558. Substantial Shifts in Ranking of California Hospitals by Hospital Onset Methicillin-Resistant Staphylococcus aureus (HO-MRSA) Infection Following Adjustment for Hospital Characteristics and Case Mix
Session: Oral Abstract Session: Advancing Public Reporting and Surveillance of HAIs
Thursday, October 3, 2013: 2:00 PM
Room: The Moscone Center: 200-212

Background: States have established public reporting of hospital associated infections including those of MRSA. However, these reports do not account for hospital case-mix or post-discharge infections.

Methods: We conducted a retrospective cohort study of CA hospitals during 2009-2010. We used mandatory discharge data and home zip code characteristics from US Census data to generate hospital and patient population characteristics summarized across all admissions. To evaluate facilities' proportion of HO-MRSA admissions, we examined all MRSA pneumonia or septicemia cases which occurred pre- or post-discharge (coded as present on admission and within 30 days of previous hospitalization). Facility level factors associated with hospitals' proportions of HO-MRSA infection admissions were identified using multivariate Poisson regression models. We fit a Poisson-Gamma hierarchical generalized linear model with saturated random effects to create adjusted hospital rankings based on HO-MRSA infection proportions and compared them to raw rankings using Kendall's Tau concordance statistic.

Results: Among 323 CA hospitals, a median of 16 cases (range: 0-102) per 10,000 admissions were identified as HO-MRSA infections and were associated with facility level predictors (Table). Concordance between raw and adjusted hospital rankings was only 0.47 (95% CI: 0.42-0.52). For any given pair of hospitals, 27% of hospitals switched their relative ranking with regard to one another after adjustment.

Conclusion: When adjusting for non-modifiable facility characteristics and case mix, hospital rankings of HO-MRSA pneumonia and sepsis substantially changed. Quality indicators for hospitals require adequate adjustment for patient population characteristics for valid inter-hospital comparison.



Estimated Multiplier

95% CI




1.39, 1.77


Renal Failure


1.07, 1.39


Metastatic Cancer


1.05, 2.00




1.07, 22.9


Discharged to Home


0.81, 0.92


Admitted from low education zip codes


1.00, 1.15


1Per 10% increase in hospital admissions involving patients with respective characteristics there is a multiplicative change in the proportion of HO-MRSA infection admissions


David M. Tehrani, MS1, Chenghau Cao, MPH1, Michael Phelan, PhD2, Rupak Datta, PhD MPH1, John Billimek, PhD3, Hoanglong Nguyen, MD1, Homin Kwark1 and Susan S. Huang, MD, MPH, FIDSA1, (1)Division of Infectious Diseases and Health Policy Research Institute, University of California Irvine School of Medicine, Irvine, CA, (2)Department of Statistics, University of California Irvine, Irvine, CA, (3)Health Policy Research Institute, University of California Irvine School of Medicine, Irvine, CA


D. M. Tehrani, None

C. Cao, None

M. Phelan, None

R. Datta, None

J. Billimek, None

H. Nguyen, None

H. Kwark, None

S. S. Huang, None

Previous Abstract | Next Abstract >>

Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 2nd with the exception of research findings presented at the IDWeek press conferences.