Letter to the Editor: Pivoting Care Coordination Efforts Based on Risk and Return

Providing nursing care coordination to patients transitioning from hospital to home is common, ^{1 –3} but does it result in improved outcomes for all patient risk levels? At Froedtert and the Medical College of Wisconsin we pivoted away from providing care coordination to patients at low predicted risk for readmission with the intent of providing an alternate digital solution; however, the latter was delayed. This presented an opportunity to examine the impact of care coordination on this population. We looked at utilization for patients discharged with either inpatient or observation status in a pre- and post-pivot comparison. Low risk for readmission was defined as a risk score for unplanned readmission of <11 (inpatient status) or for a hospital admission or emergency department (ED) visit of <21 (observation status). ⁴ The pre-pivot group had documentation of a care coordination episode in the electronic health record (EHR) during the time frame of August 2018 through June 2019. The post-pivot time frame was August 2019 through June 2020.

Post pivot, those discharged from inpatient status had similar inpatient excess and observation days and less ED utilization (3.7% vs 5.4%, P = 0.029). Those discharged from observation status had similar findings, including less ED utilization (2.1% vs 5.6%, P = 0.006). The pre-pivot group patient demographics were significantly greater in age, male gender, Medicare Advantage as the payer, and more likely to be discharged from the Froedtert Health community hospitals. This pre-pivot group also had significantly lower risk of hospital admission or ED visit scores (for those discharged from inpatient but not observation status) and general adult risk scores. Removing patients during the months potentially influenced by COVID had no impact on findings.

Our study transpired in a singular health system (1 urban academic medical center and 2 rural community hospitals), a practice of 2500 providers (3/4 specialty, ¼ primary). As such, our analysis of low-risk patients may not be transferrable to other health systems.

There is always a concern regarding “stopping care,” especially interventions that are perceived to be effective. These data can help articulate how we can shift important and limited care management resources to populations who will derive the most benefit. However, care coordination intervention could have helped in unmeasured outcomes. At the time of the pivot, we had intended to provide a technological solution to support low-risk populations, but this was delayed because of issues regarding EHR integration. This technological solution is designed to provide a safety net and to drive patient engagement through “check-ins” and alerts for escalation to care coordinators. We have now implemented this solution and will be analyzing its effectiveness.

We believe nursing care coordination need not be broadly applied to discharged patients who are at low risk for readmission. Little is published regarding how to optimize limited care coordination resources amid expanding population health strategies or what happens when you do less rather than more. Tailoring how health systems layer human and digital population health support at reasonable expense is ripe for further study.