Lessons from the hospital readmissions reduction program

A commentary from Sheehy et al. (2023) argues that the hospital readmissions reduction program has been a disappointment.

“If you can’t measure it, you can’t improve it”—a quote featured prominently on the Centers for Medicare & Medicaid Services (CMS) Measures Management website.3 But like any initiative intended to improve quality, evaluation, and iteration is key to assessing and improving delivery. More than a decade later, the question now must be flipped to ask “We have measured it, have we improved it?”

One reason for this is that the increasing use of observation hospitalizations is distorting the HRRP metrics. Observation hospitalizations, often indistinguishable from normal inpatient stays, are excluded from the HRRP measure (both as index admissions and as rehospitalizations). Being hospitalized “under observation” means that patient’s stay is covered under Medicare Part B (vs. Part A) and the coverage for post-acute skilled nursing facility (SNF) coverage is no longer included. One study (Sabbatini et al. 2022) concluded that “including observation stays in both the index hospitalization and 30-day rehospitalization counts more than halved the perceived readmission improvements in targeted conditions.”

Hospitals are increasingly providing care in the outpatient setting. MedPac (2018) reports that outpatient hospital services increased by 49.0% while inpatient discharges decreased by 21.8% between 2006 and 2016.

The solution to the problem is not clear. One option would be to include observations stays are part of the HRRP calculation but not merge them within the inpatient perspective payment system. Another approach would disallow observation stays as a separate type of hospital visit, but this is unlikely to happen as shifting more care to the outpatient setting is likely a positive trend (HRRP notwithstanding). Note that while CMS will pay for a new readmission diagnosis-related group (DRG) for Medicare fee-for-service patients, most Medicare Advantage plans will not pay for a new DRG for a readmission for the same type of hospitalization as the index.

The authors also note that HRRP has not adequately accounted for social determinants of health.

The authors argue that tracking readmissions for quality measurement purposes but not linking to reimbursement could be a viable solution, particularly to incentivize hospitals to care for disadvantaged populations.

This is a difficult issue and the authors bring up some key challenges for hospital quality measurement.