Utilisation by homeless people of acute hospital services in London.

OBJECTIVES--To estimate the numbers and distribution of homeless people in London; to quantify the utilisation of acute inpatient services by homeless people in two health authorities; and to predict the total numbers of admissions in homeless people in district health authorities across London. DESIGN--Data were collected from various sources on the distribution of homeless people across London boroughs. All unplanned acute inpatient admissions during November 1990 to relevant hospitals were identified. SETTING--Bloomsbury and Paddington and North Kensington, two former inner London district health authorities. SUBJECTS--Homeless people in London residing in bed and breakfast and private sector leased accommodation, residing in hostels, and of no fixed abode. MAIN OUTCOME MEASURES--Number and cost of acute unplanned admissions in homeless people in two health authorities in November 1990; predicted number of such admissions each year in district health authorities in London. RESULTS--There were at least 60,000 homeless people in London in March 1990. The majority were housed in temporary accommodation (55,412). There were at least 3295 hostel dwellers and 651 people sleeping rough. Homeless people accounted for 105 (8%) of the 1256 acute unbooked admissions in residents of Bloomsbury and Paddington and North Kensington health authorities in November 1990. Considerable variations in the pattern of acute unplanned admissions in homeless people were observed in the two districts with respect to housing status and specialty of admission. The total number of acute unplanned admissions in homeless people across London each year was estimated at 7598, ranging from 38 in Bexley to 1515 in Parkside. CONCLUSIONS--The results have fundamental implications for resource allocation across London. Allocation must take better account of the heterogeneity, uneven distribution, and extra health needs of homeless people.     

Bayesian methods for multiple capture-recapture surveys

To estimate the total size of a closed population, a multiple capture-recapture sampling design can be used. This sampling design has been used traditionally to estimate the size of wildlife populations and is becoming more widely used to estimate the size of hard-to-count human populations. This paper presents Bayesian methods for obtaining point and interval estimates from data gathered from capture-recapture surveys. A numerical example involving the estimation of the size of a fish population is given to illustrate the methods.     

Impact of an integrated service system on client outcomes by gender in a national sample of a mentally ill homeless population

Background: Homeless populations with concurrent mental illness have a complex array of service needs that are often addressed in a haphazard or uncoordinated manner. Information is lacking about the effectiveness of programs and public policy in women who are both homeless and mentally ill.Objective: This study assessed the impact of gender on the outcomes of a multisite public policy intervention that implemented components of an integrated service delivery system to address the needs of mentally ill homeless men and women.Methods: A secondary analysis was performed using longitudinal data obtained from homeless men and women in the Access to Community Care and Effective Services and Supports prospective study, which was conducted at 18 sites across the United States. Outcome data from baseline, 3, 12, and 18 months across 5 annual cohorts were analyzed to determine the impact of gender on 6 clinical outcomes, including days of housing, relationships, victimization, social support, alcohol use, and drug use, in homeless men and women with mental illness. The analysis plan included multilevel modeling of longitudinal data.Results: Data were analyzed from a total of 7229 participants, including 4502 men (62%)) (mean age, 38.2 years) and 2727 women (38%) (mean age, 37.2 years). After 18 months of follow-up, women had significantly better outcomes in terms of family relationships (estimated mean score increased 0.100), victimization (score decreased 0.164), and social support (score increased 0.363) than did men (aall, P < 0.001). Being accompanied by children was significantly associated with less change in drug use among women compared with men (P < 0.01). These outcomes were the same across all 18 program sites.Conclusion: Although the addition of services for the homeless in conjunction with clinical and support services offered by case managers had generally positive effects in this study's population, a gender-specific substance abuse treatment intervention should also be included in a comprehensive program for homeless women with mental illness.     

Census of single homeless people in Sheffield.

OBJECTIVE--To determine the number of single homeless people in Sheffield and to examine their demography and social and medical details. DESIGN--Census carried out between 8 am and 8 pm on one day. Participants completed a questionnaire designed to provide data relating to employment history, contact with welfare and health services, social state, prison history, medical history, and health state. SETTING--Sites in Sheffield identified by local workers as being places of residence of homeless people. SUBJECTS--340 single homeless people. MAIN OUTCOME MEASURES--Self reported history of alcohol or drug misuse, existence of a chronic medical condition, and use of general practitioner and hospital services. RESULTS--The mean age of the population was 42.5 years and a quarter of the population were aged less than 30; there were 48 women. Significant differences were noted between men and women with respect to self reported psychiatric illness (77/266 men v 27/42 women), self reported alcoholism (83/273 v 4/44), prison history (152/255 v 8/41), and registration with a general practitioner (73/275 v 38/46). Various chronic medical conditions were reported, and the perceived health state of the population was low; 129 claimed to have been admitted to a psychiatric hospital. 220 people were registered with a general practitioner, and 179 claimed to see their doctor. Sixty five had attended or been admitted to a general hospital in the month preceding the study, 45 for accident and emergency services. CONCLUSIONS--The homeless in this population were younger than those found in previous studies. The prevalence of psychiatric illness was high in the population, and the overall health state was poor. Most subjects obtained health services from general practitioners.     

Multilist population estimation with incomplete and partial stratification

Multilist capture-recapture methods are commonly used to estimate the size of elusive populations. In many situations, lists are stratified by distinguishing features, such as age or sex. Stratification has often been used to reduce biases caused by heterogeneity in the probability of list membership among members of the population; however, it is increasingly common to find lists that are structurally not active in all strata. We develop a general method to deal with cases when not all lists are active in all strata using an expectation maximization (EM) algorithm. We use a flexible log-linear modeling framework that allows for list dependencies and differential probabilities of ascertainment in each list. Finally, we apply our method of estimating population size to two examples.     

Reaching impaired populations with HIV prevention programs: a clinical trial for homeless mentally ill African-American men.

This study tested an intervention to reduce sexual risk behaviors in a high risk impaired population: homeless African-American, Caucasian and Hispanic men with mental illness. In a comparison group clinical trial, men were assigned to an experimental cognitive-behavioral or a control intervention and followed up over 16 months. Men were recruited from a psychiatric program in two shelters for homeless men in Nashville, Tennessee. An ethnically mixed cohort of subjects (54% African-American, 42% Caucasian and 4% Hispanic) were included in the study. Most had a chronic psychiatric disorder and a co-morbid substance abuse disorder. The 257 participants who were sexually active (130 experimental, 127 control) prior to the trial were the main target of the intervention. An experimental intervention (SexG), adapted from Susser and Associates (51), comprised 6 group sessions. The control intervention was a 6-session HIV educational program. Sexual risk behavior was the primary outcome. The experimental and control groups were compared with respect to the mean score on a sexual risk index. Complete follow-up data were obtained on 257 men (100%) for the initial six-month follow-up. These individuals have been followed for the remainder of the 16-month follow-up. This intervention, (SexG), successfully reduced sexual risk behaviors of homeless mentally ill African-American, Caucasian and Hispanic men. Similar approaches may be effective in other impaired high-risk populations.     

An exact test for all-way interaction in a 2M contingency table: application to interval capture-recapture estimation of population size

An exact conditional test for an M-way log-linear interaction in a fully observed 2M contingency table is formulated. From this is derived a procedure for interval estimation of the total count N in a 2M contingency table, one of whose entries is unobserved. This procedure has an immediate application to interval estimation of the size of a closed population from incomplete, overlapping lists of records, as in capture-recapture analysis of epidemiological data. Data on the prevalence of spina bifida in live births in upstate New York in 1969-1974 illustrate this application.     

Health status of the temporarily homeless population and residents of North West Thames region.

OBJECTIVES--To survey the health status of the temporarily homeless population of North West Thames region and make comparisons with regional residents. DESIGN--Direct interview with standardised questionnaires. SETTING--Temporarily homeless people resident in hotels in the London boroughs in the North West Thames region and a random sample of regional residents. SUBJECTS--137 hotels thought to be providing accommodation to homeless people selected at random from a list of 295. 113 (82%) participated in the study, and 319 (61%) of 522 homeless people approached participated. The study was restricted to adults aged 16 and over selected at random. RESULTS--The homeless population was predominantly female (195/319; 61%), young (229 (72%) aged 16-34), and poor, 54% (172/319) receiving income support. 207 subjects (65%) had dependent children aged 16 and under. Rates of acute illness among homeless people (32 cases; 10%) were similar to those reported by regional residents. The prevalence of longstanding limiting illness (108 cases; 34%) was similar to that for regional residents, but the prevalence of mental morbidity was twice that for the region as a whole (145 cases (45%) v 1485 (18%)). Utilisation of general practitioner services, accident and emergency departments, and inpatient admission was much higher by the homeless population than by regional residents. General practitioner registration rates were above 90% for the homeless sample. CONCLUSIONS--Survey data provide empirical evidence about the nature and characteristics of the temporarily homeless population. The high service utilisation recorded may, in part, have resulted from the higher morbidity in this sample of homeless people. The concentration of homeless people into specific locations may suggest that additional funding should be provided to the district which provides care to this group. However, such funding should not necessarily be used for additional acute care but should be used to purchase appropriate services which meet the health needs of this very young, poor and vulnerable group.     

Semiparametric estimation of animal abundance using capture-recapture data from open populations

A semiparametric partially linear model for the size of an open population is proposed and inference is conducted using weighted martingale estimating equations. This extends a previous nonparametric approach to modeling capture-recapture data for open populations with frequent capture occasions. Analytic expressions for the large sample variances are derived and these are confirmed in a simulation study. The method is illustrated on monthly penguin banding data collected over 6 years.     

What happens to homeless mentally ill people? Follow up of residents of Oxford hostels for the homeless.

OBJECTIVES--To follow up severely mentally ill residents of hostels for the homeless to determine their social and psychiatric outcome. DESIGN--Follow up at 18 months of hostel residents previously assessed with psychiatric and behavioural measures. SETTING--Two Oxford hostels for the homeless. SUBJECTS--48 hostel residents previously identified as disabled by mental illness. MAIN OUTCOME MEASURES--Current housing, admissions to psychiatric hospital, violent or antisocial behaviour, and score on standardised behavioural rating. RESULTS--45 of the 48 residents were traced. 27 had remained in the hostels; only 10 had been rehoused, mainly in bedsits or with their families. 16 had a poor outcome as judged by death (four subjects), lengthy hospital readmission (two), marked deterioration in behaviour (six), sleeping rough (one), or disappearance (three). CONCLUSION--More effort is needed to provide suitable housing for homeless mentally ill people.     

A mobile surgery for single homeless people in London.

Little is known about the social and medical characteristics of people who regularly sleep rough, or whether medical care can be targeted at these people. In 1987 a mobile surgery was used to provide primary health care at two sites in central London where many single homeless people sleep outdoors. One hundred and forty six patients were seen with illnesses ranging from scabies to osteomyelitis and tuberculosis. Sociodemographic data showed the patients to be generally an isolated group with deprived and unstable backgrounds, often compounded by alcohol abuse. Over a third of the patients from one site attended a drop in surgery for homeless people in Soho within a month after seeing a doctor in the mobile surgery. This suggests that the project can be a first step in integrating this isolated group with health care facilities.     

Estimating the Rate of Transfer to a Voluntary Improvement or Rehabilitation Programme Following Capture

A method is developed for determining the effectiveness of counselling individuals to undertake a voluntary improvement or rehabilitation programme following capture. This is achieved by estimating the rate of transfer. Estimates are also obtained of the number who have undertaken the programme previously and those who have not undertaken the programme. For a closed population, the estimated rate of transfer is shown to be unbiased and its variance derived. The estimate is obtained under general conditions and, in the special case, when deaths of drug addicts occur, it is shown that the estimate is approximately unbiased. The theory could also be used to estimate the tag losses/gains due to migration in biological populations when capture-recapture techniques are used.     

Estimating Population Size for Capercaillie (Tetrao urogallus L.) with Spatial Capture-Recapture Models Based on Genotypes from One Field Sample

We conducted a survey of an endangered and cryptic forest grouse, the capercaillie Tetrao urogallus, based on droppings collected on two sampling occasions in eight forest fragments in central Switzerland in early spring 2009. We used genetic analyses to sex and individually identify birds. We estimated sex-dependent detection probabilities and population size using a modern spatial capture-recapture (SCR) model for the data from pooled surveys. A total of 127 capercaillie genotypes were identified (77 males, 46 females, and 4 of unknown sex). The SCR model yielded atotal population size estimate (posterior mean) of 137.3 capercaillies (posterior sd 4.2, 95% CRI 130–147). The observed sex ratio was skewed towards males (0.63). The posterior mean of the sex ratio under the SCR model was 0.58 (posterior sd 0.02, 95% CRI 0.54–0.61), suggesting a male-biased sex ratio in our study area. A subsampling simulation study indicated that a reduced sampling effort representing 75% of the actual detections would still yield practically acceptable estimates of total size and sex ratio in our population. Hence, field work and financial effort could be reduced without compromising accuracy when the SCR model is used to estimate key population parameters of cryptic species.     

The elderly mentally ill--whose responsibility?

The elderly mentally ill make considerable demands on health and social services. To evaluate the need for these services a one-day census of all people aged 65 and over was carried out in an area containing 220 000 people (40 000 over 65). Data were obtained on the nursing needs and psychiatric state of the 2162 elderly people in hospital or local authority residential homes for the elderly, or living at home receiving care from the community nursing service. One-third were classified as having psychiatric problems, more than half of them being outside hospital. Residential homes and community nurses play a significant part in caring for the elderly mentally ill, and an integrated but flexible manpower policy is important.     

Evidence That Head and Body Lice on Homeless Persons Have the Same Genotype

Human head lice and body lice are morphologically and biologically similar but have distinct ecologies. They were shown to have almost the same basic genetic content (one gene is absent in head lice), but differentially express certain genes, presumably responsible for the vector competence. They are now believed to be ecotypes of the same species (Pediculus humanus) and based on mitochondrial studies, body lice have been included with head lice in one of three clades of human head lice (Clade A). Here, we tested whether head and body lice collected from the same host belong to the same population by examining highly polymorphic intergenic spacers. This study was performed on lice collected from five homeless persons living in the same shelter in which Clade A lice are prevalent. Lice were individually genotyped at four spacer loci. The genetic identity and diversity of lice from head and body populations were compared for each homeless person. Population genetic structure was tested between lice from the two body regions and between the lice from different host individuals.We found two pairs of head and body lice on the same homeless person with identical multi locus genotypes. No difference in genetic diversity was found between head and body louse populations and no evidence of significant structure between the louse populations was found, even after controlling for a possible effect of the host individual. More surprisingly, no structure was obvious between lice of different homeless persons.We believe that the head and body lice collected from our five subjects belong to the same population and are shared between people living in the same shelter. These findings confirm that head and body lice are two ecotypes of the same species and show the importance of implementing measures to prevent lice transmission between homeless people in shelters.     

“Because Somebody Cared about Me. That's How It Changed Things”: Homeless,Chronically Ill Patients’ Perspectives on Case Management

Background

Case management programs for chronically ill, homeless people improve health and resource utilization by linking patients with case managers focused on improving management of medical and psychosocial problems. Little is known about participants’ perspectives on case management interventions.

Methods

This qualitative study used in-depth, one-on-one interviews to understand the impact of a case management program from the perspective of participants. A standardized interview guide with open-ended questions explored experiences with the case management program and feelings about readiness to leave the program.

Results

Four recurrent themes emerged: (1) Participants described profound social isolation prior to case management program enrollment; (2) Participants perceived that caring personal relationships with case managers were key to the program; (3) Participants valued assistance with navigating medical and social systems; and (4) Participants perceived that their health improved through both the interpersonal and the practical aspects of case management.

Conclusions

Chronically ill, homeless people enrolled in a case management program perceived that social support from case managers resulted in improved health. Programs for this population should consider explicitly including comprehensive social support interventions. Further research on case management should explore the impact of different types of social support on outcomes for homeless chronically ill patients.     

Bayesian inference for small-sample capture-recapture data

We consider data on the survival of a population of Cephalorhynchus hectori, Hector's dolphins, in a marine area of New Zealand. To estimate survival probabilities of animal populations, a multiple capture-recapture sampling scheme can be used. In this paper, we propose a practical methodology to derive approximations to posterior distributions based on Laplace methods. We show how to calculate Bayes estimates and credible intervals in this setting.     

Obstetric outcome in homeless women.

OBJECTIVE--To characterise the pregnant homeless population booking and delivering at St Mary''s Hospital, London, and ascertain whether their obstetric outcome was adversely affected by their homeless condition. DESIGN--Retrospective comparison of demographic characteristics of 185 homeless women booking for delivery with those of housed women booking in the same period and with the population of North West Thames region; comparison of obstetric performance of homeless women with subgroup of the housed population (group matched for age, parity, and ethnic origin). SETTING--Consultant obstetric unit, St Mary''s Hospital, London. SUBJECTS--All women booking between April 1987 and March 1988 who subsequently had a registrable birth. MAIN RESULTS--185 (8%) Of the 2308 women studied were homeless. Compared with the housed population, they had a larger proportion of young women, women of high parity, and Indo-Pakistani women and a smaller proportion of primiparas. Homeless women booked later and had had more previous obstetric problems than housed women. Pregnancy outcome (assessed by birth weight and prematurity rates) was worse than that of both women housed locally and the regional population. Antenatal attendance, complications, intrapartum performance, and perinatal outcome of homeless women did not differ from those in the control group. CONCLUSIONS--This study has been unable to show any significant differences in the outcome of pregnancy in homeless women that can be directly attributed to living in bed and breakfast accommodation, but these women have sociodemographic characteristics and obstetric risk factors that contribute to a poorer outcome in pregnancy than for the general population.     

Non-invasive genetic monitoring of wild central chimpanzees

Background

An assessment of population size and structure is an important first step in devising conservation and management plans for endangered species. Many threatened animals are elusive, rare and live in habitats that prohibit directly counting individuals. For example, a well-founded estimate of the number of great apes currently living in the wild is lacking. Developing methods to obtain accurate population estimates for these species is a priority for their conservation management. Genotyping non-invasively collected faecal samples is an effective way of evaluating a species'' population size without disruption, and can also reveal details concerning population structure.

Methodology/Principal Findings

We opportunistically collected wild chimpanzee faecal samples for genetic capture-recapture analyses over a four-year period in a 132 km² area of Loango National Park, Gabon. Of the 444 samples, 46% yielded sufficient quantities of DNA for genotyping analysis and the consequent identification of 121 individuals. Using genetic capture-recapture, we estimate that 283 chimpanzees (range: 208–316) inhabited the research area between February 2005 and July 2008. Since chimpanzee males are patrilocal and territorial, we genotyped samples from males using variable Y-chromosome microsatellite markers and could infer that seven chimpanzee groups are present in the area. Genetic information, in combination with field data, also suggested the occurrence of repeated cases of intergroup violence and a probable group extinction.

Conclusions/Significance

The poor amplification success rate resulted in a limited number of recaptures and hence only moderate precision (38%, measured as the entire width of the 95% confidence interval), but this was still similar to the best results obtained using intensive nest count surveys of apes (40% to 63%). Genetic capture-recapture methods applied to apes can provide a considerable amount of novel information on chimpanzee population size and structure with minimal disturbance to the animals and represent a powerful complement to traditional field-based methods.