National Health Expenditures, 1996

Highlights

NHE in 1996 reached $1.035 trillion, a modest increase of 4.4 percent from 1995 levels. On a per person basis, health spending rose $126, from $3,633 per person in 1995 to $3,759 in 1996.

The rate of growth in spending for the last 4 years was the slowest experienced since 1960 (Figure 1). The growth in NHE has continued to decelerate since 1990. In the history of tracking NHE, this was the first instance where growth in NHE decelerated for 6 consecutive years. This was also the first time in more than three decades that NHE growth was in the single digits for 6 successive years. These growth rates contrast sharply with the period 1980-90, when both private and public sectors averaged double-digit annual growth rates (11.3 and 10.5 percent, respectively). This period of high growth preceded the most recent call for health care reform.

Health care spending as a percent of gross domestic product (GDP) stabilized over the 1993-96 period at 13.6 percent, as health spending and GDP grew at roughly the same average annual rate. This is the first time in the 37-year history of the current NHE series that the share of GDP has been stable for more than 3 consecutive years.

The portion of NHE funded through government programs continued to increase in 1996, rising 0.8 percentage points to 46.7 percent. The public share has steadily increased for 7 consecutive years, from 40.4 percent in 1989 to 46.7 percent in 1996. This resulted from public sector health spending (up an average of 9.7 percent) growing at almost twice the rate of private sector spending (5.8 percent on average) over that period.

Factors Accounting for Growth

Personal health care expenditures (PHCE) grew 4.4 percent in 1996. This year's growth rate continued a trend of deceleration in PHCE and in many specific health service categories as well. To better understand the causes of this deceleration, PHCE growth can be separated into four factors—economywide inflation, medical-specific price inflation above and beyond economywide inflation (referred to as “excess medical inflation”), population change, and changes in the use and intensity of services.

In 1996 price increases accounted for more than one-half of PHCE growth: Economywide inflation drove 56 percent of health spending growth, while medical price inflation in excess of overall inflation accounted for only an additional 5 percent. Growth in intensity of services accounted for 19 percent and population growth for 21 percent (Figure 2). Excess medical price inflation steadily declined over the past few years, and in 1996 represented the smallest share since 1979.¹ Measurement of medical inflation in this analysis was substantially affected as a result of using newly released Producer Price Indexes (PPIs) to deflate selected categories of nominal PHCE (explained later in this section).

After adjusting for population, nominal PHCE per capita increased 3.5 percent in 1996. When economywide and excess medical price inflation of 2.6 percent were taken into account, real services purchased per person increased only 0.8 percent. This residual amount measures the change in intensity of service use per person, reflecting the quantity of services delivered, age/sex composition of the population, technological improvements that may allow for more efficient use of services, and the cumulation of any measurement errors in expenditures, inflation, or population.

Several revisions were added to the methodology for calculating factors accounting for health expenditure growth this year. The first revision affected the service category of non-durable medical products, which was split to distinguish growth for prescription drugs from growth in non-prescription drugs and sundries (Table 1). The CPIs for prescription drugs and over-the-counter drugs were applied to each category to determine price impact on growth. Applying more narrowly defined CPIs to prescription drugs and non-prescription drugs demonstrated no effect on this analysis.

The second revision was to the service categories of physician services and nursing home care. In the past, these series were deflated using the CPI for physician services and HCFA's nursing home input price index. The Bureau of Labor Statistics (BLS) started publishing a PPI for offices and clinics of doctors of medicine in 1994, which measures changes in actual transaction prices for unique sets of services; in 1995 an index for skilled and intermediate care facilities was added. Both of these indexes were incorporated into the factor analysis in 1996.

Using the price proxies employed in the recent past, the contribution of medical price inflation above and beyond economywide inflation to overall growth in PHCE would have been 13 percent in 1996. By incorporating PPIs that better measure actual transaction prices, the impact of excess medical price inflation was lowered to 5 percent of total growth.² Furthermore, the effect of overall price inflation (including both economywide and excess medical price inflation) dropped from 68 percent of total growth in 1996 using previously available price proxies to 61 percent of growth after incorporating PPIs. The difference between the two methods (7 percentage points) is attributable to the use of revised PPI price proxies.

Personal Health Care

Americans spent $907.2 billion for PHCE in 1996, an increase of 4.4 percent from 1995. All segments of PHCE decelerated in growth with the exception of non-durable medical products. In 1996 hospital and physician services were among the slowest growing segments of PHCE. Because hospital and physician services account for most of PHCE (61.8 percent), their slow growth affected the overall growth of health care expenditures. Non-durable medical products consist of over-the-counter medicines and medical sundries, and prescription drugs. In 1996 both components grew at faster rates than overall PHCE (4.7 percent and 9.2 percent, respectively).

Private health insurance (PHI) and public programs attempt to control health care costs by targeting those providers consuming the greatest expenditures—hospitals and physicians. Some of the most significant changes in health care are seen in these two providers. Hospitals and physicians were subjected to restrictions on payments by prospective payment systems and fee schedules, and to scrutiny by managed care organizations (MCOs) on the type and number of services provided. In response, hospitals devised ways to increase their revenues or improve their market position by expanding their lines of business, specializing, and merging with other hospitals; physicians are affiliating with larger practices and physician networks to increase their competitiveness. Even some of the smaller expenditure categories that have registered rapid growth in recent years (such as prescription drugs and home health care) are being examined closely by insurers for ways to restrain growth.

As a result, the distribution of PHCE has changed in recent years. The slow growth of expenditures for hospital and physician services caused these categories combined to drop almost 4 percentage points as a share of PHCE since 1990 (Table 2). Hospital expenditures as a share of PHCE went from 41.7 percent in 1990 to 39.5 percent in 1996; similarly, the share of PHCE for physician services declined from 23.8 percent in 1990 to 22.3 percent in 1996. Share declines in hospital and physician service expenditures were offset by increases in other services, particularly home health care and other personal health care (OPHC). Both of these service categories measure spending for alternative, lower cost services³ that can substitute for more traditional physician and hospital services.

Hospital Services

Hospital care expenditures, 39.5 percent of PHCE, grew to $358.5 billion in 1996. The small 3.4-percent increase in spending halted a 5-year trend in which growth decelerated from 10.7 percent in 1990 to 3.3 percent in 1995. Community hospital expenditures are responsible for 89 percent of all hospital spending. Growth in expenditures for community hospital inpatient services⁴ have also decelerated since 1990. In 1996 inpatient days declined 3.6 percent, the result of a decline (-0.4 percent) in admissions—the first decline in 3 years—and in length of stay (-0.2 days). Across the United States, hospital occupancy rates have fallen from 64.5 percent in 1990 to 58.7 percent in 1996, despite the 7-percent decrease in number of hospital beds during that period. Although also decelerating, continued strong growth in community hospital outpatient services has helped to offset decelerating growth in inpatient hospital expenditures. Hospitals succeeded in aggregate in reducing expense growth to rates approximately equal to revenue growth, thus maintaining their profit margins (American Hospital Association, 1996).

These statistics reflect the changing environment in which hospitals operate, as their traditional line of inpatient hospital services shrinks under pressure from insurers to reduce expensive inpatient hospital stays. To remain financially viable, hospitals have sought ways to improve their competitiveness, improve their negotiating positions with insurers, increase their risk-bearing ability, control a greater share of the market, and generate greater revenues. Hospitals continue to adapt to the managed care environment in a number of ways, characterized by: (1) increased merger and joint-venture activity, (2) specialization in profitable product lines, and (3) diversification into new lines of business.

Merging with existing multihospital organizations or developing collaborative agreements among facilities helps hospitals control expenses in the face of decelerating revenue growth. Increased size gives hospitals greater control over negotiating with insurers, contracting for the purchase of goods and services, and generating capital required for survival in this increasingly competitive market, while increasing hospitals' risk-bearing ability. Some observers raise concerns, however, that the merger phenomenon may be disadvantageous for consumers in markets dominated by large hospitals and may lead to reduced competition, higher prices, and/or reduced services (Alpha Center, 1997).

One type of specialized merger is the purchase of non-profit hospitals by for-profit organizations. In some cases, a for-profit conversion is the only way to maintain hospital services in a community when its nonprofit hospital, unable to meet expenses, faces closure. Communities worry, however, that other indirect benefits of non-profit facilities, such as charity care, research, physician education, public health functions, and accountability may shrink (Langley and Sharpe, 1996; Altaian and Shactman, 1997).

Hospitals may also collaborate with other health providers to improve their competitiveness. To increase their ability to attract new customers, hospitals have joined with physicians and other hospitals to form physician-hospital organizations (PHOs), networks, health care systems, and joint ventures. These arrangements make it easier for hospitals to diversify their services, allowing them to offer comprehensive ranges of services to insurers and gain guaranteed access to patients.

Hospitals also have adapted to the new competitive hospital environment through specialization. By identifying high-profit services and developing efficient and effective treatment protocols for a narrow range of services, hospitals can effectively negotiate with insurers for a large share of specialized services. Cardiac care, cancer treatment, rehabilitation services, or women's health care are often identified as high-profit niches for hospital specialization. Facing competition from specialized hospitals, more traditional hospitals complain that revenues from their most profitable services erode while they must continue to provide less profitable services (such as trauma centers and neonatal intensive care units) and a larger proportion of care to uninsured patients (Myerson, 1997).

Finally, as the share of revenues from inpatient services declined, hospitals expanded their range of services, particularly into post acute care, where insurer payments were less restrictive. Expenditures for nursing home care and home health care services provided by hospital-based facilities, shown later in this article, have exhibited strong growth for the last 6 years. Hospital-based nursing home care and home health care expenditures accounted for 5 percent of all non-Federal hospital spending in 1996, up from just over 2 percent in 1990. This diversification of services allowed hospitals to capitalize on Medicare's payment regulations to maximize their revenues. For example, because Medicare compensation is based on a patient's diagnosis-related group (DRG), hospitals can increase profits by shortening inpatient stays and discharging the patient to a hospital-based skilled nursing facility (SNF), where the hospital can also receive cost-based payment for the patient's SNF care. A recent U.S. General Accounting Office (GAO) (1996) report found that “hospitals with SNF units saw larger decreases in the average patient length of stay than did hospitals without SNF units.” The GAO also reported that the number of hospital-based SNFs increased more than 80 percent between 1980 and 1996, from 1,145 to 2,088 (U.S. General Accounting Office, 1997d).

Nursing Homes and Home Health Facilities

For the first time, estimates of expenditures for services provided to all patients by hospital-based nursing homes and home health facilities were developed and presented in this report to provide a more comprehensive picture of spending for nursing home and home health care than is available in these NHE service categories. Expenditures reported in the nursing home and home health care service components of the NHE measure free-standing facility expenditures only. Expenditures for hospital-based facilities are included with hospital care in the NHE and are not readily identifiable. Information contained on cost reports submitted to the Health Care Financing Administration (HCFA) by hospitals, nursing homes, and home health agencies were used to develop these hospital-based expenditure estimates (Jing Xing Health and Safety Resources, Inc., 1997).

Spending for freestanding nursing home care amounted to $78.5 billion in 1996, implying an estimated cost of a 1-year stay in a freestanding nursing home in excess of $47,000. An additional $9.0 billion was spent for nursing home care provided in hospital-based facilities (Table 3). Growth in spending for nursing home care at all sites slowed from 7.1 percent in 1995 to 5.3 percent in 1996. The reason for this deceleration is, in part, a slowdown in the growth of nursing home input prices; i.e., the cost to facilities for providing services (Sensenig, Heffler, and Donham, 1997).

Expenditures for freestanding home health care reached $30.2 billion in 1996, and an additional $7.8 billion was spent for home health care provided by hospital-based home health agencies (Table 4). Annual spending growth for home health care delivered from all sites decelerated for the fourth consecutive year, from 23.4 percent in 1992 to 9.5 percent in 1996. Data from the Medicare program, the largest single payer for home health care, showed a dramatic slowdown in the growth of average number of visits per person served and persons served per 1,000 enrollees in 1996. Since 1993 Medicare has placed additional restraints on the growth in per visit payments (Sensenig, Heffler, Donham, 1997).

Annual expenditures for hospital-based nursing homes and home health care facilities grew more rapidly than their freestanding counterparts each year from 1990 through 1996. Although growth rates slowed in 1996, expenditures for care provided by these hospital-based facilities continued to increase faster than any of the national health expenditure service categories. For example, expenditures for hospital-based home health care grew 24.5 percent in 1996, compared with growth of 6.2 percent for freestanding facilities.

A growing number of hospitals are expanding their lines of business to include hospital-based nursing home and home health facility care (Jing Xing Health and Safety Resources, Inc., 1997; American Hospital Association, 1989-96). These facilities allow hospitals to provide a continuum of professional care to discharged patients not yet fully recovered from their illnesses. Moreover, such facilities enable hospitals to expand their revenue base in three ways: providing ready access to patients requiring these services (a competitive advantage in attracting patients), attracting patients that are more likely to be covered by Medicare or PHI (both reimburse at higher rates than Medicaid), and maximizing Medicare payments (skilled nursing home services are reimbursed on a reasonable cost basis).

Medicare's share of funding for home health care has been increasing steadily since 1988. In 1996 almost one-half of spending for home health care (49.4 percent) delivered from all sites was funded by Medicare. Because of the sustained high growth rates, Medicare spending for home health care provided to Medicare beneficiaries has been under scrutiny by the U.S. Department of Health and Human Service's Office of the Inspector General (OIG). As a result of problems detected through OIG's and HCFA's fraud and abuse activities, President Clinton declared a moratorium on licensing and certification of all new home health agencies (HHAs), effective September 15, 1997, until stricter requirements for Medicare participation have been instituted (Goldstein, 1997).

Physician Services

Expenditures for physician services rose to $202.1 billion in 1996, up 2.9 percent from the previous year. Growth in expenditures for physician services has decelerated steadily since 1990 and has remained in the single digits since 1992. The slowdown in growth is in part the result of the expanding role of managed care and changes in physician practices.

Managed care continued to be a major player in the physician marketplace. From 1990 to 1996, the percent of physicians with managed care contracts grew from 61 percent to 88 percent (Figure 3). The amount of overall physician revenues from managed care contracts more than doubled from 17 percent in 1990 to 39 percent in 1996 (Emmons and Wosniak, 1997). For roughly the same time period (1990-95), median physician net income grew at an average annual rate of 4.2 percent, a lower rate than the previous two 5-year periods (Moser, 1997).

In response to pressures in the physician marketplace, the composition of physician specialities changed. These changes were partly driven by managed care and the introduction of the Medicare payment system based on a resource-based relative value scale (RBRVS). Since 1985 the percent of office-based physicians with medical (including family practice) and hospital-based specialties grew, while those with surgical specialties declined (National Center for Health Statistics, 1997).⁵

Physician practice characteristics also changed in recent years. The size of physician practices grew from an average of 10.6 physicians in 1990 to 14.5 in 1996 (Emmons and Kletke, 1997). The proportion of physicians in solo practices or self-employed in group practices declined (Moser, 1996), while the proportion of employee physicians grew from approximately 32 percent in 1990 to 42 percent in 1996 (Emmons and Kletke, 1997). The increased size of practices enables physicians to control costs by operating more efficiently, spreading risk, and meeting the demands of MCOs or employer contracts. In the future, this trend is expected to continue as changes in the Federal Trade Commission antitrust guidelines give more leeway for physician practices to merge and create networks to contract with health care purchasers (Kuttner, 1997).

The share of physician services funded by public sources increased from 29.7 percent in 1992 to 32.9 percent in 1996. Medicare expenditures, 21.1 percent of physician service expenditures, continued to grow at rates faster than overall physician services, despite decelerating in 1996. In recent years, Medicare implemented the Medicare Fee Schedule and Volume Performance Standards (VPS).⁶ VPS is designed to reward or penalize physicians for changes in aggregate per capita utilization patterns. The reward or penalty is incorporated into the physician payment 2 years after the change in utilization pattern occurs. Because volume in 1994 grew faster than specified targets, there was a penalty incorporated into the 1996 payment rates, which accounts, in part, for the deceleration in Medicare spending for physician services.

Prescription Drugs

Spending for non-durable medical products is the second-fastest-growing segment of PHCE, increasing 7.7 percent in 1966. The category of non-durable medical products consists of over-the-counter drugs and medical sundries, as well as prescription drugs. Americans purchased $62.2 billion in prescription drugs in retail outlets in 1996, an increase of 9.2 percent over 1995 (Table 5). Most third-party payers witnessed continued high expenditure growth over the past several years, rather than the decelerated growth seen in most other service sectors. For PHI and all public programs in aggregate, prescription drug expenditures increased 13.3 percent in 1996. At the same time, out-of-pocket payments continued to grow very slowly, at 1.9 percent.

Prescription drugs in 1996 continued the 3-year trend of increases in utilization (as measured by the number of prescriptions dispensed), overshadowing prices as the primary factor accounting for growth. According to several surveys, increases in utilization ranged from 4.3 to 5.8 percent in 1996, up from a historic average growth of about 2 percent (IMS America, 1996a; Schondelmeyer, 1997). Although price increases, measured by the CPI for prescription drugs at 3.4 percent, remained relatively small, they too showed an acceleration from the 1.9-percent increase observed a year earlier (U.S. Department of Labor, 1997).

Managed care and improved drug therapies are important sources of expenditure growth in prescription drugs. As evidence of this trend, insurance plans paid 34 percent more per member per month for pharmacy benefits between 1993 and 1996 (Ukens, 1997). The substitution of drugs for other types of health care was especially evident in utilization increases in two categories of drugs: antidepressants (up 52 percent) and narcotic analgesics (up 13 percent) (Ukens, 1997). Antidepressant drugs substitute for more expensive psychotherapy and inpatient mental hospital stays. Narcotic analgesics are used in conjunction with surgery, enabling patients to avoid or shorten inpatient hospital stays. In addition, people discharged from the hospital with prescriptions for narcotic analgesics fill these prescriptions at local retail pharmacies (Vecchione, 1997); these expenditures are captured in the nondurable medical product category of NHE, rather than under NHE hospital expenditures as would have been the case had the patient received the drugs as part of an inpatient hospital stay.

The drug utilization incentives for physicians in managed care can act in opposing directions. On one hand, data on the number of pills per managed care prescription versus FFS prescription suggests physicians are keeping patients out of the office by increasing the number of pills per prescription (IMS America, 1996b). On the other hand, MCOs attempt to constrain pharmacy costs by holding physicians at risk for exceeding a pre-established drug budget (Johannes, 1997). Increases in drug expenses were one of the biggest factors in the sluggish 1996 earnings reports of health maintenance organizations (HMOs). Drugs accounted for 10 percent of HMOs' medical budgets last year but 50 percent of their cost increases (Johannes, 1997).

Another possible source for growth in this category may be direct-to-consumer advertisements (DTCA). Spending for DTCA doubled in 1996 from $0.3 billion to $0.6 billion and is expected to exceed $1 billion in 1997 (Ukens, 1997). The cost of advertising is a factor in price increases, adding to the growth of prescription drug expenditures. DTCA also places added pressure on physicians when advertisements encourage consumers to demand costly drugs (Vecchione, 1997).

Another source of price and utilization increase is the rise in the number of new drugs introduced during 1996. The Food and Drug Administration approved a record 53 new molecular entities in 1996 (Figure 4). New costlier drug therapies, such as two new protease inhibitors, a new class of drugs for asthma, a new treatment for multiple sclerosis, and a new schizophrenia medication, caused expenditures to rise. Net increases in product mix and the effect of new costlier therapies also added to prescription drug expenditure growth. In fact, the Federal Government credited the decrease in the number of new acquired immunodeficiency syndrome (AIDS) cases reported and a drop in AIDS-related death rates to new therapies such as the “AIDS cocktails,” widely available for the first time in 1996 (Associated Press, 1997).

Whether or not this sort of growth can be sustained is questionable. Many drugs created during a pharmaceutical research boom in the 1970s and 1980s are approaching the end of their patents, resulting in a record number of profitable drugs coming off patent in a short period of time (Tanouye and Langreth, 1997). The number of patent expirations may hold down pharmaceutical price increases and total expenditure growth for years to come.

Health Services and Supplies

Of the $1.0036 trillion spent for health services and supplies (HSS) in 1996, $540.9 billion (53.9 percent) resulted from private sector expenditures, mostly through PHI (33.6 percent) and out-of-pocket spending (17.1 percent) (Table 6). Another $462.7 billion (46.1 percent) came from government expenditures. Medicare and Medicaid alone accounted for more than three-quarters of all public expenditures for health care.

Since 1990 the share of total HSS spending paid for by the Federal Government has grown steadily. Between 1990 and 1996, the Federal share rose from 27.5 to 33.5 percent, while the private share fell from 60.0 to 53.9 percent. State and local governments were responsible for 12.6 percent of health expenditures in 1996, approximately the same proportion that they paid in 1990. The shift in the shares of spending between Federal and private funds is primarily the result of a dramatic slowdown in the growth of private sources of funding, principally PHI and consumer out-of-pocket payments. Growth in Federal funding, driven mostly by increases in Medicare and Medicaid payments, also decelerated somewhat, but remained well above private growth levels during the 1990-96 period.

The effects of health system changes are evident in the contrast between private and public sector financing responsibilities (Figure 5). From 1960 to 1990, growth in spending by both the private and public sectors were similar, with only two notable exceptions: the period 1966-67, when Medicare and Medicaid were introduced, and the period 1974-75, which recorded the effects of the 1973 expansion of Medicare to cover the disabled population. Each of these major expansions in public program coverages produced offsetting, step-wise shifts in public and private financing responsibilities, with the share shouldered by the public sector increasing. Although the number of people covered by Medicaid did increase rapidly between 1990 and 1994, the continued shift toward a larger public share since 1994 has not been driven, as it was in earlier periods, by public sector initiatives to add new populations or expand services. Public sector expenditure growth has continued at a slightly slower average annual rate since 1990 (9.4 percent) than it did between 1980 and 1990 (10.7 percent). However, average annual growth in private spending decelerated markedly between 1990 and 1996 to 4.9 percent, from the 13.1-percent average annual growth experienced during the 1980-90 period.

Private Health Insurance

In 1996 PHI premiums equaled $337.3 billion, up just 3.2 percent from 1995. This is the fourth consecutive year of decelerating growth and the sixth year of single-digit growth. The recent deceleration in premiums coincides with the dramatic shift by the health insurance marketplace away from traditional FFS indemnity insurance toward managed care.

The proportion of workers enrolled in managed care plans has skyrocketed in recent years, reaching virtually three-quarters of the enrolled workforce in 1995 (Jensen et al., 1997). As managed care plans have proliferated, price competition for market share has also increased. This competition greatly benefited employers, who found ways to exert additional pressure on insurers to limit plan price increases, including analyzing and negotiating premiums, forming purchasing coalitions, and limiting the number of insurers with which they contract (U.S. General Accounting Office, 1997b), as well as shopping for lower premiums.

In addition to efforts to control health care prices, aggregate spending on PHI premiums is affected by trends in enrollment. During the 1990s, the total number of persons with PHI appears to have leveled off,⁷ which can be partly attributed to the concentration of job growth in the service sector, where workers are less likely to be offered health insurance (KPMG Peat Marwick, 1992-96). Other factors include employers discontinuing insurance coverages, or employees dropping their employer-sponsored coverage because of increased employee costs and changes in benefits (U.S. General Accounting Office, 1997a).

Over the last few years, employers shifted more of the health insurance cost burden to their employees by requiring employees to pay an increasing share of the premium, particularly for dependent coverage. Between 1988 and 1993, the average employee contribution rose 64 percent for single coverage and 79 percent for family coverage (U.S. Bureau of Labor Statistics, 1994). In addition, changes to employee benefit packages may have discouraged enrollment. These changes included the introduction of flexible benefit plans, as well as basing premium rates for family health policies on family size (U.S. General Accounting Office, 1997a). Flexible benefit plans, or accounts, allow employees to allocate a fixed level of employer dollars or credits to a diverse menu of benefits such as child care, life insurance, health insurance, retirement savings accounts, vacation days, or cash payments, based on their personal needs. Thus, these plans can create an incentive for healthier workers to forgo health insurance altogether or simply to eliminate more expensive dependent health coverage. Similarly, linking the required employee premium for health insurance to family size increases the costs to employees with large families and may discourage dependent coverage.

From 1993 to 1995, the number of persons over age 65 covered by both Medicare and PHI policies declined slightly (National Center for Health Statistics, 1997). Many Medicare enrollees own PHI policies, commonly referred to as “medigap” policies, to cover expenses not paid by the Medicare program. However, the increasing costs of medigap insurance coupled with the availability of Medicare managed care plans may be affecting PHI enrollment for the elderly (Jeffrey, 1997). Medicare HMOs generally offer a rich benefit package and have low out-of-pocket costs. For example, Medicare HMOs typically offer outpatient prescription drug coverage and require only small copayments to utilize plan benefits.⁸ From 1992 to 1996, enrollment in Medicare HMOs almost doubled, from 6 percent of all enrollees to 11 percent (Health Care Financing Administration, 1997).

In 1996 Americans used $292.3 billion in PHI benefits, up only 3.4 percent over 1995. Controlling costs has been the primary focus of the PHI industry over the last few years. By negotiating prices and services with providers, emphasizing outpatient services rather than inpatient hospital care, providing more preventive services to enrollees, and utilizing less expensive treatment options to help patients avoid surgery and other more costly medical care procedures (Tanouye, 1997), the insurance industry has been able to slow down benefit cost growth.

Net cost of insurance (the difference between premiums earned and benefits incurred) was $45 billion in 1996. Since 1993 growth in benefits incurred has outpaced the increase in premiums earned. Competition for market share and the willingness of both employers and individuals to switch health plans to save money has forced insurers to keep premium increases below the growth in benefits (Ginsburg, 1997). Therefore, the net cost, which includes the cost of administering a health plan and plan profits, was squeezed. One report cites that only 35 percent of HMOs were profitable in 1996, down from 90 percent in 1994 (Center for Studying Health System Change, 1997).

Out-of-Pocket Spending

Out-of-pocket spending includes any payments for HSS made by individuals and not covered by health insurance. Such spending would include any copayments or deductibles that were paid by the privately insured population as a condition for receiving covered services. The share of HSS spending from private out-of-pocket sources continued to fall for the 11th straight year, reaching a low of 17.1 percent in 1996. Between 1990 and 1996, the growth rate in out-of-pocket spending dampened considerably, remaining well below the growth rate of PHI spending. Part of the slowdown was the result of low medical price inflation. Medical care price growth, as measured by the CPI, decelerated to 3.5 percent in 1996, the result of slower price growth across health care sectors and payers (Sensenig, Heffler, and Donham, 1997).

Another factor influencing out-of-pocket spending was slow growth in the aggregate copayments and deductibles required by third-party payers (KPMG Peat Marwick, 1992-96), which failed to keep pace with third-party payments. As more privately insured persons moved from traditional FFS to managed care plans, they faced flat out-of-pocket charges per visit or copayment rates that were frequently smaller than a percentage copayment required by traditional insurance plans.

Paradoxically, HMO plans, which have traditionally minimized enrollee out-of-pocket payments, recorded the largest increases of any plan type in cost-sharing in 1996. The percentage of employer-sponsored HMO plans with no copayment decreased from 15 percent in 1995 to 10 percent in 1996, while the percentage of HMOs requiring $5 copayments increased from 18 percent in 1995 to 22 percent in 1996 (KPMG Peat Marwick, 1996). For managed care plans requiring deductibles, notable changes were found only in preferred provider organization (PPO) plans, where deductibles rose for both in-network and non-network physicians (KPMG Peat Marwick, 1996).

Medicare

The Medicare program is the largest public payer for health care services and supplies. In calendar year 1996, Medicare's Hospital Insurance and Supplementary Medical Insurance programs financed $203.1 billion of spending for the health care of its 38.1 million aged and disabled enrollees.

Annual growth in Medicare spending slowed from 10.6 percent in 1995 to 8.1 percent in 1996. This deceleration reflects, in part, slowing medical prices, legislated limits that restrain the growth in Medicare payments to providers, penalties in the form of stricter limits on the growth in physician fees imposed on physicians for exceeding the Medicare volume performance standards in 1994 (as discussed previously), provider reaction to OIG fraud and abuse detection activities, and decelerating growth in the population eligible for enrollment in Medicare.⁹

Medicare payments to managed care plans increased from 4.8 percent of total Medicare expenditures in 1990 to 10.1 percent in 1996 (Figure 6). As this proportion grew, using the managed care methodology of distributing capitation payments based on FFS information led to greater distortions in overall Medicare expenditures by service. A new NHE methodology was developed for estimating the distribution of Medicare capitation payments to the services based on Adjusted Community Rating (ACR) forms submitted to HCFA.

As a result of this methodological change, the portion of Medicare spending allocated to hospital care decreased 2.7 percentage points, while the physician share increased 0.5 percentage points, and expenses for administration and the net cost of insurance increased 1.0 percentage points. For the first time, this report presents Medicare spending for dental services and for drugs and other medical non-durables.

The most notable differences in the 1996 distributions of FFS expenditures and capitated payments were in spending for physician services and for hospital care. Physician services consumed 38.4 percent of capitated payments and only 19.0 percent of FFS expenditures. Conversely, hospital care consumed 41.8 percent of capitated payments, compared with 60.1 percent of FFS expenditures. Administration (and the net cost of insurance) accounted for 9.1 percent of capitated payments but only 1.9 percent of Medicare FFS spending. In addition, 3.5 percent of capitated payments were for prescription drugs, a service not widely covered by the Medicare FFS program (Table 7).

Medicaid

Combined Federal and State Medicaid spending accounted for 14.7 percent of total HSS in 1996 and largely funded institutional services. In 1996 hospital care and nursing home care accounted for 61.2 percent (35.8 percent and 25.4 percent, respectively) of the $147.7 billion in combined Federal and State Medicaid spending for HSS. Medicaid is the largest third-party payer of long-term care, financing 47.8 percent of nursing home care in 1996. In fiscal year 1996, there were 36.1 million persons who received some type of Medicaid. Nearly one-half of all Medicaid recipients were children covered under Aid to Families with Dependent Children (AFDC) (16.7 million).

Medicaid is funded jointly by Federal and by State and local governments. For a State to receive Federal matching funds, it must adhere to minimum requirements for eligibility and services set by the Federal Government. Within this broad framework, State governments are afforded considerable flexibility in designing the total scope of their programs within the constraints of the State budgetary process. One way States employ this flexibility in Medicaid program design is through Medicaid waivers. There are two types of Medicaid waivers: program waivers (including home and community-based service waivers and freedom of choice waivers) and research and demonstration waivers. Home and community-based waivers (section 1915 (c) of the Social Security Act) allow States to place Medicaid-eligible persons into alternative, non-institutional settings for certain types of medical and personal care. Freedom of choice waivers, authorized under section 1915 (b) of the Social Security Act, allow States to place Medicaid beneficiaries into a mandatory managed care plan (where beneficiaries have a choice of a minimum of two providers). Research and demonstration waivers (section 1115 of the Social Security Act) allow Federal Medicaid requirements to be waived in order to conduct experimental, pilot, or demonstration projects.¹⁰

During most of the 1980s, the combined share of HSS financed by Federal and State Medicaid expenditures remained fairly steady, accounting for approximately 10 percent of the total. However, beginning in 1990, the Medicaid share of HSS began to grow rapidly, increasing by 3.5 percentage points in just 4 years, to 13.9 percent in 1993. By contrast, over the last 3 years, the pace of growth decelerated considerably, with the Medicaid share increasing just under 1 percentage point, to 14.7 percent in 1996.

Remarkably, between 1989 and 1993, Medicaid spending was growing at an average annual rate of 18.0 percent. Since 1993 Medicaid spending has risen at an average annual rate of just 7.0 percent. The reasons for the rapid acceleration in the early 1990s are: (1) the rapid escalation of disproportionate share hospital (DSH) payments, (2) increases in the cost of providing services to beneficiaries (the prices paid for services and the average cost per beneficiary), and (3) the growing number of program beneficiaries (U.S. General Accounting Office, 1997c).

The 1996 growth in Medicaid spending was 5.3 percent, the lowest annual growth since the inception of the program. The slow growth rate of total Medicaid spending mirrored the overall slow growth in health care spending nationwide in 1996. Several factors accounting for the slow 1996 growth in Medicaid spending are attributable to the generally favorable economic conditions that prevailed in 1996. Low unemployment rates reduced the number of people receiving welfare and consequently the number of Medicaid-eligible persons in many States. Similarly, historically low rates of medical price inflation held down increases in the cost of medical goods and services, which in turn dampened the growth in nominal spending per enrollee.

The Federal share of Medicaid spending grew very slowly in 1996, masking considerable variation in the growth rate of Medicaid spending among the States. In fiscal year 1996, growth rates among States varied from an increase of 25 percent to a decrease of 16 percent. States often have large changes in spending growth from year to year because of major program changes or accounting variances that change the fiscal year in which a portion of the expenditure is reported (U.S. General Accounting Office, 1997c).

Viewed in this context, the slowdown in State Medicaid spending can be attributed primarily to three factors that affected different States' expenditures in distinct ways. These factors are not common to all States and cannot necessarily be expected to recur. First, decreases in DSH funding as a result of program caps enacted by Congress in 1991 and 1993¹¹ slowed spending in some States. Second, slowdowns in State-initiated eligibility expansions substantially lowered the growth in Medicaid spending in several States. For example, although Hawaii, Oregon, and Tennessee implemented eligibility expansions in 1994, the increased expenditures associated with these expansions had leveled off by 1996. Finally, because Congress was considering legislation that would have established aggregate spending levels based on 1995 expenditures (as part of the block grant proposal), a few States accelerated payments into fiscal year 1995 that would ordinarily have occurred in 1996. These accelerated payments in effect shifted payments to 1995 and therefore lowered the rate of growth in 1996.

Medicare Revisions

The methodology for allocating Medicare spending to service categories has been revised to more accurately represent services funded by Medicare for beneficiaries enrolled in Medicare managed care plans. All Medicare enrollees receive coverage for a standard package of benefits. Medicare managed care enrollees may be covered for a wide variety of additional services such as routine physicals, preventive care, and prescription drugs. Medicare managed care funding was categorized in previous NHE estimates based on provider billings for FFS enrollees. The revised methodology provides separate service allocations for Medicare FFS and capitated payment expenditures (Table 7).

Before this project, financial information available from HCFA's Office of Managed Care reported total Medicare payments to managed care plans and separate amounts for services covered by the Hospital Insurance (Part A) and the Supplementary Medical Insurance (Part B) parts of the program. In 1996, 55 percent of the $20.6 billion Medicare paid to managed care plans covered Part A services, and the residual 45 percent covered Part B services. All of the Part A share was classified as hospital care in the national health accounts, and the Part B share was split, with three-quarters going to physician services and one-quarter to hospital care.

The revised methodology allocates Medicare managed care payments to both services and administrative expenses in the NHE estimates. These estimates represent a portion of expenditures funded by Medicare through capitation payments to managed care plans for Medicare beneficiaries who choose to enroll in managed care. Comprehensive statistics on specific services used by managed care enrollees are not reported to HCFA. Therefore, the service distribution of Medicare capitated payments was estimated from ACR forms submitted to HCFA annually by risk-type managed care plans. These forms are submitted for approval of the monthly premiums that the plan intends to charge and the services it intends to deliver to Medicare enrollees for the upcoming year.

Medicaid Revisions

Unlike Medicare, Medicaid has no national data sources useful for estimating Medicaid-purchased PHI service distributions. PHI payments continue to be allocated to services based on FFS Medicaid. Revisions incorporated this year include refinements to this FFS methodology.

The first refinement is in the estimation of the administrative portion of Medicaid PHI payments. Administrative costs were estimated by multiplying Medicaid capitation and insurance premium payments by cost-to-premium ratios calculated from PHI industry data. The resulting administrative costs were then removed from estimates of Medicaid insurance payments and added to estimates of Medicaid administration, prior to distribution to services.

The second refinement respecified the services to which Medicaid PHI¹² payments, coinsurance, and deductibles would be allocated. These refinements more accurately identified services likely to be purchased by Medicaid managed care and PHI.

Census Revisions

Expenditure estimates for nursing home care and each of the professional service components of personal health care (physician services, dental services, other professional services, and home health services) were revised to incorporate changes resulting from more recent information received from the Census Bureau. Estimates of spending for these service categories are based on business receipts of service establishments collected in the Census Bureau's quinquennial Census of Service Industries. Establishments are classified by industry according to the 1987 Standard Industrial Classification Manual (SIC) (Executive Office of the President, 1987). The most recent data available from the 5-year census are for calendar year 1992.

Information from the Census Bureau's Services Annual Survey (SAS) supplement data from the quinquennial census. The SAS provides estimates of year-to-year change in business receipts of firms by SIC classification. Annually, the SAS surveys a sample of service businesses. Every 5 years the sampling frame is changed to reflect the industry composition of each SIC component from the most recent quinquennial census. The sampling frame was revised in 1996 to reflect SIC industry compositions determined from the 1992 quinquennial census. Preliminary results from the 1996 survey altered growth rates obtained from earlier SAS surveys. To reflect these changes, the NHE estimates for 1993-95 were revised. A discussion of the SAS sample design and the sampling frame are presented in the 1996 Services Annual Survey (U.S. Bureau of the Census, to be published).