Housing Instability and Type 2 Diabetes Outcomes.

Importance: Housing instability may worsen control of type 2 diabetes outcomes. Objective: To estimate the association between stable vs unstable housing and type 2 diabetes outcomes. Design, Setting, and Participants: This cohort study analyzed electronic health record data of adults with type 2 diabetes from US community-based health centers from June 2016 to April 2023. Targeted minimum loss estimation was used to examine the association between having vs not having housing instability and subsequent type 2 diabetes outcomes, adjusting for age, date of housing instability assessment, sex, race and ethnicity (social constructs that may indicate the experience of racism), language, comorbidities, health insurance, income, and census-tract level social vulnerability index. Analyses were conducted from July 2023 to September 2024. Exposure: Report of housing stability or instability. Main Outcomes and Measures: Mean hemoglobin A1c (HbA1c) level was the primary outcome; secondary outcomes included systolic and diastolic blood pressure (SBP and DBP) and low-density lipoprotein (LDL) cholesterol. The primary time point was 12 months after initial assessment, with secondary time points at 6, 18, 24, 30, and 36 months. Results: A total of 90 233 individuals were included (mean [SD] age, 55.4 [13.7] years; 50 772 female [56.3%]; 25 602 Black [28.4%], 27 277 Hispanic [31.4%], 51 720 White [57.3%]); 28 784 individuals (31.9%) had a primary language other than English, and 15 713 (17.4%) reported housing instability. Prior to first housing instability assessment, mean (SD) HbA1c was 7.64% (1.94%), mean (SD) SBP was 130.0 (13.5) mm Hg, mean (SD) DBP was 78.2 (7.8) mm Hg, and mean (SD) LDL cholesterol was 101.1 (35.2) mg/dL. We estimated had all individuals experienced stable housing, compared with unstable housing, mean HbA1c would have been 0.12% lower (95% CI, -0.16% to -0.07%; P < .001), SBP would have been 0.77 mm Hg lower (95% CI, -1.14 mm Hg to -0.39 mm Hg; P < .001), and DBP 0.27 mm Hg lower (95% CI, -0.49 mm Hg to -0.06 mm Hg; P = .01), but LDL cholesterol would not have been lower (estimated difference, -1.46 mg/dL, 95% CI, -2.96 mg/dL to 0.03 mg/dL; P = .05) at 12 months, with numerically similar results at other time points. Conclusions and Relevance: In this cohort study, housing stability was associated with small differences in type 2 diabetes outcomes; combining housing stability interventions with other diabetes interventions may be needed to improve type 2 diabetes outcomes more substantially.