443 The Impact of Observation Services for Medicare Beneficiaries With Chest Pain

Abstract

Study Objectives: Multiple factors have led to growing utilization of observation services: a focus on reducing readmissions, the growth of emergency department (ED)-based observation units, and greater scrutiny on short-stay admissions. There was a 2008 payment rule change that packaged all claims for observation services with the preceding services (ie, the ED visit), and eliminated both restrictions for specific diagnoses and the 48-hour maximum for length of stay (LOS). In this project, we assessed care patterns and 30-day outcomes among Medicare beneficiaries presenting to the ED with chest pain, which was the most common diagnosis for patients receiving observation services in 2006 and 2009, to explore changes in service use, any revisit within 30-days, and 30-day mortality. Methods: We conducted a retrospective cohort study of chest pain visits in Medicare beneficiaries, using a 20% nationally representative sample from CY 2006 and CY 2009. Visits were identified using the following primary and secondary ICD-9 codes: 78659 Chest Pain NEC, 78650 Chest Pain NOS, 78651 Precordial (chest wall) Pain, 4139 Angina Pectoris NEC/NOS, and 78652 Painful Respiration. The study was conducted at the beneficiary level and data between January 1 and December 1st of the given year were used when assessing 30-day revisits (ie, any revisit to a hospital including an ED visit, observation stay, or inpatient admission) and 30-day mortality. We tested differences across years using Fisher's exact test and multivariable logistic regression analysis within year to see if observation v. inpatients had any differences in revisit rates. Results: We studied 230,732 beneficiaries in CY 2006 and 219,636 in CY 2009 with chest pain. A higher number and proportion of beneficiaries were treated for chest pain in observation status in 2009 (n=51,493, 23.4%) than in 2006 (n=35,496, 15.4%), p<0.01 and fewer were treated as inpatients in 2009 (n=44,371, 20.2%) than 2006 (n=70,505, 30.6%), p<0.01. The average LOS in observation increased from 25.9 hours to 27.2 hours (p<0.01) from 2006 to 2009 while the average inpatient LOS decreased from 2.5 days to 2.2 days (p<0.01). In a logistic regression model, controlling for demographic characteristics and chronic conditions, the odds ratio for any revisit comparing index inpatient stays to index observation stays was 1.41 (CI 1.29-1.54) in 2006 and 1.41 (CI 1.36-1.44) in 2009. There were no meaningful differences 30-day mortality between CY 2006 and CY 2009 which remained constant at 1%. Conclusion: Shifts in payment policies have been associated with an increased use of observation services over inpatient hospitalizations for Medicare beneficiaries presenting to the ED with chest pain. Despite the increasing use of observation services, the likelihood of a revisit comparing observation v. inpatient admissions has remained relatively constant, after adjusting for other factors, and mortality has not changed.