Ursa Health recently worked with an integrated delivery system to reduce readmissions from skilled nursing facilities (SNFs). The geriatrician charged with leading the effort had summary data from the CMS QIO about the hospital’s readmission rates, but these did not help her to identify specific changes or to engage her colleagues, the hospitals, and the SNFs.
Six weeks to insight
The geriatrician was able to develop a targeted dashboard within six weeks using Ursa Studio, Ursa Health’s fully no-code analytics development platform . The dashboard included targeted, custom measures that reflected the local nature of how her system relates to its partner SNFs.
The first step was to accurately identify cases in as close to real time as possible to facilitate the system’s continuous quality improvement project.
Ursa Studio blended out-of-the-box, foundational logic with local knowledge:
- Accountable care organization (ACO) payer claims data to evaluate SNF bed-days, readmissions to hospitals within and outside the system, and spend across a custom SNF episode
- EMR encounter data from the system’s hospitals.
- Automated abstraction from case managers’ discharge planning documentation, which was as the most reliable source of whether a SNF was involved and which SNF
Because 12 of the SNFs have a closer affiliation with the system, these needed to be categorized separately.
Ursa Studio enabled the rapid development of a readmission diagnostic instrument—custom segmentation to profile cohorts of discharged patients and to stratify their readmission risk—based on concepts that the system already used:
- By payer
- Rich condition profiling
- CMS readmission category
- LACE score (a readmission predictive model)
- Age category
- Destination SNF
- Established primary care patients
Intervention 1: ensuring hand-off back to primary care
Although transition from the hospital to a SNF receives plenty of attention, the stratified readmission diagnostic identified a previously unknown problem: transition from a SNF to home.
The team created an Ursa Studio measure for the percentage of SNF discharges for patients who are established with the system’s PCPs prior to hospitalization with a PCP visit within 35 days of discharge, using the following data objects:
- “Established patients”—attribution rule for a historical PCP visit pattern
- “SNF discharges”—based on EMR hospital discharges with a case manager disposition of the SNF
- “PCP visit”—attribution rule based on EMR ambulatory encounters
Using the measure, the system determined that it had a low overall rate of “successful” transitions back to PCPs. Hospitalist-staffed SNFs showed worse performance than geriatrician-staffed SNFs.
To address this issue, the system resourced a new FTE as a “SNF coordinator,” which doubled the follow-up rate to PCPs after discharge to a SNF.
Intervention 2: improving SNF performance and patient satisfaction
The readmission diagnostic identified an opportunity: that because patients and families may dislike a SNF, they demand to go back to the hospital.
The team created an Ursa Studio measure for the percentage of readmissions from a SNF where the patient’s second hospital discharge was to a different SNF and discovered the following:
- Nearly a quarter of all readmissions from SNFs were because the patient/family wanted a different SNF.
- This was a much bigger problem for some SNFs than others; the measure ranged from 8% to 57% from the initial hospitalization.
To address this issue:
- Health system leadership met with SNFs with high rates (i.e., weaker performance) in this measure to press for changes.
- For patients known to be going to a SNF, patients and families were allowed to visit the facility prior to discharge.
Effective, targeted change leads to substantial savings
Examining a 30-day all-cause readmission rate from SNFs:
- The baseline, pre-intervention rate was 23.5%.
- Post-intervention, the rate was 18.8%.
In addition, 954 bed days were saved over the course of 9-month improvement project.
