Convert your FHIR JSON -> XML and back here. The CDA Book is sometimes listed for Kindle here and it is also SHIPPING from Amazon! See here for Errata.

Monday, October 18, 2010

Analysis of Proposed ISDS Standards for Disease Surveillance

The executive summary:  For Meaningful Use Biosurveillance requirements, you should use HITSP C39 for now, and when this work is complete, and if adopted, you will be pretty darn close.  Why we couldn't have adopted the HITSP C39 for now is quite beyond me.

Here is what I found comparing the two:

What's added to the ISDS specification:
Facility Identifier and Address -- Not specified HITSP specification, and there is a clear need for it.
Visit Identifier -- Not specified in the HITSP Specification
Patient Country -- Not specified in the HITSP Specification

Patient Class -- The ISDS specification does not constrain the vocabulary to Emergency, Inpatient and Outpatient like HITSP did.  Instead they recommend that constraint.  It's not clear why this is done, but I can imagine that it might contain useful information for public health if a larger value set was used.

Date/Time of Message -- ISDS puts this in the EVN segment (in EVN-2), but the specification also says this is the time of the message transmission which HL7 Version 2 clearly states goes into MSH-7 message creation time.  I think the concern here is that an intermediary will report a different time when it forwards a message than what the facility provides.  So, put your date time in EVN-2 and MSH-7.

What was in HITSP that isn't in the ISDS specification any more:
Date of Birth -- Explicitely excluded, most likely due to privacy issues.
Diagnosis Date / Time -- The ISDS specifiction doesn't include this data element.  It seems as if it would be useful, so I'm not sure why they didn't include it.
Discharge Disposition -- Now uses UB04 instead of UB92 codes, but otherwise the same.  This would have been fixed in HITSP had we ever gotten the change to update the V2 specifications to the new data architecture.
Heart Rate -- Not included in the ISDS specification, not sure why.
Blood Pressure -- Not included in the ISDS specification, not sure why.
Nurse / Triage Notes -- Not included in the ISDS specification.

Altogether, I'm pretty happy with how well the work matches what was already done in HITSP.  I'm extremely dismayed that the developers didn't report these deltas in their own efforts, because it is pretty clear that they could have rather easily.  I did it in about 4 hours of effort, mostly because I had to copy/paste tables from the Adobe PDF into a spreadsheet in order to complete the analysis.  That was a total waste of time.  Once that was done the analysis took about 15 minutes.  At least with the HITSP work I had the original source for the tables.

When this work is done, I want it to be available as an HL7 Version 2 Profile -- Enough of these PDF tables.  Also, I want to be able to post comments on this work.  It's fine to get comments from the Joint Public Health Information Task Force, but you also need feedback from implementors of these specifications, or as I said previously, the process is broken.

The specification needs to be a heck of a lot easier to read.  Those tables were a nightmare.
1.  Put the information in the tables in the order it appears in the message.
2.  Show the message segment layout first
3.  Put table stuff in tables and notes below in text for each field, just like HL7 does for the standard.  It's a lot easier to get through that way.
4.  Separate fields and segments and defines lengths for each.


  1. Keith, can you post a link to the ISDS specification that you used for this comparison? Thanks for your blog. Bryan