|
BACKGROUND/RATIONALE:
The number of infections caused by antibiotic-resistant bacteria continues to
increase in United States hospitals. One specific multi-drug-resistant bacteria, methicillin-resistant Staphylococcus aureus (MRSA), is prevalent in hospitalized patients and give rise to excess morbidity, mortality and costs. Patient-to-patient is the most significant causal factor in the continued increase in MRSA infections. Importantly, a large number of patients who are colonized with MRSA do not develop clinical infection, and therefore remain unrecognized but serve as important sources of patient-to-patient transmission. Active surveillance programs, which utilize surveillance swabs to purposefully detect previously unrecognized, colonized patients and isolate them, are effective, yet costly. Clinical prediction rules have been developed for a variety of medical conditions to identify high-risk patients for therapy or screening. Factors from patients' past medical history have been identified as risk factors for MRSA infection and could be incorporated into easily implemented prediction rules. These prediction rules could then be used to identify high-risk patients for MRSA colonization at
admission to VA acute care hospitals to target for surveillance culturing that would be cost-effective and improve patient safety.
OBJECTIVE(S):
The results of this three-year study will provide specific guidance as to where in Veteran's Affairs Hospitals it would be most cost-effective to implement prediction-rule guided MRSA active surveillance. The long term objective is the creation of a VA-specific, yet flexible mathematical models of antibiotic resistant transmission to decision makers at individual Veterans Affairs Hospitals in identifying cost-effective control points to limit the spread of multi-drug-resistant bacteria.
Our Objectives are (1) Enroll a 2-year prospective cohort of patients admitted to the Baltimore VA Medical Center and collect anterior nares and wound cultures and data on past and current medical history using questionnaires and the VAMC Computerized Patient Records System (CPRS), (2) Assess the unadjusted and
adjusted for demographics, comorbidity and other confounders, statistical associations of potential predictors with colonization of MRSA at the time of hospital admission among high-risk individuals using data from the first 12 months of the cohort, (3) Validate the created prediction rule using the standard validity measures of sensitivity, specificity, and positive and negative predictive values in a new cohort of patients upon admission to the general medicine and surgical wards during the second year of the cohort, (4) Create a hospital-wide MRSA differential equation model incorporating data collected under the Specific Aims of Hypothesis 1, existing hospital administrative and microbiology databases and published literature regarding MRSA transmission and active surveillance/patient isolation effects, (5) Create an MRSA-specific Markov model and combine it with the model created in the prior aim to determine the effectiveness and costeffectiveness of the created prediction rule directed active surveillance and isolation, and (6) Perform sensitivity analyses of the combined models to gauge the generalizability of the model and allow individual VA hospitals to determine if prediction rule directed active surveillance should be implemented in their unique hospital.
METHODS:
The proposed 3-year project includes both a 2-year prospective cohort study design and a cost-effectiveness analysis (CEA) based on mathematical modeling. All patients admitted to the general medical and surgical wards of the acute care Baltimore VA Medical Center will be approached for enrollment in this study. If they consent to participate, an in-person interview will be administered and swab cultures of the anterior nares (nasal) and sites of skin breakdown (e.g wounds) will be obtained. In addition, evidence of past exposures in the year preceding the current admission will be collected from CPRS data extraction and minimal chart review, including antibiotic exposures, comorbid conditions, and previous admission history. A prediction rule for the presence of MRSA colonization will be created using collected data. The effectiveness and cost-effectiveness of prediction-rule guided active surveillance will be assessed with mathematical models. Sensitivity analysis will be completed to determine the robustness and generalizability of the findings by varying key parameters of the model.
FINDINGS/RESULTS:
Study is still ongoing. We have enrolled approximately 400 patients into the prospective cohort. Preliminary findings were presented in abstract form at the 2008 Society for Healthcare Epidemiology of America (SHEA) national meeting in Orlando (April 2008). An additional abstract is accepted for presentation at the 2008 Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) in Washington DC (October 2008). Patient self-report of having received antibiotics in the past year was the most sensitive predictor for MRSA colonization (75%). A prediction rule using self-report of hospitalization or receiving antibiotics in the past year would have identified 92% of patients colonized with MRSA.
IMPACT:
VHA acute care hospitals need to know the most cost-effective way to control MRSA in their unique specific environment. If targeted active surveillance (guided
by prediction rule) of patients at high risk for MRSA upon hospital admission is found to be cost-effective, then targetted surveillance programs to control MRSA could be implemented and patient outcomes could be improved with reduced costs compared to mandatory hospital-wide surveillance as currently mandated VHA-wide.
PUBLICATIONS:
None at this time.
DRA:
Acute and Traumatic Injury, Chronic Diseases, Health Services and Systems
DRE:
Prevention, Diagnosis and Prognosis, Quality of Care
Keywords:
Acute illness, Screening
MeSH Terms:
none
|
