Ann M Borzecki1, Qi Chen2, Hillary J Mull2, Michael Shwartz2, Deepak L Bhatt2, Amresh Hanchate2, Amy K Rosen2. 1. From the Center for Healthcare Organization and Implementation Research (CHOIR), Bedford VAMC Campus, Bedford, MA (A.M.B.); Department of Health Policy and Management, Boston University School of Public Health, MA (A.M.B.); Department of Medicine (A.M.B., A.H.) and Department of Surgery (H.J.M., A.K.R.), Boston University School of Medicine, MA; CHOIR, Boston VA Campus, MA (Q.C., H.J.M., M.S., A.H., A.K.R.); Boston University Questrom School of Business, MA (M.S.); Brigham and Women's Hospital Heart & Vascular Center, Boston, MA (D.L.B.); and Department of Medicine, Harvard Medical School, Boston, MA (D.L.B.). amb@bu.edu. 2. From the Center for Healthcare Organization and Implementation Research (CHOIR), Bedford VAMC Campus, Bedford, MA (A.M.B.); Department of Health Policy and Management, Boston University School of Public Health, MA (A.M.B.); Department of Medicine (A.M.B., A.H.) and Department of Surgery (H.J.M., A.K.R.), Boston University School of Medicine, MA; CHOIR, Boston VA Campus, MA (Q.C., H.J.M., M.S., A.H., A.K.R.); Boston University Questrom School of Business, MA (M.S.); Brigham and Women's Hospital Heart & Vascular Center, Boston, MA (D.L.B.); and Department of Medicine, Harvard Medical School, Boston, MA (D.L.B.).
Abstract
BACKGROUND: The 3M Potentially Preventable Readmissions (3M-PPR) software matches clinically related index admission and readmission diagnoses that may signify in-hospital or postdischarge quality problems. To assess whether the PPR algorithm identifies preventable readmissions, we compared processes of care between PPR software-flagged and nonflagged cases. METHODS AND RESULTS: Using 2006 to 2010 national VA administrative data, we identified acute myocardial infarction and heart failure discharges associated with 30-day all-cause readmissions, then flagged cases (PPR-Yes/PPR-No) using the 3M-PPR software. To assess care quality, we abstracted medical records of 100 readmissions per condition using tools containing explicit processes organized into admission work-up, in-hospital evaluation/treatment, discharge readiness, postdischarge period. We derived quality scores, scaled to a maximum of 25 per section (maximum total score=100) and compared cases on total and section-specific mean scores. For acute myocardial infarction, 77 of 100 cases were flagged as PPR-Yes. Section quality scores were highest for in-hospital evaluation/treatment (20.5±2.8) and lowest for postdischarge care (6.8±9.1). Total and section-related mean scores did not differ by PPR status; respective PPR-Yes versus PPR-No total scores were 61.6±11.1 and 60.4±9.4; P=0.98. For heart failure, 86 of 100 cases were flagged as PPR-Yes. Section scores were highest for discharge readiness (18.8±2.4) and lowest for postdischarge care (7.3±8.1). Like acute myocardial infarction, total and section-related mean scores did not differ by PPR status; PPR-Yes versus PPR-No total scores were 61.2±10.8 and 63.4±7.0, respectively; P=0.47. CONCLUSIONS: Among VA acute myocardial infarction and heart failure readmissions, the 3M-PPR software does not distinguish differences in case-level quality of care. Whether 3M-PPR software better identifies preventable readmissions by using other methods to capture poorly documented processes or performing different comparisons requires further study.
BACKGROUND: The 3M Potentially Preventable Readmissions (3M-PPR) software matches clinically related index admission and readmission diagnoses that may signify in-hospital or postdischarge quality problems. To assess whether the PPR algorithm identifies preventable readmissions, we compared processes of care between PPR software-flagged and nonflagged cases. METHODS AND RESULTS: Using 2006 to 2010 national VA administrative data, we identified acute myocardial infarction and heart failure discharges associated with 30-day all-cause readmissions, then flagged cases (PPR-Yes/PPR-No) using the 3M-PPR software. To assess care quality, we abstracted medical records of 100 readmissions per condition using tools containing explicit processes organized into admission work-up, in-hospital evaluation/treatment, discharge readiness, postdischarge period. We derived quality scores, scaled to a maximum of 25 per section (maximum total score=100) and compared cases on total and section-specific mean scores. For acute myocardial infarction, 77 of 100 cases were flagged as PPR-Yes. Section quality scores were highest for in-hospital evaluation/treatment (20.5±2.8) and lowest for postdischarge care (6.8±9.1). Total and section-related mean scores did not differ by PPR status; respective PPR-Yes versus PPR-No total scores were 61.6±11.1 and 60.4±9.4; P=0.98. For heart failure, 86 of 100 cases were flagged as PPR-Yes. Section scores were highest for discharge readiness (18.8±2.4) and lowest for postdischarge care (7.3±8.1). Like acute myocardial infarction, total and section-related mean scores did not differ by PPR status; PPR-Yes versus PPR-No total scores were 61.2±10.8 and 63.4±7.0, respectively; P=0.47. CONCLUSIONS: Among VA acute myocardial infarction and heart failure readmissions, the 3M-PPR software does not distinguish differences in case-level quality of care. Whether 3M-PPR software better identifies preventable readmissions by using other methods to capture poorly documented processes or performing different comparisons requires further study.
Authors: Hillary J Mull; Laura A Graham; Melanie S Morris; Amy K Rosen; Joshua S Richman; Jeffery Whittle; Edith Burns; Todd H Wagner; Laurel A Copeland; Tyler Wahl; Caroline Jones; Robert H Hollis; Kamal M F Itani; Mary T Hawn Journal: JAMA Surg Date: 2018-08-01 Impact factor: 14.766
Authors: Rozalina G McCoy; Stephanie M Peterson; Lynn S Borkenhagen; Paul Y Takahashi; Bjorg Thorsteinsdottir; Anupam Chandra; James M Naessens Journal: Med Care Date: 2018-08 Impact factor: 2.983