Ingber MJ, Feng Z, Khatutsky G, Wang JM, Bercaw LE, Zheng NT, Vadnais A, Coomer NM, Segelman M
Abstract
Nursing facility residents are frequently admitted to the hospital, and these hospital stays are often potentially avoidable. Such hospitalizations are detrimental to patients and costly to Medicare and Medicaid. In 2012 the Centers for Medicare and Medicaid Services launched the Initiative to Reduce Avoidable Hospitalizations among Nursing Facility Residents, using evidence-based clinical and educational interventions among long-stay residents in 143 facilities in seven states. In state-specific analyses, we estimated net reductions in 2015 of 2.2–9.3 %age points in the probability of an all-cause hospitalization and 1.4–7.2 %age points in the probability of a potentially avoidable hospitalization for participating facility residents, relative to comparison-group members. In that year, average per resident Medicare expenditures were reduced by $60–$2,248 for all-cause hospitalizations and by $98–$577 for potentially avoidable hospitalizations. The effects for over half of the outcomes in these analyses were significant. Variability in implementation and engagement across the nursing facilities and organizations that customized and implemented the initiative helps explain the variability in the estimated effects. Initiative models that included registered nurses or nurse practitioners who provided consistent clinical care for residents demonstrated higher staff engagement and more positive outcomes, compared to models providing only education or intermittent clinical care. These results provide promising evidence of an effective approach for reducing avoidable hospitalizations among nursing facility residents.
Insights Results
Overview of article/program
This article report results of an evaluation of the Initiative to Reduce Avoidable Hospitalizations among Nursing Facility Residents based on analyses of quantitative and qualitative data for 2014 and 2015 for long-stay residents
CMS launched the Initiative to Reduce Avoidable Hospitalizations among Nursing Facility Residents to test the effectiveness of evidence-based clinical and educational interventions in reducing potentially avoidable hospitalizations. The Initiative consists of 7 Enhanced Care and Coordination Provider (ECCP) models in 7 states (Alabama, Indiana, Missouri, Nebraska, Nevada, New York, and Pennsylvania) aimed at improving the overall health and healthcare of participating long-stay residents of nursing facilities. Primary goals were to reduce potentially avoidable hospitalizations, improve quality of care, and decrease healthcare spending. Each ECCP had the flexibility to select and implement specific interventions with the inclusion of a few required elements (e.g., hiring on-site facility staff to focus on improvements)
All ECCPs chose to use a home quality improvement program, Interventions to Reduce Acute Care Transfers (INTERACT), a program that aligns to the Initiative’s goals through early identification, assessment, and documentation of and communication about changes in the status of residents in skilled nursing facilities. More specifically, the ECCPs used one of more of INTERACT’s tools including: 1) The Stop and Watch Early Warning Tool – assists staff who witness a change in a resident’s condition to notify a nurse; 2) “Care Paths” – symptom specific decision support tool for nurses; 3) The Situation, Background, Assessment, and Recommendation (SBAR) tool – template for nurses to use in assessing a resident’s condition before notifying a physician of a change in that condition; 4) The Nursing Home-Hospital Transfer form – organizes resident records to simplify communication between staff at 2 different institutions; and 5) the Quality Improvement Tool for Review of Acute Care Transfers – facilitates review of acutely ill residents transferred to the hospital by facility staff members
Methods of article
Authors used a mixed-methods approach to evaluate the effects of the Enhanced Care and Coordination Provider interventions on resident outcomes. Authors conducted a quantitative analysis of secondary data (claims and assessments) using difference-in-differences multivariate regression models to compare resident outcomes between the ECCP intervention group and a matched comparison group in each state
Qualitatively, authors collected data annually in the period 2013–15 from in-person annual site visits with the leadership of each ECCP and a rotating selection of 4 participating facilities in each state, annual telephone interviews with one or more facility staff members in all facilities not visited in person, and an annual web-based survey of administrators from all participating facilities. A one-time web-based survey of comparison-group facilities was conducted in 2015
Matching at comparison sites was based on characteristics such as staffing levels and mix, bed size, payer mix, chain membership, ownership type, resident case-mix, and presence of special care unit
Outcomes measured ECCP intervention effects for 4 outcome variables: any all-cause hospitalization, any potentially avoidable hospitalization, and Medicare expenditures associated with each type of hospitalization
Results
Across all 4 outcomes and 7 states, the estimated effects of ECCP interventions had negative signs in 2014 and 2015, which is evidence of a relative reduction in hospitalizations and related Medicare expenditures in each year as a result of the Initiative to Reduce Avoidable Hospitalizations among Nursing Facility Residents. However, the consistency and strength of these effects varied by state, outcome and year
In general, this analysis indicates consistently significant favorable effects of the Initiative in Missouri and Indiana and less significant in Pennsylvania, Nevada and nearly nonsignificant positive effects in Alabama, New York and Nebraska
Study found that the implementation experience varied substantially across the seven states and across facilities within states. Model components frequently were modified or implemented at different speeds and on varying timelines. In some instances, components were introduced in a different order, depending on characteristics of individual participating facilities—such as their degree of engagement with the Initiative to Reduce Avoidable Hospitalizations among Nursing Facility Residents or staff turnover. Examples include a 2-year delay in Pennsylvania for the implementation of telemedicine in rural facilities with wireless Internet connectivity problems, and the decision of chain facilities in Alabama to modify standard INTERACT tool
Overall, models in which nurses provided education had smaller and less consistent effects, compared to models in which nurses provided regular hands-on clinical care. More hands-on models also demonstrated greater changes in facility culture, greater support for the need to reduce avoidable hospitalizations and greater overall buy-in to the initiative from facility staff
Key takeaways/implications
The findings show the importance of building relationships between ECCP nurses and staff and primary care providers. Buy-in from leadership is also important, however, this may be undermined by staff turnover or lack of support from individual physicians who were unwilling to include outside nurses. Also models of this magnitude require fundamental cultural change for introduction and maintenance of the initiative
ECCP models faced difficulty in adoption of specific interventions generally related to information technology issues and lack of appropriate infrastructure for implementation
Limitations to the study include 1) Use of a small comparison group in Nevada; 2) Co-occurrence of the Imitative with other initiatives also focused on reducing hospital admissions; and 3) Inability to parse the effects of specific components of ECCP
