Impact of Nonpayment for Preventable Infections on Billing Rates for Central Line Associated Bloodstream Infections and Catheter-Associated Urinary Tract Infections
Methods: We used an interrupted times series design to assess whether the HAC policy was associated with changes in the intercept and trend of CLABSI and CAUTI billing rates. We included adult Medicare patients from 57 California, 10 Massachusetts, and 32 New York hospitals in the State Inpatient Databases. Data through September 2011 were included, starting with January 2008 for CLABSI analyses and January 2007 for CAUTI analyses. We estimated rates of infection (per 100,000 discharges) using ICD9 codes and present-on-admission indicator variables. We utilized mixed effects logistic regression with clustering at the hospital level. Independent variables included policy period, time, policy period*time, state, number of beds, ownership type, teaching status, and percent Medicare admissions.
Results: Prior to the HAC policy, billing rates for CLABSI and CAUTI increased (pre-intervention OR per quarter for CLABSI: 1.17, 95% CI 1.11, 1.23; OR for CAUTI: 1.13, 95% CI 1.05, 1.22). The policy was associated with an immediate drop in the billing rate for CLABSI (OR for change at HAC policy=0.75, 95% CI 0.70, 0.82), but not for CAUTI (OR for change at HAC policy =0.93, 95% CI 0.70, 1.24). Following the policy, slight declines or no changes in billing rates were seen for CLABSI and CAUTI (post-intervention OR per quarter for CLABSI=0.98, 95% CI 0.97, 0.99; OR for CAUTI=0.98, 95% CI 0.94, 1.10).
Conclusion: The HAC policy was associated with an immediate decline in billing rates for CLABSI and with a change in the trend in billing rates for CLABSI and CAUTI. This contrasts with a study that found no impact of the HAC policy on CLABSI and CAUTI rates using standardized National Healthcare Safety Network definitions that use clinical and laboratory data. The impact of the HAC policy on CLABSI and CAUTI may be overestimated when measuring infection rates in billing data, highlighting the importance of the data used for assessing policy impacts.
A. T. Kawai,
S. Soumerai, None
L. E. Vaz, None
M. Rett, None
G. Lee, None
See more of: Oral Abstract Session