BACKGROUND: Avoidable readmissions to the hospital present a significant challenge for health systems with an estimated $41.3 billion per year in additional healthcare spending attributed to unnecessary rehospitalization. Existing interventions targeting readmissions shows mixed evidence of effectiveness and context dependent success for certain strategies. This study is a pragmatic evaluation of the effectiveness of a hospital-wide readmission reduction initiative at Boston Medical Center (BMC), a large safety-net hospital, with the goal of advancing a Learning Health System model.
METHODS: Adult patients admitted to BMC were risk stratified using a proprietary algorithm into one of four risk groups: Low Risk (LR), Moderate Risk (MR), High Risk (HR) and Super Utilizer (SU). The MR, HR and SU groups were each assigned to receive a different bundle of evidence-informed readmission reduction interventions. We used a quasi-experimental design combining principles of Regression Discontinuity and Difference-in-Difference to estimate the effect that each of the three bundles had on a 30-day readmission outcome. Patient visits from February 2015 to January 2016 were included in the pre-implementation period and visits from November 2016 to October 2017 in the post-implementation period.
RESULTS: There were 18,634 patient visits included in the pre-intervention period and 10,714 observations in the post-implementation period. We found no significant effect for any of the three bundles of interventions with adjusted pre-post changes in 30-day readmission of 2.67% (95%CI: -1.27, 6.61) for the MR group, 1.02% (95%CI: -4.65, 6.68) for the HR group and 8.07% (95%CI: -4.33, 18.46) for the SU group.
DISCUSSION: The interventions in the BMC readmissions reduction initiative were not successful in reducing readmission rates for any of the targeted risk groups. Further work is needed to identify specific factors in the design and implementation of the interventions that limited their effectiveness. However, the results of this evaluation can be used to guide iterative improvement for future readmission reduction efforts. Additionally, the analytic strategy used in this study provides a model for hospitals to develop Learning Health System capabilities that can be applied to targets beyond hospital readmissions. / 2020-10-23T00:00:00Z
| Identifer | oai:union.ndltd.org:bu.edu/oai:open.bu.edu:2144/32693 |
| Date | 23 October 2018 |
| Creators | Cordella, Nicholas |
| Contributors | Carey, Kathleen |
| Source Sets | Boston University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis/Dissertation |
Page generated in 0.0023 seconds