Reducing potentially preventable hospital transfers: results from a thirty nursing home collaborative

Nursing home (NH) residents experience frequent hospital transfers, some potentially avoidable. The objective of this report is to describe a replication of the Interventions to Reduce Acute Care Transfers program among member facilities of a New York City area NH provider association (INTERACT NY) and estimate its effect on hospital transfers.

INTERACT is a program that provides tools and strategies to assist NH staff in early identification, communication, and documentation of changes in resident status. Funding was obtained from a New York State health workforce training grant to conduct 13 INTERACT education and training sessions in 2010-2011. INTERACT NY session topics included the implementation process; use of its simple standardized communication tools, advance care planning tools, care paths, and change in condition support tools; quality review of hospital transfers; exercises for refining clinical skills; teamwork; and lessons learned. Sessions engaged NH executives, department heads, front-line nursing staff and their labor union, and staff from NHs' partner hospitals. Pre-/post- INTERACT NY hospitalization rates per 1000-resident days were compared using paired t-tests, stratified by level of facility engagement with the program and by baseline hospitalization rates.

All 100% of participating NHs were non-profit or public. Those with complete evaluation data had 377 beds on average. There were a total of 333 attendees of the program (mean 25.6 per session; mean 11.1 per facility over the course of the program; range 1-44 per facility). The most common attendees in order of frequency were (1) nurse administrators, (2) unit-based nurses, (3) medical directors and attending physicians, (4) nursing home administrators, (5) certified nursing assistants, and (6) case managers and social workers. Sixteen nursing homes implemented at least one INTERACT tool. Overall, there was a nonsignificant 10.6% reduction in hospital admissions from 4.07 to 3.64 per 1000 resident-days from pre- to post-INTERACT NY (P = .332). Among nursing homes with high engagement there was a nonsignificant 14.3% reduction in hospital admissions from 4.19 to 3.59 per 1000 resident-days (P = 0.213). Finally, among nursing homes in the highest tertile of baseline (pre-INTERACT NY) hospital admission rate, there was a nonsignificant 27.2% reduction in hospital admissions from 7.32 to 5.33 per 1000 resident-days (P = .102). Planning and implementation lessons from INTERACT NY leaders and participants are reported.

INTERACT NY, a novel collaborative training program, resulted in good uptake of the INTERACT tools and processes among its member nursing homes. Changes in hospitalization rates associated with INTERACT NY were similar to those observed in previous implementations of INTERACT. The program addresses a growing interest in reducing potentially preventable hospital admissions among nursing home residents and providing alternatives to hospital care through standardized approaches to communication, early identification of clinical issues, decision-support, and support for partnerships between acute and post-acute care providers.