PREDICTION MODEL TO IDENTIFY PATIENTS DEVELOPING MRSA INFECTION IN AN ORTHOPAEDIC UNIT

Abstract

Objective: To identify institution specific risk factors for developing MRSA surgical site infection (SSI) and develop an objective mechanism to estimate the probability of MRSA infection in a given patient admitted to the orthopaedic unit.

Design: A cohort study was performed to identify risk factors in all patients who had MRSA infection during admission on the orthopaedic unit between January 2002 and December 2004. Logistic regression was used to model the likelihood of MRSA. A stepwise approach was employed to derive a model. The MRSA prediction tool was developed from the final model.

Results: Of the 11 characteristics included in the logistic regression, the features that strongly predicted a MRSA infection were ASA grade, patient’s residence and reason for admission.

110 had MRSA infection in their surgical wound. 83 of 110 (75.5%) patients were non-elective admissions, of which 49 (60%) were proximal femur fractures. 20% of proximal femur fractures admitted from nursing home and 7.8% from their own homes developed SSI with MRSA. This cohort of SSI with MRSA had an average of 5.7(1–18) previous admissions. 25 (23%) had been previously colonised with MRSA. Majority of them (76%) were between 70–90 years old and were ASA grade 3–4.

Conclusion: Through multivariate modelling technique we were able to identify the most important determinants of patients developing SSI with MRSA in our institute and develop a tool to predict the probability of MRSA in a given patient. This knowledge can be used to guide the use of appropriate prophylactic antibiotic and to take other required measures to avoid the SSI with MRSA.