White Papers

White Papers September 2011

Picture of health: New XDS-I registry gives Quebec physicians access to digital medical images

Daniel Cote, Yvan Foster and Jody Lynch


Better use of healthcare resources and access to the right information from anywhere at any time by the right person has been the “Holy Grail” quest for many healthcare organizations throughout Canada.

Sharing of medical imaging data to optimize the use of scarce resources and increase patient safety has been a goal of the Canadian Healthcare system since the introduction of Digital Medical Imaging.  The millions of dollars Canada Health Infoway invested into the system to digitize Medical Imaging has certainly moved Canada much closer to this goal.  However, even with large regional implementations of Medical Imaging Systems,  there is still a strong demand to continue to expand the sharing and interoperability of medical imaging systems to a provincial level and beyond so the system can operate as one provincial system.   This is a daunting task, complicated by a plethora of standards and competing approaches.  However, by turning to frameworks and standards that have been supported internationally, tested through a rigorous annual process, and actually implemented by commercial vendors the goal of one provincial Medical Imaging system can be achieved.


“Continuity of care”, “collaboration”  and “information sharing” are words that have been used over and over again for the past 10 years in healthcare.  It is an acknowledged fact the healthcare industry is lagging well behind other industries when it comes to Information Technology (IT) adoption and that healthcare reform can not happen without increased use of IT.
The Federal Government decided to create Canada Health Infoway  in 2001, a not for profit organization tha

has been acting as a strategic investor of funds for the development of Canada’s Electronic Health Record  (EHR)  infrastructure. Besides managing 2.1B dollars, Infoway was also responsible for providing a technology blueprint in order to insure inter-Province compatibility and solution reusability. On top of that Infoway prescribed communication and technology standards that would support secure and accurate sharing of patient information. While every Province could propose a different approach in EHR implementation, they had to align with Infoway’s componentized acquisition approach, where each new component was a new building block which had to build upon existing investments (EMPI, HIAL, etc…)

Medical imaging is one of the clinical domains identified by Infoway as part the EHR infrastructure. However most Provinces’  healthcare organizations quickly realized an architecture blueprint and communication standards were only part of the solution to sharing of medical imaging data. Putting all the pieces together to provide a sound structure that could be adapted and adjusted to the new Canadian healthcare business requirements was not trivial. Moreover the EHR paradigm was a new concept and there were no successful “recipes” that could be used by Canadians as an implementation blueprint.

One of the options was to use the model defined by the IHE organization (Integrating the Healthcare Enterprise). IHE has been very active in promoting the use of existing healthcare standards to build a workflow that would support processes in the healthcare organization, especially in Radiology where frameworks have been in place for 15 years and supported by a robust international community of contributors and supporters that use a very successful change process to evolve the framework. In reality there were few other alternatives that looked as compelling as the IHE framework and for that reason Infoway decided to promote adoption of the IHE framework for the EHR imaging domain.

The Province of Quebec EHR medical imaging team has chosen to apply the IHE framework model as prescribed by Infoway Canada by developing the first Cross Enterprise Document Sharing for Imaging (XDS-I ) compliant infrastructure in Canada.

The requirements

Medical Imaging Data is one of the core clinical information needs which must be shared by healthcare providers in the context of a jurisdiction or territory. The primary goals are to:

  • Combine the imaging data with other clinical information for a complete assessment of the patient situation
  • Provide better diagnostics based on historical data
  • Share workload among the radiologist community in order to optimize resource utilization especially in remote regions
  • Avoid exam duplication that generates unnecessary costs

There were two types of users considered by Quebec’s EHR Imaging team:  the EHR users which are primarily family physicians; and the radiologist community. However the final solution needed to cover the needs for all users with the same infrastructure.
The family physicians, emergency physicians and even some specialists needed to share and access images and diagnostic reports to help them determine a course of action for their patients. The EHR imaging domain was meant to act as a large virtual medical imaging archive to which all users (with access rights) could access the right patient information at the right time from any location.
In another context, the radiologist community required access to all medical images to share the workload or to have access to prior medical imaging exams that could be relevant for diagnostic purposes. 

The solution also needed to be scalable to support the large amount of storage that would be required to store all medical images that would be generated by the different Province of Quebec hospitals and private clinics. It was estimated by the customer 10 million imaging exams would be generated each year producing close to 70 million images. Other very important considerations like the use of a generic viewer, consent compliance, audit capability, data security and indexing capacity were critical in the design of the final solution.

The solution

While Infoway promoted adoption of the EHR blueprint and communication standards which provided a high level of guidelines, the details of the final solution where left to the individual Provinces except for the imaging domain for which Infoway strongly suggested the adoption of the IHE model as shown in the following picture. The IHE model specifies in the form of profiles and transactions how to use the different standards to enable workflow related to a given clinical domain within an enterprise.

At a very high level, the components of an XDS solution are easy to understand

  • Patient Identity Source: provides the information to uniquely identify every patient
  • Document source: generates the data that will be stored in the Document repository. To simplify the XDS transaction model in the context of a medical imaging domain we will accept that, the PACS will act as the source. However in reality the PACS and DI-r do not communicate with XDS-I transactions and the DI-r plays the role of both Document Source and Repository.
  • Document Repository: will store the images and other documents. In the context of medical imaging domain we call it Digital Imaging repository DI-r which basically is PACS technology expanded to support image archiving. For each new document stored, the DI-r will send metadata to the registry.
  • Document registry: will store the metadata generated by the DI-r and attach it to the proper patient unique identifier. The metadata are well defined piece of information such as the patient identifier,,the date, the type of exam, the location of the image in the DI-r, etc…
  • Document consumer: In the context of a medical imaging domain the consumer will be any EHR image viewer or radiologist diagnostic viewer. The consumer will first query the registry for all available documents by extracting the metadata. Then the user will choose in the list which document is of interest and will select the required document. The viewer will use the location information of the document and extract it from the DI-r

The Province of Quebec chose to apply the IHE model to support the two communities of users. It was decided the Province of Quebec territory would be divided in 4 regions that would each be supported by the 4 University Hospitals ( CHU Sherbrooke, CHU Laval, CHU McGill and CHU Montreal) who would also act as advisor to the solution. Finally three Digital Imaging repositories (DI-r) would support the entire Province of Quebec. Laval and Sherbrooke would be separated while McGill and Montreal would be combined. All medical images would be stored locally and a copy would also be sent to one of the 3 DI-r for sharing and back-up purposes.

XDS-I registry role

The XDS-I registry acts as an index for all queries and stores the meta data that is sent from the DI-r. The richness of the meta data will greatly influence the “intelligence” of the system and is instrumental to generate the workflow required to support the users processes. The XDS-I registry became a cornerstone of the solution and a lot of development has been done to provide additional features that were required to deal with some issues related to the complexity of the requirements.
Consent is one of the issues that required long hours of discussion. While the EHR infrastructure

rovided a consent application it did not cover all the use cases proposed by the customer. It was decided the solution would be to use pre-processing connectors available from the registry to add a second layer of consent validation to the process to deal with exceptions when a physician works from different hospitals

The indexing role of the registry has been also carefully studied to save time for radiologists when they are looking for relevant priors that might be useful in analyzing a new image. It has been decided the anatomic information (body part) would be part of the metadata stored in the registry. With this information, the system can specify which exam is relevant and then reduce the list of documents that will be presented to the radiologist.
As of December 2010 Quebec’s Medical Imaging

omain is operational and it is the first to implement a fully compliant XDS-I structure. The success of this project is closely related to the involvement of the medical imaging users, the customer’s subject matter expert  and the decision to use models that were proven and available. A lot of energy has been spent on developing used cases with the users and the approach was to adapt the technology to the requirements, not the opposite.


While the previous diagram shows very simple transactions between the components, the final solution was carved out of thorough analysis, workshops and collaboration between the vendors and the customer.

What makes a big difference in the final outcome of this project was the fact the customer and implementation teams had the IHE framework they could use as a basis and that both teams could count on skilled integration resources.

The customer had carefully laid out the preliminary requirements , but a methodology was put in place to deal with gaps that consisted in simple but efficient steps.

  1. Collect the requirements and continuously performing a gap analysis against IHE frameworks to determine if the requirements were implementable with existing standards and tools.
  2. If there were gaps discuss the approach to either a) changing requirement b) building a custom extension c) pushing a change to the standard/framework through the system.


In Canada, Infoway laid the foundation by providing Provinces with a blueprint and a plethora of standards. Despite that, many projects wandered through the I/T technology sphere until they realized late in the process that they had to go back to the drawing board and review the choice of components  and solution design which eventually lead to “proprietary” solutions.

The XDS framework promoted by IHE organization clearly provides an additional layer of instructions on how to assemble the components based on the analysis of many architects and healthcare subject matter experts from the healthcare vendor community who have worked together to define the XDS profiles. There is also a very strict process called “connectathon” during which vendors can validate their compliance to the IHE profiles. This 1 week activity is performed every year in the US, Europe and Japan. While XDS is still a work in progress it was certainly a key factor in the success of Quebec’s medical imaging project because this framework has broad support from vendors.

Let’s face it, when we started talking about the EHR back in 1990 we had only a vague idea of the challenges ahead of us. There are two very important stakeholders that need to be involved in such a complex project --  the end user team and the I/T team, in order of importance. We have seen too many projects where I/T architects have attempted to build new solutions based on their own interpretation of the users’ needs, spending more time analyzing the different technology options than discussing the value for the end user. The success of a project as complex as an EHR is closely related to the detailed analysis of users’ needs. Assuming this first step has been properly covered, the I/T team then will have to decide on architecture, standards and undoubtedly a framework .

We must accept that any standard or framework is only going to partially meet requirements of a project this big and complex. For that reason there has to be a team and process in place with the ability to adapt, extend, and or change the selected frameworks and standards to perfectly fit the requirements.

Sharing clinical information in a healthcare enterprise is a very wide statement. Sharing information in a meaningful way is much tougher to accomplish because it requires questioning how the process are performed today and what should be done to support processes that will be dictated by the new healthcare enterprise business model. These questions should always be covered at the very beginning of the project keeping in mind that the infrastructure should support the processes, not the opposite. Building “use cases” is a very good way to provide guidelines to the implementation team.


The thought leadership of the customer was a key driver to the success of Quebec’s EHR imaging Healthcare domain as they were able to clearly express the detailed processes that needed to be supported. The decision to use of the XDS-I framework as prescribed by Infoway avoided wandering through major architecture decisions providing the implementation team with more time to concentrate on how to fill gaps. In the end the solution was successfully delivered on time and within budget. So far, the Quebec imaging EHR solution provides complete sharing of information using multiple DI-r. Eventually  the infrastructure in place will be able to connect to any other XDS compliant components provided by the multiple vendors that adopted IHE.

As a final thought, it is clear that XDS could be applied to other EHR domain and help accelerate EHR implementation in Canada by avoiding proprietary or hybrid solutions! IHE is a worldwide initiative and and other countries including the Europe medical community are specifying the IHE framework as a requirement for their EHR infrastructure further promoting the value of this model.

About the Author(s)

Daniel Cote is the lead architect responsible for the development of IBM’s XDS-I solution, coted@ca.ibm.com
Yvan Foster is the business development manager for IBM’s healthcare consulting services, yfoster@ca.ibm.com
Jody Lynch is IBM’s Canadian healthcare solution architect, jlynch@ca.ibm.com


Be the first to comment on this!

Note: Please enter a display name. Your email address will not be publically displayed