Hospitals vary in their procedures for assigning patients to “observation” status (OBS). This outpatient designation has a number of implications including reduced hospital payments, lower patient copayments, and often results in OBS time not being included in inpatient risk adjustment and hospital performance measures. Thus, variation in OBS use may bias interhospital comparisons. Our objective was to determine variation in OBS use and evaluate how this correlates with hospital and patient outcomes.
Using Veterans Affairs (VA) administrative data files from 20092010, univariate analyses compared facility rates of OBS use to reported Observed Minus Expected Length of Stay (OMELOS). Multivariate models then examined how individual patient assignment to OBS, percentage of medicine patients at the facility admitted to OBS, and the interaction between these two variables related to hospital LOS and 30day allcause readmission rates in a subcohort of patients with a cardiac symptom DRG (303, 305, 310, 311, 312, 313, 316).
Across 128 VA hospitals, there was significant variation in OBS use. In 2009, the mean percentage of medicine patients first admitted as OBS was 10.3% (SD=11.7), range 0 51%, with 30 hospitals admitting < 1% of patients to OBS; rates were similar in 2010. Between the two years, 9 hospitals decreased OBS by > 5% and 16 hospitals increased OBS by > 5%. Regression analyses indicated that for every additional 1% of the patient population placed under OBS status, OMELOS decreased by 0.024 days (95% CI: 0.034 0.014, p<0.001). For patientlevel analyses, facilities were classified into five categories based on percentage of patients first admitted to OBS during the year (<1%, 1 10%, 10 20%, 20 30%, = 30%). Table 1 presents the mean LOS and odds ratio (OR) of readmission for a 60 yearold white male admitted for chest pain. Compared to admitting all patients as acute, OBS led to shorter LOS and reduced odds of readmission. However, this benefit decreased and odds of readmission eventually exhibited an increased risk as the percentage of OBS increased at the hospital.
There was considerable variation in OBS which biased hospital OMELOS. This bias however was not associated with improved patient outcomes. Instead, as hospitals increased their use of OBS, those patients experienced increased LOS and odds of readmission. This likely reflects that there are no process benefits to OBS, instead the difference between categories reflects a changing patient case mix. These factors need to be carefully considered and evaluated when determining how to handle OBS patients in interhospital comparisons.
Table 1Mean LOS and odds ratio (OH) of readmission for a 60 yearold while male chest pain admitted as either an acute or OBS patient in each of the 5 facility categories
To cite this abstract:Smith A, Glasgow J, Kaboli P, Johnson S. Observation Status: Impact on Hospital and Patient Outcomes. Abstract published at Hospital Medicine 2012, April 1-4, San Diego, Calif. Abstract 97656. Journal of Hospital Medicine. 2012; 7 (suppl 2). https://www.shmabstracts.com/abstract/observation-status-impact-on-hospital-and-patient-outcomes/. Accessed March 20, 2019.