Choice of antibiotic in open heart surgery

Abstract

It is the purpose of this paper to investigate antibiotic prescribing for patients undergoing open heart surgery. Patient data were collected for approximately 2100 patients, with a focus on infection and antibiotic orders. Physicians were separated into high infection rate and low infection rate groups with 6% infection as the cutpoint. Market basket analysis and kernel density estimation were used to investigate the prescribing practices. The analysis indicated that there is a difference between the high infection and low infection groups in the amount and kind of antibiotic prescribed. Low infection physicians use more of the antibiotic, Cipro while high infection physicians use Levaquin. Once the difference is found, other statistical methods are used to examine details, including text analysis. A data mining analysis of medication prescriptions can lead to improved decision making and patient outcomes.