Risk of nursing home admission in association with mental illness nationally in the department of veterans affairs

Objective: We sought to determine whether patients with mental health diagnoses in the Department of Veterans Affairs (VA) are more likely to be admitted to nursing homes and to identify socio-demographic, utilization, and clinical characteristics, especially indicators of mental illness severity, associated with nursing home admission among mentally ill patients. Methods: Patients receiving treatment in the VA system nationally during FY 2000 and having no evidence of nursing home utilization during FY 1999 or FY 2000 were followed through FY 2003 using administrative claims data. Three-year incidence rates and unadjusted odds ratios were estimated for each diagnosis. Logistic regression was used to examine the correlates of admission among mentally patients, including analyses stratified by age. Results: Of 3 952,229 VA patients with no prior nursing home use, 15.2% received a mental health diagnosis, of which, 4.6% were eventually admitted to a nursing home. Among mentally ill patients, risk of admission was highest for those with any inpatient medical/surgical days (odds ratio [OR] = 2.28), followed by 3+ outpatient medical visits (OR = 1.48), inpatient mental health days (OR = 1.31), and outpatient mental health visits (OR = 1.09). Patients diagnosed with dementia were 58% more likely to be admitted. Patients diagnosed schizophrenia (OR = 1.26), other psychosis (OR = 1.15), and personality disorder (OR = 1.14) had the next highest probabilities. Elderly patients with bipolar disorder (OR = 1.28) were also more likely to enter. Conclusion: Although factors leading to nursing home entry among the mentally ill are similar to those driving entry in the general population. those with more severe mental health problems are still more likely to be admitted.