570. Recent Patterns in Nosocomial and Community-Associated Infection Rates: Insights from a Nationwide Electronic Health Outcomes Database
Session: Poster Session: Hospital-acquired and Transplant Infections
Friday, October 30, 2009: 12:00 AM
Room: Poster Hall A
Background: The Nosocomial Infection Marker (NIM) and Community Infection Marker (CIM) are validated electronic, laboratory-based markers used by an automated infection prevention surveillance service to identify nosocomial infections (NIs) and community-associated infections (CAs), respectively. Changes in NIM, MRSA NIM, blood NIM and analogous CIM rates over a 2-year period were studied in 171 US hospitals.
Methods: De-identified data from the MedMinedTM Health Outcomes Database were analyzed from 8/2006 - 10/2006 and 8/2008 - 10/2008. Mean and median relative changes in NIM and CIM rates between periods were calculated. Rate differences of NIMs and CIMs specifically marking MRSA and blood infection were also determined. Analysis was limited to hospitals with ≥ 2 years of continuous data ending 10/2008. NIM rates were defined as the number of NIMs detected in each hospital per the number of inpatient admissions in that hospital. CIM rates were defined similarly.
Results: 1,200,000 admissions from 171 hospitals were analyzed. The median number of monthly hospital admissions was 1,145 [IQR: 483, 1765]. Two hospitals had < 50 monthly admissions and 2 had > 4,000. Between 2006 and 2008, median changes in NIM, MRSA NIM and blood NIM rates were -19% [IQR:-34%, -4%], -27% [IQR:-46%, 1%], and -23% [IQR -43%, 1%]. Mean rate changes were significantly lower at -14.5% for NIMs [SE: 2.7%], -17% for MRSA NIMs [SE: 3.9%] and -15% for blood NIMs [SE: 3.5%]. Median changes in CIM, MRSA CIM and blood CIM rates mostly trended upward and were 10% [IQR: -4%, 37%], 13% [IQR: -16%, 62%] and -3% [IQR: -31%, 29%]. Mean changes in CIM, MRSA CIM, and blood CIM rates were 32% [SE: 6%], 40% [SE: 7%] and -3% [SE: 6%].
Conclusion: Automated hospital-wide, electronic infection surveillance enables rapid, wide-scale study of NI and CA epidemiology. In a large nationwide sample, use of validated highly sensitive and specific infection markers suggests a recent decrease in NI rates but an increase in CA rates.
Stephen Brossette, MD, PhD, Cardinal Health, Birmingham, AL, Emily Falk, MSPH, Cardinal Health - MedMined (TM) Services, Birmingham, AL, Patrick Hymel Jr., MD, Cardinal Health MedMined (TM) services, Birmingham, AL, Richard Johannes, MD, MS, Cardinal Health MedMined (TM) services, Marlborough, MA and  S. E. Brossette,
Cardinal Health MedMined (TM) services Role(s): Employee, Received: Salary.
E. A. Falk,
Cardinal Health MedMined (TM) services Role(s): Employee, Received: Salary.
R. S. Johannes,
Cardinal Health MedMined (TM) services Role(s): Employee, Received: Salary.
Brigham and Women’s Hospital, Division of Gastroenterology, Harvard Medical School Role(s): Employee, Received: Salary.
P. A. Hymel Jr.,
Cardinal Health MedMined (TM) services Role(s): Employee, Received: Salary.