services declined from 15 percent to 5 percent. The number of workers having to pay less than 20 percent coinsurance dropped from 80 percent to 41 percent in the same period.74 Although the most prevalent deductible in employer-sponsored insurance remained $100 from 1980 to 1988, the number of participants with a deductible of $150 or more rose from less than 10 percent to 40 percent.75 Many would applaud these benefit changes, on the grounds that they encourage employees to be more responsible in the use of health services. Moreover, these changes were accompanied by an additional benefit: increases in the number of workers in plans that limit potential out-ofpocket costs for enrollees, enhancing "catastrophic" protection. 76

More troubling is increasing reliance on a second. way of shifting costs to employees: reducing the employer's contribution to the basic insurance premium. The proportion of employees paying $30 or more per month for individual coverage rose from 3 percent in 1986 to 13 percent in 1988. For family coverage, the proportion paying $100 or more per month went from 5 percent in 1986 to 16 percent in 1988. Average employee payments for their own coverage rose 32 percent between 1988 and 1989, while the workers' share for family coverage rose 20 percent.77

Employers' cost saving efforts have begun to arouse employee resistance and have become a major source of labor unrest. Health care concerns were at the root of an estimated seven out of eight of the nation's labor-management disputes in the last two years. 78 This trend can be expected to continue.

"Ultimately, health care costs cannot be controlled at the bargaining table. Thus CWA and the IBEW also proposed that [AT&T] work with us to promote and achieve health care reform."

-James Irvine, vice president, Communications Workers of America

Growing anxiety about the burden of retiree health benefits compounds the pressure employers face. Many businesses provide full health coverage to retirees who have not yet reached age 65 and supplemental coverage to retirees who are eligible for Medicare. But future health benefit obligations do not have to be fully funded or carried as a liability on corporate accounts, unlike pension benefits.

“We are a large company, with more than 300,000 employees. The CWA and the IBEW are large unions with hundreds of thousands of members. Out of this last round of bargaining came the realization that, as large as we are, the health care dilemma is bigger than AT&T and its unions. It's truly a national issue." -Charles Brumfield, director, labor relations, AT&T

The General Accounting Office (GAO) has estimated that, as of 1988, the unfunded health benefit liability for persons already retired was $221 billion, and the potential liability if corporations were to keep their promises of retiree health benefits for their active employees was $402 billion.79 The Financial Accounting Standards Board (FASB), whose minimum requirements for corporate accounting are widely used, has decided that the future obligation for retiree health benefits should be shown as a liability on corporate accounts. If the FASB's decision is followed, and these enormous obligations must be added to corporate accounts, some of America's largest businesses could become technically insolvent-a threat that increases the pressure on employers to find some way of reducing their health care liabilities.

Both labor and management, then, must view rising health care spending as an increasing threat to their economic well-being.

Are We Getting Value for Our Dollars?

The United States spends a greater proportion of its wealth on health care than any other country. United States health care expenditures as a percentage of gross domestic product (GDP) have been substantially higher than in any of the other 23 members of the Organization for Economic Cooperation and Development (OECD), which includes other Western industrial nations. In the United States, health expenditures as a percent of GDP were 10.8 percent in 1987.80 In other countries, health care spending as a percentage of GDP ranged from a low of 5.3 percent in Greece to a high of 9.0 percent in Sweden. Canada and France each spent 8.6 percent. On a per person basis, the United States spent $2,051, while Canada, the next highest, spent $1,483.81

With costs so high and continuing to rise, Americans have come to question whether we are getting our money's worth. Even with such high health care

spending the United States falls behind many countries that spend much less on health care, as measured by infant mortality, life expectancy, age-adjusted mortality, and other basic health indicators. In 1986, for example, the infant mortality rate for the United States was higher than that of 21 other countries, including Singapore, Spain, Italy, and the German Democratic Republic. 82 In 1986, the United States ranked nineteenth in life expectancy at birth for males and fifteenth for females, behind such countries as Japan, Sweden, Canada, Greece and Spain. 83 Among the OECD nations, the United States ranked twelfth in total age-standardized death rates per 100,000 for the early 1980s, behind countries like Japan, Sweden, Canada, Iceland, and Greece. 84

Mortality statistics for impoverished populations within the United States suggest that pockets of Americans fare worse than citizens of some of the world's poorest nations. Bangladesh is one of the lowest-income countries. Nonetheless, a recent study found that black men born in central Harlem, a lowincome area in New York City are less likely to reach age 65 than men in Bangladesh. Women in Bangladesh, if they survived past the age of 10, also could be expected to live longer than women in Harlem. The high mortality in Harlem was partly due to the large amount of violent crimes and substance abuse in that community. But the greatest single source of the high mortality rate was cardiovascular disease, suggesting that lack of access to services contributed to these statistics. 85

Care in the United States may be superior in many ways that are not measured in these gross statistics. And these statistics reflect many factors other than medical care. Still, as U.S. health care expenditures continue to rise, many question whether we are using our money wisely. Increasingly, these questions focus on the necessity and appropriateness of the medical care most Americans receive.

What Is Appropriate Care?

There is no single prescription for high-quality health care, and a wide range of practices may be accepted for specific conditions. Thus, a lot in medicine falls in the gray areas of uncertainty where a physician's judgment becomes critical.

The most striking evidence of the effects of uncertainty in medicine is found in research on geographic variations in practice patterns, or what is called "small area analysis." These studies have found significant geographical variations in average length of hospital stays, hospital discharge rates, and surgical

procedures. Researchers discovered, for example, that the likelihood that a woman in Maine would have a hysterectomy by the time she reached 70 years of age ranged from a low of 20 percent in one community to a high of 70 percent in another.86 In comparing the practice of medicine in Boston and New Haven (cities otherwise similar in medical resources and population health status), researchers found that Boston doctors were twice as likely to perform carotid endarterectomies but only half as likely to do coronary bypass surgery.87 Boston physicians were much more likely to hospitalize Medicare patients than were their colleagues in New Haven.88 Areas with a high frequency of procedures are not necessarily those providing a lot of inappropriate care. Sometimes, a small number of physicians in high use areas are performing a large number of procedures, judged to be appropriate given patient indications. 89

These variations cannot be fully explained by differences in the health status of the populations or by standard physician characteristics such as age, specialty, or medical school. The most important factor seems to be differences in the practice styles of physicians.90 While the resulting care may be appropriate, often the procedure chosen is not superior to alternative treatments in producing net benefits for the patient.

Which practice styles best produce desired patient outcomes is the central concern of appropriateness research (which is sometimes done within small area studies). This research uses a number of methods for identifying inappropriate use, including audits of medical records, second opinions for surgical procedures, and examination of pathology specimens after surgery, 91

Such research has determined that common procedures such as angioplasty, coronary artery bypass surgery, cardiac pacemaker implantation, and Caesarean section deliveries, are often used without producing any medical benefit for the patient.92 For example, 32 percent of carotid endarterectomies, 17 percent of coronary angiographies, and 17 percent of upper gastrointestinal endoscopies have been found to be inappropriate. 93 A review of the appropriateness literature by the GAO found that the inappropriate use of surgical procedures ranged from 14 percent to 32 percent, and inappropriate hospital admissions ranged from 7 percent to 19 percent. 94

Unnecessary and inappropriate services not only waste the health care system's limited resources and deprive others of needed care but also pose risks to patients. An unnecessary Caesarean section, for example, can lead to infections, a prolonged hospital stay,

and psychological distress for the newborn's mother. Inappropriate use of antibiotics may expose patients to unjustified side effects and result in the proliferation of antibiotic-resistant organisms.95 The Food and Drug Administration (FDA) reported that in 1986, adverse drug reactions caused more than 1,300 deaths and over 4,400 hospitalizations. 96

Analysts also differ about how much of medicine is of unambiguous value to patients. Some have estimated that as little as 15 percent to 20 percent of all contemporary clinical interventions are supported by objective scientific evidence that they do more good than harm. 97

The validity of the evidence supporting various procedures has also come under question. The Office of Technology Assessment (OTA) has estimated that only 10 percent to 20 percent of medical practices are supported by randomized controlled trials.98 A review of statistical methods in the published medical literature concluded that as many as half the articles that used statistical methods did so incorrectly. Another study found that the methodologies reported in 39 scientific papers on the use of percutaneous transluminal angioplasty were deficient for evaluating the procedure. None of the papers considered outcomes that patients might think important, such as relief from pain and continued ability to walk.99

Not all quality problems reflect overuse of service. Underuse is a quality problem as well. Some underuse reflects barriers to access. Underuse of services that is not due to direct access barriers may be classified as underdiagnosis or undertreatment. 100 undertreatment. 100 Malpractice suits frequently allege that the physician or provider underdiagnosed or undertreated by missing a diagnosis or failing to follow-up on an abnormal test result. 101

Medicare Peer Review Organizations (PROS) include various screens for undertreatment in their quality reviews. For example, each PRO has to screen a sample of medical records to determine whether a patient was discharged prematurely or received an adequate discharge plan. These screens have detected few cases of underservice of Medicare patients, however. More prevalent, it would appear, are other types of underservice, such as less complete diagnostic services and fewer rehabilitation services than experts judge the elderly need. 102

Bad medicine resulting from errors, negligence, incompetence or unprofessional behavior also jeopardizes the quality of care. Knowledge of such problems results from existing quality assurance mechanisms— such as Medicare Peer Review Organizations, the Joint Commission on Accreditation of Healthcare

Organizations, utilization management entities that oversee the quality of care for the nation's larger employers, and inspections by state survey and certification teams. 103 Bad medicine can take many forms, including patients being discharged prematurely from hospitals; errors in diagnosis, medication, and treatment; faulty laboratory tests; hospital-induced infections and other complications resulting from medical treatment; and similar misadventures.

Malpractice Litigation: Protection or Problem?

Malpractice litigation provides consumers with a private course of action when they believe that the quality of the care required falls below acceptable standards. 104 At the level of the individual patient, poor quality medicine is often the catalyst leading to such a suit. But the evidence is strong that malpractice litigation does not adequately protect against or compensate for poor quality care.

A recent study of malpractice claims in New York showed that adverse events due to negligence occurred in 1.0 percent of all hospital discharges. Eight times as many patients suffered an injury from negligence as filed a malpractice claim, and 16 times as many patients suffered an injury from negligence as received compensation from the tort liability system. 105 This is consistent with evidence from the mid-1970s suggesting that at most one in 10 persons injured as a result of negligence actually filed a claim, and evidence from claims data of the early 1980s that one in five persons claiming negligence brought suit. 106, 107

The same study also found that fewer than half the claims made by patients lead to payments by physicians or their insurers-suggesting that negligence did not occur.

Not only does malpractice litigation provide limited protection; it creates cost, quality, and access problems of its own. The costs of malpractice premiums raise the costs of medical care. And the risk of litigation may distort physicians' practice patterns.

Total malpractice premium costs of doctors and hospitals were over $6 billion in 1987, about 2 percent of dollars spent on these services. 108 Professional liability premiums paid by self-employed physicians nationwide averaged $15,900. There is a great deal of variation in premium by specialty and by geographic location. Of the surveyed specialties, obstetricians/ gynecologists averaged the highest premiums ($35,300), while psychiatrists averaged the lowest

out of particular geographic locations. This problem is seen most sharply in studies of obstetrical services in low-income areas, where physician clinics are reducing this part of their practice, they say, due to the combined pressure of Medicaid cost containment and rising malpractice premiums. 110

A recent Institute of Medicine study concluded that, because of liability concerns, significant numbers of each of the physician groups examined were reducing obstetrical services to high-risk women, eliminating obstetrical practice altogether, or dropping it earlier in their careers than they otherwise would. Similarly, the American College of Obstetricians and Gynecologists (ACOG) in a 1987 survey of its membership found that its members were reducing or eliminating obstetrics as a result of the risk of malpractice. 111

It is difficult to measure how much defensive medicine exists or how much it costs. By one estimate, about $11.7 billion out of the $82.8 billion paid to physicians in 1985 was attributable to service-increasing defensive medicine.112 These numbers may overestimate the amount of potential harm, of course, because much of what might be labeled defensive medicine is good medical practice. But it is significant that a large portion of the medical profession believes that physicians have modified their practice patterns, and in some cases left the profession, in response to both malpractice premiums and their perception of the legal consequences of their medical decisions.

UNDERSERVED AREAS AND SPECIAL POPULATION GROUPS

The Commission's deliberations on health care focused chiefly on the financial barriers to access. But there is more to access than simply paying the bills. Even if health care were financed in full for everyone, some people would still have trouble obtaining adequate care for a variety of reasons-geographic, systemic, cultural, and personal. Like financial barriers to access, these nonfinancial barriers often result in people getting care later or less often than needed. As a result, their medical conditions become more severe and require more acute, and therefore more costly, treatment.

Nonfinancial barriers to access apply primarily to those living in medically underserved areas-such as rural communities or inner cities-and to members of population groups with special needs-such as the poor, substance abusers, the mentally or physically handicapped, and pregnant teenagers.

Medically Underserved Areas

Both rural and inner city areas have trouble recruiting or maintaining physicians and other health care professionals-in part because of the perceived undesirability of living or working in these areas, and in part because of the economics of maintaining a practice in places with large numbers of uninsured.113 In 1988 there were 33 million people living in federally designated health manpower shortage areas (HMSAs). This population was about equally divided between urban and rural areas. 114

Rural Areas-In 1985, there were 97.9 patient-care physicians per 100,000 population in rural counties, compared with 174.7 patient-care physicians per 100,000 population in urban counties. In rural counties with fewer than 10,000 population there were only 53 physicians per 100,000; and in counties with fewer than 2,500 residents there were only 29.9 per 100,000.115 Although physician supply has increased recently in the more populated rural areas, it is expected to decline in the less populated areas due to a projected 25 percent rate of physician retirement in rural areas and severe cutbacks in the National Health Service Corps (NHSC) program, which supplies physicians to HMSAs. Even areas that can hold on to their physicians may be deprived of some essential services.116 The problem of too few providers in rural areas is compounded by a general decline in the growth of primary care doctors, who are uniquely suited for rural practice and the shortage of nurses and other midlevel practitioners in rural areas. 117, 118

Problems facing rural hospitals and health centers aggravate the situation. Low patient volume (because patients are so widely scattered in rural areas), Medicaid and Medicare reimbursement policies, rising numbers of patients unable to pay, and personnel shortages have forced an increasing number of rural hospitals to close their doors-all trends expected to continue.119 Community and migrant health centers-important sources of primary care in some rural communities-have been facing increasing financial strains due to the failure of federal funding to keep pace with inflation, the increase in patients without insurance, rising malpractice premiums, and NHSC's cutbacks. 120

Inner Cities-In 1988, 17 million residents lived in medically underserved urban areas-primarily inner cities. These areas are unattractive to many physicians and may not be able to support an economically viable medical practice. 121 The shortages are particularly acute for office-based primary care physicians. One study found that, between 1963 and 1980, the

availability of such physicians declined in poverty areas, despite growth in the overall number of practicing patient care physicians. 122 This puts even greater burdens on the hospitals, which are beginning to close emergency rooms and trauma centers to avoid the increasing demand for uncompensated care. Finally, community health centers in urban underserved areas face the same financial and personnel strains as their rural counterparts.

Special Needs of Special Populations

Even those with the means to pay for care and providers to supply it may not have access to care for a variety of logistical, physical, mental, cultural, or personal reasons. The following are just a few examples of groups that face particular hurdles in obtaining care, although each of the barriers could apply to anyone. 123

The Poor-Even if their health care is paid for, persons in poverty may face significant logistical barriers to care. For those who cannot afford to maintain a car, take public transportation, or pay taxicab fare, travel to medical care is a serious problem. When public transportation is available and affordable, the effort required to get from home to the source of care may be extraordinary. 124 Providers' limited office hours can pose another obstacle for the working poor who cannot take time away from their jobs.

Minorities-Minorities are more likely than average to be poor and thus to face the access barriers associated with poverty. They also face additional barriers arising from residence in the inner city, discrimination, and a variety of socioeconomic and cultural factors that make some individuals less disposed to seek necessary services even when those services are available. 125

High-Risk Pregnant Women and Their Infants-An estimated one-third of all mothers receive insufficient prenatal care. Of these, the highest proportions are black, low-income, unmarried, or still children themselves. 126 Even those with insurance coverage may not obtain services because of child care or transportation problems, ignorance of the importance of prenatal care, or inability to find an acceptable provider.

The Geographically Isolated-Despite their access to transportation, people may have trouble getting to the source of health care in a timely manner if they are geographically isolated. Living in isolated areas also makes it more difficult to obtain home care and social support services, such as counseling, 127