Open source and open standards in healthcare:
Advantage Europe

Source: HealthcareITNews.EU http://healthcareitnews.eu/
Date: 7 th March 2007
Written by Heinz Joerg Schwartz, Guest Columnist   

Europe has had a leading role in adopting and popularizing open source technologies. But open source and open standards are still rare sights in healthcare IT, both in Europe and in the United States. Why? Let's look at the status of these projects and address some of the barriers to implementation.

The success of open source technology in recent years is undeniable. Many companies use servers with the Linux operating system to run applications, relational databases like PostgreSQL and MySQL have many followers in the commercial world, and companies generate hundreds of millions of dollars with services revenue associated with open source projects.

The European Commission supports open source and open standards actively in a program called Interoperable Delivery of European eGovernment Services (IDABC). The reasons are obvious: open standards enable heterogeneous information systems to exchange data –a requirement for cooperation and services exchange among institutions in various member states.

While some local standards might allow exchange of healthcare data within a local region or nation, common standards are required to exchange data across borders and language barriers. One day it should be possible for a citizen of Europe to choose a medical provider in the Czech Republic or Spain based on medical outcomes, price, or other preferences, and transfer pertinent data that allows continuity of personal care records.

Maybe the best existing standard for electronic medical records is currently CDA2 (clinical document architecture) within HL7 version three. At a recent roundtable discussion in Finland, I was surprised to learn that many electronic medical record systems offered in that country adhere to CDA2. The adoption of standards seems to be more actively promoted in Europe than it is in the United States, where many commercial EMR systems still do not adhere to CDA2.

If commercial EMR vendors can supply CDA2 compliant data structures in Finland, why can't they in other countries in Europe, or in the United States? One answer could be that they don't have to. If customers don't ask for common, open data standards, they won't get them. Unless there was an alternative, which brings us to the second part of the topic: open source software.

While other commercial sectors are quite eager to leverage open source software in ecommerce (such as online advertising, web hosting) or high performance computing (seismic analysis, drug research), healthcare has been reluctant to support and implement. The open source project tracker of the IDABC project maintains a Web site with case studies and an open source application portfolio. Among the case studies is only one hospital in Ireland –a project that merely covers office productivity, not core medical applications. The portfolio of truly medical applications contains several projects in alpha and beta states addressing peripheral needs.

The heavyweight in the open source EMR scene in certainly VISTA, a package developed over the years by the federally funded U.S. Veterans Administration. The VA, without many followers for VISTA within the United States, now offers the code to the global healthcare IT community.

So why is there not more activity in this sector? Why is healthcare IT still characterized by proprietary standards and application software?

One reason is the sheer complexity of building a clinical applications package. SAP took years to develop a new application that is now being deployed at several European hospitals. Open source activists would have to build working groups for different medical specialties, assess workflow requirements, and develop application logic that uses data architecture compatible with CDA2 or another standard. Various other complex tasks have demonstrated that this can be done –but for some reason it isn't.

The second major barrier is regulatory requirements that call for rigor in systems specification and documentation. The dilemma is that building an EMR can hardly be justified as a research project, especially in IT, in which middleware projects are preferred to applications.

Open source EMRs won't mushroom overnight. The best bet in this regard indeed seems to be VISTA, which must be enhanced and adapted to European standards. Also, the European Commission would have to be more involved in this endeavor than they have been.

The gain from doing this would be twofold. Developing countries, within the EU and outside, could both cut the problematic cycle of proprietary software with proprietary standards and implement a powerful EMR with interoperable data standards.

However, another element of the open source economy remains missing. Healthcare IT departments are accustomed to full service vendors that implement point solutions or component systems. For open source software to be successful in healthcare, an equivalent service provider infrastructure would be required. Examples for this exist in the general IT world, such as with MySQL.

In a previous research paper I discussed in detail that services provisioning would be absolutely compatible with the open source community.

In healthcare, it is “ condition sine qua non”.