Health Status of Homeless and Marginally Housed Users of Mental Health Self-Help Agencies

Abstract

The study discussed in this article investigated the health status of 310 homeless and marginally housed people to determine the usefulness of mental health self-help agencies (SHAs) in addressing their physical health needs. The study compared self-reported health problems among SHA users with similar reports and clinical assessments of other homeless or marginally housed populations. Findings indicate that frequencies of health problems among respondents were similar to those of other homeless or marginally housed groups and that the study group had a higher prevalence of HIV infection and tuberculosis than the general population. Because this hard-to-reach group actively seeks SHAs, these organizations may be uniquely suited to health outreach, education, testing, and treatment.

The high prevalence of severe health problems and elevated death rates among people with severe mental disabilities is well documented (Segal & Kotler, 1991). State mental hospitals in the past provided comprehensive health care for this population. The shift to community mental health care and the concomitant decentralization and diffusion of responsibility has placed people with severe mental illness at increased risk of not receiving adequate health care. These problems may be particularly severe for the segment of this population that is poor and lives in marginal and temporary housing or alternates between such housing and homelessness. These individuals, often alienated from the traditional mental health system, which is frequently unable to serve them, have begun increasingly to use easy-access, low-self-disclosure settings such as self-help agencies (SHAs) (Segal & Baumohl, 1988). SHAs, founded on principles of mutual assistance by and for consumers during the past 15 years, have become inexpensive alternatives by which financially strapped mental health systems can provide social services to people who are homeless and have severe social, mental, and physical disabilities.

With the rise in managed care organizations and increasing restrictions on public health spending, vendorship and efficient siting of health services has become a critical issue. If significant numbers of people who frequent SHAs are found to have serious and chronic health conditions along with low utilization rates of necessary health care provided at traditional sites, the use of mental health SHAs could reduce the use of expensive general hospital emergency rooms for physical health care. This article identifies the health problems of homeless and marginally housed people seeking help at four mental health SHAs, discusses the extent of their use of health care services, and addresses the question of whether these agencies might be appropriate service settings for targeted interventions.

SHAs are usually multiservice settings offering survival resources such as food and shelter referrals, as well as advocacy, peer counseling, and housing search assistance. A major component of the SHA is a drop-in facility, in which members can socialize, drink coffee, and play table games. SHAs are run by and for the consumers of their services. Past and current users make up the majority of frontline service providers. Users may often achieve their original service objectives but continue to come to the SHA for companionship or additional services or to volunteer.

The demographics of individuals served by SHAs are similar to those of most poor homeless or marginally housed populations (Segal, Silverman, & Temkin, 1995). Their high-risk health status can best be determined by comparing the prevalence of physical health problems among SHA users with that of other homeless or marginally housed groups. Since the early 1980s, a great deal of research has focused on these homeless groups. Much of this research attempted to obtain demographic descriptions (Bassuk, 1984; Bassuk, Rubin, & Lauriat, 1986; Fisher, Breakey, Shapiro, Anthony, & Kramer, 1986; Roth, Bean, Lust, & Saveanu, 1985) or examined mental health status in depth (Bachrach, 1984; Farr, Koegel, & Burnam, 1986; Gelberg, Linn, & Leake, 1988; Segal, Baumohl, & Johnson, 1977; Susser, Struening, & Conover, 1989).

Far less attention has been given to the physical health status of such groups (Struening & Padgett, 1990; Wright, 1990; Wright & Weber, 1987; Wright, Weber-Burdin, Knight, & Lam, 1987). Some studies have identified various health conditions, such as respiratory, gastrointestinal, and musculoskeletal problems and cutaneous infections (Noble, Scott, Cavicci, & Robinson, 1985) or noted higher incidences of disorders such as scabies and lice infestation, hypertension, tuberculosis, peripheral vascular disease, nutritional disorders, and disorders resulting from exposure or trauma (Brickner, Scharer, Conanan, Elvy, & Savarese, 1985).

The results of the study discussed in this article provide a validation of the high-risk status of the mental health SHA user population, giving a description of their health problems and a comparison of these to what is known about the health problems of other poor homeless or marginally housed adults.

Method

The study is part of a larger investigation of mental health SHA users and their organizations. We have previously reported analyses related to psychiatric diagnosis and patterns of use of SHAs (Segal et al., 1995).

Participants

Participants were long-term adult users of four client-run mental health SHAs in the San Francisco Bay area. Data were gathered from 1992 to 1993. Only participants who attended the SHA at least 12 times over a period of at least three months (inclusive of staff and volunteers, themselves clients) were included in the study. Virtually all staff and volunteers and a random sample of other long-term users (N = 323) were approached, because these individuals would be most likely to make use of health services provided at the SHA and thus were seen as the most appropriate for risk status assessment. Overall, 96 percent (n = 310) of those asked to participate in the study agreed to do so.

Study Sites

All four SHAs had drop-in programs and conformed to the general SHA model described earlier. All are located in urban settings—three in densely populated, low-income, ethnically diverse areas and the fourth in a developing residential area of single-family homes and duplexes. The latter site is, however, within walking distance of the urban downtown area and serves individuals who frequent this downtown area. Many agency members, including those who were homeless, lived or had lived in the areas where the agencies are located.

Data Collection

Data were collected by interviewers, both former mental health clients and mental health professionals, trained by the Center for Self-Help Research. Interviews assessed a number of variables, including demographic indicators (such as gender, ethnicity, income, marital status, and educational background), housing status, employment, and self-reported health status. The Diagnostic Interview Schedule (DIS) (Robins, Helzer, Cottler, & Goldring, 1988) was also administered as part of the assessment.

Measurement of Health Status

The study assessed the health status of the 310 participants with a general health status self-report question and the Health Problems Checklist, which was developed for the study. The checklist obtains self-reports on experiences with 34 possible health problems. Health access was assessed by asking respondents where, if at all, they normally received health care, the last time they had seen a doctor or other health practitioner, and how their health care had been paid for.

RESULTS

Participant Characteristics

The mean age of participants was 37, the median 38. Of particular note was the high proportion of African Americans (64 percent, n = 198) in the sample and the entire sample’s marginal housing status. If only the homeless population is considered, 46.1 percent (n = 143) of respondents lived on the streets or in a shelter. The remaining 53.9 percent (n = 167) were often precariously housed; 18.4 percent (n = 57) had to vacate their residence within two months of the start of the study and almost half (49.1 percent) within two weeks. Of those who had to leave within two months, 61.4 percent (n = 35) did not know where they would live next. Thus, 59.0 percent (n = 183) of the study group had no idea of where they would be living within the next two months and consequently were deemed homeless. In addition, 78.1 percent (n = 242) of respondents had been homeless at least once in the past five years—often for considerable periods of time. The median time homeless was a little more than two years. Ten percent (n = 31) had been homeless for five years.

Eighty-seven percent (n = 270) of the respondents had confirmed DSM-III-R (American Psychiatric Association, 1993) diagnoses. Using the standard of moderate to severe substance abuse/dependence, one-half (n = 155) had a dual diagnosis; an additional 20.0 percent (n = 62) had a diagnosis of only substance abuse/dependence (Segal et al., 1995).Thus, 70 percent (n =217) had a diagnosis of alcohol or drug abuse/dependence, possibly complicating their health status.

Health Status

When asked to evaluate their own health status, 42.9 percent (n = 133) reported being in fair or poor health. Thirty-two percent (n = 99) rated their health as good and 23.9 percent (n = 74) as excellent. Only 10 percent (n = 31) reported none of the 34 listed health problems during the preceding six months. The most commonly reported problem was trouble with teeth or gums (41.6 percent, n = 129) followed by back or spine problems (38.4 percent, n = 119) (Table 1).

Table 1.

Reports of Health Problems among Long-Term Homeless or Marginally Housed Users of Self-Help Agencies (N = 310)

Health Problem	%
Anemia	13.6
Arthritis or rheumatism	22.3
Asthma	12.0
Back or spine pain	38.4
Bowel or urination problem	15.8
Breathing trouble	27.7
Cancer	0.6
Chronic bronchitis	13.5
Cough (frequent) or chronic cold	20.0
Diabetes	4.5
Epilepsy, fits, seizures, or other similar disorders	7.4
Fainting	12.6
Foot cramps (frequent)	22.3
Gall bladder problem (chronic)	2.3
Headaches (frequent)	28.7
Heart trouble	11.0
Hepatitis	4.2
Hernia or rupture	3.2
HIV infection	4.5
Hypertension	24.5
Joint and muscle ache	33.5
Leg cramps (frequent)	23.2
Other injury or chronic condition	22.3
Pain in heart or chest	24.5
Paralysis (any)	3.5
Parkinson’s disease	0.3
Skin problem	26.8
Stomach or duodenal ulcer	10.5
Stomach pain (frequent)	17.1
Stroke	1.6
Swollen ankles	11.9
Teeth and gum problem	41.6
Tires very quickly	29.7
Tuberculosis	4.9

Respondents reported high prevalence of serious infectious diseases, most notably tuberculosis, 4.9 percent (15 of 310). We did not assess whether any of the cases were of active tuberculosis. Estimated tuberculosis infection among the national population is about 9 per 100,000 and in urban populations about 19 per 100,000 (Wright, 1990). The data for tuberculosis prevalence in the three geographic locales where our four SHAs operate justifies concern about the health risks of the homeless population. In 1993, the year we gathered data, San Francisco had 47.6 cases, Berkeley 11.4 cases, and Alameda County 12.8 cases of tuberculosis per 100,000, respectively (California Department of Health Services, 1994). We do not know whether the participants in the study over- or underreported the actual prevalence of tuberculosis. However, if we accept the number at face value, the long-term users of SHAs have a prevalence rate of 4,839 cases per 100,000. This rate is approximately 100 to 430 times larger than the base prevalence rates in these three locations. Another extremely worrisome result was the 4.5 percent of HIV infection reported by respondents: 14 cases among 310 respondents (or a rate of 4,501 cases per 100,000). This rate is, however, somewhat less than the 8.5 percent found among the general homeless population in San Francisco (Zolopa et al., 1994).

The reported rate of hypertension, 24.5 percent (n = 76) (not significantly different for African Americans and other races in the sample), is very similar to research results of ongoing studies of hypertension in the homeless carried out at St. Vincent’s Hospital in New York, which found abnormally high blood pressures in 25 percent to 33 percent of the patients (Kellogg et al., 1985). Results of the current study on the rates of anemia (13.6 percent, n = 42), bowel and urination problems (15.8 percent, n = 49), and neurological difficulties including epilepsy and other seizure disorders (7.4 percent, n = 23) conform with earlier research in showing greater prevalence in the adult homeless population (Harris, Mowbray, & Solarz, 1994; Ritchey, La Gory, & Mullis, 1991; Wright, 1990).

We reviewed conditions not listed in other studies that showed a high frequency of occurrence: frequent leg cramps (23.2 percent, n = 72), frequent foot cramps (22.3 percent, n = 69), and tiring very quickly (29.7 percent, n = 92).

Debilitating Characteristics.

Major debilitating characteristics that might influence the frequency of reported health problems are housing status, dual diagnosis, major mental disorder (meeting the DIS criteria for DSM-III-R diagnoses of schizophrenia and major affective disorders) and substance abuse of moderate or severe dependence.

No differences in reported health problems were observed between the homeless and the housed populations other than the frequency of foot cramps (29 percent for homeless people compared with 18 percent for housed people, p = .024), a fact probably attributable to the need of homeless individuals to be on their feet for long periods of time. The absence of differences between these groups also validates the decision to treat the sample as a unitary population of “poor often homeless/marginally housed” adults for the purpose of understanding their health problems.

Individuals with a dual diagnosis (that is, an Axis I or II DSM condition and substance abuse of moderate or severe dependence) reported eight conditions with greater frequency than those without such diagnoses: anemia (17 percent compared with 10 percent, p = .047), frequent cough or chronic cold (25 percent compared with 14 percent, p =.017), fainting (19 percent compared with 6 percent, p = .001), frequent headaches (34 percent compared with 23 percent, p = .039), heart trouble (30 percent compared with 19 percent, p = .016), joint and muscle ache (42 percent compared with 25 percent, p = .002), skin problems (32 percent compared with 21 percent, p = .022), and teeth and gum problems (50 percent compared with 33 percent, p = .003). Seven of these problems were reported with greater frequency among those classifiable by DIS criteria as suffering from schizophrenia or from a major affective disorder: anemia, frequent cough or chronic cold, fainting, headaches, joint and muscle ache, skin problems, and teeth and gum problems. In addition, this group reported greater frequency of bowel and urination problems (24 percent compared with 12 percent, p = .007) and back pain (50 percent compared with 33 percent, p = .004).

Five of the seven problems reported for those classified with dual diagnosis were reported with greater frequency among those with substance abuse problems of severe or moderate dependence irrespective of any other diagnosis: frequent cough or chronic colds, fainting, joint and muscle ache, skin problems, and teeth and gum problems. In addition, individuals with substance abuse problems reported a higher frequency of back and spine pain than respondents without this problem (44 percent compared with 32 percent, p = .039). The nature of these health problems may relate to substance abuse side effects. They also appear to involve persistent pain and conditions treated with over-the-counter medications in the general population. The latter facts may support the hypothesis that substance abuse is in part a form of self-medication among the homeless population in the absence of easily accessible (nonbureaucratically provided) prescribed medications.

Demographic Characteristics.

Differences in the frequency of reported conditions by gender appear to be attributable in part to an interaction of interviewer-respondent gender. The interviewers were mostly women; apparently, men reported fewer health problems when they were interviewed by women. The opposite was not true of women, a fact perhaps attributable to their greater openness in reporting health problems (Ritchey et al., 1991). Women reported eight health problems with greater frequency than men: anemia (29 percent compared with 8 percent, p = .000), asthma (24 percent compared with 7 percent, p = .000), breathing trouble (40 percent compared with 23 percent, p = .002), chronic bronchitis (22 percent compared with 10 percent, p = .009), frequent cough or chronic cold (28 percent compared with 17 percent, p = .019), frequent headaches (34 percent compared with 21 percent, p = .013), leg cramps (33 percent compared with 19 percent, p = .01), and frequent stomach pains (27 percent compared with 13 percent, p = .002).

Most differences in reported frequency of health problems occurred between African Americans and other races. Individuals from other races reported nine health problems with significantly greater frequency than African Americans: bowel and urination problems (21 percent compared with 12 percent, p = .038), breathing troubles (36 percent compared with 23 percent, p = .012), chronic bronchitis (22 percent compared with 9 percent, p = .001), frequent cough or chronic colds (28 percent compared with 15 percent, p = .008), hepatitis (9 percent compared with 5 percent, p = .001), HIV infection (10 percent compared with 1 percent, p = .000), skin problems (35 percent compared with 21 percent, p = .008), frequent stomach pains (23 percent compared with 14 percent, p = .038), and tuberculosis (9 percent compared with 2 percent, p = .007). Although these differences may be attributable to underreporting by African American respondents, we could not ascertain this to be the fact. On the contrary, other studies of the homeless show similar differences by race (First, Roth, & Arewa, 1988; Davis & Winkleby, 1993). Researchers further indicate that there is no evidence that African Americans over- or underreport risk factors to a greater degree than white people and have found inconclusive evidence regarding reporting by Hispanics (Davis & Winkleby, 1993).

Finally, we found few differences related to work status: Unemployed SHA users were more likely than staff and volunteers to have problems with swollen ankles (15 percent compared with 5 percent, p = .016) or with feeling tired when there was no reason to do so (33 percent compared with 21 percent, p = .044).

Health Care Utilization

The overwhelming majority of respondents reported some place that they could go for health care. Only 8.1 percent (n = 25) reported they had no place to go. However, accessibility was not the same thing as regular use or preventive care. Twenty-seven percent (n = 84) had received their care only from a hospital, presumably from the emergency room. Twenty-two percent (n =68) had not seen a doctor in more than a year for any reason, and 49 percent (n = 152) said they saw a medical doctor about once a year. Even smaller numbers got routine dental care. Fifty-nine percent (n = 183) had not seen a dentist within the past year.

Of those who had received medical care during the preceding year, 41.2 percent (n = 128) had their health care paid for by Medi-Cal—undoubtedly reflecting the SHAs’ focus on helping their clientele get certified for Supplemental Security Income. Fifty-four percent (n = 167) received at least some of their medical care free. The study did not ascertain the exact sources of the free care, but some respondents received it at free clinics, from Health Care for the Homeless vans making biweekly visits to the SHAs, or from volunteer medical personnel at the centers. Some of the respondents had received their care in prison or jail.

Thus, the majority of this homeless and marginally housed population maintained some connection to medical care and did not rely only on hospital emergency rooms. However, given the health problems reported by these individuals, they did not see doctors or other medical personnel with sufficient frequency.

Discussion

The long-term users of SHAs, at least in the San Francisco Bay area, suffer serious, debilitating physical health problems at a greater rate than the general population and at a rate similar to that found in other homeless adult populations. Subgroup differences in the reported frequency of specific problems are generally consistent with individual status and disability-associated needs.

The high prevalence of HIV infection and tuberculosis found in the self-report data is supported by several recent research reports that used “hard” tests, such as serum antibody status and tuberculin skin test reactivity, to confirm infection among the homeless population and more particularly that subgroup of homeless people who have severe mental illnesses (Empfield et al., 1993; Meyer et al., 1993; Susser, Valencia, & Conover, 1993; Zolopa et al., 1994).

The first and currently only representative study of HIV and tuberculosis infection among the homeless in a major U.S. city—San Francisco—has found an HIV prevalence of 8.5 percent and a tuberculosis prevalence of 32.0 percent (Zolopa et al., 1994). Three other recent studies, evaluating the HIV status of the homeless people with severe emotional disturbance, have found prevalence rates ranging from 5.8 percent to 6.4 percent for inpatient homeless people (Empfield et al., 1993; Meyer et al., 1993) to 19.4 percent for psychiatric patients in a large urban shelter (Susser et al., 1993). These somewhat higher prevalence rates suggest that the self-report data in our study are not biased by overreporting.

All these studies also confirm the difficulty of finding and treating this very vulnerable population. These studies support our observations that homeless and marginally housed people with mental disabilities are likely to be disaffiliated from social services and too disorganized to gain access to the mental health system; they apparently have little contact with outpatient medical settings. Moreover, the studies cited found that homeless people with mental illness may not go to public shelters, which often provide on-site medical and psychiatric evaluation.

Other studies have found that homeless individuals postpone seeking health care because of significant barriers posed by the system (Liebow, 1993; Snow & Anderson, 1993): There may be long waits, available sites may be located at a distance, and the care providers may not be sensitive to the particular limitations of a homeless existence (for example, prescribing bed rest or unaffordable medication). Furthermore, some homeless individuals may have been banned from public shelters where health care may be available because of their inability to deal with shelter rules.

Given the constellation of risk factors present—high prevalence of HIV and tuberculosis infection and overlapping infections and lack of access to medical services—it is probable that the health problems of homeless people will escalate as a public health problem in the United States. The potential implications of the situation go well beyond those who do not have their own homes (Zolopa et al., 1994). What have not been identified yet are convenient sites that could be used for providing the outreach, education, testing, and treatment necessary for battling the two very serious infectious disorders and other health conditions found among this extremely vulnerable group.

The data in our study corroborates the fact that the four San Francisco SHAs serve and are voluntarily frequented by this specific population. Because this population is less likely to deal with the established systems of physical and mental health care, these SHAs may be the most appropriate alternatives for delivering health care to the homeless population with severe mental illness. These agencies are run by their members and therefore may be uniquely positioned to reach and involve in treatment members who may be suffering from these ailments. One of the SHAs in the study has hired an AIDS educator. Another offers confidential testing on-site. The effectiveness of these services has yet to be evaluated.

Limitations of the Study

The SHAs that participated in the study may serve a unique group of self-helpers. Segal, Silverman, and Temkin (1995), however, documented that the SHAs’ respondent characteristics in the study were similar to those of other marginally housed or homeless adults found in urban settings. Health status assessment was based on self-reports, presenting issues as to the accuracy of the respondents’ answers. Usually these difficulties relate to either the underreporting or overreporting of facts or, as in the present study, actual health conditions. Clinical data are far superior and should be used whenever possible. The results of the study, however, compare well with studies using clinical assessments. Furthermore, two important studies that examined the validity of self-reported health data showed unambiguously that in relation to chronic health conditions such as HIV infection and tuberculosis, particularly HIV seropositivity, the tendency is for respondents to underreport (Lindan et al., 1994; U.S. Department of Health & Human Services, 1987). Thus, the most troubling findings, reported HIV and tuberculosis status, may be underreported. Future studies of this group should use standard clinical tests.

The data in the study are limited to 34 health issues, some not necessarily indicative of serious health problems. We, for example, did not inquire about eye or vision problems, sexually transmitted diseases, hearing difficulties (for example, ear infections), or female reproductive system-related health problems (Harris et al., 1994). Finally, caution must be taken in comparing self-report data in this study with published findings from other studies in which data were gathered in some other manner.

Conclusion

The findings in this article document the high prevalence among homeless and marginally housed people of serious and chronic life-threatening health conditions that require immediate treatment. If these findings can be further corroborated by more rigorous studies, then SHAs might have a vital role to play in the future delivery of essential health care services to this severely underserved population. Because these hard-to-serve clients attend these agencies voluntarily, the value of such agencies as potential sites for medical treatment has special importance. Since the early 1990s, a dramatic increase in the discussion about and the call for involuntary treatment of noncompliant tuberculosis patients has arisen among health professionals (Bayer & Dupuir, 1995). For example, New York City plans for its health department to detain tuberculosis patients who do not cooperate in their treatment (Navarro, 1993). There have also been discussions about the usefulness of implementing such policies in other cities (see Arax, 1994; Turner & Reichman, 1995).

We suggest, therefore, that it would be fruitful to inquire further into the role self-help agencies can play in increasing access to health care and in engaging clients in voluntary treatment. Several of the agencies provided occasional on-site clinical services through volunteer medical personnel and Health Care for the Homeless medical vans. All made referrals to the public clinics in the area. One had added an AIDS educator during the study period. Additional work is required to document whether SHAs can be used effectively as points of access to health services for individuals who might otherwise not get health care except in emergencies or who may be forced to undergo involuntary treatment for the severe health conditions found to be highly prevalent among homeless people.