Computer surveillance of hospital-acquired infections and antibiotic use

• HIT Description: Computerized medical record, which contained a computerized knowledge-base that analyzed all test results from the microbiology laboratory More info...
  
• Purpose of Study: To improve the identification of hospital-acquired infections and associated antibiotic use
  
• Years of study: 1984
  
• Study Design: Cohort
  
• Outcomes: Identification of infections and utilization of services
  

• Settings: Academic-affiliated 520-bed hospital
• HIT System: The HIT system was locally developed, beginning in the 1970s and had been in place for 15 years at the time of study. The system was known as ÒHealth Evaluation through Logical Processing.Ó It is a computerized medical record with medical decision logic. In this study, the authors developed a computerized knowledge base that analyzed all test results from a microbiology laboratory in order to identify patients 1) with hospital-acquired infections, 2) not receiving antibiotics to which potential pathogens are susceptible, 3) who could be receiving a less expensive, let equally efficacious antibiotic or 4) receiving surgical prophylaxis longer than 48 hours without evidence of infection.
• Comparison of HIT system with traditional surveill: There were 4,679 patients discharged from the hospital during the 2-month study period, and a total of 217 patients with suspected hospital-acquired infections were identified by either the HIT system or the traditional method of surveillance (infection control clinicians). Study physicians determined that 155 of these 217 patients met the criteria established for the diagnosis of hospital-acquired infection. The HIT system identified 140 (90%) of the confirmed cases, whereas the infection control clinicians identified 118 (76%). False-positive diagnoses were made by both methods at nearly the same frequency, about 20%. During the 2-month study period, the HIT system generated 108 antibiotic-related alerts, of which 40 were determined to be an error. Lastly, the computer correctly identified 142 of the 156 patients who were receiving prophylactic antibiotics longer than necessary.
• Utilization of services: The authors report that the traditional method of infection surveillance required 138 hours of the practitioners time, whereas the HIT system required only 8.6 hours plus 15 minutes of physician reviewer time for 45.5 hours of total time, in order to do the same work but with increased sensitivity.