Eric's BizTalk & HL7 blog

Event Type: Error
Event Source: BizTalk Accelerator for HL7
Event Category: None
Event ID: 8706
Date:  20/08/2008
Time:  12:55:30
User:  N/A
Computer: IN12620
Description:
Unable to log message "An empty batch was received with fr..." to the event log due to exception:
"No connection could be made because the target machine actively refused it 127.0.0.1:4000".

Username: IN12620\BizTalk
Time: 20/08/2008 12:55:30

The issue, is obvious: The HL7 Logging Service is not running:

A friend recently asked if it was ABSOLUTELY necessary to have the BHS/BTS segments if batching was enabled.

I asked some friends and the HL7 DASM demands those segments:

[FHS] (file header segment)

{ [BHS] (batch header segment)

{ [MSH (zero or more HL7 messages)

....

....

....

] }

[BTS] (batch trailer segment)

}

[FTS] (file trailer segment)

The FHS segment is used to head a file (group of batches), that is why HL7 DASM always expect BHS segment after FHS segment.

The only way is to create a pipeline component that would live in the Decode stage that inserts those segments into the batched message.

Microsoft has put a lot of effort into making your environment secure, however, that still does not absolve you of some things that you still have to configure.

Here is an example (using the examples in the SDK) of showing PHI:

Notice that you can essentially see EVERYTHING in that message, including that ever-so-important PHI.

There are a few things that you can to to remove/restrict access to that PHI.

Because there is the ability to see PHI in non secure places, Microsoft has a couple of places that you use to secure that data, based on your security policy. In some instances, the security stops at the firewall, if an intruder gets past the firewall, they have the keys to the kingdom. This security policy is fine for most instances, (thank goodness)! In my case, the last thing I want to do is have to peel off the layers of security to simply figure out that an element is missing from the HL7 message. At 2:00 am I simply want to log in and find out what is going on, what the error is, and fix it.

However, there are some out there that either do not trust their network admins, want to make troubleshooting harder, don’t trust their co workers, or were forced to implement additional security around PHI. For those, here are a few things that are available to increase security around PHI>

The first would be to turn off Event Logging and possibly choose SQL where more restrictions can be managed.

(Remember to restart the Audit and Logging Service)

Another would be to simply restrict who can have access to the event log in the first place, this can be done by changing the following key in the registry (did you see my disclaimer on the right side of this blog?)

Hive: HKEY_LOCAL_MACHINE
Key: SYSTEM\CurrentControlSet\Services\EventLog\Application
Name: RestrictGuestAccess
Type: REG_DWORD
Value: 1 Restrict access to Application log

But if you really want to get fancy, you can simply add the following entry into the BizTalk Server Operators group in Active Directory and restrict everyone else: (click to enlarge)

There is the issue of the behavior of the mapping process creates invalid XML.

The input instance looks like this:

<xml>
 <loopA />
 <loopB />
 <loopA />
 <loopB />
</xml>

However, when using the mapper, you create your output and it ends up looking like this:

<xml>
 <loopA />
 <loopA />
 <loopB />
 <loopB />
</xml>

How to get this to work (only in R2) is to open up the btm file and change the following attribute from its default value of No to Yes in the mapsource element

PreserveSequenceOrder="No"

<mapsource Name="BizTalk Map" BizTalkServerMapperTool_Version="2.0" Version="2" XRange="100" YRange="420" OmitXmlDeclaration="Yes" TreatElementsAsRecords="No" OptimizeValueMapping="No" GenerateDefaultFixedNodes="Yes" PreserveSequenceOrder="No" CopyPIs="No" method="xml" xmlVersion="1.0" IgnoreNamespacesForLinks="Yes">

Here is the new code:

<mapsource Name="BizTalk Map" BizTalkServerMapperTool_Version="2.0" Version="2" XRange="100" YRange="420" OmitXmlDeclaration="Yes" TreatElementsAsRecords="No" OptimizeValueMapping="No" GenerateDefaultFixedNodes="Yes" PreserveSequenceOrder="Yes" CopyPIs="No" method="xml" xmlVersion="1.0" IgnoreNamespacesForLinks="Yes">

I have just installed an online support link on the right of the blog. Seeing how it works. Try it out!

I have just installed an online support link on the left of the blog. Seeing how it works. Try it out!

To extend what I wrote about for the Encoding for HL7 messages post. There are two values that are used for the Message Encoding context properties. It isn't obvious but follow the logic:

According to Extended Encoding Support page in the Note section of step 5 you can only choose Western European or UTF8, so if I look at the Encoding Class documentation, the table below are the values you can use:

In a follow up to my initial discovery of the new context property (largely undocumented as far as I can tell) called BTAHL7Schemas.MessageEncoding we were converting a working process from 2004 to 2006 R2 and we started getting this error:

Event Type: Error
Event Source: BizTalk Server 2006
Event Category: BizTalk Server 2006
Event ID: 5720
Date: Date
Time: Time
User: N/A
Computer: ComputerName
Description:
There was a failure executing the send pipeline: "BTAHL72XPipelines.BTAHL72XSendPipeline, BTAHL72XPipelines, Version=1.3.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" Source: "BTAHL7 2.X Assembler" Send Port: "MLLPSendPort" URI: "localhost:11000" Reason: Object reference not set to an instance of an object.

So we put the MessageEncoding in and then started getting the following error:

Event Type: Error
Event Source: BizTalk Server 2006
Event Category: BizTalk Server 2006
Event ID: 5720
Date: Date
Time: Time
User: N/A
Computer: ComputerName
Description:
There was a failure executing the send pipeline: "BTAHL72XPipelines.BTAHL72XSendPipeline, BTAHL72XPipelines, Version=1.3.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" Source: "BTAHL7 2.X Assembler" Send Port: "HL7_SP" URI: "C:\Labs\HL7_Outbound\%MessageID%.txt" Reason: Specified cast is not valid.

Low and behold, hot fix 947261 that deals with these two error message. If you check version the following file Microsoft.solutions.btahl7.hl72fasm.dll, ours was at 2.0.302.0

Another one of the gotchas (yes, my blog seems to be full of them) is the 'association' of the acknowledgment with the original message that  is generated from the BTAHL7 receive pipeline. In the BTAHL7 Configuration explorer, by default the Acknowledgment settings are as follows:

Notice that the MSH15 and MSH16 overrides are set to AL- Always generate the System and Application Acknowledgment, and that below the Route ACK to send pipeline on request-response receive port is checked. This means that the default behavior of the pipeline is to generate two acknowledgments, and route one of them back along the receive location's send pipeline. What is it supposed to do with the second acknowledgment that it created?

That is where the error message is generated: Event ID 5778: The Messaging engine failed to process a message submitted by adapter: MLLP Source URL:{whaver server:port} Details: The published message could not be routed because no subscribers were found. This error occurs if the subscribing orchestration or send port has not been enlisted, or if some of the message properties necessary for subscription evaluation have not been promoted.

The resolution: a simple solution is to create a send port for this data and drop it off as a file.

Not quite the elegant solution that I want to deal with, as now there is a maintenance aspect of making sure that this folder is purged.

Let's just not create the application acknowledgment. You do this by changing the parties definition from Original to Enhanced mode and turning the application acknowledgment to NE (never).

Is anyone going there this year?

Let me know.

I was helping a friend with a HL7 question about how to change the MSH1, and I decided to make sure that my syntax was correct so I looked at the help and

BAM

I noticed that there is a new context property

DoorbellFinalMessage(BTAHL7Schemas.MessageEncoding) = 65001;

A question was recently posted on the forums on how you can use the default pipelines in multiple applications.

I actually was stumped when I started testing the beta of BizTalk 2006 also.

The steps are pretty easy:

In your new application, right click the application and in the context menu click properties

In the properties click on references and Add BizTalk Application 1 (which is where the pipeline is installed to)

Now you can access the pipeline within your own application.

Stephen Thomas hit me up to write an article for his newsletter (if you aren't automatically getting it, you should sign up for it here).

I wrote about how to automatically uninstall a BAM definition as part of the MSI uninstall. You can read about it here: http://www.biztalkgurus.com/newsletter/TheBizTalker-Volume-11.html

The one dilemma I have is that I cannot seem to figure out how to get a batch script to resolve the installation path by using this code in a batch script:

for /f "skip=2 tokens=2,*" %%a in ('reg query "HKLM\SOFTWARE\Microsoft\BizTalk Server\3.0" /v "InstallPath"') do SET BTSbm="%%~bTracking\bm.exe remove-all -DefinitionFile:"

I am looking into it and will update this blog entry when I finally do figure it out.

If you are installing the HL7 accelerator on a 64 bit development environment, make sure that you install both BizTalk 2006 R2 and the HL7 components in the \Program Files (x86)\ folder.

If you don't the HL7 projects (BTAHL72X Common, ...) will not show up as available as BizTalk projects. You will still be able to create HL7 schemas using the new button'd dialog box, but the underlying projects are not available.

One of the most noticeable differences between the 2006 and the 2006 R2  is the dialog on creating new schemas. Here is what it looks like in 2006

I never quite got that pressing Finish created the schema, and Cancel closed the dialog box. In my meetings with the developers, I asked if that could be changed.

Viola!

I have updated my custom trigger message entry here since a friend tried unsuccessfully to follow my instructions.

I have added a few more pictures and have noted some gotcha's.

If you are interested in learning about the RHIO HIE space, my business partner Mark Singh keeps an active blog about the things that are going on in the physician's world.

The company site is here and his blog is here.

Go take a look and let me know what you think!

Again, what I know, I think everyone else knows. In this case, a question on the forum raised my level of awareness  of things that I take for granted.

A question was asked how to change table values, where to look at what is there, and how to change them.

I will start by looking at the ORM_O01 message and we are going to be looking at the Gender element (yes I could use the real word, but I don't want to end up on a blocked site list!)

By clicking on the PID_8 in the properties you see the information you need in the Base Data Type

So we go to the tablevalues_{version}.xsd and look for that table (fortunately we only have to look at the first one!) and click on Enumeration:

Now you can add/remove the values that you need.

Note: The thing you need to make sure is that you will need to redeploy all of the schemas to make it work again.

This is what I would like to see in HL7 v.Next. This entry I hope serves to be launch pad for other entries. I have spent a great deal of time thinking over the following guideline in the HL7 standards, which are not followed in the BTAHL7 accelerator. 

Ignore segments, fields, components, subcomponents, and extra repetitions of a field that are present but were not expected.

Let's take a look at the current situation. You have all of these fields that are defined in the schema as required, and they are at various levels in the schema.

Here there is a obviously wrong implementation of the standard. If I don't care about this, why is this element required, and why might it have to match a pattern and possibly have an enumeration that it needs to match against?

What I would like to do is four things initially:

1. When I start up a new project where I am going to need to create a new trigger message for a data source, I don't want to have to hack into the schemas to change the target namespace for the base schemas (datatypes, segments, and tablevalues). Why can't there be a way from within the schema to define a party that this message is defined to, and it have it derive a new definition off of the base schemas for that party? In the BTAHL7 Configuration explorer, it will automatically make the targetnamespace coincide with the party name. Something that looks like this:
   
   which would automatically change the Target Namespace to this when I choose the Partner URI in the schema
   
   
2. I would like to open up the schema and be able to do a 'Select All' and then with the CTRL button go and de-select the fields I care about, and with the right click make all of the fields optional and override the data type to ST or TX with no enumeration or patterns.
   
3. I would like to be able to see the enumeration lists for the base schemas even though I am looking at an inherited schema. When I click on PID 2.5.0 in the ORM^O01 schema, because it has imported that definition, I don't see the enumeration. I have to go and open up the tablevalues schema and search thru it to change those values to have to recompile and redeploy. Why can't I first see the values and second override the existing values with values I know are going to come in, sometimes less, sometimes more values.
    
   
4. The last thing is that every segment, sub element should have the optional Trailer to consume additional data that might be there:
   

It should be pretty easy to accomplish, shouldn't it?

This is just something that I thought of this week while at MS-HUG. I do not doubt that there are many issues with my theory, but those three things would make implementing HL7 with BizTalk SO MUCH easier and bring Microsoft's HL7 component much closer to the standard.

Send me an email thru my contacts page if you have more ideas.

Well, unlike day 1, I was unable to dedicate a whole lot of time to the convention.

The only thing that I can attest to is that my discussion with Mark Singh brought some interesting ideas for the components of SEMRHIO

I also met with Eric Battalio again where I was grilled for more about an hour and a half. "Hands around my neck squeezing hard and forcing me to tell how documentation can be better." Or something like that :)

Actually it was a great discussion, we discussed things that I did not like about the documentation, above and beyond my Documentation Rant. Hopefully something will come out of that, maybe some enhancements to the core product, some changes to the core MSDN site, and other things. I hope that documentation will no longer be a second rate citizen in the eyes of Microsoft. My feelings that if there is better documentation (some documentation is better than others granted), but how we, the unenlightened, actually interact with the documentation that more resources of MS could be dedicated to making better software and not so many resources dedicated to supporting it.

I once again brought up that I use Google almost exclusively compared to Live Search because of the hit ratio.

Wanted to recap the list of things I walked away with today from the conference at MS-HUG:

Keynote address

Bill Crounse, MD, Director Worldwide Health Microsoft Corporation

Well, I got to listen to Bill (no not that Bill) and it was interesting, I learned today that CMS will significantly curtail the payments of procedures for in hospital infections. This will have rippling effects for the providers, and will spurn the pay for performance initiatives and quality of care initiatives as hospitals will now have to foot the bill for those issues.

Winning Trust, Minimizing IT Resources: Key to Forming RHIOs

Mark Singh President Clinicore Solutions

Kathleen Sullivan CEO Salient Health Solutions/SEMRHIO Consultant

Being the architect I am very much aware of what is going on (I had better!), however it was interesting because most of the questions were focused on the business model, the revenue generation process. This is the next phase in our process, but having a working process in place I feel is the key to success as we now have 3 hospitals up with minimal IT resources.

Reference Implementations for Healthcare: HCE and iHE XDS.b (sorry the link does not work as of yet)

Roberto Ruggeri Senior Technical Strategist for Microsoft Worldwide Health Microsoft Corporation

Randy Fusco Technical Strategist for Microsoft US Health Providers Microsoft Corporation

This was a great presentation where I learned about how much Microsoft is looking for us to help. For those who write blogs and to those who read blogs (you know who you are) it was a call to action! Yes you!

On http://solshare.net you can download the code and start to try it out and using their forum, you can submit bugs, ask for enhancements, etc.

I asked if there was a list of known issues, in which Roberto told me that no, this was a feature, that we get to start our own list and he will see if we really know what we are doing! (j/k) Seriously, he said that there is no existing list of issues, so I would assume that you can post those known issues/fixes directly on the solshare.net site.

Lunch

I went out to eat with Eric Battalio, Muhammad, and Liza from the documentation team, and wow, I learned a lot! The first and most important thing I learned is that those comments that you can leave on the documentation site actually go to someone. My suggestion is that if you want to start a dialog with the documentation team (which by the way, they would LOVE), it would be nice to have some type of post submission question asking if you (the feedbacker) would like to be contacted by Microsoft and what email they can contact you.

In the interim, just put your email address in the comments so they can contact you directly, I assume that there will be something in the works addressing this issue. Below is an example of something I have had on the back of my mind recently. 

Andrew Kirby Director of Solutions Development Center Microsoft Corporation, UK

Like the CHF, this is very much a work in progress and since it is very early on in the development, it is more of a guideline and they need as much feedback as possible. However there are no immediate plans for a reference implementation to build off of other than the first implementation in England.

What’s New in BizTalk Server 2006 R2

Stuart Landrum Program Manager, BizTalk, Microsoft Corporation

Well, it was the first time I had met Stuart face to face, had been on some conference calls with him, so it was good to meet him. He explained the new features with the HL7 accelerator (now officially branded 2.0). There was very little pictures of the product so the discussion went rather quickly and ended early with few questions, I had to ask a few questions to get people in the asking mood, but it did not go much further than that.

Closing Keynote: Using IT to affect efficiency and safety changes in the erioperative process

Conclusion 

Paul St. Jacques, MD Associate professor and Director of Perioperative informatics Vanderbilt University

This doctor was actually kind of funny. I am not sure if he meant to be, but if it was not for his phones (yes plural) that added a little static to his voice, or his low ranking on the pay for performance ratings that he refuse to show on the list of many doctors ranked (it was too small to see), it was a good presentation that ended on a high note, even though he essentially had to ask us to laugh, even though there wasn't really anything funny about his presentation. He did use one of my all time favorite phrases, while describing the hospital system that he works in, he described it as a level 1 trauma facility that is "soup-to-nuts" - where did that come from and should we be saying that phrase as professionals?

The interesting thing that I thought that he was going to joke about was his workflow from a patient pre-registration process through surgery completion. look closely for anything wrong with this picture

nope, not the fact that they are using SQL Server 2003 (must be a special version of SQL Server that I am not aware of) 

The real problem with this picture is that there is a little girl that starts the surgery process, and when the workflow is complete, the girl is actually Henry Kissinger! Either that process took a really long time (and something major wrong happened), or this is a plastic surgeon's billing dream come true!

Talk to you all tomorrow!

In the documentation for HL7 there is an interesting chapter called Using Dynamic Data Validation and it is located here. To explain how this is to be done you first have to go into the HL7 configuration explorer and disable the data validation.

This will allow data to arrive without being checked against its Enumeration:

This however will not disable the structure validation that is the pain point of most HL7 implementations.

Once data is structurally validated, then the message arrives into the orchestration you then start to use the Business Rules engine to validate the data

This is the starting point of how you enable Dynamic Data Validation using the HL7 accelerator

Let me know...