|
The Availability of Health and Mental Health Providers by Population DensityLetter to the Field No. 11by Charles E. Holzer III, Harold F. Goldsmith, and James A. Ciarlo Table of Contents This Letter examines the availability and geographical accessibility of mental health services for those living in rural areas and those living in areas designated as frontier due to their extremely low population per square mile. Our initial assumption, which is largely born out in the following analyses, is that persons living in frontier and isolated rural areas have fewer mental health services available and accessible to them either through the general medical sector or through specialty mental health sectors. In addressing this question we define availability of mental health services as the presence and the number of those psychiatric and/or psychological providers or services that traditionally address the needs of those with various types of psychological or psychiatric disorders. Area Resource File. The primary source of data for
addressing the availability of services is the Area Resource File (ARF) of the Department
of Health and Human Services, Bureau of Health Professions, Office of Research and
Planning (Quality Resource System, 1996). Most of the data included in the file is drawn
from governmental agencies such as the National Center for Health Statistics and the U.S.
Census, or from private agencies such as the American Medical Association (AMA), and the
American Hospital Association (AHA). They collect data either from administrative records
of the agencies or from surveys of hospitals or other facilities. Public Use Microdata 5% Sample (PUMS) - 1990.
Because the Area Resource File focuses primarily on professionals in specified health
settings, it contains only minimal data on psychologists and social workers. In order to
gather additional information on these professions, we conducted supplementary analyses of
data from the Public Use Microdata 5% Sample (PUMS) of the 1990 U.S. Census. This contains
a sample of approximately 5% of the U.S. population, particularly those who filled out the
Census long form questionnaire. Because the PUMS is a sample of individual records,
confidentiality is protected by limiting geographic identification to areas corresponding
to about 100,000 persons. For purposes of the present analyses we aggregated sub-county
areas up to the county level, and for areas containing multiple (small) counties we
allocated the record data to the counties contained within it, proportionally by county
population, and then adjusted the marginals of tables to match available published Census
tabulations (STF files). This procedure is described elsewhere, but has the effect of
approximating the county information in a manner that should not be unduly biased in
subsequent tabulations. Definitions of Rural and Frontier Our primary measure of the urban-rural dimension is
population density. This is the number of persons residing in a county divided by its land
area. We have adopted the Frontier Mental Health Services Resource Network's definition of
"frontier" as counties with less than 7 persons per square mile. To provide a
continuum, we have further divided this density category into "very frontier" (0
to 1.9 persons per square mile) and "frontier" (2.0 to 6.9). Availability of Mental Health Services The Area Resource File provides counts for a number of
different types of health and mental health providers by the setting in which they
practice. Rather than attempting to present too many subtypes, we have selected for
presentation a few of the major types of providers. These include psychiatrists, child
psychiatrists, psychologists, social workers, family practice physicians, and all
physicians. We have also presented selected information about psychiatric hospitals and
psychiatric beds in general hospitals. Unfortunately, the ARF does not identify or
differentiate free-standing Community Mental Health Centers or the non-physician staff who
work there. For each of the selected provider types we have presented
the availability of the provider by population per square mile categories. In each figure
the legend identifies the population density category and the number of counties in that
density category. For each variable the percentage of counties with any providers in the
category is given. This is important because it shows that many counties have no service
providers of the particular type. Also present in each figure is the number of providers
or units of service per 100,000 persons living in the designated county area. Thus one can
see when there are not only fewer providers, but also fewer providers relative to the size
of the population. Availability of Psychiatrists. Figure 1 presents
the availability of psychiatrists for each population density category. This variable is
reported from the AMA survey of medical specialists, and does not include those employed
by the federal government. The county for the psychiatrists in this source appears to
correspond to the office location rather than his/her residence. In the lowest density
category, from 0 to 1.9 persons per square mile, there are 129 counties that are
designated "very frontier". In the next category, with densities from 2.0 up to
7.0 persons per square mile, there are 296 counties identified as "frontier". Availability of Child Psychiatrists. Figure 1 also
presents the availability of child psychiatrists in 1994, who are not federal employees,
from the AMA Physcian Master File by density category. This figure shows that child
psychiatrists are not present in any of the lowest density counties and are found in only
0.7 percent (i.e., only two) of the remaining frontier counties. Less than 10% of the
counties with 7 through 99.9 persons per square mile have any child psychiatrists. Only in
the counties with over 100 persons per square mile does the percentage with even one child
psychiatrist rise to 58%. Similarly the availability of child psychiatrists by rate
increases with population density, but then only to an average of 1.5 per 100,000 persons
for the densest areas. The maximums in a few areas are much higher, but those are
primarily in counties with major medical schools. Availability of Family Practice Physicians. According to the de facto model (Regier, et al., 1978), as much as half of mental health care may be obtained in the general medical sector. Much of this care will be from physicians in family or general primary care practices. We have included family practice physicians to identify the availability of physicians who could provide mental health care in the absence of mental health specialists. It should also be noted that many family practice physicians obtain additional training in psychiatry, not only as a general requirement of their residencies, but because of interest in mental health in family settings. Figure 2 presents the 1994 distributions for MD's in non-federal family practice, as obtained from the AMA Physician Master File by population density. Of the counties with the lowest density, 0-1.9 persons per square mile, only 33.3% have MD's in family practice. Noteworthy, this reflects an average of only 0.7 per county. With a rate of 14.3 family practice physicians per 100,000 persons, an average practice size would be nearly 7,000 if patients were not accessing other forms of care. Clearly a medical practice this large would not allow a great deal of time for providing mental health services.
For the counties with densities of 2.0 to 6.9 persons per
square mile (frontier), the availability of family practice physicians increases to 68.9
percent, or an average of two per county. The availability of family practice physicians
per 100,000 population is 25.8. This is not only a substantial increase from the very
frontier counties, but is actually higher than for any of the more densely settled areas.
This finding may indicate that while it takes a minimum population density to maintain a
practice, the family practitioner is a common form of medicine in frontier areas.
Additionally, as will be noted later, after our examination of psychologists and social
workers, the family practitioner is also the most common medical practitioner who is
likely to provide mental health services. Availability of All Physicians. In order to address
the question of whether the reduced availability of specialty mental health providers or
family practice providers is part of a general pattern of availability of physicians, we
have included figures on the distribution of all MD's engaged in patient care. Figure 2
also presents the availability of all physicians engaged in patient care by population
density. This variable includes all MD's who are active and providing patient care,
including office-based, hospital residents, clinical fellows, and hospital-based full-time
staff. It excludes those primarily involved in administration, teaching, or research. It
was drawn from the AMA Physician Master File for 1994. Even when all types of physicians are considered, less
than half of the very frontier counties have any physicians, although this jumps to over
85% for even the frontier counties. The rate per 100,000 is 30.7 for the very frontier
counties. The range is 53.4 to 68.7 per 100,000 for all the intermediate categories, and
then jumps to 179.8 per 100,000 for the highest density counties/cities. Clearly,
physicians are concentrated in the cities. Moreover, in comparison with psychiatrists and
child psychiatrists, it is clear that in the more densely populated areas there is more
opportunity for choice of a mental health provider who is in the specialty sector or the
medical sector. Availability of Psychiatric Hospitals. Returning to the specialty mental health sector, the next set of figures presents the availability of psychiatric hospitals. The variable is the number of psychiatric hospitals within the group of long-term stay hospitals in the American Hospital Association County Hospital File for 1993. Long-term hospitals are identified as having an average length of stay of 30 days or more, although that applicability designation is not well documented in the source file. Figure 3 presents the availability of (long-term) psychiatric hospitals by population density. Psychiatric hospitals as identified in this file are relatively rare, with only 20.5% of even the high density counties having one or more. The data show a few psychiatric hospitals located in middle density areas. However, except for very frontier, for all of the density categories the rate of such hospitals is only 0.1 per 100,000 population. This data does not speak to size; large state hospitals have historically been located in rural areas.
Number of Inpatient Days for Psychiatric Hospitals. Although the number of psychiatric hospitals is very low, an alternative means of examining hospitalization is to look at the number of psychiatric hospital bed days being utilized. To interpret the meaning of these data, one has to consider that utilization is limited by availability as well as access. In particular, it becomes relevant that these results are reported by the location of the hospital rather than the county of residence (or origin) of those hospitalized. The ties of long-term residents of psychiatric hospitals to their community of origin are often lost over time. The second half of Figure 3 presents psychiatric long term
hospital inpatient days by population density. The percentage of counties reporting any
bed days closely parallels the availability of psychiatric hospitals, confirming that
these figures are based on hospital location. Although the rate per 100,000 is zero for
the very frontier areas, the rate for other population densities is relatively even,
ranging from 7,814 to 13,014 per 100,000, with the highest figure in the densest areas.
This is consistent with a smaller number of large psychiatric hospitals being located in
low density areas, as was the historical pattern. One may ask whether the newer pattern of
deinstitutionalization in favor of outpatient and community-based treatment has been more
feasible in the more densely populated areas. If this were so, then why would the number
of hospital days be higher in the dense areas? Psychiatric Services in Short Term General Hospitals. The use of small psychiatric units and scatter beds in short-term general hospitals has been an alternative to the use of long term psychiatric hospitals. Figure 4 presents psychiatric beds in short stay general hospitals by population density. Once again there is a tremendous differential in the availability in frontier versus high density areas. Less than 1 percent of very frontier and 1.4 percent of frontier counties have any psychiatric beds of this type. Availability increases with population density, but this type of service is available primarily in higher density areas. The availability of these beds as a rate per 100,000 population is also extremely low in frontier and other low density areas.
The above pattern suggests that while the number of
hospitals with this service does not increase much in the high density areas, the urban
hospitals must be considerably larger to account for the differential in psychiatric bed
days used. It may also be that only the larger hospitals can afford to differentiate
psychiatric beds from other types of beds. Examination of the number of general hospitals
versus their total number of beds, found elsewhere in the ARF file, shows that while even
low density areas tend to have hospitals, they are generally much smaller, and presumably
less specialized in their services. This results in equivalent urban-rural differentials
for other identified specialized services such as substance abuse beds and child
psychiatry beds (in additional tables available from the author). Supplemental Analyses from U.S. Census PUMS Data Rationale. In preliminary work, we examined the distribution of psychologists. We discovered during this work, however, that the available counts in ARF are limited to psychologists working full- or part-time in hospitals, whether short- or long-term facilities. The numbers of psychologists identified by that means was extremely low, with only a small proportion of counties having any psychologists identified at all. Even in metropolitan central cities, only 66% had even one psychologist. Although this pattern could be consistent with psychologists being in individual or group private practice outside hospitals, or working in nonhospital-based health care settings, it appeared to greatly understate the availability of psychologists. In order to obtain a more complete picture of the distribution of psychologists, we examined the U.S. Census Public Use Microdata Sample for 1990. We were able to identify psychologists, social workers, and clergy as potential sources of mental health care, but these titles are non-specific and do not identify function. Because many with these occupational titles are engaged in work totally unrelated to mental health, we placed two additional restrictions on the persons to be tabulated. First, we limited the count to persons with at least a master's degree. This is the usual level at which one can engage in independent mental health practice, although there is variability in state statutes regulating private practice. Second, we limited the count to persons who are identified as working in any kind of health-related industry, including offices and clinics of physicians, health practitioners, hospitals, nursing and personal care facilities, health services not otherwise classified, residential care facilities, and miscellaneous professional and related services. We think this would include psychologists in private practice. This reduced the numbers of M.A. or higher psychologists by about 30-40% from their total and M.A. level social workers by 60-70%. Figure 5 presents the number of M.A. or greater psychologists working in health-related settings by population per square mile. Because these figures are based on 1990 rather than 1993 data, the distribution of counties by population per square mile is slightly different. In the areas with less than 2 persons per square mile only 13.3% of counties have psychologists in health care settings, although that percentage would increase to nearly 31.1% if all settings were included. The difference might be influenced by psychologists working primarily in schools who may have both mental health and educational roles. For population densities of 2.0 through 6.9, about 43.1% of counties have psychologists in health care settings. This increases gradually and then jumps to 79.5% for high density counties. In a parallel way, the number of psychologists per 100,000 population is lowest (13.0 per 100,000) in the least densely populated counties and is greatest in the densest counties (28.9 per 100,000).
Figure 5 also presents the population density distributions of social workers with
master's degrees or greater who work in health-related settings. Only 18.5 percent of
counties with less than 2 people per square mile had social workers of this type, and only
23.4 percent of the remaining frontier counties had comparable social workers. In
contrast, 73.2 percent of the most densely populated counties have social workers at this
level. As a population rate, the low density counties had about 12.8 social workers of
this type per 100,000 population, as compared to 23.6 per 100,000 for the high density
counties. The rates for the intermediate counties were even lower than for the frontier
counties. For reference, social workers in health settings represented about a third of
all social workers with comparable education levels. Other settings in which social
workers are found include schools and social service agencies without a health focus. In the analyses above, we have presented data on the
availability of selected types of specialty mental health services drawn from the Area
Resource File. Initial analyses showed tremendous differentials in the availability of
psychiatrists and child psychiatrists by population density, where psychiatrists were less
available in frontier areas. Differentials for family practitioners were much smaller
overall. Long-term psychiatric hospitals were found to be few in number but found occasionally in all but the most rural areas. This appears to reflect the historical practice of locating mental hospitals in the bucolic countryside, where the air was fresh and the land cheap. Examination of the number of inpatient days by area suggests that they are mostly large facilities, to which patients are brought on a regional rather than a county basis. Thus, they are "local" resources only to those who happen to live nearby or remain in the area after treatment. Psychiatric beds in short-term general hospitals contrast with the long-term psychiatric facilities in that they are more concentrated in higher density areas. This may reflect a minimum hospital size or urbanization necessary before specialization into psychiatric beds occurs. The reasons for and consequences of the differentials in
service availability by population density are not directly addressed in the ARF data set.
Yet it seems reasonable to consider several possibilities that have been widely discussed
in the literature. At the crux of the question is whether the patterns are driven by the
provider or the consumer side of the equation. On the provider side is a myriad of factors influencing
whether organizations with the resources to provide services choose to be present or are
placed in rural settings. Some organizations and individual providers are market-driven
and will move into a rural area if and only if the "market" can sustain the
organization. Others are driven by governmental or public sector concerns and are less
concerned about economics. There are other less tangible factors influencing provider's
choice of location. These range from particularistic concerns such as a potential provider
being from the local area or liking rural life through more pragmatic concerns such as the
availability of good schools, housing, etc. Also involved are professional concerns such
as the availability of colleagues to provide backup or vacation coverage. Although we may
have little apparent influence over these concerns, many rural communities have
experienced difficulties because of them in recruiting health and mental health
professionals of various kinds. On the consumer side of the equation are issues of
identification and classification of a problem, which may or may not receive a mental
health label, and the choice of sectors for obtaining help with the problem. The de facto
mental health services system as described by Regier and colleagues (Regier, Goldberg,
& Taube, 1978; Regier, et al., 1993) includes the use of the specialty mental health,
primary/general medical care, and general human services sectors. A particular emphasis
was placed on the wide use of the general medical sector for mental health problems, which
may be increasing in the present world of managed care. In these areas we have
demonstrated substantial differentials between rural and urban areas in the availability
of these general medical providers. In the more densely populated areas there is more
opportunity for choice among specialists in organized multi-specialty settings. Regier, et al. (1978, 1993) identified the substantial use
of local resources such as family, friends, and self-help groups. Optimal use of local and
community-based providers of services has received great emphasis in recent publications
on rural mental health services (Hill & Fraser, 1995; Merwin, Goldsmith, &
Manderscheid, 1995; Bergstrom, 1982). Some have emphasized expansion of the roles of the
general medical sector in dealing with mental health. Other approaches have emphasized
development of specialty mental health providers such as psychologists to work in these
areas (Hargrove & Breazeale, 1993). Ultimately, the choice should lie with the
consumer. Clearly the Area Resource File has limited current information about types of specialty mental health services. Conspicuously absent is information about organized community mental health centers and other community-based free-standing mental health facilities. The distribution of psychiatrists may be a rough surrogate for that distribution, but it can only do so minimally due to the large proportion of mental heath services offered by psychologists, counselors, and other mental health workers. A second limitation of the present analyses is that the providers of services are identified by the county in which they are located rather than differentiating the counties that they serve. The ARF identifies contiguous counties but provides no means of allocating services identified to neighboring populations. A third limitation is that it does not take into account the ability or willingness of rural residents to access services beyond their county of residence. Finally, access to services, even when present, can be limited by economic, social, and psychological barriers to access. Bergstrom, D.A. (1982). Collaborating with natural helpers for delivery of rural mental health services. Journal of Rural Community Psychology, 3, 5-26. Hargrove, D.S., & Breazeale, R.L., (1993). Psychologists and rural services: Addressing a new agenda. Professional Psychology: Research and Practice, 24, 319-324. Hill, C.E., & Fraser, G.J. (1995). Local knowledge and rural mental health reform. Community Mental Health Journal, 31, 553-568. Merwin, E.I., Goldsmith, H.F., & Manderscheid, R.W., (1995). Human resource issues in rural mental health services. Community Mental Health Journal, 31, 525-537. Quality Resource Systems, Inc. (1996, February). Area Resource File [CD-ROM]. Bureau of Health Professions, Office of Research and Planning. Regier, D.A., Goldberg, I.D., & Taube, C.A.(1978) The de facto US mental health services system: a public health perspective. Arch Gen Psychiatry, 35, 685-693. Regier, D.A., Narrow, W.E., Rae, D.S., Manderscheid, R.W., Locke, B.Z., & Goodwin, F.K. (1993). The de facto US mental and addictive disorders service system. Epidemiologic catchment area prospective 1-year prevalence rates of disorders and services. Arch Gen Psychiatry, 50, 85-94. U.S. Census of Population and Housing. (1992).1990, Public Use Microdata Samples, United States [Technical Documentation prepared by the Bureau of the Census]. Washington, D.C.: Author.
Frontier Mental Health Resource Network |
|||
The first data set presents the
percentage of those counties identified as having any non-federal psychiatrists,
regardless of type of activity. Less than 1 percent of the very frontier counties had any
psychiatrists, and only about 10% of the frontier counties had any psychiatrists. This
contrasts sharply with 30.6% for counties with more than 15 persons/sq.mi., and with 91%
for counties with over 100 persons/sq.mi. The next data set presents the number of
psychiatrists per 100,000 population. This figure takes into account the smaller
population of the frontier counties. Even as a rate of psychiatrists per 100,000
population, the availability of psychiatrists in frontier counties is almost nonexistent
and much lower (0.1/100,000 and 1.3/100,000 respectively) than the 10.5/100,000 found for
the most densely settled counties.
Bed Days for Psychiatric Subacute Care in Short-Term General
Hospitals. In order to extend our understanding of the use of psychiatric beds in
short-term general hospitals, Figure 4 reports psychiatric (subacute care) bed-days of
reported utilization of short-term general hospitals for 1993. Consistent with the number
of beds reported above, the bed days reported for very frontier areas is negligible, but
increases with population density to 4,950 bed days per 100,000 for high density areas. 
