, 1990). In contrast to cost sharing, which attempts to restrict health care use by influencing patient behavior, UM seeks to influence physician behavior. The mechanism of influencing physician decisions is simple and direct: denial of payment for services deemed unnecessary.
UM is related to the unit of payment in the following way: Whoever is at financial risk (see Chapter 4) performs UM. Under fee-for-service reimbursement, insurance companies perform UM to reduce their payments to hospitals and physicians. The DRG system induces hospitals, at risk for losing money if their patients stay too long, to perform UM. Under an HMO capitation contract with a primary physician group, the physician group conducts UM so that it does not pay more to physicians than it receives in capitation payments. If an HMO pays a hospital a per diem rate, the HMO may send a UM nurse to the hospital each day to review whether the patient is ready to go home.
Micromanage, Inc., performs UM for several insurance companies. Each day, Rebecca Hasselbach reviews the charts of each patient hospitalized by these insurers to determine whether the patients might be ready for discharge. Usually, if the attending physician wants the patient to remain in the hospital, his or her opinion is honored. By pushing for early discharges, Ms. Hasselbach, her Micromanage colleagues around the country, and the medical director save their insurers about $1,000,000 each year. The annual cost of administering UM is $900,000.
There is little evidence that UM yields substantial savings, particularly when the overhead of administering the UM program itself is taken into account (Wickizer, 1990). UM would appear to be a painless form of cost control because it intends to selectively reduce inappropriate or unnecessary care. However, reviewers often make decisions on a case-by-case basis without explicit guidelines or criteria, with the result that decisions may be inconsistent both between different reviewers for the same case and among the same reviewer for different cases (Light, 1994).
UM has come under fire as a process of micromanagement of clinical decisions that intrudes into the physician–patient relationship and places an unwelcome administrative burden on physicians and other caregivers. Physicians in the United States have been called the most “second-guessed and paperwork-laden physicians in western industrialized democracies” (Lee & Etheredge, 1989). Substantial physician time goes into appealing denials and persuading insurers about the appropriateness of services delivered. A physician and public backlash to UM forced health insurance plans to relax their UM activities in the late 1990s. However, many plans reintroduced UM around 2003 as costs escalated (Mays et al., 2004) and it remains prevalent in 2015.
Several approaches to UM have been developed that attempt to avoid some of the onerous features of case-by-case utilization review. Practice profiling, rather than focusing on individual cases, uses summary data on practice patterns to identify physicians whose overall use of services significantly deviates from the standards set by other physicians in the community. These outlier physicians can be made subject to strict UM monitoring with denials.
The strategy of “narrow networks” takes UM to its logical conclusion: denying any payments to providers with high-cost profiles by eliminating them from the insurer’s provider network. HMOs are the original narrow network, restricting their members to physicians and hospitals within the HMO. Now, many insurers offering policies within the Affordable Care Act’s insurance exchanges have narrowed their provider networks by excluding high-price providers. While narrow networks may control costs, a concern has arisen that the networks may lower health care quality by excluding high-quality providers (Corlette et al., 2014).
Bob is a patient in the Canadian province of Alberta. He develops back pain, and after several visits to his family physician requests an MRI of his spine to rule out disk disease. His physician, who does not suspect a disk herniation, agrees to place him on the waiting list for an MRI, which for nonurgent cases is 5 months long.
Rob lives in Alberta, and after lifting an 80-lb load at work, experiences severe lower back pain radiating down his right leg. Finding a positive straight-leg-raising test on the right with loss of the right ankle reflex, his family physician calls the radiologist and obtains an emergency MRI scan within 3 days.
Supply limits are controls on the number of physicians and other caregivers and on material resources such as the number of hospital beds or MRI scanners. Supply limits can take place within an organized delivery system such as an HMO in the United States, or for an entire geographic region such as a Canadian province.
The number of elective operations and invasive procedures, such as cardiac catheterization, performed per capita increases with the per-capita supply of surgeons and cardiologists, respectively (Bodenheimer, 2005). This phenomenon is sometimes called “supplier-induced demand” (Evans, 1984; Rice & Labelle, 1989; Phelps, 2003). Controlling physician supply may reduce the use of physician services and thereby contribute to cost containment.
Supplier-induced demand pertains to material capacity as well as to physician supply. Per-capita spending for fee-for-service Medicare patients is over twice as high in some regions of the United States than in others (Gawande, 2009, www.dartmouthatlas.org). This remarkable cost variation is not explained by differences in demographic characteristics of the population, prices of services, or levels of illness, but is due to the quantity of services provided. Residents of areas with a greater per-capita supply of hospital beds are up to 30% more likely to be hospitalized than those in areas with fewer beds (Fisher et al., 2000). The maxim that “empty beds tend to become filled” has been known as Roemer’s law (Roemer & Shain, 1959). Conversely, strictly regulating the number of centers allowed to perform heart surgery establishes a limit for the total number of cardiac operations that can be performed. In situations of limited supply, physicians must determine which patients are most in need of the limited supply of services. Ideally, those truly in need gain access to appropriate services, with physicians possessing the wisdom to distinguish those patients truly in need (Rob) from those not requiring the service (Bob).
Although there may not always be a directly linear relationship between supply and use of services, there are clear instances in which limitations of capacity restrain use. For example, international comparisons in 2013 demonstrate large variations in use of coronary revascularization procedures (coronary artery bypass surgery and angioplasty), with the United Kingdom’s rate of these procedures only 57%, and Canada’s 78%, of the US rate (OECD, 2013). These rates correspond to the degree to which these nations regulate (minimally in the case of the United States) the number of centers performing cardiac surgery. In spite of doing more procedures, US coronary heart disease mortality is slightly higher than that of the UK and Canada (OECD, 2013).
A “natural experiment” provides an illustration of how restricting the supply of a high cost resource may be implemented in a relatively painless manner for patients’ clinical outcomes. A US hospital experiencing a nursing shortage abruptly reduced the number of staffed intensive care unit beds from 18 to 8 (Singer et al., 1983). For patients admitted to the hospital for chest pain, physicians became more selective in admitting to the intensive care unit only those patients who actually suffered heart attacks. Limiting the use of ICU beds did not result in any adverse health outcomes for patients admitted to nonintensive care unit beds, including those few nonintensive care unit patients who actually sustained heart attacks. This study suggests that when faced with supply limits, physicians may be able to prioritize patients on clinical grounds in a manner that selectively reduces unnecessary services. Establishing supply limits that require physicians to prioritize services based on the appropriateness and urgency of patient need represents a very different (and less intrusive) approach to containing costs than UM, which relies on external parties to authorize or deny individual services in a setting of relatively unconstrained capacity.
Controlling the Type of Supply
A specific form of supply control is regulation of the types (rather than the total number) of providers. Chapter 5 explored the balance between the number of generalist and specialist physicians in a health care system. Increasing the proportion of generalists may yield savings for two reasons. First, generalists earn lower incomes than specialists. Second, and of greater impact for overall costs, generalists appear to practice a less resource-intensive style of medicine and generate lower overall health care expenditures, including less use of hospital and laboratory services (Bodenheimer & Grumbach, 2007).