Abstract

Background: The transition to a skilled nursing facility (SNF) after an acute hospitalization is one of the most perilous times in the life of an older Veteran. Veterans undergo more than 250,000 transitions between hospitals and SNFs annually, but more than 1 in 4 is readmitted to the hospital from SNF and less than half have returned to the community by 100 days following hospital discharge. Although the intent of SNF care is to allow recuperation and rehabilitation, Veterans who do not successfully recover are commonly placed in institutional long-term care at significant cost to themselves and to the VA, which spends more than $7 billion annually on this care. However, SNFs vary widely in their rates of community discharge and costs of care delivered. It is unclear how to identify “high-value” SNFs (those that deliver the best community discharge rates at lowest cost) for Veterans since existing public quality metrics do not include VA SNFs, do not list Veteran- specific outcomes, and do not include costs. Similarly, it is unclear how much matching individual Veteran needs with particular SNF characteristics might improve value. The VA as both payer and provider of SNF care has the unique opportunity to develop an optimal SNF network to drive high-value care. Significance/Impact: This work aligns with VA priorities to develop an integrated, high-performing network for Veterans as part of the MISSION Act and positions the VA as a leader in delivery of post-acute care. There are more than 4 million Veterans currently over age 65, making it imperative to improve outcomes and lower costs in SNFs as more Veterans transition out of the hospital to this care setting. Innovation: The approach uses novel data sets and methods drawn from health economics, big data, and systems engineering to provide new insights. To our knowledge, there are no published studies describing the outcomes of Veterans in post-acute care, identifying characteristics of high-performing facilities, nor establishing how matching patient to post-acute care provider characteristics affects outcomes. Specific Aims: Our Specific Aims are to: 1) Compare outcomes (successful discharge to the community) and costs (Federal dollars) across the population of Veterans discharged from a VA hospital to the three most common post-acute care settings where Veterans receive SNF care: CLCs, CNHs, and non-VA SNFs. 2) Evaluate the effect of matching individual subpopulations of Veterans (e.g., by risk for adverse outcome) to SNF type (CLCs, CNHs, or non-VA SNFs) and SNF star rating on outcomes and costs. 3) Compare the effects of consolidating SNF referrals to the SNF type with best outcomes and lowest costs (Aim 1) or matching individual Veteran characteristics to different SNFs (Aim 2) on Veteran outcomes, overall costs of care, and SNF capacity. Methodology: This proposal uses advanced statistical techniques (such as instrumental variable and machine learning methods) and a unique dataset (the 2014-18 Residential History File, which concatenates VA, fee- basis, Medicare, and Medicaid data into longitudinal episodes of care for individual Veterans) to accomplish our Aims. Implementation/Next Steps: The results of this work will be disseminated to VA Geriatrics and Extended Care and Office of Community Care leadership, who have been involved in the development of the proposal, as well as VISN and VA facility leadership through two tools that can be used 1) at the bedside to optimize SNF choice and 2) at a leadership level to help shape the SNF network to maximize value.

Dimensions for VA

Dimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.

Learn more about Dimensions for VA.

VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address.     Search Dimensions for this project

DRA: Aging, Older Veterans' Health and Care

DRE: TRL - Applied/Translational

Keywords: Care Coordination, Outcomes - Patient, Practice Patterns/Trends

MeSH Terms: None at this time.