S. J. Baker1,2, J. Richman1,2, E. Dasinger1,2, T. Wahl1,2, L. Graham3,4, K. Itani5, H. Mary3,4, M. Morris1,2 1University Of Alabama at Birmingham,General Surgery,Birmingham, Alabama, USA 2Birmingham Veterans Affairs Medical Center,General Surgery,Birmingham, ALABAMA, USA 3Stanford University,General Surgery,Palo Alto, CA, USA 4VA Palo Alto Healthcare Systems,Palo Alto, CA, USA 5VA Boston Healthcare System,General Surgery,West Roxbury, MA, USA
Introduction: The Veterans RAND-12 Health Survey (VR-12) is used within Veterans Affairs’ hospitals to measure health related quality of life. In this study we explored if specific VR-12 questions could predict post-operative readmissions. We hypothesized that patients who reported lower overall health or depressive symptoms would have higher readmission rates.
Methods: Patients undergoing general, vascular, or thoracic surgery at 4 Veterans Affairs (VA) Medical Centers, August 2015-June 2017 with a post-operative hospital stay over 48 hours were prospectively enrolled. Trained interviewers assessed patient’s health status on the day of discharge using the VR-12 survey consisting of 12 questions corresponding to eight principal physical and mental health domains: general health perceptions, physical functioning, role limitations due to physical and emotional problems, bodily pain, energy-fatigue, social functioning, and mental health. Each item is scored on a 5 point Likert scale. Unplanned readmissions within 30 days post-discharge were identified using VA records with a telephone interview at day 30 to identify readmissions to non-VA hospitals. Relationships between individual VR-12 responses and readmission were evaluated using bivariate tests and logistic regression models with and without adjustment for clinical and demographic covariates.
Results: We recruited 736 patients with a median age of 67 (IQR: 61-71), 96% (n=692) were male, and the majority were Caucasian (84%, n=607). Patients with and without readmission were similar in age (65 vs. 67, p=0.11), need for emergent procedure (3.6% vs. 2.8%, p=0.79), operative time (hours 4.01 vs. 4.45, p=0.28), and surgical procedures (p=0.51) but differed in Charlson comorbidy index (CCI) and functional status. Readmitted patients had higher CCI (5.4 vs 3.9, p<0.001) and were more often classified as dependent functional status (13% vs. 5%, p=0.004). A model that included the two VR12 questions which approached significance [(self-reported general health (ANOVA p<0.01) and ‘How much of the time in the past 4 weeks have you felt downhearted and blue?’ (ANOVA p=0.06)] was adjusted for specialty, age, operative time, CCI, dependent status, and emergency procedure. In this model, self-reported health approached significance (ANOVA p=0.06) driven by those reporting ‘poor’ health trending towards more frequent readmission (OR 3.80, 95% CI 0.97-14.88); feeling blue was associated with readmission (ANOVA p=0.01), in particular, patients who reported feeling blue ‘a good bit of the time’ were more likely to be readmitted (OR 3.47, 95%CI 1.45-8.31).
Conclusion: Patients who report feeling ‘downhearted and blue’ at the time of discharge are more likely to be readmitted and could be identified with a single question. This information can be used to efficiently identify patients at risk for readmission and should be validated through other studies.