ISSUE: Data mining is the process of data exploration, selection, and conversion. Data mining has developed as an important tool in healthcare. Hospital databases are rich data resources, which may be transformed into powerful information for use to enhance patient care. The Infection Control Department at St. Luke’s Episcopal Hospital (SLEH) is committed to providing quality health care to the community we serve and so we takes advantage of modern technology to collect, store, and data mine to this end.

PROJECT: The Infection Control Department at SLEH evaluates various data mining software products with the following goals in mind:

RESULTS: Commercial Data Mining software products are available but at significant cost. We evaluated the resources we currently have and discovered, that with a lab interface to our current database, we could have a system capable of data mining, which meets all of our requirements.

Similar to other data mining products, our system is capable of house-wide or targeted surveillance, generate monthly reports, provide patient admitting diagnosis and culture history, offer unit-specific rates, and provide rapid identification of patients with multi-drug resistant organisms.

Additionally, our system offers excellent custom report-writing features, which are highly effective with outbreak investigations and supports research projects. Furthermore, this system can be set up to benchmark against the National Healthcare Safety Network (NHSN).

LESSONS LEARNED: Evaluating data mining software products begins with set goals. Custom report-writing features are beneficial and can assist in outbreak investigations as well as research. Although weekly trend reports are optimal, data mining software must provide, at a minimum, monthly trend reports to track infection-related events in a healthcare facility—graphing capabilities are a significant asset. The capability to detect a sudden increase in any specific organism and the ability to monitor a known problematic microorganism is a necessity. Purchasing new software products can be expensive. Sometimes, current resources can be utilized to meet data mining needs.