ASSOCIATION RULES AND DATA MINING IN-HOSPITAL INFECTION-CONTROL AND PUBLIC-HEALTH SURVEILLANCE

Objectives: The authors consider the problem of identifying new, unexp ected, and interesting patterns in hospital infection control and publ ic health surveillance data and present a new data analysis process an d system based on association rules to address this problem. Design: T he authors first illustrate the need for automated pattern discovery a nd data mining in hospital infection control and public health surveil lance. Next, they define association rules, explain how those rules ca n be used in surveillance, and present a novel process and system-the Data Mining Surveillance System (DMSS)-that utilize association rules to identify new and interesting patterns in surveillance data. Results : Experimental results were obtained using DMSS to analyze Pseudomonas aeruginosa infection control data collected over one year (1996) at U niversity of Alabama at Birmingham Hospital. Experiments using one-, t hree-, and six-month time partitions yielded 34, 57, and 28 statistica lly significant events, respectively. Although not all statistically s ignificant events are clinically significant, a subset of events gener ated in each analysis indicated potentially significant shifts in the occurrence of infection or antimicrobial resistance patterns of P. aer uginosa. Conclusion: The new process and system are efficient and effe ctive in identifying new, unexpected, and interesting patterns in surv eillance data. The clinical relevance and utility of this process awai t the results of prospective studies currently in progress.