The overarching objective of the CDC Epicenter is to conduct research and lead programs that reduce the burden of healthcare-related infections and other types of adverse events. The Center’s underlying theme is the transformation of the practice of infection control and healthcare epidemiology through effective use of health information technology. Our premise is that real-time surveillance and computerized decision support, when implemented within a framework of system change can drive improved prevention and management of health care-related infections.
The Center brings together notable strengths in infection control, patient safety, medical informatics, quality improvement, infectious disease epidemiology, behavioral sciences, and public health. The health care and academic institutions that comprise the Center are internationally recognized for their leadership in these areas. Projects which have been developed to bridge distinct clinical entities will demonstrate the generalizability of the evaluated strategies, not just within Utah, but across other geographic regions [as well]…Our target audience includes infection control practitioners, healthcare epidemiologists, clinicians, public health personnel, hospital administrators, and chief financial officers of healthcare institutions. Our ultimate aim is to serve as a powerful instrument of change toward safer, more efficient, and effective delivery of care. Substantially improved patient outcomes through prevention of healthcare-associated infections and other types of adverse events is a reachable goal.
Urinary Catheter Duration Study: The urinary catheter alert study involves a computer algorithm to flag those patients who have a urinary catheter in place which might no longer be needed and provide an electronic alert. This study was implemented in November 2008 in two Intermountain Healthcare hospitals. A nurse survey/questionnaire to assess user impressions and feedback about the system is also part of this study.
MRSA De-colonization Study: This pilot MRSA decolonization study tests the feasibility of implementing a protocol for hospital-wide use of intranasal mupirocin and topical chlorhexidine to treat all MRSA carriers detected by active surveillance. This protocol, developed for a 30 VA hospital cluster-randomized trial, is being conducted at the Salt Lake VA and the Pittsburgh VA.
MRSA Electronic Algorithms Study:
MRSA algorithms are being developed and evaluated to classify distinct categories of healthcare associated infection (HAI) on the basis of electronic health record data. The initial implementation of these algorithms uses local VA SLC data. The algorithms are undergoing additional refinement and testing. Manual chart review is used to create a reference standard to evaluate algorithm performance against a reference standard as well as infection preventionist (IP) surveillance.
MRSA Modeling Study:
An agent-based model of MRSA infection as seen in the hospital setting is being developed and validated. The model encompasses pathways for transmission mediated by environmental reservoirs or by healthcare workers or both. We have used the simulation to explore factors that contribute to the effectiveness of an active surveillance strategy, and to compare this strategy with alternative surveillance approaches, such as surveillance that relies on the targeting of patients who are at high risk of colonization at the time of hospital admission.
Additionally, our work on dynamic models of MRSA infection supports statistical estimation of transmission parameters using data collected from intervention studies. These methods involve application of the Markov Chain Monte Carlo procedure to fit Bayesian models. The advantage of these models is that they account for missing data and imperfect diagnostic tests. We have used synthetic data sets generated from the simulation model to show that dynamic-based models are less biased than conventional statistical techniques.
Inter-rater Reliability Study:
Literature shows that variability of subjective interpretation of HAI case criteria impacts the validity of publicly reported healthcare-associated infection rates. This project provides a method for estimating the magnitude of the variability, using a study design in which infection preventionists (IPs) from different VA facilities interpret records on the same patients.
Recent project accomplishments include: 1) pilot testing of the Web-based HAI assessment tool identified a need to revise the automated de-identification system; 2) a CLABSI reference standard using an objective algorithm was established for all 120 patient records; and 3) 20 VA IPs were recruited, trained, and are in the process of completing their manual reviews.
Statewide Implementation of a Simplified BSI Algorithm Study:
The goal of this project is to evaluate statewide implementation of simplified, objective criteria for BSI surveillance: IPs have disagreed with algorithm decisions in approximately 10% of cases, and overall responded favorably to a web-based survey focused on acceptance of the algorithm. Individual level comparisons will be made between reported cases and BSI episodes identified by an electronic algorithm at hospitals in the University Health Care and Intermountain Health Care systems.
Surveillance of CDAD Study:
C. difficile infection (CDI) study: Both University Health Care and Intermountain Health Care are participating in the retrospective validation of C. difficile infections using an automated algorithm. University Health Care is also contributing to the prospective validation of the automated algorithm and in a CDI sentinel surveillance network by providing patient specimens for typing.
Antibiotic Utilization Study: Data has been aggregated and compiled by the CDC Epicenters’ antibiotic utilization workgroup which includes four hospitals. Information is being studied to evaluate the barriers and lessons learned on the collection and analyses of antibiotic data from the four different hospitals. Additionally, a comparison of antibiotic utilization between the ICUs at the four hospitals is being performed.
Ventilator Associated Pneumonia Study: Electronic data on 200 VAP patients has been extracted. Data items included ventilator setting, complications, and IP verified VAP information. Analysis includes comparing VAP and ventilator associated complications with respect to outcomes.