In a healthcare system that increasingly relies on hospital readmission rates as a quality indicator, a critical omission may be dramatically skewing the data, finds a study published Thursday in the New England Journal of Medicine.

A review of nine years’ worth of hospital admissions finds that calculating outcomes of “observation” patients — who are technically outpatients — would erase most progress recently touted about lowering readmissions.

Though the study focuses on the pressures borne by hospitals, skilled nursing providers are also being incentivized to reduce readmissions and use their success to attract new provider partners and residents.

“Policies seeking to reduce readmission may not be as effective as we thought,” said study co-author Amber Sabbatini, M.D., MPH, assistant professor of emergency medicine at the University of Washington School of Medicine.

Observation status has grown in use has grown in use among commercially insured beneficiaries, an average of 0.3% annually since 2007, according to researchers. By 2015, data from Truven Health Analytics MarketScan Commercial Claims and Encounters database showed that 14% of patients with “index” emergency department visits were hospitalized, and 43% percent of those were hospitalized under the observation category.

Observation stays do not count toward the three-night minimum required for Medicare-covered skilled nursing services.

While inpatient readmission dropped by 2.3% during the study period, the researchers found the rate of readmission after an observation stay increased by 3.9%. Many patients also returned for repeat observation stays.

“Because of the expectation that observation stays last less than 48 hours, patients hospitalized under observation may be discharged at an earlier stage in the resolution of their illness than inpatients are,” the study reports.“Finally, under current policies, hospitals face no repercussions for unsafe or poorly handled discharges from observations care that may lead to repeat hospitalizations.”