Nebraska Experience w. automated ELR to track antimicrobial resistance: examples of HL7 2.5.1 culture and susceptibility test, and sample dataset

In 2010 we began to experiment with making antimicrobial susceptibilities reportable on bugs that were on our reportable disease list.  At that time we focused on Strep pneumo, Staph aureus, Acinetobacter, and enterococcus—bugs that were on the CDC’s NHSN  MDRO module.  In 2017 we upgraded our regs

and added several bugs, along with language:

So this gets us the complete antibiogram on any of our reportable bugs (including TB, salmonella, etc), if the antibiogram was performed and if the lab is doing automated ELR.  We added a handful of bugs this past revision so now our list includes all isolates of Pseudomonas, Klebsiella, Enterobacter, E coli, and a few others---see our regs for all the bugs that are required to be reported.  We request both the MIC and the interpretation.  Some labs (e.g., LabCorp) are sending only the interpretation, but I think that could be negotiated to get the MIC. 

Our state population is 1.8 million.  We have maybe 25-30 labs to onboard for this kind of reporting.  Still a work in progress.  We have built a pretty good template or highway for this to happen, so now it’s a matter of getting other labs to conform to the model. 

All this data flows to us as part of the automated ELR data stream in HL7 format.  We get some labs doing HL7 2.3.1—most do HL7 2.5.1.  This process works for both formats but we'd like to move to 100% HL7 2.5.1.  I have attached a document that shows two lab reports.  One has just one isolate with its susceptibilities. The other is polymicrobial: it has 3 bugs isolated from one culture, plus the susceptibilities for the three bugs.  HL7 accommodates this just fine.  I have included screen shots to show how this appears in our NEDSS Base System application. 

We rarely look at these susceptibilities on an individual basis.  The goal is to have the data flow into a datamart where we can study trends in resistance, look for clusters, etc.

EXAMPLE OF HL7 FORMAT FOR ANTIMICROBIAL SUSCEPTIBILITY NEBRASKA NEDSS BASE SYSTEM.docx 362 KB • Download

 

SAMPLE DEIDENTIFED ANTIBIOGRAM DATA SET FOR BASECAMP 3.xlsx 248 KB • Download

We are planning to get an automated e-mail alert set up for resistance patterns needing immediate attention, such as CRE.  Just haven’t had time to do that yet.

I am attaching a small sample of our datamart created from processing the incoming data.  I’ve deleted certain fields—personal identifiers etc.  This took some serious SAS processing to clean up that data to the point where we have a pretty clean line listed data set.  You can see this is pretty much a dream data set for epi purposes: each isolate is one row/record, with the range of antibiotic susceptibilities as columns.  Info about dates and facilities and specimen sources, etc etc are included---anything that can come from within an HL7 ELR report. 

We have made minor forays into getting additional info about the patient through the PV1 segment (Patient Visit 1).  This segment has info on date of admission and date of discharge and patient location within the facility.  I have attached a document that describes the kind of data stored in the PV1 segment.  This has potential for studying hospital-acquired infections, but needs work and national balloting etc to have it become part of the national implementation guide for ELR.

We are working on importing this data into an application called WHONET (http://www.whonet.org/software.html).  John Stelling in Boston has worked in this space for a long time and has consulted with WHO.  We are new to this application and are exploring how it could work for us---still a work in progress.  We are able to import our spreadsheet into the application and have succeeded in generating some tables, graphs and charts.  Very early on at this point: unable to provide much comment on how this will work out…

We are happy to answer questions, discuss, show-and-tell, webinar etc if anyone is interested.  We put in an abstract to discuss this at CSTE Boise---hope it gets accepted and hope to see y'all there!!!  Tom Safranek, State Epi Nebraska  402-471-0550---desk phone     tom.safranek@nebraska.gov