Maya Reddy1, Susan Schneiders-Rice1, Casey Pierce1, Garrett Fitzmaurice1, Alisa Busch1. 1. The authors are with the Department of Psychiatry, McLean Hospital, Belmont, Massachusetts (e-mail: mnreddy@partners.org ). Dr. Fitzmaurice is also with the Department of Biostatistics, Harvard School of Public Health, Boston. Dr. Busch is also with the Department of Psychiatry, Harvard Medical School, Boston.
Abstract
OBJECTIVE: This study examined how accurately inpatient case managers predicted 30-day readmission and whether objective patient characteristics improved prediction accuracy. METHODS: In this prospective study, inpatient case managers at a psychiatric hospital rated their concern (1, not concerned; 5, very concerned) about readmission after discharge of 282 privately insured patients. Sensitivity and specificity of the ratings were calculated. Logistic regression identified whether patient characteristics that could affect 30-day readmission improved prediction accuracy. RESULTS: Concern levels ≥3 yielded 86% sensitivity, 37% specificity, and a positive predictive value (PPV) of 13%; levels ≥4 yielded 39% sensitivity, 78% specificity, and a PPV of 17%. Concern level independently predicted readmission; appointments within seven days postdischarge further improved model accuracy (p=.03) (area under the curve=.67, 95% confidence interval=.58-.78). CONCLUSIONS: Although not highly accurate, case manager concern identified some patients at higher risk of 30-day readmission. Appointments within seven days of discharge improved prediction accuracy.
OBJECTIVE: This study examined how accurately inpatient case managers predicted 30-day readmission and whether objective patient characteristics improved prediction accuracy. METHODS: In this prospective study, inpatient case managers at a psychiatric hospital rated their concern (1, not concerned; 5, very concerned) about readmission after discharge of 282 privately insured patients. Sensitivity and specificity of the ratings were calculated. Logistic regression identified whether patient characteristics that could affect 30-day readmission improved prediction accuracy. RESULTS: Concern levels ≥3 yielded 86% sensitivity, 37% specificity, and a positive predictive value (PPV) of 13%; levels ≥4 yielded 39% sensitivity, 78% specificity, and a PPV of 17%. Concern level independently predicted readmission; appointments within seven days postdischarge further improved model accuracy (p=.03) (area under the curve=.67, 95% confidence interval=.58-.78). CONCLUSIONS: Although not highly accurate, case manager concern identified some patients at higher risk of 30-day readmission. Appointments within seven days of discharge improved prediction accuracy.
Authors: Ejemai Eboreime; Reham Shalaby; Wanying Mao; Ernest Owusu; Wesley Vuong; Shireen Surood; Kerry Bales; Frank P MacMaster; Diane McNeil; Katherine Rittenbach; Arto Ohinmaa; Suzette Bremault-Phillips; Carla Hilario; Russ Greiner; Michelle Knox; Janet Chafe; Jeff Coulombe; Li Xin-Min; Carla McLean; Rebecca Rathwell; Mark Snaterse; Pamela Spurvey; Valerie H Taylor; Susan McLean; Liana Urichuk; Berhe Tzeggai; Christopher McCabe; David Grauwiler; Sara Jordan; Ed Brown; Lindy Fors; Tyla Savard; Mara Grunau; Frank Kelton; Sheila Stauffer; Bo Cao; Pierre Chue; Adam Abba-Aji; Peter Silverstone; Izu Nwachukwu; Andrew Greenshaw; Vincent Israel Opoku Agyapong Journal: BMC Health Serv Res Date: 2022-03-12 Impact factor: 2.655