Reconciliation feature

[MatiasTech]

Hi,

Is there a reconciliation feature in dcm4chee that would check the incoming DICOM data against MWL entries during the storing of DICOM studies?

Reconciliation feature would check if the incoming data is correct and if not it would redirect the incoming data to different database to be fixed before storing it to the main database.

Best regards,

Matias

[Vrinda Nayak]

Perhaps you mean coerce attributes using MWL. See Overwrite received DICOM object's attributes from previously received HL7 Order Messages in https://github.com/dcm4che/dcm4chee-arc-light/wiki/Coerce-received-DICOM-objects-attributes-by-XSLT#use-cases

[MatiasTech]

Hi,

That is not the one that I am looking for.

I'm looking for some kind of validation when the DICOM study arrives. Preferably checking that the SUID exists in the Modality worklist + checking that the patient is valid.

Doesn't the IHE SWF and IHE PIR profiles describe some validation? These seem to be marked as supported in the DCM4CHEE IHE integration statement.

https://www.ihe.net/uploadedFiles/Documents/Radiology/IHE_RAD_TF_Vol1.pdf

I found this wiki page about validation but not sure is it related:
https://github.com/dcm4che/dcm4chee-arc-light/wiki/Validate-objects-on-storage

Br,

Matias

[Vrinda Nayak]

Yes, archive does support Patient Information Reconciliation (PIR) profile. However, the steps you've mentioned are similar to the use case described in Overwrite received DICOM object's attributes from previously received HL7 Order Messages, wherein with help of configured coercion rule archive checks for previously received MWLs with matching Study Instance UID or Accession Number and merges attributes from MWL into the incoming objects' attributes.

Whereas, the Patient Information Reconciliation (PIR) profile workflow steps are different from your defined use case. In PIR, it is required that modality sends Study UID information in MPPS messages to Order Filler or Department System Scheduler when it also stores study of patient (with temporary patient information details) to the archive. The patient information is later manually reconciled on Order Filler or Department System Scheduler and the corresponding correct patient information is sent to the archive via HL7 patient update / merge (ADT^A40) messages. Furthermore, after this patient information has been updated, Order Filler or Department System Scheduler sends HL7 Procedure Scheduled order message with the Study UID it previously received from modality in MPPS. Attached test case from IHE.

[Vrinda Nayak]

https://github.com/dcm4che/dcm4chee-arc-light/wiki/Validate-objects-on-storage is not related directly to your use case, unless your application has some external provider which shall authorize the objects being stored to the archive. This service is also independent of MWL as long as external service has the rules concerning authorization of objects' attributes.

In the use case explained in Overwrite received DICOM object's attributes from previously received HL7 Order Messages, default configured mwl2series.xsl stylesheet merges the MWL attributes (Patient Identifiers + Patient Demographics + Requested Procedure + Scheduled Procedure) on each of the incoming object's dataset. You may choose to customize this stylesheet based on your application needs.

[MatiasTech]

Hi,

Thank you for the quick responses. Here are some clarifications on what I meant and what is the use case.

What I'm looking for is to have a list of studies with problems that need manual action from an admin:

• study created manually and it has bad patientId or name info or ac number / SUID doesn't match any MWL entry

  • The UI would then have some list the admin user could look into to see that these studies are missing orders and someone needs to create them or correct some data etc.

Attribute Coercions:

• I think this doesn't match my use case fully since Attribute Coercion would require the study to exist in the MWL, wouldnt it? What if it is an emergency trauma case which was created without an MWL entry filling the info manually. Would attribute coercion mark the study as "unreconciled" or something like that if it doesn't find the entry in the MWL?

• Or perhaps have I misunderstood something - could the attribute coercion rule check patientId doesn't exist in MWL or SUID doesn't exist in MWL and therefore the study is directed to a second DICOM database?

  • Then the admin user could check that this second dicom database has new objects which didn't pass the coercion rules and need manual action?

Patient Information Reconciliation (PIR):

• It sounds like this PIR flow could work for the above emergency use case. "Order Filler or Department System Scheduler sends HL7 Procedure Scheduled order message with the Study UID it previously received from modality in MPPS."

  • Figure 4-4-3 shows Image manager (DCM4CHEE archive) sending RAD-7 (MPPS) to Order filler (RIS I guess)

  • Can DCM4CHEE be configured to do this, any pointers on how?

• So then it kind of depends on the RIS is it able to receive the MPPS and does it trigger someone to take manual action to create the order message afterwards with the SUID it received in the MPPS

Best regards,

Matias

[Vrinda Nayak]

My comments inline

What I'm looking for is to have a list of studies with problems that need manual action from an admin:

• study created manually and it has bad patientId or name info or ac number / SUID doesn't match any MWL entry

  • The UI would then have some list the admin user could look into to see that these studies are missing orders and someone needs to create them or correct some data etc.

Using archive UI, in Navigation - Studies, you can search for studies not connected to MWL by selecting Unscheduled in Unscheduled Studies dropdown (available when you expand More block)

Attribute Coercions:

• I think this doesn't match my use case fully since Attribute Coercion would require the study to exist in the MWL, wouldnt it? What if it is an emergency trauma case which was created without an MWL entry filling the info manually. Would attribute coercion mark the study as "unreconciled" or something like that if it doesn't find the entry in the MWL?

"an emergency trauma case which was created without an MWL entry filling the info manually" case exactly matches the case explained in Patient Information Reconciliation (PIR)  profile - 4.4.5 Case 5: Image Acquisition Completed Without Scheduling at Department System Scheduler/Order Filler. 

"Would attribute coercion mark the study as "unreconciled" or something like that if it doesn't find the entry in the MWL?" - No! If there was no MWL created and sent to the archive, there is no MWL record in the archive's database to search for. This is independent of whether an attribute coercion rule is configured or not. For a given Study UID / Accession Number

- if there is no attribute coercion configured to search for MWL with Study UID / Accession Number

OR

- if an attribute coercion is configured to search for MWL (in archive's DB) with Study UID / Accession Number, but a corresponding MWL entry is not found (in archive's DB) - emergency trauma case -

then such studies' objects' dataset do not contain Request Attributes Sequence (0040,0275).

Note : Attributes of Request Attributes Sequence are copied from MWL attributes for normal scheduled workflows.

• Or perhaps have I misunderstood something - could the attribute coercion rule check patientId doesn't exist in MWL or SUID doesn't exist in MWL and therefore the study is directed to a second DICOM database?

  • Then the admin user could check that this second dicom database has new objects which didn't pass the coercion rules and need manual action?

No! If there was no MWL created and sent to the archive, there is no MWL record in the archive's database to search for.

Patient Information Reconciliation (PIR):

• It sounds like this PIR flow could work for the above emergency use case. "Order Filler or Department System Scheduler sends HL7 Procedure Scheduled order message with the Study UID it previously received from modality in MPPS."

  • Figure 4-4-3 shows Image manager (DCM4CHEE archive) sending RAD-7 (MPPS) to Order filler (RIS I guess)

IHE Patient Information Reconciliation (PIR) test case refers to 4.4.5 Case 5: Image Acquisition Completed Without Scheduling at Department System Scheduler/Order Filler. Figure 4.4-5 slightly differs with its description. Figure 4.4.5 and its description differ slightly with IHE Patient Information Reconciliation PIR test case step showing Modality sending MPPS to DSS / OF only.

"A temporary ID and name are entered by the technologist at the Modality and conveyed to the Department System Scheduler/Order Filler and to the Image Manager."

whereas, the figure shows modality sending RAD-7 MPPS to Image Manager (archive) and archive further sending the same MPPS to DSS / OF.

• • Can DCM4CHEE be configured to do this, any pointers on how?

If the modality in your application setup would follow the steps as shown in Figure 4.4.-5 instead of as described in the description above, you may configure MPPS Forward rules in archive to forward the received MPPS to interested MPPS destinations.

• So then it kind of depends on the RIS is it able to receive the MPPS and does it trigger someone to take manual action to create the order message afterwards with the SUID it received in the MPPS

Yes, it is a manual reconciliation step on the DSS / OF system. DSS / OF should be capable of :

- receiving MPPS either directly from modality or from archive or both.

-  manually reconcile patient information and send corresponding ADT patient update / merge messages to the archive

- identifying and using Study UID it receives from MPPS to be used further in the order message to be sent to the archive.