Housing affordability and psychological distress: Household measures mirror individual estimates but miss one in three at-risk adults.
Conventional household-based housing affordability measures assume equal burden-sharing, potentially diluting or obscuring health risks of individuals with disproportionate burdens. To examine the implications of this assumption in research and preventive policy, we investigated the respective associations of household and individual housing affordability with psychological distress.
Analyzing data from a web-based cross-sectional survey conducted in November 2025 (6619 Japanese adults), we assessed classification discordance between household and individual housing unaffordability (cost-to-income ratio ≥ 30%) and examined their associations with psychological distress (Kessler Psychological Distress Scale ≥5). Modified log Poisson regression estimated prevalence ratios by housing tenure, using multiple imputation.
Adjusted prevalence ratios (aPR) were similar for the household (1.15; 95% confidence interval [CI]: 1.05, 1.26) and individual (1.17; 95% CI: 1.06, 1.28) indicators. Stratification by housing tenure showed strong associations for both indicators among homeowners (both aPR = 1.33), whereas no significant associations were observed among renters. However, the household measure overlooked about a third of individuals with unaffordable housing.
While the assumption of equal burden-sharing appears to be sufficient for research, it may obscure specific vulnerabilities when used for screening. Preventive strategies could be refined by incorporating individual indicators to capture these hidden risks.
Analyzing data from a web-based cross-sectional survey conducted in November 2025 (6619 Japanese adults), we assessed classification discordance between household and individual housing unaffordability (cost-to-income ratio ≥ 30%) and examined their associations with psychological distress (Kessler Psychological Distress Scale ≥5). Modified log Poisson regression estimated prevalence ratios by housing tenure, using multiple imputation.
Adjusted prevalence ratios (aPR) were similar for the household (1.15; 95% confidence interval [CI]: 1.05, 1.26) and individual (1.17; 95% CI: 1.06, 1.28) indicators. Stratification by housing tenure showed strong associations for both indicators among homeowners (both aPR = 1.33), whereas no significant associations were observed among renters. However, the household measure overlooked about a third of individuals with unaffordable housing.
While the assumption of equal burden-sharing appears to be sufficient for research, it may obscure specific vulnerabilities when used for screening. Preventive strategies could be refined by incorporating individual indicators to capture these hidden risks.