nous (IV) drug users in that city, compared to the predominance of gay men in San Francisco. Female IV drug users and women infected by their sexual partners who are IV drug users represent a much larger proportion of AIDS cases in New York than in San Francisco.

AIDS cases also vary significantly by race and risk group. San Francisco's AIDS patients are more likely to be white, whereas New York's are more likely to be black or Hispanic. This reflects general demographic differences in these cities. Compared to all AIDS patients in the US, patients in San Francisco are much more likely to be homosexuals, whereas patients in New York consist of a larger proportion of IV drug users.

Heavy incidence among IV drug users affects another patient group-children. In New York, where IV drug use and sharing of contaminated needles is more common than in the rest of the country, 79% of children with AIDS were presumably infected during pregnancy by mothers who were either IV drug users themselves or who had been infected by IV drug users.16 The proportion of IV drugrelated AIDS cases among children is likely to rise because of the continued spread of AIDS among drug users and the decline in the number of cases transmitted through contaminated blood and blood products (as a result of improved screening of the nation's blood supply).

Two patient groups remain hospitalized longer than others-IV drug users and children with AIDS. Both groups are concentrated in New York City. There have been 2,570 AIDS cases among IV drug users and 166 cases of pediatric AIDS reported in New York City through 1986. This represents 52% and 40% of the nation's cases in the respective groups. Only 33 IV drugrelated and seven pediatric cases were reported in San Francisco during this time. (The figures for both cities do not include approximately 8-13% of cases among homosexual or bisexual men who also report using IV drugs.)

The municipal hospitals in New York are overwhelmed with IV drug users, a group which comprises 60% of all AIDS patients hospitalized there. These patients are generally indigent, have poor underlying health status, have fewer sources of social support, and are more likely to be homeless than other patient groups. The primary AIDS diagnosis for many IV drug users is Pneumocystis carinii pneumonia (PCP). In general, this opportunistic infection is more debilitating and requires longer and more frequent hospital stays than other AIDS-related conditions, the most common of which is Kaposi's sarcoma, which is often treated on an outpatient basis.17 In a study of AIDS patients at San Francisco General Hospital (SFGH), the average length of stay for AIDS patients with Kaposi's sarcoma was 7.6 days, compared to 18.1 days for those with PCP. Furthermore, during 1984, Kaposi's sarcoma accounted for 13.3% of the total AIDS admissions at this hospital, as compared to only 3.5% of AIDS admissions in the public hospitals in New York City. All these factors increase the likelihood of a longer hospital stay and an increased direct cost of care for AIDS patients in New York. Hospitalization of children with AIDS has not been a major problem in San Francisco, since there have been so few cases. New York, by contrast, has had to address issues arising from pediatric AIDS, including the unusually long length of stay for these patients. As discussed above,

more than three quarters of the pediatric AIDS patients in New York have parents who are or have been IV drug users. Many of these parents may be ill, and some have died from AIDS. In addition, there is a lack of alternative facilities to care for children outside the hospital, and it has been extremely difficult to place children with AIDS in foster care. Again, this is consistent with the situation confronting adult AIDS patients, who, in New York's tight housing market, also face difficulty in finding available space in nursing homes, and thus are likely to have prolonged hospital stays. A one-day survey conducted in New York's municipal hospitals found nine pediatric AIDS patients awaiting nonhospital placement. All were homeless, and all were candidates for home care. Three of these patients had been on an alternate level of care status (ie, those patients who no longer require acute care hospital services and are ready for placement in nonhospital settings) for 12, 14, and 18 months. 18 Thus, overall, the AIDS epidemic is quite different in New York than in San Francisco. New York is confronted with approximately three times the number of cases, and the patients are more heterogeneous than those in San Francisco with respect to race, age, sex, income, and major risk factors. In effect, the differences in scale and diversity of patients in New York, when compared to San Francisco, present local policy makers with two quite different epidemics of the same disease.

Preexisting Social and Political Structures. The local responses to the epidemic did not occur in a vacuum. They grew out of and were influenced by the social and political structures already existing in each community. Each city's historical development-in size, number, and type of institutions providing health services; methods of payment for health care; and the roles of public agencies-set the stage for what could happen in the epidemic. We differentiate two components in the preexisting social and political structures: one is related to the community and its institutions, especially the structure of health and human service organizations; the other is related to the social and political structures of those persons in the major risk groups.

The scale and complexity of health and human service organizations is far greater in New York than in San Francisco. The difficulties associated with a larger scale in New York range from the simple logistics of epidemiologic case monitoring to the institutional complexities of developing a response in 11 acute care municipal hospitals and over 70 other private and voluntary hospitals scattered unevenly throughout five boroughs. This contrasts with the one city health care facility and 13 other hospitals in San Francisco. In addition, the existence of a single medical school in San Francisco has fostered a relatively unified response to the epidemic, while in New York there is a history of competition in selected areas among the seven medical schools, making a similar response much more difficult.

The cities are similar in having a complex role in the financing of health care for part of their populations. These roles have been illuminated by the AIDS epidemic and the enormous resources needed to cope with it. One of the major differences in the two cities is how local tax monies have been allocated. One report estimated that

22-690 0 - 90 - 2

MAY 1987/NEW YORK STATE JOURNAL OF MEDICINE 267

more than 90% of New York City AIDS funds are spent on inpatient hospital care, largely through the city government's 25% share of Medicaid expenditures, and direct subsidies to the NYCHHC and voluntary hospitals providing indigent care.19 This results in sizable expenditures of local funds in New York because of the large number of AIDS patients who are enrolled in the Medicaid program. In comparison, in San Francisco only approximately 25% of municipal AIDS funds are spent on inpatient care. The large difference is due in part to San Francisco's local government not contributing to the Medicaid program. This reduces the cost of hospitalization to the city, while increasing the cost to the state of California.

These differences point out the importance of localstate relationships in responses to AIDS, especially in funding. In New York, the state health department has increased its influence vis-a-vis the New York City Health Department through its statutory responsibilities in Medicaid and in voluntary hospital financing in general. In addition, following the fiscal crisis of the 1970s, the state assumed other responsibilities that were formerly the province of city agencies. For example, administrative control over substance abuse programs was shifted to the state during the late 1970s, and it is these agencies that are trying to cope with the growing epidemic among IV drug

users.

20

Both cities help pay for the costs of hospitalizing medically indigent patients who are not covered by any thirdparty payers. This has beer an especially acute problem with AIDS patients, many of whom are treated with experimental drugs because no efficacious drug regimen is available. Third-party payers, both private and public, have been reluctant to pay for hospital care when experimental drugs are used in treatment. Thus, both local governments have been forced to subsidize treatment costs. Even more important for the cities' budgets is the fragmented mix of policies that have resulted in a large number of AIDS patients joining millions of other Americans who "fall through the cracks" and have inadequate health insurance coverage. It is estimated that more than 55 million people in this country have little or no health insurance at all. These patients, who are generally confined to municipal hospitals, are subsidized by local tax revenues in both cities.

The payer mix for AIDS patients in the municipal hospitals of each city is quite similar (Table IV). In the mu

nicipal hospitals, those patients not covered by private insurance, Medicaid, or Medicare usually have no other insurance coverage and are subsidized by local government. In San Francisco's nonmunicipal hospitals, Kaiser Permanente, a health maintenance organization, covers at least 50% of those patients in the "other" category, reducing the overall level of uncompensated care to approximately 13% of the total AIDS inpatient caseload in that city.

The budgets of the cities in the years preceding the epidemic also may have influenced the extent to which each committed financial resources to the epidemic. The ability of local government in San Francisco to respond to the AIDS crisis with substantial funds as early as 1983 was made easier by the existence of a budget surplus that existed during the early 1980s (but is now gone). In contrast, New York was just coming out of a protracted period of financial crisis and near bankruptcy in the late 1970s.

Another factor that existed before the epidemic and influenced local responses was the different roles of the local health departments. In San Francisco, the health department incorporated traditional public health functions, outpatient clinics, and San Francisco General Hospital under one bureaucracy. This strengthened local public health decision makers in their position that the health department should take the lead in responding to the AIDS epidemic through direct action and active coordination, funding, and support of other organizations. In New York, the health department's functions were separate from hospital services, which fell within the bureaucracy of the NYCHHC. Thus, the health department in general "should provide those services that others have not, will not, should not, or cannot provide," 22 ie, its role is to fill the gaps in the system. In an early response to the AIDS epidemic, the department called a meeting of all interested parties in the city and defined its role as "seeking not to direct, but to provide a neutral meeting ground." 22

A further difference lies in the difficulties of coordinating the diversity of services that can be used to address the AIDS epidemic. In New York, three large, separate city bureaucracies, the Health Department, the Health and Hospitals Corporation, and the Human Resources Administration, each developed programs to deal with AIDS, with little formalized central coordination. In San Francisco, these functions were under the authority of one agency, and this allowed extensive health department organizational resources to be used to encourage the integration of community-based, ambulatory, and institutional programs.

The preexisting social and political structures of persons in the major risk groups are a second component that helps explain the differences in the two cities. In San Francisco, AIDS has predominantly affected white, middle-class, homosexual men. In New York, at least half the cases are reported among minorities, and almost a third are known IV drug users-an impoverished, highly stigmatized, and politically powerless stratum of society. In a 1984 survey, the number of openly homosexual or bisexual men in San Francisco was estimated to be approximately 69,000, or 24% of all men aged 15 years and older.23 The homosexual community was also found to be relatively affluent and well-educated, with 44% having

pre-tax earnings of $25,000 or more, 57% having college degrees, and 21% having postgraduate degrees. A preliminary survey in New York indicates a similar socioeconomic background for homosexual men.24 However, given the distribution of cases in other patient groups (eg, IV drug users) and by race, AIDS has afflicted a poorer and less educated population in New York.

The roles of the risk groups in the economic and political life of these communities are markedly different. The homosexual community is a well-organized, visible, and powerful political force in San Francisco, with openly gay representatives holding elective office in local government since the late 1970s. There is no comparable group of gay political power in New York. Thus, in San Francisco, the major identified risk group has been represented by organized spokesmen who were an integral component of the local political establishment even before the AIDS epidemic began.

Such visibility and political power for the major risk group in San Francisco is probably reflected in the differences in newspaper coverage of AIDS by The New York Times and The San Francisco Chronicle. Although these two papers play somewhat different roles, in that the Chronicle is viewed as more locally oriented than the Times (and therefore would be expected to put more emphasis on a locally important issue), the more extensive AIDS coverage by the Chronicle is still noteworthy. Between June 1982 and June 1985, 442 articles about AIDS appeared in the Chronicle, of which 67 were front-page stories. This compares to 226 articles published in the Times, of which seven were page-one news. The extensive nature of coverage by the Chronicle, aside from providing a degree of health education not found in New York, helped to sustain a level of political pressure on local government and health officials to respond to the AIDS crisis. The difference in coverage is also probably a reflection of the relative importance of the AIDS epidemic among other local issues. In San Francisco, AIDS was clearly a top issue; in New York, it had to compete for attention with a multitude of other issues.

Thus, the preexisting social and political structures of the institutions and major risk groups reinforced the differences in the epidemiology of AIDS in the two cities. In New York, the complexity of coping with AIDS was heightened by the heterogeneity of the risk groups and their diffuse political and social organization, by the number and diversity of organizational actors in the health and human services sector, by the roles of public agencies in financing health care and in responding to the epidemic, and by the relative importance of AIDS as a local policy issue. In San Francisco, on the other hand, both the epidemiology and the social and political factors put fewer constraints on those agencies that responded to the epidemic. Indeed, the conditions were that a well-educated, highly organized, and politically active group of citizens were at risk of AIDS in a city with relatively well-coordinated relationships among the major public health care institutions. The epidemiology and the social and political factors contributed, directly and indirectly, to some of the differences between the two cities through their influence on the cities' policy responses.

Policy Responses in New York and San Francisco. In

both cities, the AIDS epidemic has generated a largely unanticipated need for a variety of medical, public health, social, and educational resources. For example, active disease surveillance and epidemiologic investigations have been necessary to track the epidemic, to establish routes of transmission, and to plan for future health care resources to service the needs of those afflicted. AIDS patients require hospital care, outpatient services, social support and, in many cases, housing and food. Members of highrisk groups need public health education, counseling, and specific guidelines regarding behavior to prevent transmission of the disease. The population outside the main risk groups needs education to understand the risks and how the disease is transmitted in order to reduce unwarranted fears, to promote support for effective programs, and to prevent further spread of the epidemic. The policy responses in the two cities, as reflected in agency behavior, followed paths that were consistent with the differences in epidemiologic factors and social and political structures. In 1982, the San Francisco Department of Public Health (SFDPH), through its Office of Lesbian and Gay Health, began to coordinate efforts to plan and develop services to deal with AIDS. By 1983, in efforts to avoid duplication of services by a growing number of organizations and to coordinate the city and county's responses, a separate AIDS Activity Office was established within the SFDPH with the following specific purposes: to better coordinate and link the continuum of services related to the AIDS epidemic; to identify service gaps and develop plans for addressing these needs; to oversee, monitor, and support AIDS-related contract services; to anticipate funding requirements and justify new funding requests; and to develop and maintain the department's liaison with certain external funding and service agencies, community groups, and education programs. However, significant local funding for AIDS programs in San Francisco did not begin until FY 1984, when expenditures increased to $4.3 million from $180,000 the previous year.8

During the early years of the epidemic in New York (1981-1984), there was no centralized governmental response. Although an Office of Gay and Lesbian Health existed within the Department of Health since 1982, it was an advisory body with no policy-making authority. More importantly, it had less political power than the comparable office in San Francisco due to its location within the health department and its inability to influence other city bureaucracies that deal with AIDS in New York.

In November 1984, at the direction of the mayor, an AIDS Policy and Planning Committee was formed. Its first order of business was to send a fact-finding delegation to San Francisco (January 1985) to study that city's response and to make recommendations for improving New York's services to AIDS patients.

The delegation's foremost recommendation was to improve administrative coordination among all city agencies dealing with the AIDS crisis through the Policy and Planning Committee. This body was formalized in March 1985, and its charge was to oversee planning, coordination, monitoring, and evaluation of all the AIDS services provided by the city directly or by contract. Other recommendations called for increased resources for the surveil

lance efforts of the health department, and enhancement of acute and chronic medical services (including outpatient services), the provision of housing and social services, and educational programs. Most of these recommendations and other AIDS programs were implemented by the beginning of 1986, with a large increase in city appropriations during FY 1986.9

In San Francisco, the existence of an office within the health department devoted to gay and lesbian health issues provided a natural organizational focus for AIDS issues as the epidemic developed. The relatively small scale of the city, the recognized importance of homosexuals in local politics, the relatively high educational and income levels of the gay community, and its social organization all contributed to the development of an integrated system of outpatient and community-based services for AIDS patients. These services not only provide an important part of the care for AIDS patients, they have a financial impact on the local health care system by allowing many patients to remain outside the hospital for longer periods and, once admitted, to be discharged earlier. The range of community services available include outpatient clinics, public health education, hospice care, legal services and entitlements advocacy, psychosocial counseling, substance abuse services, and practical support and housing for AIDS patients.

While these services now exist in each city, the timing of their development and sources of funding differed. For example, outpatient clinical services for AIDS patients have been available at SFGH since September 1981. In the fall of 1982, SFGH established a special multidisciplinary clinic (infectious diseases, oncology, psychiatry, psychology, nursing, dermatology, and social work). At the University of California in San Francisco (UCSF), another outpatient clinic for AIDS patients opened in August 1984. During 1985, the clinics at SFGH and UCSF averaged approximately 1,000 patient visits per month.8

In New York City, the first outpatient clinic for AIDS patients was opened at Kings County Hospital in early 1984. During 1985, it was open two afternoons per week and handled between 120 and 175 patient visits per month. The first full-time AIDS ambulatory care clinic was opened in Manhattan in the spring of 1985. Slowness in providing comprehensive AIDS outpatient services in New York has probably increased the level and expense of hospitalization there.

The diversity of AIDS patients in New York makes the process of developing a useful range of services far more difficult. Developing services appropriate for both IV drug users and children with AIDS has obviously been much more of a problem for New York than for San Francisco. On the other hand, the highly successful, specialized inpatient AIDS ward at San Francisco General Hospital, where most patients are middle class gay men, has benefited from the homogeneity of its patients and extensive support from the gay community. Attempts to develop such a ward in a setting where the majority of patients are substance abusers would obviously entail difficulties. Other community-based programs in San Francisco, such as the Shanti AIDS Residence Program, which provides longterm housing for AIDS patients (and thereby helps keep patients out of the hospital), are not designed to accom

modate patients with substance abuse problems, and, in fact, Shanti does not accept such patients into its program. Thus, the ability of San Francisco to develop a range of successful community-based services is due not only to good planning and a commitment of local resources, it is a reflection of the strength and importance of the gay community, whose members comprise the majority of patients and those at risk.

To support its community-based services, New York City has relied more on state and private funding and less on local funds compared to San Francisco. This may be due to the transfer of responsibility for various services from New York City to New York State government after the fiscal crisis of the 1970s, or it may be due to a paucity of local political power among the groups at risk. It may also reflect the enormous management and planning difficulties-rooted in the complexity of the epidemic in New York of local government officials as they attempted to develop programs. For example, a city-funded program for home attendant services to persons with AIDS provided care to only a fraction of those in need. An audit of this program found that from its inception in December 1983 through February 1985 a total of only 80 clients had received service, although it was designed to serve a 200client caseload.25 The report elaborated a series of reasons for the program's failure to meet its goals, including poor contract planning, lack of needs assessment, enrollment limited to those who were eligible for Medicaid, excessive processing time, lack of housing, and poor community outreach.

Nevertheless, community-based AIDS organizations have played a key role in responding to the epidemic in both cities. They have provided an important and otherwise missing dimension to patient care and have been instrumental in developing and disseminating risk reduction strategies. At the same time, they have had an important financial impact on the local health care systems by keeping patients out of the hospital.

One reason government support for community-based organizations is a cost effective strategy is the heavy reliance on volunteer labor in these groups, which allows for a greater production of services per dollar expended than would be possible if government provided these services directly. However, a significant level of financial support is still necessary to develop administrative structures and to pay staff who can recruit, train, supervise, and support volunteers. According to one national survey, 26 nearly 80% of services provided by community-based AIDS groups around the country were performed by volunteers. In New York and San Francisco, the magnitude of donated labor is enormous, conservatively estimated at more than 100,000 hours in each city (Table V). The amount of donated labor at the Gay Men's Health Crisis in New York is similar to the combined total of the Shanti Project and the San Francisco AIDS Foundation. However, the ratio of unpaid staff hours to paid staff hours is twice as large in New York, indicating a greater reliance on volunteers and lesser funding by local government there.

CONCLUSIONS

The large number of AIDS cases in New York and San Francisco has had a significant impact on the local health

care systems. It is likely that as the epidemic spreads it will place increasingly greater strains on other health care systems in cities and counties with growing numbers of cases, as well as on public and private sources of health care financing. While the absolute number of AIDS cases continues to grow in metropolitan areas, there is convincing evidence that the disease is spreading across the country. According to the Centers for Disease Control, the four cities with the largest number of AIDS cases-New York, San Francisco, Los Angeles, and Miami-accounted for 73% of all reported cases in September 1982. This figure declined to 65% by September 1983, and, by February 1987, their proportion of the nation's total had fallen to 50%. In other words, the proportion of total AIDS cases not based in these four metropolitan areas has increased by more than 85% during this time period.

The availability of outpatient and community-based care facilities and programs and the relatively low charges per AIDS admission in San Francisco may give an important clue to the direction other communities should follow in a rational planning policy for the treatment of AIDS patients. The average length of hospitalization for AIDS patients in San Francisco is far shorter than in New York and in most other cities. This is due in part to the variations in patient groups, case mix, and practice patterns that are found in different regions of the country. However, an integrated system of health care delivery which is subsidized by local government in San Francisco and includes outpatient clinics, home health and hospice care, housing, and other social support services, allows patients to be discharged from the hospital earlier than in other cities where such services are not as readily available. The success of San Francisco's response to the epidemic does not, however, mean that other cities should attempt to respond in exactly the same way. New York's example illustrates how one city, confronted with complex problems, considered a variety of local conditions rather than merely adopting another city's model. Each community should develop its own response in light of its own epidemiologic circumstances and social and political structures, learning from the experiences of San Francisco and New York City.

The investment of public funds in community-based services affords better quality care for AIDS patients. It is also a rational fiscal response that helps reduce the economic impact of the epidemic by reducing the need for inpatient care. This lowers private health insurance expenditures, Medicaid outlays, and local tax revenues that

must be spent when Medicaid or other third-party reimbursement is unavailable or below costs.

Aside from financial support from the public sector, the viability of community-based AIDS organizations depends on a large, steady stream of unpaid labor. If the patient mix shifts further away from homosexual men towards IV drug users, as it slowly appears to be doing, it is unclear whether the level of voluntarism can be maintained. In low incidence regions, there may not be an identified at-risk population from which to draw volunteers. Thus, unless further financial support is forthcoming, gaps in services to AIDS patients, already experienced by many of these communities, may become more severe as the epidemic continues to spread geographically.

New York City and San Francisco illustrate that the AIDS epidemic may manifest itself quite differently in various cities and can trigger diverse responses based on local conditions. With the likely prospect of other cities facing large increases in the number of those afflicted with AIDS, the experiences of New York and San Francisco may prove useful. They indicate that each community's response to AIDS will probably reflect the underlying social, economic, and political characteristics of AIDS victims and the existing structure and organizational roles of traditional health care and community-based service providers. They also indicate the intrinsic limits to the current dependency on unpaid labor and the contributions from private charity and local government, increasing the pressure for additional state and federal support for community-based services and the care of AIDS patients.

Acknowledgment. The authors thank the many people in New York City and San Francisco who generously provided their time and expertise in helping to gather and interpret the data from each community.

2.

REFERENCES

1 Arno PS: Statement before the Subcommittee on Health and the Environment, Committee on Energy and Commerce, US House of Representatives, July 22, 1985 Hardy AM, Rauch K, Echenberg D, et al: The economic impact of the first 9,000 cases of acquired immunodeficiency syndrome in the United States. Presented at the International Conference on AIDS, Atlanta, Georgia, April 17, 1985. 3 Shilts R: AIDS crisis hits an unprepared New York. San Francisco Chron icle, February 1985, 14:1. 4

Leishman K: San Francisco: A crisis of public health. Atlantic Monthly, October 1985, pp 18-41.

5. Boufford JI: Statement before Intergovernmental Relations and Human Resources Subcommittee of the Committee on Government Operations, US House of Representatives, September 13, 1985.

9

Sencer DJ, Botnick VE: Report to the Mayor: New York City's Response to the AIDS Crisis. Office of the Mayor, City of New York, December 1985. 10 West Bay Hospital Conference: Quarterly AIDS Utilization Report, San Mateo, California, March 5, 1986.

11 Scitovsky AA, Cline M, Lee PR: Medical care costs of patients with AIDS in San Francisco. JAMA 1986; 256:3103-3106.

12. Knickman J, Foltz AM: Regional differences in hospital utilization. How much can be traced to population differences? Med Care 1984; 22:971-986. 13. Gay Men's Health Crisis: 1984 Annual Report, New York, 1985. 14.

Arno PS: The nonprofit sector's response to the AIDS epidemic. Community-based services in San Francisco. Am J Public Health 1986; 76:1325-1330. 15. Centers for Disease Control: AIDS Weekly Surveillance Report. US AIDS Program, Center for Infectious Diseases, Atlanta, Georgia, January 5, 1987. 16. New York City Department of Health: AIDS Surveillance Update. Office of Epidemiologic Surveillance and Statistics, January 1987.

17. Volberding PA: The clinical spectrum of the acquired immunodeficiency syndrome. Implications for comprehensive patient care. Ann Intern Med

1985; 103:729-733.

MAY 1987/NEW YORK STATE JOURNAL OF MEDICINE 271