Does a Complex Care Discharge Planning Initiative Reduce Unplanned Hospital Readmissions? (G78717-Readmits)

Overview

Unplanned hospital readmissions are associated with increases in morbidity, mortality, cost and patient dissatisfaction,. Policymakers continue to seek effective policy solutions to avoid readmissions in order to improve quality of care and reduce unnecessary expenditures,. One attempt to reduce readmissions was implemented on June 1 2012, when the Specialist Services Committee of British Columbia (a partnership of Doctors of BC and the Ministry of Health) introduced the new "G78717" fee code for physicians. The objective of the fee code was to create a financial incentive for physicians to provide a point-of-care supplemental discharge summary to patients and their primary care providers prior to discharge from hospital. Initially, only urgent hospital admissions were eligible for this incentive payment but on Nov 1 2015 the incentive was extended to include elective admissions as well. The other eligibility criteria remained unchanged. The effectiveness and cost-effectiveness of the fee code intervention is unknown. This study will address important questions relevant to this policy intervention using rigorous methods and empirical data. This study will employ two methods for measuring changes in readmission risk. First, we will use interrupted (multivariate) time series to measure whether there was a temporal change in provincial readmission risk associated with the implementation of the new fee code. We will complement the above analyses with a stronger design, comparing hospitalizations for which the fee code was charged (intervention group) with a cohort of clinically similar hospitalizations for which the fee code was not charged (control group). For this approach, multivariate logistic regression will be the primary statistical method. Using this analytic strategy, 30-day readmission risk between the intervention and control group will be measured over time, adjusted for patient-, provider-, and hospitalization-level covariates.

Conditions