EXAMINING THE EFFECTIVENESS OF DIFFERENT ELEMENTS OF A READMISSION PREVENTION BUNDLE

Purpose: Interventions to reduce preventable hospital readmissions have included multiple steps, but it is unclear which elements of these bundled interventions are most critical. We sought to evaluate the effect of individual elements within our health system’s readmission prevention program.

Method: We examined hospital admissions from September 2013 – February 2014 at one academic medical center that was implementing an evidence-based intervention to prevent readmissions for moderate and high risk patients. Risk was determined by previous admissions, chronic conditions, and number of medications, using a previously validated model. The bundled intervention was implemented through a 20-item checklist, completed by the care team during the admission and in the immediate post-admission period. Checklist items included assessment by care management; pharmacy review, medication reconciliation and counseling; nursing self-care education and teach back; direct communication between inpatient and outpatient care managers, pharmacists, and physicians; scheduling of a timely follow-up appointment; and pre-discharge “purposeful pause” to ensure readiness. We also recorded whether a post-discharge transition call was completed and whether a follow-up visit actually occurred. The main outcome was 30-day readmission to the same academic medical center, determined by chart review. We used chi-squared tests and logistic regression to examine differences in readmission based on individual checklist item completion.

Result: We examined 1737 admissions, of which 486 were readmitted within 30 days (28.0%).  Checklists were started for 1339 admissions (77.1%). Completion rates for individual items ranged from 38.8% to 78.2%. Mean number of completed checklist items did not differ by readmission status (8.5 with no readmission; 8.1 with readmission). The only item for which completion was associated with a statistically significant lower readmission rate was  inpatient pharmacist communication to the outpatient provider describing significant medication changes or concerns (25.2% vs. 29.8%, OR 0.78, 95% CI 0.62, 0.99). In addition, completion of a timely follow-up visit was strongly associated with lower readmission risk (24.8% vs. 38.5%, OR 0.52, 0.41, 0.67)

Conclusion: Documented completion of inpatient-delivered evidence-based care processes had little effect on readmission rates. In contrast, the presence or absence of an outpatient follow-up visit was strongly associated with readmission. These findings suggest decision makers should focus readmission improvement efforts on ensuring that patients have a timely post-discharge follow-up visit.