Development, implementation, and evaluation of a computerized pediatric infection surveillance system

This project had two main objectives. The first objective was to determine if a computerized expert system could improve the detection of infections in hospitalized pediatric patients, and consequently help with the surveillance of hospital-acquired infections. The second objective was to analyze if computer-generated alerts and reminders about infections could influence clinicians (staff physicians, physicians' assistants, and nurse practitioners) practicing patterns. To accomplish these objectives, a computerized expert system (denominated COMPISS) using boolean rules was developed and validated in a retrospective study. During the retrospective study, newborn patient data for a two-year period from the LDS Hospital database were processed to validate COMPISS. During the retrospective validation COMPISS had a sensitivity of 84.5% and specificity of 92.8%. After the validation study, COMPISS was evaluated in two prospective studies. In the first prospective study, COMPISS was implemented into Primary Children's Medical Center (PCMC) Infection Control Program. COMPISS results were compared to the results produced by the manual surveillance done by an Infection Control Coordinator (ICC). COMPISS had a sensitivity of 100.0% for bloodstream infections and 57.1% for surgical wound infections. The ICC manual surveillance had a sensitivity of 42.6% for bloodstream infections and 47.6% for surgical wound infections. The first experiment demonstrated that COMPISS could help with the surveillance of the two studied infections and could also reduce the time required for the detection of these infections. In the second prospective study, the conclusions produced by COMPISS were presented directly to the attending clinicians in some units at PCMC. Clinician interventions were compared before and after the implementation of COMPISS. The results obtained in the second experiment failed to demonstrate a statistical difference between the clinicians' practice patterns before and after COMPISS implementation. The findings suggest that a larger sample size was required to demonstrate the small effect of the reminders. The clinicians and nurses' opinion about COMPISS were assessed through a paper questionnaire following of the second experiment. The questionnaire results demonstrated that the respondents appreciated COMPISS alerts and reminders and agreed that they should be continued and implemented in other units of the hospital.