Carbapenem-resistant Enterobacteriaceae (CRE) are extensively drug-resistant organisms (XDROs) that pose a public health threat. CRE can spread between healthcare facilities through movement of CRE-colonized patients. Inter-facility communication of patient CRE status enables regional CRE control by allowing healthcare facilities to enact appropriate infection control precautions at the time of admission, but routine information sharing is inconsistent.
In November 2013, the Illinois Department of Public Health and Chicago CDC Prevention Epicenter launched the XDRO registry; CRE must be reported to the registry and authorized providers could manually query the registry at the time of admission to determine if an individual patient was previously reported as CRE-positive. In 2015, we tested automation of CRE alerts at three facilities; facilities received an alert anytime a patient from the XDRO registry was admitted to their facility.
METHODS:
Pilot hospitals sent inpatient admission data (encrypted patient name and birth date) to the XDRO registry as often as every hour. When a full match (based on birth date, last name, first name) or partial match (birth date, last name, first initial) between hospital and registry identifiers was detected, a generic e-mail alerted the facility’s infection preventionist to view the patient’s identifiable information in the XDRO registry’s secure website.
RESULTS:
Three hospitals were connected to the automatic alert system (first facility January 2015, others July 2015); they received 83 alerts for 45 patients through 12/15/2015. Seventy-nine (95%) alerts were full matches, 80 (96%) identified the correct patient, and three (4%) were false partial matches for one patient who visited the same hospital multiple times.
Among 80 true matches, the facility did not know CRE status for 35 (44%) patients, of which 23 (66%) were not in contact precautions at the time of the alert. Even among the 45 patients for whom facilities had prior CRE information, 16 (36%) patients were not in contact precautions, likely because admitting personnel were unaware of prior CRE status. When a facility had knowledge of CRE status before the alert, the knowledge was due to previous information at that same facility (n=45; 100%) rather than inter-facility communication.
CONCLUSIONS:
Patient-specific CRE information can be successfully shared between healthcare facilities in an automated fashion using a public health registry. Such a system can improve situational awareness of patient CRE status at the time of admission to support timely implementation of infection control measures.