A new architecture for EHRs?

Amidst the efforts to develop the elusive electronic health record (EHR) and to modernise our health care systems are many technical challenges that need to be addressed.  One significant challenge is integrating information from a variety of difference sources (i.e., lab systems, pharmacy systems, scheduling systems, directories, nursing and physician notes, etc) and to present this information in a coherent, user friendly fashion.  In the past, I’ve participated in some of the technical discussions regarding potential solutions and let me tell you that there are no easy solutions.  We’re talking about massive amounts of data needing to be transferred between different points.  Then we also need to talk about how to take this data and make the information accessible to the health care provider (and ultimately the patient) at the time and place needed.

From my understanding, much of this integration work has gone along the path of developing systems that function as "dashboards" by collecting and aggregating information from various sources.  I discussed this concept with Khaled El-Emam in the past and he indicated that database technology isn’t yet sophisticated enough to do massive real-time push updates.  It can be done on a small scale, but the model doesn’t scale.  I noted this point when the University Health Network (UHN) started moving towards an enterprise clinical data ware-house.  The original thought was to update all of the hospitals clinical systems in real time as new patient data was created.  Unfortunately, practical limitations forced the real-time aspect to be abandoned.  I wasn’t privy to the final solution, but I’m told that the solution that was agreed upon was batch updates nightly for things that were time dependent and weekly (or even monthly) updates for other types of data.  We’re talking about a single hospital corporation having difficulty updating its own systems – granted, UHN is one of Canada’s largest hospital corporations and has three main hospital sites with over 10,000 staff, so we’re not talking about a small organization.

I was notified of a proposed alternative method of integrating information within an electronic health record by Vitaly Latush.  He outlines:

"an alternative approach to implementation of a nation-wide easily accessible electronic health record solution based upon the "publish-discover" paradigm successfully used on the global scale to manage immense volumes of non-integrated information available through the Internet"

His main argument against the current architecture/model is that integrating data from all of the different data sources is far too complex because it is based on exchanging data based upon standardized rules.  Latush suggests that a model based on how the Internet currently operates is far more efficient and ultimately more effective.

In this alternative model, he suggests that data sources, be it physicians in an office, labs, or hospitals should only be responsible for "announcing" that new data is available and then providing access to this information.  I would assume that some sort of RSS technology (or a derivative) would be used.  Instead of waiting for some program or system to take this data and integrate it into a dashboard, users would search for information, like they do on Google or any other type of search engine.  Health information would be aggregated by some sort of unique patient identifiers or other keywords.

Latush suggests the following advantages of this model over the existing architecture:

  1. Minimum integration effort – searching for information is now the main function
  2. Based on *proven* Internet principles
  3. Does not require new technologies or skills
  4. Can be built on top of existing health information systems very easily
  5. Facilitates incremental incremental improvements of the functional EHR
  6. Suited for unstructured data manipulation
  7. Cost-efficient in terms of up front cost and maintenance
  8. Self-organizing

 After reading the white paper, I have to admit that the idea is interesting and deserves some investigation.  Here are some thoughts I have about this idea:

  • This alternative architecture for the EHR requires a secure "health information ‘Internet’ with restricted access".  In Ontario, the Smart Systems for Health Agency (SSHA) has been doing just that – developing a health specific network with only authorized access.  I’m not sure if other jurisdictions are doing anything similar or would even be interested in such a development.
  • Access control – By switching a publish-search-discover type model, I wonder how access will be controlled.  Will audit-logs be automatically generated to track who access information?  This issue may need to be further addressed given the privacy legislation surrounding health information.  I would assume that local agents would be responsible for identifying what types of information will be "published" and made available.
  • What about the patient? – What role will the patient have in this new model?  I can understand the need to focus on getting health professionals up and running first, but there doesn’t seem to be any mention (or allowance) for patient involvement.  In the US, there is talk about personal health records that patients maintain and control.  Will this model allow for the two systems to interact?
  • How will this "EHR discovery center" develop – In my understanding of this concept, we will need a health specific "Google" that functions to identify new information and then present information to the searches.  As a theoretical concept, I can see how it works, but I think we need some sort demonstration of this concept in practice.
  • The model seems very flexible – This Internet-based model seems very flexible.  Much like the Internet today, the device(s) used to access the information will be responsible for the presentation of the data.  As long as the data is stored in standardized ways, users should be able to use a smartphone, computer, or other device to access the information and have it presented in a format of their choosing.

Again, this idea sounds very interesting.  From my understanding of things here in Ontario, I believe that some of these concepts may be adopted already or at the very least being discussed.  As the health system moves toward regionalized and integrated models, the need to share information becomes greater.  Regardless of your thoughts on this specific proposal, I think the value is in the discussion and generation of new ideas.  As well all know, health care is important but sometimes takes a back-seat to more "sexy" things like new technologies or recent events.

I highly recommend you read Vitaly Latush’s blog post on this topic or read his white paper, titled "EHR 2.0 – new Electronic Record concept".

Powered by Qumana






4 responses to “A new architecture for EHRs?”

  1. […] to the health care provider (and ultimately the patient) at the time and place needed.” Article Hans Oh’s eHealth Blog, 5 February […]

  2. […] A new architecture for EHRs? I can understand the need to focus on getting health professionals up and running first, but there doesn’t seem to be any mention (or allowance) for patient involvement. In the US, there is talk about personal health records that … […]

  3. Yaron Derman Avatar
    Yaron Derman


    how does the ‘publish and announce’ model align to the IHE XDS architecture?

  4. Hans Avatar

    Hi Yaron,

    By IHE, I hope you mean “Integrating the Health Enteprise” effort (www.ihe.net).

    Honestly, I didn’t really connect to two concepts together. From my understanding of IHE, the effort is based on a series of use-cases and business processes (i.e., clinical work flows). Once you have a specific clinical process in mind, say cardiac care or imaging, then the collaborative group’s effort goes to examining all of the data transactions and system interfaces to ensure a high degree of certainty that the hardware tested *works* and can be recommended for use by others.

    I wasn’t too familiar with XDS (Cross Enterprise Document Sharing), so I looked it up on Google and came across this very interesting link: http://www.itl.nist.gov/div897/docs/XDS.html. From the sounds of this description, XDS sounds like it would be related to Latush’s idea about “publish and announce”.

    In my opinion, the “publish and announce” model is more of a conceptual model of how information could be shared within the context of a health system with multiple data sources and multiple access points. The IHE XDS effort seems to be focused on the STANDARDS. We have to remember that Latush’s idea is just that – a concept that needs many more details. However, it is one that seems interesting.

    To answer your question, I wouldn’t say that the two are related necessarily, but definitely may share some similar ideas. IHE XDS seems to be focused on the development and promotion of standards for data interchange. The “publish and announce” concept is (for today) a thought experiment of applying the model used in the Internet to health care.

    Please note that my technical background has limits. While I do understand the technologies at a conceptual level, I am by no means an expert. Someone with either more expertise or experience on these topics (IHE XDS) would be able to give you a more definitive answer.

    Thanks for dropping by. Great question!