of either measure. In the meantime we have experimental HMOs developed with and without public funds and we have the Social Security Amendments of 1972 making HMOs eligible for Medicare and Medicaid beneficiaries.

Even though HMOs have been promoted with great logic and persuasion, there remains considerable disagreement about their characteristics, merits, and mechanics. An examination of the dozen public and private HMO developments in Chicago suggests that most of the community-based efforts are spinning their wheels because only a citywide enrollment drive through employer and labor groups can yield the necessary number of subscribers to make an HMO viable. With the exception of a couple of plans it is impossible to identify all the providers, the population targeted, the guarantees against biased selection, the mechanics of marketing the plan, the Medicaid-Medicare mix of the enrollees, the peer review strategy, and the duration of the contract.

The Health Care Planning Corporation of Greater Chicago, spurred by the field office of Group Health Association of America, has produced a creditable metropolitan areawide HMO plan. Its strategy for delineating 6 service regions and locating 21 primary health care centers is critically examined, and a simple heuristic algorithm model of optimum location is proposed and applied.

Metropolitan areawide HMO marketing efforts by Chicago groups like GHAA, Blue Cross, CNA, the teaching hospitals, and the state and county medical societies may provide viable models, but even these are too fragmented. The thresholds of minimum enrollment and maximum travel time for primary care centers are too demanding to justify the extravagance of a dozen or even a half-dozen competing HMO developments. The major contenders should pool their efforts into a consortium such as the Rochester, New York pilot project started last year under the sponsorship of GHAA and the National Blue Cross Association and Blue Shield Plans.

V. PHYSICIAN MIGRATION FROM INLAND

TO COASTAL STATES*

Some states and urban areas in this country have two to three times as many physicians per 1,000 population as other states and urban areas. Results of a recent, unpublished study by the author revealed that the gap between doctor-rich and doctor-poor states has greatly widened since 1960, with Chicago and Illinois joining the ranks of doctor-poor communities and states that are rapidly losing ground to more glamorous centers on the East and West coasts. These huge differences in physician distribution are most commonly explained by local differences in population density, affluence, and medical school production of new doctors.12

Indeed, the nation's 300 metropolitan areas, the dense, affluent population centers in which most medical schools are located, have a physician/population ratio more than twice that of nonmetropolitan areas. In 1970, the ratio of physicians to 1,000 population in metropolitan areas was 1.73; for nonmetropolitan areas it was .80; and for the nation as a whole, it was 1.48. Yet when the physician/population ratios of the nation's 20 largest metropolitan areas are compared, it becomes apparent that disparities even larger than those found among the states exist and that there is practically no association between physician numbers and population density and affluence.

Table 3 shows that nine of the 10 cities with the highest physician/population ratios are located on the East and West coasts, with the cosmopolitan centers of Boston, New York, and San Francisco leading the way. Nine of the 10 lowest ranking cities are inland, with Chicago's industrial neighbors, Detroit and Milwaukee, at the bottom of the list. Boston, which ranks first in physician/population ratio, ranks 8th in population size and 9th in average family income, while Detroit, ranking 20th in physician/population ratio, ranks 5th in population size and 3rd in average family income. Cleveland and Anaheim, the exceptions in the list, are parts of larger metropolitan regions. Cleveland provides health care for eleven counties and four metropolitan areas (Cleveland-Akron-Canton-Youngstown), whose aggregate ratio is 1.44 physicians per 1,000 population, while Anaheim falls within the shadow of Los Angeles.

An examination of the ten states with the highest physician/population ratios confirms the strong attraction that East and West Coast centers of glamor and good climate exert on physicians (Table 4).

Locational Preferences

Traditional explanations of physician location were probably too simple-minded even for the more free-market locational forces of the early 1960s. They are certainly inadequate explanations of present-day locational forces, particularly those operating among large urban centers.

The analytical complexities are of two orders-the partially inconsistent economic and social goals that motivate physician location, and the mixed private-public market for physician services that reduces traditional dependence on a geographic private office practice.

We can identify four types of life-style goals capable of motivating physicians to locate in certain states and urban centers (interstate choices) and in certain communities within a state or urban center (intrastate choices):

(1) Income-maximizing goals that stress consumption aspirations, or locating a practice to get the most material benefits out of one's medical skills. These goals would favor intrastate locations that are accessible to affluent patients, that minimize distances to the hospitals they use, and that are not overly saturated with competing physicians. Interstate locations are more dependent on the total physician supply of these larger areas and should result in approximately equal physician-population ratios among urban centers of equal per capita income.

*This case study evaluates the effects on the distribution of medical manpower of a program a few years after its introduction. It is adapted from Working Paper 11.20 (May 1972) and a published version appearing in the February 1973 issue of the Journal of Medical Education.

'S. Joroff, and V. Navarro, "Medical Manpower: a Multi-Variate Analysis of the Distribution of Physicians in Urban United States." Medical Care, 9-428-438, 1971.

*R. Scheffler, "The Relationship Between Medical Education and the Statewide Per Capita Distribution of Physicians." Journal of Medical Education, 46:995-998, 1971.

41-860-75 pt. 3 - 6

TABLE 3. TWENTY LARGEST METROPOLITAN AREAS IN THE UNITED STATES RANKED BY NUMBER OF PHYSICIANS PER 1200 POPULATION”

*Sources are Distribution of Physicians in the United States 1970, Chicago American Medical Association, 197) and the 1970 US Census of Population.

TABLE 4. TEN STATES WITH THE HIGHEST NUMBER OF PHYSICIANS

(2) Family-oriented goals stressing the best environment for bringing up children and satisfying the social daytime needs of the mother. Intrastate locations would favor communities of single-family housing, with good schools, and an active community life. Newer affluent dormitory suburbs are targets for this kind of mover. Interstate locations would favor good climate states with year-round outdoor recreation.

(3) Social prestige goals stressing a life style and environment consistent with the social status of physicians. Since physicians are well near the top of the social status ladder in the United States, these goals would result in intrastate locations in the traditional affluent, social register communities, with a preference for those that cater to executives and professional workers who commute downtown as opposed to country club or polo club types of suburbs for the idle rich. Interstate locations would favor cosmopolitan centers of intellectual and academic excellence, with a preference for Europe- and Asia-oriented centers on the East and West Coasts as opposed to Middle American provincial capitals.

(4) Professional interaction goals stressing professional and social interaction among people sharing common skills, values, and interests. Physicians, for example, would tend to maximize interspecialty contacts at the professional level, and intraspecialty contacts at the social level. These goals partly overlap with income-maximizing and social prestige goals, but they also reflect the special social-psychological needs of an elite steeped in arcane skills and vocabulary that cannot be shared easily with members of other elite groups. These goals favor intrastate locations accessible to the concentrations of specialists' offices downtown and in the large university medical centers. Interstate locations would favor regional centers with the largest number of specialists and the largest and most prestigious medical schools and basic science schools.

These four goals are somewhat overlapping and somewhat conflicting. The first goals would tend to equalize physicianpopulation ratios among states and communities, but the second, third, and fourth goals would tend to generate gravitation to large centers of amenity, social, and professional excellence. The principle of cumulative advantages of large medical concentrations is somewhat similar to the phenomena of retail gravitation evident in the regional location of large department stores and of the very large-scale economies found in basic steel production and automobile assembly. There is an important difference in terms of the public or consumer interest, however. Consumers get considerable benefits from commercial and industrial concentrations in the form of better accessibility and lower prices. But the public or consumer interest is very poorly served by medical gravitation that is motivated by provider preferences for amenity locations, and which results in poor accessibility and high prices in doctor-poor areas. The continuing shifts of physicians at both intrastate and interstate scales from doctor-poor to doctor-rich areas suggest a weakening of the constraint of income-maximizing goals, permitting trade-offs from the physician-population equalizing goals to the gravitational goals, either because of a shift in goal prefer. ences or because of a substantial increase in economic demand.

It is hypothesized that the tremendous increase since 1965 in public payments for physician services has in effect raised the potential economic demand for medical services of most communities and states by about a fourth. In the absence of a corresponding increase in medical manpower, physicians find their locational options greatly increased by being able to move to amenity locations at least until full saturation of the potential new market. If this did in fact happen, we would expect to find that these shifts were made without the expected result of lowering medical price levels in the receiving amenity areas and we would also expect to find considerable geographic disparities in physician participation in new public programs corresponding to the traditional locational preferences of physicians.

Income

Fein's economic diagnosis of the physician shortage has led some to consider local variations in physician earnings as a reason for the wide variations in physician-population ratios. Yet the play of the private market should actually help prevent the uneven geographic distribution of physicians, that is, given reciprocal licensure and free physician movement among states, huge disparities in physician population between states of similar population density and income would not develop. Otherwise doctors in states with lower ratios would be able to earn more money than those in states with high ratios.

For example, with minor regional variations, American communities spent last year an average of $90 per person for physicians' services, yielding an average gross income of $60,000 per physician. If physicians were dependent on private patients only for their practice, average physician income in Boston, New York City, and San Francisco would be about half of what physicians could earn in Detroit, Milwaukee, and Pittsburgh.

Actual average physician income was estimated by state and region by projecting 1966 data based on random samples of physician income tax returns. These projections were compared with 1969 and 1970 data from the Medical Economics continuing survey."," The projections revealed a slight tendency for average physician income to vary inversely with the physician/population ratios of the nation's major regions (Table 5). The Northeast and West regions rank highest in physicianpopulation ratios and lowest in average physician income. However, while the Northeast region's physician/population ratio is

'R. Fein, The Doctor Shortage: An Economic Diagnosis. Washington, D.C.: The Brookings Institution, 1967.

*R. Schaffner, and I. Butter, “Geographic Mobility of Foreign Medical Graduates and the Doctor Shortage." Inquiry, 9:24-33, 1972. 'L. Reed, Studies of the Incomes of Physicians and Dentists. U.S. Department of Health, Education, and Welfare. Washington, D.C.: U.S. Government Printing Office, 1968, p. 96.