Developing Indicators of Access to Care: The Case for Migrants and the Homeless
Joanne E. Lukomnik 1
Developing indicators that can monitor progress in ensuring equitable access to health care services is especially difficult when considering some of the most vulnerable populations in the United States. Many of the indicators developed in the main body of this report depend on data reported from national surveys; these data are coded by geographic area and then ascribed to specific communities. Certain population groups, however, are either systematically underrepresented in most national surveys or have other characteristics that make it difficult to track their access to services. Among these populations are migrant farmworkers and their families, the homeless, undocumented workers, and others whose employment or other life circumstances necessitate frequent movement and residential shifts.
This review concentrates on existing knowledge regarding access barriers and the consequent health status of migrant farmworkers and the homeless, two examples of such populations. For these groups, knowledge regarding access to services, barriers to access, and the health consequences of utilization (or the lack of utilization) may depend on our ability to perform periodic special surveys rather than on indicators derived from already existing data bases. This review relies on the sparse published literature as well as program data and experience generated by the Migrant Health (Public Health Service Act, Section 329) and Health Care for the Homeless
(Public Health Service Act, Section 340) programs. The last portion of this paper proposes a series of indicators of barriers to access for these groups and a set of special studies that would provide data on access for these populations that are otherwise lacking today.
The homeless and the migrant farmworker population share a distinctive characteristic in our society: they survive without a permanent home address, a fixed locality where mail can be sent, phones can be installed, and the census bureau can locate them. For example, the National Health Interview Surveys and many studies on access, including the Robert Wood Johnson Foundation (1986) survey on access, rely on telephone contacts to enter respondents into the studies. Such surveys in specific geographic areas frequently also rely on the telephone (e.g., Hubbell et al., 1991). Other surveys use lists of randomly generated household addresses. Because of the lack of a fixed address and the related lack of telephone service, migrant and seasonal farmworkers and the homeless are often excluded from these types of studies. The National Health Interview Survey (NHIS) requires that at least one household member in a respondent household be an English speaker. Because many migrant and seasonal farmworkers are Hispanic or members of recently arrived, non-English-speaking immigrant groups, they may be excluded from the NHIS and other surveys requiring English. Tragically, migrants and the homeless share more than the lack of a fixed address: members of both groups live in extreme poverty, have less than the national average level of education, and have a greater burden of illness, higher rates of infant mortality, and shorter life expectancies than Americans as a whole (National Migrant Resource Program, 1990; Wright and Weber, 1987).
Migrant farmworkers and their families, the single adult living on the street, and homeless families in shelters periodically become "visible." For brief periods, the news media and policymakers focus on the problems of homelessness or the plight of migrants. These news stories include references to the poor health indices of both groups. Despite individual studies and some targeted surveys, however, few systematic national efforts have been made to monitor the health status of members of either group and their access to quality health care. Interestingly, far more is known about the health status and illness patterns of the homeless than about the comparable status of migrants, even though homelessness is only a decade-old phenomenon in its most recent manifestation and migrant farmworkers have been an essential component of agribusiness for more than half a century. Several possible explanations can be postulated. The creation of a class of people known as the homeless was an inadvertent offshoot of other social policies and programmatic decisions. The homeless are highly visible in the nation's cities and media centers, which are also the epicenters of medical and health services research. During the 1980s, as the number of
homeless multiplied, attention and new programs proliferated (specifically, the Robert Wood Johnson Foundation's Health Care for the Homeless program and its Public Health Service successor, the McKinney Health Care for the Homeless). In contrast, migrants are seen as a necessary part of the agricultural work force. They are hidden from the view of most people, and their stories seldom capture the attention of the public. Furthermore, there is an assumption that such mechanisms as the migrant health clinics and occupational health and sanitation laws "took care of their problems" long ago. Renewed efforts to improve the health status of migrants and their access to health care services might prove to be quite expensive, requiring reorganization of agricultural labor patterns. These reasons make it easy to postulate a reluctance to focus on migrant health and access issues.
Available information about health status, access to health services, and utilization of the medical system by migrants and seasonal farmworkers and the homeless is generally derived from surveys and evaluations of specially targeted health care delivery systems (e.g., the Migrant Health Clinics and Health Care for the Homeless programs). These studies provide important information, but by definition they are concentrating on that proportion of these populations who have, in fact, gained access, at least briefly, to the medical care delivery system. Because access and utilization cannot be assumed to be synonymous, indicators of access must look beyond utilization figures. Utilization of health care services is a function both of individual attributes of the patient and organizational factors, including the availability and accessibility of health care services. Among individual attributes, the severity of a person's health problem, his or her perception of vulnerability, cultural and psychological attitudes toward physicians and the health care system, and the perceived costs and benefits involved in seeking care will all influence utilization behaviour (Aday and Anderson, 1975).
Organizational factors that affect utilization include the economic cost, availability, distance, and location of health care services, appropriate linguistic services, and other economic, ecological, and organizational aspects of the health services themselves (Aday and Anderson, 1984). For the vulnerable populations in question, the homeless and migrant and seasonal farmworkers, measuring access becomes more complicated because the traditional organization of care may fail to meet their individual or organizational needs, despite relatively poor health and a documented need for services.
The development of indicators to monitor access to health care services for these populations should be a priority not only in and of itself but also because these groups represent an extreme along a spectrum of vulnerable populations. These populations are known to be at risk for poor health outcomes, which are partially attributable to limited access to appropriate health care services. Although the homeless and migrant populations constitute
two groups with known poor health outcomes, they are not, in fact, totally distinct groups. Migrant workers are frequently homeless, particularly between picking seasons. Homeless persons, particularly single men, may enter the migrant stream briefly or intermittently.
More importantly, neither the migrant nor homeless population is stagnant. During the 1980s the nation saw an explosion in the number of poor who found themselves without housing for at least part of the. Families living in poverty found themselves vulnerable to the threat of homelessness even when they did not experience the actuality. Many people become vulnerable to homelessness through the loss of jobs and income, exhaustion of family support systems, or other tragedies that may precipitate them into homelessness. (Hopper and Hamberg, 1984; U.S. General Accounting Office, 1985). Likewise, the boundary between migrant farmworker, seasonal agricultural worker, and the unemployed often blurs, particularly between agricultural seasons or when bad weather, other natural disasters, or poor economic times limit harvesting of crops. The rural poor may participate in seasonal agricultural work even when not joining the migrant streams.
Although both groups are a heterogeneous population of black, white, Hispanic, Haitian, and other ethnic backgrounds, minorities, who are already at risk for poor health, are disproportionately represented among both the homeless and migrants. Undocumented workers, recent immigrants, and their families often join the migrant streams where few questions about immigration status may be asked. Undocumented workers and recent immigrants may also become homeless because they may not be eligible for welfare or unemployment benefits. Individuals who suffer from mental illness, alcoholism, and drug addiction are represented among the homeless in numbers far exceeding their proportions in the general population (Institute of Medicine, 1988). The living and working environments of migrants may also create or select for individuals with these conditions.
The Institute of Medicine's Committee on Monitoring Access to Personal Health Services defines access as "the timely use of personal health services to achieve the best possible health outcomes." The health care needs of both the migrant and homeless populations are enormous and complex, as a consequence of poverty, environmental and occupational risks, mental health needs, and the living conditions that define and determine the existence of these groups. In turn, monitoring access, evaluating the barriers to health care, and assessing the appropriateness of services for the homeless and for migrant populations are essential to understand the complex relationships among health status, health care utilization, and outcomes. Ultimately, this knowledge must be joined with the political will to guarantee access to high-quality care for these interrelated and most vulnerable of populations.
MIGRANT HEALTH STATUS
Estimates of the number of migrant and seasonal farmworkers and their dependents vary between 2.7 million and 5 million people (National Migrant Resource Program and the Migrant Clinicians Network, 1990). Determining the actual number of migrants is complicated by the number of different federal and state agencies involved in collecting data, the transient nature of the population, systematic undercounting of workers by agribusiness, and the desire of many workers themselves (particularly if they are documented or undocumented immigrants) to avoid contact with any government agencies. The Office of Migrant Health (1992) of the Health Resources and Services Administration quotes a figure of approximately 4 million migrant and seasonal farmworkers and their dependents; this statistic is derived by defining a migrant or seasonal farmworker as ''an individual whose principal employment within the last 24 months is in agriculture on a seasonal basis." (Office of Migrant Health, 1992). In this definition, the only difference between a migrant and a seasonal farmworker is that the migrant travels and establishes a temporary abode for employment purposes (Office of Migrant Health, 1992). The Public Health Service divides this number into 1.7 million migrant workers and their dependents and 2.5 million seasonal farmworkers and their dependents. The methodology used to calculate these numbers is complex and includes, in addition to reporting from each state, calculations that estimate the number of person-hours required to harvest the acreage undercultivation.
The Bureau of Labor Statistics uses a narrower definition (employed farmworkers over the age of 14) and notes a steady decline between the 1950s and the early 1980s; it gives a relatively steady current figure of 2.7 million. By this accounting, migrant farmworkers number only 200,000 (not including dependents) with seasonal agricultural workers and other employed farmworkers making up the remaining 2.5 million (U.S. Bureau of the Census, 1990). The Department of Labor and the Department of Agriculture count only employed farmworkers in a given moment in time as reported by employers. Each of these departments uses different methods to calculate the number of employed farmworkers. Neither department estimates the number of dependents or differentiates among farmworkers, crew chiefs, managers, and the like.
All of these estimates are subject to undercounting because of the difficulties inherent in quantifying a work force based on daily hire, the use of crew chiefs who receive the pay for a group of workers, and the still common, although illegal, custom of using children under age 14 (who are excluded from counts by definition) in the fields. The transient and seasonal nature of the work force, and the undocumented movement among Mexico, Central America, Jamaica, and other Latin countries, further complicate
the ability to accurately estimate the size of the migrant and seasonal farmworker population.
The variations and imprecision in population estimates make it extremely difficult to calculate vital statistic rates. A number of sources quote a life expectancy of 49 years, as compared with the national average of 75 years, for migrant farmworkers, and an infant mortality rate that is 125 percent above the national average (National Migrant Resources Program and the Migrant Clinicians Network, 1990). However, literature searches commissioned by the Department of Health and Human Services in 1984, the Farmworker Justice Fund in 1985 and 1988, and one performed by a migrant health physician in 1990 revealed no published studies that included specific mortality or survival data (Rust, 1990). A literature search performed for this paper and personal communications with personnel from the Office of Migrant Health also failed to discover data on life expectancy, age-specific mortality, or crude death rates.
Recent information about perinatal outcomes is equally hard to obtain. Infant mortality and low birthweight rates of women using migrant health centers (see the later discussion) have not been calculated. Most studies of birth outcomes are at least 15 years old. In 1978, a study that relied on the mother's recall questioned 132 women in Wisconsin (Slesinger and Christensen, 1986). The authors reported an infant mortality rate of 29 per 1,000 and a mortality rate of 46 per 1,000 children up to the age of 5. Infant mortality among Mexican American farmworkers in Colorado was reported to be 63 per 1,000 in 1971. This rate was three times the national rate for the period (Chase et al., 1971).
The Office of Migrant Health records that 500,000 migrant and seasonal farmworkers and their dependents annually use the 102 migrant health centers located in 43 states and Puerto Rico (These figures are derived from unpublished data for calendar year 1990–1991 from the Health Resources and Services Administration's Bureau of Health Care Delivery and Assistance.) These users represent approximately 17 percent of all migrant and seasonal farmworkers and their families, if one's calculations employ the Public Health Service's number of eligibles as the denominator. We cannot know how representative the users of migrant health centers may be of the total population of migrant and seasonal farmworkers; nevertheless, some information regarding their health status is available and may indicate some general trends.
In a survey of migrant health centers (with a 49 percent response rate from centers representing 54 percent of all patients served nationally), the migrant health centers identified the following as the most common conditions among their maternity and pediatric patients: malnutrition, anemia, hypertension, gestational diabetes, and infection among the pregnant women. For children, the most commonly reported conditions were lack of
immunizations, the need for routine exams and dental care, developmental disabilities, dysentery, malnutrition, infectious and parasitic disease, skin disorders, hypertension, fever, measles, and anemia (National Association of Community Health Centers, 1991). Unfortunately, although the list of commonly reported diagnoses is useful, the survey methodology does not permit the calculation of incidence rates.
Another analysis of the migrant health center data used diagnostic codes recorded for all visits during appropriate three-month periods when migrant workers were employed in four migrant centers in three states (Michigan, Indiana, and Texas) that are in the midwestern migratory "stream" (Dever, 1990). The Texan migrant health centers are considered to be "home base" or "downstream" sites, whereas the Indiana and Michigan centers are "nonhome base'' or "upstream" centers. The study also looked at the demographics of the counties in which the centers were located. The data in the next two paragraphs are drawn from this study (Dever, 1990).
Unfortunately, the study draws conclusions using a mix of data from the counties' demographics and data from the actual encounters at the migrant centers, thereby raising certain methodological questions. Still, the findings indicate an overall trend: using major diagnostic groups (after all the diagnoses were coded according to International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM], categories), disorders of the newborn, burns, ear/nose/throat (ENT) conditions, infectious/parasitic disease, injury and poisoning, and eye disorders all exceeded the reported U.S. indices by ranges of from 25 percent to 150 percent. The most common principal diagnoses for all age groups were diabetes mellitus (8.3 percent of total diagnoses), well-child care services including immunization (6.7 percent), otitis media (5.9 percent), pregnancy (5.5 percent), upper respiratory infection (4.5 percent), essential hypertension (4.2 percent), contact dermatitis and other eczema (2.5 percent), and hard tissues of the teeth disease (2.2 percent). Beginning with children ages 10–14 and continuing through older age groups, significant dental disease was noted, especially among ages 15–19 for whom hard tissues of the teeth disease accounts for 6.3 percent of all visits, indicating a lack of appropriate dental care at earlier ages. In addition, beginning with adolescence, diseases related to agricultural work begin to appear, especially contact dermatitis, parasitic diseases, sprains and strains, and injury. By late adolescence, ages 15–19, visits for diabetes mellitus begin to be more common. For females in this age bracket, diabetes is the third most common reason (4.6 percent) for seeking care, following only pregnancy and dental disease. Diabetes accounts for an increasing proportion of visits for women throughout all adult age groups. Beginning in their thirties, diabetes becomes an increasingly frequent diagnosis for men. By ages 45–64, the top four diagnoses (diabetes, hypertension, arthropathies, and soft tissue diseases) account for 50 percent of all visits.
When the author of this study compared the top 20 principal diagnoses from the four migrant health centers examined with those reported from the National Ambulatory Medical Care Survey (NAMCS), he found overlap among only 8 diagnostic categories. Twelve of the diagnostic categories noted in data from the migrant health centers did not appear as common visits in the NAMCS data, which represented all U.S. physician visits. These 12 categories included infectious, nutritional, and occupational (including contact dermatitis and eczema) diagnoses. Visits for diabetes were 338 percent above the NAMCS figures. Visits for otitis media and acute respiratory infection were also overrepresented in the migrant health centers (138 percent and 97 percent greater, respectively). As the author correctly notes, using proportions of clinic visits for specific diagnoses fails to provide information regarding disease incidence or prevalence; however, the variation in visits by principal diagnoses between migrants and seasonal farmworkers and their dependents and the general population suggests that migrant and seasonal agricultural workers suffer from different health problems and a greater burden of chronic diseases at a younger age than do most Americans.
Other studies also point to increased health risks among migrant and seasonal farmworkers. Earlier reports documenting the most common diagnoses (at rates far above the national averages) confirm that the major reasons for seeking health care are diabetes, hypertension, and cardiovascular disease. Infectious diseases, especially parasitic diseases, account for a relatively higher proportion of visits than is found among the general population (Health Care Resources, Inc., 1984). The U.S. Bureau of Labor Statistics estimates that there are 12.7 cases of injury and illness per 100 full-time workers per year and 1,700 work-related deaths (52 per 100,000 workers). This makes agriculture the nation's most hazardous occupation. A population-based, cross-sectional study of migrant farmworkers in eastern North Carolina revealed that 8.4 percent (24 of 287 interviewed) had reported an occupational injury during the previous three years (Ciesielski et al., 1991). Another survey reported that 44.5 percent of farmworker households have a disabled individual (Inter America Research Association, 1974).
In addition to the injuries and illnesses attributed directly to agricultural work, chronic low-level pesticide exposure carries potential risks, including teratogenesis and carcinogenesis (Rust, 1990). Farmworkers also appear to be at greater risk of acquired immune deficiency syndrome (AIDS) and other infectious diseases, including tuberculosis. A study by the Centers for Disease Control (CDC) in 1988 found a 0.4 percent prevalence of human immunodeficiency virus (HIV) among farmworkers who sought care for any condition; however, a more recent study of farmworkers in the migrant camps of southern New Jersey found a 3.2 percent rate of seropositivity, eight times the national rate (Lyons, 1992).
If knowledge of the health of migrant and seasonal farmworkers and their dependents is limited, an understanding of their health care utilization and the access barriers they actually experience is even more general and inferential. Ninety percent of all migrant families have family incomes below the federal poverty level, and the per-capita income in communities heavily populated by migrant families is half the U.S. average (National Migrant Resource Program, Inc., undated). In 1985, the average migrant farmworker earned only $3,295 per year from farm labor; his or her total income from all sources was only $6,194. (Rust, 1990).
Despite their poverty and the virtual absence of private insurance, migrants and seasonal farmworkers experience more barriers to obtaining Medicaid than other low-income groups. A survey of migrant health centers, conducted by the National Association of Community Health Centers in the spring of 1991, documented the remaining barriers, from the providers' perspective, after the Medicaid expansions mandated by Congress in 1989 and 1990 (National Association of Community Health Centers, 1991). Additional barriers to receiving Medicaid benefits would surely emerge if migrants and seasonal farmworkers were queried directly. The survey's most important findings include (1) the difficulty migrants have in establishing state residency and completing the application process before they must move on (these difficulties have persisted despite 1979 Health Care Financing Administration [HCFA] regulations that attempt to ease residency requirements for migrants); (2) the problems migrants experience in retaining coverage and satisfying periodic redeterminations once they receive benefits; (3) the barriers created by documentation and application procedures (because of the time and level of paperwork involved); and (4) the language and cultural barriers inherent in the application process, including the unavailability of forms and translators for non-English-speaking migrants. Forty-three percent of respondents reported mobility-related problems, 43 percent reported language barriers, and 77 percent reported documentation problems. These problems were almost equally present for pregnant women and children—notwithstanding the elimination of categorical eligibility limitations and the liberalization of financial eligibility requirements, which should mean that nearly all pregnant women and children in migrant families would be able to meet Medicaid eligibility standards. Other problems noted by the migrant health centers in this survey included the migrant's inability to comply with face-to-face interview requirements; states' continuing to require permanent residence, despite HCFA guidelines to the contrary; states' denial of benefits to lawful residents because of misapplication of federal alienage standards; and the failure to have hours and locations that are accessible to migrants. These difficulties in obtaining Medicaid benefits were reported by migrant health centers that are presumably highly motivated to help migrants; consequently, the difficulties faced by migrants who
are outside the system of migrant health centers can be assumed to be much greater.
Medicaid coverage is only one of the potential factors that enable migrants and seasonal farmworkers to obtain access to necessary health services. A study of health care utilization in Wayne County, New York, performed during the summer of 1982 questioned a sample of migrants living in migrant camps rather than only surveying those utilizing health services (Chi, 1985). This study provides the data in this and the following paragraphs. The study noted that Medicaid recipients, compared with those not receiving Medicaid assistance, had a greater likelihood of visiting physicians for diagnostic and preventive health care. At least in this study, however, factors related to a person's history as a migrant played a more significant role in determining the probability of seeing a physician than did Medicaid status. Migrants who were native born and had been in the migrant stream for a greater length of time were more likely to have seen a physician in the preceding year than were recent immigrants or people new to the migrant stream. Of significance for this study, the county in which this study was conducted contained a federally funded migrant health clinic. Long-term migrants had greater knowledge of, and were more likely to use, this migrant health center. In addition, these long-term migrants were older and slightly better educated, and had a higher probability of having worked in this county or for the same employer previously. Presumably their knowledge of the migrant health center, as well as of the area, increased their ability to navigate through the health care system and gain access to care. Of those migrants who were recent immigrants, 45 percent had no knowledge of the migrant health center, as compared with only 10.4 percent of the long-term migrants. Fifty-one percent of the recent-immigrant migrants had not seen a physician in the preceding year. For all migrants, 64 percent had not seen a physician or had only seen a physician once during the previous year, a far smaller figure than the national norm. As with the general population, female migrants saw physicians more frequently than male migrants.
Equally of interest, more than 40 percent of migrant farmworkers in the sample delayed medical care or treatment for an existing medical or dental problem including such conditions as anemia, arthritis, blood in stools, hypertension, broken bones, ulcers, and chest pain. Migrants cited lack of time, followed by economic cost and lack of access (nonspecific), as the reasons cited for delaying care. Almost 25 percent listed fear or lack of confidence in the medical profession as a reason for not seeking or delaying care.
In the population-based study of occupational injuries among North Carolina migrant farmworkers conducted by Ciesielski and colleagues (1991), 11 of the 17 more seriously injured workers (65 percent) either did not receive prompt care (7/17, or 41 percent) or never received care at all (4/17,
or 24 percent). Injured farmworkers who did not receive prompt care were twice as likely to have incomplete recoveries. Refusal by the crew leader to allow workers to seek care and lack of transportation prevented 24 percent from receiving care within 24 hours and 42 percent from keeping follow-up appointments. Although this study focused on occupational injuries, an incidental finding was that 19.6 percent of the farmworkers who worked in the tobacco fields of North Carolina reported nausea, and 18.6 percent reported dizziness. None of the farmworkers reported these symptoms as injury.
Studies such as these, which survey migrants directly, can add a great deal to our knowledge of migrant health care utilization behavior and the barriers to access faced by migrants. Combined with information about self-perceived health status and health practices, these studies would allow for appropriate measures of access, as well as provide important information about disease incidence and prevalence. Unfortunately, a literature search failed to turn up additional studies of this kind.
HEALTH STATUS OF THE HOMELESS
As with migrants, there is no consensus on the number of homeless people in this country. The inability to agree on this number (the denominator) makes calculations of rates of disease and health-related problems, as well as measures of access, extremely problematic. Many observers consider the federal estimate derived from the 1990 census, 228,621, too low; these observers include local and state government officials, advocates—and even the Census Bureau itself (Noah, 1991). Estimates of the homeless population vary from this low census number to several million. According to the Wall Street Journal, in 1989 the Urban Institute estimated the homeless population at 600,000; yet estimates of homeless youths alone, age 21 and younger, from the National Network for Runaway and Homeless Youth, range from 250,000 to several million. The number of homeless families, of mentally ill homeless, of rural homeless, and of homeless elderly are also in dispute.
Despite this inability to quantify the extent of homelessness, an enormous number of studies of the health care needs and health status of homeless people have been published since homelessness emerged as a major national policy issue in the early 1980s. Pioneering work on delivering health care to the homeless and studying their health problems was done under the direction of Philip Brickner, M.D., at St. Vincent's Hospital in New York City (Brickner et al., 1990). In 1985, the Robert Wood Johnson Foundation, the Pew Memorial Trust, and the United States Conference of Mayors established Health Care for the Homeless Demonstration Projects in 19 large cities. (Wright and Weber, 1987). All of the Health Care for the
Homeless projects funded through this program participated in a major research effort directed by the University of Massachusetts' Social and Demographic Research Institute. In addition, a number of surveys and special studies have focused on the relationship between homelessness and mental illness (Blackwell et al., 1990) and homelessness and alcohol and substance abuse (Institute of Medicine, 1988).
Reviews of these studies reveal that the homeless suffer from many of the same acute and chronic illnesses that afflict people in the general population but at much higher rates (Brickner et al., 1990). Because the homeless have little or no access to adequate bathing and hygienic facilities, survive on the streets or in unsafe and generally unsanitary shelters, smoke and drink to excess, and suffer from inadequate diets, their physical health is compromised. Among the findings from the Health Care for the Homeless Demonstration Projects were that the most commonly reported acute conditions were upper respiratory infections, trauma, and skin ailments. Nutritional deficiencies were found in 2 percent of those seen. Of patients seen more than once, 37 percent had at least one chronic condition including hypertension, arthritis and other musculoskeletal disorders, dental problems, gastrointestinal and neurological disorders, peripheral vascular disease, genitourinary problems, and chronic obstructive pulmonary disease. The use of estimation techniques based on recorded diagnoses led to estimates that 38 percent of the homeless seen in the demonstration projects abused alcohol, 13 percent abused drugs other than alcohol, and 33 percent were mentally ill. During the demonstration period, the rates of tuberculosis (968 cases/100,000 population) and AIDS (230 cases/100,000) were significantly higher than those of the general population (e.g., 9 cases/100,000 for TB) (Wright and Weber, 1987). While these findings reflect the homeless population that sought care from the demonstration projects, which may inflate the burden of illness, the demographic characteristics of the patients seen in the demonstration projects did not differ significantly from those described in many ethnographic studies of the homeless (Wright and Weber, 1987).
As with the migrant and seasonal farmworker population, it is difficult to calculate vital statistic rates for the homeless. Not only is the denominator in dispute, but neither birth nor death statistics record homelessness. A study conducted in New York City compared infants born to women living in welfare hotels with infants born to women living in low-income housing projects. The babies born to homeless women living in the hotels were more likely to be of low birthweight (18 percent vs. 8.5 percent) and had a higher infant mortality rate (Chavkin et al., 1987). At the other end of the age spectrum, many who work with the homeless report that very few are over the age of 55, which suggests that the homeless die young. In support of this contention, the median age of those seen in the Health Care for the
Homeless Demonstration program was 33 (Knight and Lam, 1986). A 1984 study of Baltimore's homeless found that only 2 percent were age 65 or older, compared with 18.1 percent of the general population (O'Connell et al., 1990).
It is always difficult to disentangle the effects of access to health care, or, conversely, lack of access, on health status; for the homeless, the difficulties in doing so increase exponentially. Indeed, poor health and the resulting inability to work, often byproducts of homelessness, may also result in homelessness. Although homeless people may have the same array of acute and chronic problems as one finds in the general population, the rates are clearly higher, and the numbers of comorbidities, including alcohol/substance abuse and mental illness, are far in excess of these rates for the general population. Many of these conditions and morbidities are amenable to medical intervention; routine health care should prevent some diseases altogether and minimize exacerbations and complications of chronic diseases. Yet the personal health care services needed by the homeless may require a different organizational configuration, a different array of services, and a different mix of providers than those required for the domiciled population. These suppositions, as well as the inability of traditional clinics and hospitals to care for the homeless adequately, gave rise to the Health Care for the Homeless Demonstration Projects and the subsequent U.S. Public Health Service's McKinney Health Care for the Homeless Program. These demonstrations rely on community-based programs that are often colocated in places in which the homeless may be found in large numbers, such as congruent feeding programs and shelters.
Although the evidence that the homeless lack access to health services, except through targeted programs, is anecdotal and inferential, it is quite convincing. In the 1988 Institute of Medicine report Homelessness, Health, and Human Needs the chapter on access reviews the limitations in systems of care for the poor and medically indigent (e.g., in public general hospitals and not-for-profit hospitals serving the poor, clinics, the National Health Service Corps, categorical programs, mental health and Veterans Administration systems, and Medicaid programs) and suggests that, in general, the homeless compete with the poor for these services. In addition to general underfinancing of health care services for the poor, the report identified additional barriers to access facing the homeless: bureaucratic and scheduling issues, lack of transportation, negative perceptions on the part of providers and institutions, and the avoidance of institutions by the homeless themselves because of prior experience (Institute of Medicine, 1988). Despite the lack of quantifiable data, no one has disputed the statements made on the original brochure for the Health Care for the Homeless Demonstration Projects: "Most homeless people do not now receive needed health services. Many are afraid of large institutions, most are uninsured, and
many are perceived in some sense to be 'undesirable' as patients" (Robert Wood Johnson Foundation, 1983).
Local coalitions have attempted to document the lack of health care received by the homeless, generally by interviewing patients during a health care visit. In St. Louis, for example, reports indicate that, of the homeless seeking care, more than 70 percent had no usual health care provider and more than half had not received any health care attention in the previous year (Wright and Weber, 1987). In the study of pregnant homeless women living in New York City's welfare hotels (Chavkin et al., 1987), 56.4 percent of the women from the hotels reported three or fewer prenatal visits, compared with 22.5 percent of women in low-income housing projects and 15 percent of women citywide. The New York Children's Health Project reported that in calendar year 1988, of the 3,084 children seen, only 52 percent were adequately immunized and many were undertreated for acute and chronic illnesses. Both findings were attributed to poor or no access to health care services (Brickner et al., 1990).
Few systematic or rigorous studies have been done at shelters, on the streets, or at other gathering places. As discussed earlier, surveys on access generally rely on telephone interviews, thereby eliminating the possibility of participation by the homeless. The general agreement among policymakers and advocates that the homeless continue to have inadequate access to appropriate health care services has not been tested. Because no new federal money has been available for the McKinney Health Care for the Homeless program, few cities or rural areas have performed recent systematic health care needs assessments of this population group. Many cities and advocacy groups do report that the number of homeless continues to increase, which in turn suggests that additional service capacity is needed.
For two of the most vulnerable populations in this country—migrants and seasonal farmworkers and their families, and the men, women, and children who are homeless—neither the traditional measures of access nor the IOM committee's recommended indicators will provide the necessary information to measure and analyze either the barriers that prevent these groups from receiving appropriate, high-quality health care or the nation's progress toward ensuring equity of access. Despite some increased knowledge about the health status and health care needs of both the migrant and homeless populations (with considerably more information about the homeless), we still lack certain baseline measures regarding the access and health care needs of both populations. Although this paper is meant to discuss the development of indicators of access for these populations, indicators of
access must include an assessment of unmet health care needs. Information regarding resource allocation and effectiveness is necessary as well.
Neither the migrant/seasonal farmworkers nor the homeless are homogeneous populations. Both groups are made up of various subgroups. The health care needs and access experiences of a Chicago migrant family in the Texas Rio Grande valley will be very different from those of a single man in the East Coast migrant stream. The health care needs of a mentally ill homeless woman on the streets of New York City and her ability to maneuver through the health care system differ enormously from those of a family living in a shelter or of a young displaced worker in a rural area. Much more information is needed to understand the differing health care needs of these subgroups.
Most importantly, our need for information cannot be divorced from our commitment to provide many different types of services to these populations. The IOM study on access is limited in its scope to access to personal health care services. yet, for migrants and the homeless, the most significant access issues involve access to social, environmental, and occupational reforms that will promote health and prevent disease, as well as to personal health care services. Tracking programs designed to eliminate homelessness must accompany monitoring of the access of homeless people to health care services. Monitoring access to health care services for migrants is good health policy only if we also document migrants' access to safe drinking water and decent living conditions. We seek information from indicators of access to inform health care policy. For the migrants and the homeless, monitoring access to personal health care services is a necessary but insufficient step in understanding the interacting factors that contribute to excess morbidity and mortality in these populations.
The following recommendations address some of the gaps in our knowledge regarding access.
The Migrant Health programs (Section 329 of the Public Health Service Act) and the McKinney Health Care for the Homeless programs (Section 340 of the act) should develop data systems that include clinical information necessary to assess the health status of these populations and information regarding utilization of health care services and access barriers encountered within the health care system. Although this information will come from those migrants and homeless people who already use the two components of the health care system that have been specifically designed to minimize access barriers, much useful information can be collected, as seen by the studies done for the original Health Care for the Homeless Demonstration Projects (Wright and Weber, 1987). Particularly because both migrants and the homeless are relatively mobile populations, their experiences in obtaining health care will vary over time and by
location. Surveys of patients using Migrant Health Centers and Health Care for the Homeless programs at one point in time may provide information about their experiences in accessing other services.
Unfortunately, with the conversion of support of Health Care for the Homeless programs from foundations to the Public Health Service, neither the funding nor the commitment to maintaining this complicated information base on the homeless continued. The Bureau of Health Care Delivery and Assistance (BHCDA) of the Health Resources and Services Administration (HRSA), the agency that oversees these programs, has recently committed itself to developing a system for collecting demographic and clinical information. The necessary fiscal and personnel support must be guaranteed to ensure that the data collected provide the information needed on the health status, utilization patterns, and access issues faced by these two populations. The data collected by the BHCDA must therefore include the findings from systematic health evaluations as well as the problems individuals present in seeking help at an appointment. Longitudinal information should be sought whenever possible.
Because most indicators are derived from secondary sources or telephone/household surveys in which both migrants and the homeless are underrepresented, special, community-based surveys should be developed to provide information about health status and access. These surveys should be carried out in migrant camps, homeless shelters, congruent feeding facilities, and other such locations to gain a more complete picture than is currently available. In addition, sampling techniques need to be developed to ensure adequate representation of the subgroups that make up these populations. Surveys also need to be conducted in multiple geographic regions of the country to capture regional differences. The surveys should be patterned after the National Health Interview Survey, the National Health and Nutritional Evaluation, and the Robert Wood Johnson Foundation Access Survey in order to permit national comparisons. Supplemental questions should be developed to address the special circumstances of migrants and the homeless.
In coordination with the development of these surveys, the Federal Interagency Committee on Migrants (which is convened quarterly by the Office of Migrant Health and includes representatives from the Departments of Education, Justice, Labor, and Agriculture, and from the Environmental Protection Agency) should develop a research agenda on migrant health issues that would include specific measures of access to personal health care services, including alcohol and substance abuse treatment and mental care health programs. In addition to the measures of access to these services, the research agenda must include studies that allow for calculating vital statistic rates including maternal and infant mortality, low birthweight, and age- and cause-specific mortality.
Special studies may be needed to establish rates of hospitalization for ''ambulatory-sensitive conditions" in order to develop comparisons.
The Federal Interagency Committee on Migrants should be expanded to include representatives from state and local government and advocacy groups (e.g., the National Governors Association, the Association of State and Territorial Health Officers, the National Association of Community Health Centers, the Migrant Clinicians' Network).
The Health Care for the Homeless program should initiate a similar interagency group on the homeless. One of the immediate tasks of such a group would be the development of a research agenda. Many of the same topics, especially the need for information regarding access and the need for vital statistics, are as relevant to the homeless as they are to migrants. Additional issues related to access particularly for the homeless, include the relationships among poor health status (especially mental health), lack of access to appropriate health care services, and the precipitation or continuation of homelessness.
Special studies should be undertaken to evaluate the effectiveness of targeted initiatives in increasing access. These initiatives include, for example, changes in Medicaid and Supplemental Security Income (SSI) eligibility requirements (including guidelines for presumptive eligibility), outstationing of enrollment workers, and changes in residency requirements. Many of these initiatives were specifically designed to make enrollment and retention of Medicaid benefits easier for migrants and homeless people. Whether these initiatives have succeeded remains to be evaluated.
Special studies are needed to examine nonfinancial barriers to care, especially provider and institutional willingness to provide services to these populations, the influence of organizational structure and hours of service on accessibility, transportation and translation services, and other factors that might influence the willingness of patients to seek services. Both the migrant and homeless populations, or some subgroups among them, may also experience nonfinancial barriers that differ in kind or in scope from those experienced by other population groups.
The feasibility of designing special coding for hospital discharges and birth and death certificates to identify the homeless and migrants should be explored. Although this plan may not prove to be feasible or cost-effective nationally, a targeted study in certain regions might provide important information now lacking.
National surveys and followback studies (for example, the National Center for Health Statistics' Mortality Followback survey) should specifically ask whether the index case, or any family member covered by the survey, was ever homeless or ever worked as a migrant. Occupation is surveyed in some studies, but the question is generally asked in
relationship to work in the two weeks prior to administration of the survey; this restriction may lead to a failure to identify migrants and seasonal farmworkers. Household address is almost always requested, yet information may be lacking to indicate whether the address represents a shelter or temporary housing. Although these surveys probably do not sample enough migrants or homeless to allow for subgroup analysis, the information may prove useful, depending on the extent of the response.
HRSA has cooperative agreements with most state departments of health and supports primary care associations in many states. In developing primary care needs assessments, state departments of health should be encouraged to develop specific assessments of needs and resources for migrant and the homeless, complete with estimations of the number of individuals affected in each local geographic area. Qualitative data, including interviews with care providers, local health care institutions, and advocates and representatives of the homeless and migrants, should be included to assess the accessibility of existing services.
Many state and local health departments have developed infant mortality reviews to ascertain contributing factors. Infant mortality reviews should specifically address whether the infant in question was homeless or the child of migrants. The information gleaned from infant mortality reviews should be aggregated nationally for many purposes. With this kind of aggregation, the deaths of migrant and homeless infants could be analyzed as a separate subgroup.
Unfortunately, developing indicators for monitoring the access to health care services of migrants and the homeless cannot be divorced from the need for special studies. There exists a national commitment to providing targeted services through the McKinney Health Care for the Homeless and the Migrant Health Center programs. These programs, however, have had only small increases in funding (less than is necessary to keep up with inflation in the health sector), and few would argue that they have solved the access problems of the majority of migrants or homeless people. If we are to understand the remaining barriers and the extent to which limitations in access contribute to the reported poor health of both groups, we must commit sufficient resources to perform the necessary studies.
Aday, L. A. and Anderson, R. M. 1975. Development of Indices of Access to Medical Care. Ann Arbor, Mich.: Health Adventist Press.
Aday, L. A. and Anderson, R. M. 1984. The national profile of access to medical care: Where do we stand? American Journal of Public Health 74:1331.
Blackwell, B., et al. 1990. Psychiatric and mental health services. Pp. 184–203 in Brickner, P. W., Scharer, L. K. et al., eds. 1990. Under the Safety Net: The Health and Social Welfare of the Homeless in the United States. New York: Norton.
Brickner, P. W., Scharer, L. K. et al., eds. 1990. Under the Safety Net: The Health and Social Welfare of the Homeless in the United States. New York: Norton.
Chase, H. P., Kumar, V., Dodds, J. M., Sauberlich, H. E., et al. 1971. Nutritional status of preschool Mexican-American migrant farm children. American Journal of Diseases of Children 122:316–324.
Chavkin, W., Kristal, A., Seaborn, C., and Guigli, P. 1987. The reproductive experience of women living in hotels for the homeless in New York City. New York State Journal of Medicine 87:10–13.
Chi, P. 1985. Medical utilization patterns of migrant farmworkers on Wayne County, New York. Public Health Reports 100:480–490.
Ciesielski, S., Hall, P. and M. Sweeney. 1991. Occupational injuries among North Carolina migrant farmworkers. American Journal of Public Health 81:926–927.
Dever, A. 1990. Migrant Health Status: Profile of a Population with Complex Health Problems. Austin, Tex: Migrant Clinicians Network Monograph Series.
Health Care Resources, Inc. 1984. An Assessment of Selected Health Conditions of Migrant and Seasonal Farmworkers: Draft Report on the Literature Search. Prepared for the Bureau of Health Care Delivery and Assistance, Health Resources and Services Administration, Washington, D.C.
Hopper, K., and J. Hamberg. 1984. The Making of America's Homeless: From Skid Row to the New Poor, 1945–1984. Working Papers in Social Policy. New York: Community Service Society.
Hubbell, F. A., Waitzkin, H., Mishra, S. I., Dombrink, J., and Chavez, L. R. 1991. Access to medical care for documented and undocumented Lationos in a southern California county. Western Journal of Medicine 154:414–417, April.
Institute of Medicine. 1988. Homelessness, Health and Human Needs. Washington, D.C.: National Academy Press.
InterAmerica Research Association. 1974. Handicapped Migrant Farmworkers . U.S. Department of Health, Education, and Welfare Pub. OHD 75–25084. Washington, D.C.: U.S. Government Printing Office.
Knight, J. W., and Lam, J. 1986. Homelessness and Health: A Review of the Literature. Amherst, Mass.: SADRI.
Migrant Health Newsline. 1992. Vol. 9, No. 2. Austin, Tex.: National Migrant Resource Program.
National Association of Community Health Centers. 1991. Medicaid and Migrant Farmworker Families: Analysis of Barriers and Recommendations for Change. Washington, D.C.: NACHC.
National Migrant Resource Program. Undated. A Migrant Health Status: Profile of a Culture with Complex Health Problems. Austin, Tex.: The Program.
National Migrant Resource Program and the Migrant Clinicians Network. 1990. Migrant and Seasonal Farmworker Health Objectives for the Year 2000. Austin, Tex.: The Program.
Noah, T. 1991. Census Bureau's count of homeless fails to end debate. Wall Street Journal April 15, p. B5C.
O'Connell, J., Summerfield, J. and R. Kellogg. 1990. The homeless elderly. P. 151–153 in Brickner, P. W., Scharer, L. K. et al., eds. 1990. Under the Safety Net: The Health and Social Welfare of the Homeless in the United States. New York: Norton.
Office of Migrant Health, Health Resources and Services Administration. 1992. Farmworker Health for the Year 2000:1992 Recommendations of the National Advisory Council on Migrant Health. Rockville, Md.: U.S. Department of Health and Human Services.
Robert Wood Johnson Foundation. 1983. Health Care for the Homeless Program. Princeton, N.J.: The Foundation.
Robert Wood Johnson Foundation. 1986. Access to Health Care in the United States: Results of 1986 Survey. Special Report 2, Princeton, N.J.: The Foundation.
Rust, G. 1990. Health status of migrant farmworkers: A literature review and commentary. American Journal of Public Health 80:1213–1217.
Slesinger, D., and B. Christensen. 1986. Health and mortality of migrant farm children. Social Science and Medicine 23:65–74.
U.S. Bureau of the Census. 1990. Statistical Abstract of the United States. 110 Ed. Washington, D.C.: U.S. Department of Commerce, pp. 636–64.
U.S. General Accounting Office. 1985. Homelessness: A Complex Problem and the Federal Response. Washington, D.C.: U.S. Government Printing Office.
Wright, J. D., and Weber, E. 1987. Homelessness and Health. Washington, D.C.: McGrawHill.