Objective: Readmission rates have been proposed as a possible quality metric for inpatient psychiatry. Little is known about predicting readmissions and identifying modifiable factors that may reduce early readmissions in these settings.
Methods: We reviewed 693 medical records from our adult inpatient psychiatric unit to identify factors associated with patients' readmission within 90 days of discharge.
Results: After adjusting for all variables, and including interactions between identified factors, we found several demographic features predicting readmission, including male gender with suicidal ideation on admission (odds ratio [OR] = 13.2; 95% confidence interval [CI], 3.4-51.9), a diagnosis of a psychotic disorder with a prior medical admission (OR = 5.7; 95% CI, 1.7-20.6), and suicidal ideation with comorbid personality disorder (OR = 5.3; 95% CI, 1.4-20.6). Demographic features decreasing the odds of readmission included being non-white with homeless living situation (OR = 0.18; 95% CI, 0.04-0.82), medication changes made within 48 hours of discharge (OR = 0.44; 95% CI, 0.23-0.84), and the number of medications dispensed without documented follow-up plan or appointment (OR = 0.88; 95% CI, 0.81-0.96).
Conclusion: Future prospective studies utilizing qualitative and quantitative methods are required to more precisely define a wider array of metrics. Improved identification of demographic features associated with early readmissions may suggest areas to target as we seek to the quality of inpatient psychiatric care.