Prevalence of Cognitive Impairment Among Homeless People in Inner Sydney

Studies that have assessed cognitive impairment among homeless persons have yielded conflicting results. Using the Mini Mental State Examination (MMSE), Teesson and Buhrich (1) found that 40 percent of 65 residents at a hostel for homeless men in inner Sydney showed at least mild cognitive impairment, and one in four had severe impairment. Bremner and associates (2) sampled the cognitive functioning of 80 persons who stayed at a hostel for homeless men in London; 7 percent showed cognitive impairment on the MMSE. Current IQ was significantly lower than the authors' estimates of subjects' previous IQ, indicating deterioration in cognitive functions.

Other studies using the MMSE have found prevalence rates of severe cognitive impairment of 8.9 percent among homeless men in Munich (3), 7.8 percent among homeless men and women in Baltimore (4), and 6 percent among homeless men and women in Madrid (5). In contrast, Adams and associates (6) found no impairment among 64 homeless women as measured by the MMSE. Given the high prevalence of cognitive impairment found in the study of homeless men in one hostel in inner Sydney (1) and the fact that women were not assessed in that sample, a broader study was indicated.

The aim of the study reported here was to assess the prevalence of cognitive impairment among a cohort of homeless men and women in inner Sydney

Methods

Sydney has a population of 3.8 million. Inner Sydney is characteristic of inner-city areas of other large Western cities, with a mix of poverty and gentrification and of prostitution, alcoholism, illicit drugs, and homelessness. The homeless population of inner Sydney comprises two main groups: those who use hostels for the homeless for accommodation and those who are not residents but use the hostels for meals, clothes, and welfare services.

Seven hostels for homeless adults age 18 and over in inner Sydney participated in the study over a period of eight months (December 1996 to August 1997). The hostels offered 57 emergency beds to females and 350 emergency beds to males. Potential subjects were chosen in proportion to the number of subjects at each hostel and the number of subjects at each meal (breakfast, lunch, and dinner) from a list of computer-generated random numbers, using the numbers on the chairs in the dining hall of each hostel. Following their meal, 289 subjects were approached, of whom 79 declined to participate. Another six were excluded because they failed to complete the MMSE, could not speak English, or were deaf.

Subjects were interviewed by one of the authors (TH) using sections of the Composite International Diagnostic Interview 2.0, 12 month (CIDI), which included the alcohol use disorder sections and the MMSE. MMSE scores of 0 to 17 indicate severe impairment, and scores of 18 to 23 mild impairment.

Subjects were also screened by the same author using a psychosis screener. Anyone who screened positive, or who screened negative but who seemed odd or unusual, was referred to the psychiatrist (the senior author) to be assessed for a diagnosis of schizophrenia.

Results

A total of 204 subjects (155 men and 49 women) completed the interview protocols. Twenty (10 percent) of the subjects had cognitive impairment, using a cutoff MMSE score of 23 or less. Five of the 14 men had severe impairment, and nine had mild impairment. None of the women had severe impairment, but six had mild impairment. No significant difference was found between the men and women in the prevalence of mild or severe cognitive impairment.

Because the number of subjects with cognitive impairment was small, the data for men and women were combined. The mean±SD age of the 20 subjects was 57±16.2 years, compared with 41±14 years for the other subjects. Eleven (55 percent) of the subjects with cognitive impairment had schizophrenia, and seven (33 percent) had an alcohol use disorder, compared with 46 (25 percent) and 75 (41 percent), respectively, of the other subjects. Subjects with cognitive impairment were significantly older than those without cognitive impairment (t=4.4, df=202, p<.05). The mean±SD length of stay at a hostel for subjects with impairment was 685±1,273 days, compared with 422±1,411 days for the other subjects. However, this difference was not significant.

Logistic regression predicting cognitive impairment using the variables age, sex, length of stay at the hostel, and mental disorder found that age was the only significant predictor (p=.002); the older the subject, the more likely the subject was to have cognitive impairment.

Discussion and conclusions

The prevalence of cognitive impairment among the 204 homeless subjects in this sample was 10 percent. According to the Epidemiologic Catchment Area study, the prevalence of cognitive impairment in the general adult population is 1.7 percent. What is the explanation for this sixfold increase?

Subjects with impairment were significantly older than those without impairment. Koegel and colleagues (9) also found an increased prevalence of cognitive impairment with increasing age in a cohort of homeless people in Los Angeles. It may be that the older subjects in the impaired group have a senile dementia. However, it seems unlikely. Most subjects in our sample and in that of Koegel and colleagues were under age 65; presenile dementia is rare.

A more likely explanation relates to the duration of stay at the hostel. We have previously reported that men with cognitive impairment stayed in a hostel for a mode of 25 months, compared with 12 months for those without impairment (1). Bremner and associates (2) reported a similar finding among homeless men in London; they found that duration of homelessness was related to a deterioration in IQ. It seems likely that cognitive impairment makes it difficult for some people to move out of the hostel system. Alternatively, the impairment may make them relatively unconcerned about their situation. Either way, these findings argue that the provision of shelter alone for this group of people is not sufficient if they are expected to move out of their homeless state. Additional care and support is required.

In our study, no significant difference was found in the prevalence of alcohol use disorder in subjects with cognitive impairment compared with other subjects. Our interview protocol assessed alcohol intake only over the previous 12 months. Alcohol consumption decreases with age (10). It may be that the cognitive impairment among our subjects is the result of previous heavy drinking that has now lessened or ceased.

A trend was noted for subjects with a diagnosis of schizophrenia to be cognitively impaired. The numbers are small, and the results should be treated with caution. Nevertheless, it has become increasingly clear that people who are homeless and who have a mental illness often have cognitive deficits (11). These deficits may relate to the concurrent consumption of anticholinergic medication, although another explanation seems more likely. Among people with mental illness, cognitive deficits rather than psychotic symptoms appear to be the more detrimental in relation to loss of social contacts, ability to maintain employment, and ability to function in the community (12). These characteristics may partly explain the high prevalence of schizophrenia in the homeless population.

Acknowledgments

The study was funded by the Sydney City Mission, the Society of St. Vincent de Paul, the Salvation Army, Wesley Mission, the Haymarket Foundation, and the Centre for Mental Health of the New South Wales Department of Health.

Dr. Buhrich is clinical associate professor in the department of psychological medicine at the University of Sydney, Royal Prince Alfred Hospital, New South Wales, 2006, Australia (e-mail, [email protected]). Ms. Hodder is senior research officer in the department of mental health services at St. Vincent's Hospital in Sydney. Dr. Teesson is lecturer at the National Drug and Alcohol Research Centre at the University of New South Wales in Sydney.