Prevalence and Risk of Dementia Among Adults Who Have Experienced Homelessness, a Meta-Analysis

Abstract

Background: Understanding the prevalence and risk of dementia in homeless adults has important implications for ensuring appropriate resources are directed towards dementia prevention and healthy aging. We aimed to estimate the: (1) prevalence and (2) risk of dementia among homeless adults, and (3) risk of homelessness in adults experiencing dementia.

Methods: A systematic search was completed in MEDLINE, PsycINFO, Embase, CINAHL, Cochrane Central, Social Works Abstracts, and the gray literature from inception to September 26, 2024. Studies investigating the prevalence of dementia among individuals experiencing homelessness aged 18 years or older were included. Title, abstract, and full text screening, extraction, and risk of bias assessments were completed in duplicate. Random effects meta-analysis using generalized linear mixed models was used to estimate the prevalence of dementia. For the risk and incidence of dementia, we meta-analyzed the risk and hazards of dementia comparing homeless to housed individuals.

Results: A total of 1585 citations were retrieved, with 102 studies undergoing full-text review. Thirty-six studies were included in the final review (504,766 individuals experiencing homelessness, with the following weighted by sample size: mean age 63.08, female percentage 10.64). The prevalence of dementia was 6.81% (95% CI: 4.49, 10.19%, I ² = 99.7%). Among those who experienced homelessness compared to housed individuals, the risk ratio for dementia was 1.03 (95% CI: 0.52, 2.03, I ² = 99.7%), and the incidence (hazard ratio) for dementia was 1.54 (0.52, 4.50, I ² = 98.0%). One study showed that a younger age was associated with homelessness among adults with dementia.

Conclusions: This systematic review and meta-analysis suggests that homeless individuals have a dementia prevalence higher than population-based studies of housed adults of a similar age, where the estimated prevalence is under 3%. Research on models of care aimed at reducing homelessness in those experiencing dementia is needed.

Trial registration: PROSPERO: CRD42024595047.

Keywords: dementia; homelessness; meta‐analysis; older adults.

FIGURE 1

Forest plot for the meta‐analysis of dementia prevalence in older adults who have experienced homelessness. Meta‐analyses of the prevalence of dementia among homeless adults using generalized linear mixed models of 34 studies are shown above, with an associated prediction interval. The prevalence of dementia and the associated 95% confidence interval are shown for each study included in this meta‐analysis. The overall prevalence is indicated by the diamond. The prediction interval is shown at the bottom of the figure, which is the expected prevalence if a subsequent study were to be conducted with a similar population as the studies included in this meta‐analysis.

FIGURE 2

Sensitivity analyses of the prevalence meta‐analysis for dementia among individuals experiencing homelessness. Sensitivity analyses for the dementia prevalence estimate were completed on the categories described above, as well as their associated 95% confidence interval. The main estimate is shown first, followed by the type of sensitivity analyses completed. This includes: (1) using studies of only ICD codes, (2) excluding abstracts, (3) excluding cohort studies, (4) excluding studies at high risk of bias, (5) excluding studies that used a screening test to diagnose dementia (e.g., SPMQ), and (6) excluding studies that used a composite of cognitive impairment or dementia.

FIGURE 3

Prevalence of dementia in studies that reported a mean or median age. Study prevalence and the mean or median age of the study are plotted as shown above. The dementia prevalence among adults experiencing homelessness for each study is shown in black circles. This figure shows that there is an important proportion of homeless participants who have been diagnosed with dementia prior to the age of 65. It also shows that there is a positive relationship between age and dementia, in that as study‐level age increases, the dementia prevalence also increases.

FIGURE 4

Subgroup analyses of the prevalence meta‐analysis for dementia among individuals experiencing homelessness. Subgroup analyses for sex, sub‐type of dementia diagnoses, studies that included veterans, country, and sex are shown above, with their prevalence and associated 95% confidence interval. Each subgroup analysis can be found in an inset. Alcohol‐related dementias are studies that exclusively used alcohol‐related dementia diagnoses (e.g., Korsakoff's syndrome).