Parsing Multi-Source Microbiology Culture Data in a Regional Health Information Exchange to Describe Patterns of Infection
Methods: We built 1) an HL7v2 correction engine that deals with incorrect microbiology message structure and content. The HL7v2 correction engine parses key data elements needed for epidemiologic analyses: organism, antibiotics tested, minimum inhibitory concentrations, susceptibility interpretation, body source of the culture, and health care facility where drawn. We describe blood and/or urine cultures, for infants 0 to 91 days old, from the first 20 weeks of the new system to parse electronic microbiology culture data.
Results: Seventy-eight infants (42 girls, 36 boys) had at least one positive culture for one or more GNRs (64 urine, 9 blood, 5 infants with both). The majority of the results were Escherichia coli (54%), followed by Klebsiella pneumoniae (15%), Enterobacter cloacae (10%), Klebsiella oxytoca (6%), and various others. The age range of the 78 infants was 0 to 91 days. Three of the 78 infants had urine cultures identified as Enterobacteriaceae with extended-spectrum beta lactamase (ESBL-E), two with Escherichia coli and one with Klebsiella pneumoniae. One of these three had a first urine culture that was a GNR but not ESBL-E, followed by a urine culture of ESBL-E three weeks later.
Conclusion: This electronic microbiology parsing system shows promise, in a health information exchange, for describing infection and antibiotic resistance patterns across a region. Most patients with GNRMDRO are adults, but the system is also useful for pediatric data.
S. M. E. Finnell, None
S. Khokhar, None
A. Kho, None
See more of: Poster Abstract Session