From: Jan-Olof.Arvidsson@astrazeneca.com
Sent: Monday, May 19, 2003 2:28 AM
To: fdadockets@oc.fda.gov
Subject: Specification for Annotated Electrocardiographic Waveform Data
in Electronic Format

Regarding the ECG XML Submission Standard, AstraZeneca has encountered
problems in the ballot material as follows:

Technical issues:

Problems with Microsoft MSXML:
*	The XML file cannot be validated to XML schema with MSXML v 4,
option 'validateOnLoad' must be set false.
*	The XPATH property (which is used to locate elements) in MSXML v 4,
does not work with the HL7 namespace conventions and inlcude files.

Problems with XML schema:
*	The structure is complex and big with nested include files. The
schema should be used to validate XML files and we feel that it should be
clearly and unambigously defined. We have not found the definitions of the
items in the schema, i.e. a description of how they should be used. XML
supports such annotations.
*	The XML schema can not be opened in XML Spy without a lot of errors.
*	Validating the small example file we got in the ballot, XML Spy
takes half a minute on a 2.26 GHz P4 equipped PC running W2K.


Other issues and suggestions:

*	The use of elements in the standard is unclear, it should be more
well documented and unambigously defined.
*	The HL7 schemas seems to be big and complex, which makes everything
working with it slow and difficult to interpret. This is acceptable if the
complexity fulfils a purpose, but validation must work with standard tools
such as XML Spy and Microsoft MSXML.

AstraZeneca has implemented to the best of our knowledge (including some
guesses) an import and export function in our ECG analysis tool. The file
size becomes large, but this is ok as long as everything else works fine.
Especially annotations generate much code for little information. We feel
that there should be more practical experience and XML file interchanges
between involved organisations before releasing the standard. 


Additional elements that we suggest should be included in the XML file:

Centre ID	- This is an identification of the department/area where a
clinical study is performed
Study ID	-  Identification of study id (company specific)
Investigator ID	-  Study responsible
Subject ID	- Identification of subject (study specific coded or person
information, should support coded as well as full id)
Age
Race
Country		- Country where ECG is collected
Diagnosis	- ECG diagnosis
Technician ID		- Person that captures and stores the ECG
Treatment code		- Drug used in study
Treatment group		- Drug group 
Visit number	- Company specific, used for clinical studies
Visit day	- Date for visit
Visit name	- Clinical study specific
Source file name (when converted to XML)	- Tracability to original
file name
Comments about ECG on session or beat level	- Free comments like
quality, morphology etc.
Audit trail information 	- Audit trail information on annotations
Status on session/beat	- e.g. included or excluded in analysis



Regards

Gunnar Fager, Jan-Olof Arvidsson, Bo Skallefell
AstraZeneca R&D Mölndal, Sweden
(gunnar.b.fager@astrazeneca.com,
jan-olof.arvidsson@astrazeneca.com;bo.skallefell@astrazeneca.com)