Informatics-Review > Thoughts > The Open Source Movement

The Open Source Movement 

Linda Wedemeyer, M.D.
Kaiser Permanente Medical Group, Harbor City, CA
Veterans Health Administration, Los Angeles, CA
University of California, Davis Medical Informatics Program

I became interested in what I am calling the open source movement by following some very dynamic e-mail conversations in the clinical information systems workgroup at AMIA. Supporters speak about it with almost religious fervor.  I wanted to answer several questions for myself. What is it? What is it good for? Is anybody actually using it? My review of the literature was not very satisfying, so I ended up posing the following questions to the AMIA workgroup:  
“I am a practicing ophthalmologist in Los Angeles, and a medical informatics student at UC Davis. I have followed the conversations about open source software with interest. Recently I have been reviewing the literature on the topic. One thing is not clear to me. I see lots of software available, but IS ANYBODY ACTUALLY DOING THIS? I can see that there are bits and pieces being used here and there. Has anybody set up a whole health care institution using open source software, and if so, is it working in real life? How does the quality/cost compare IN REAL LIFE to commercial software? Does hiring commercial groups to support the products, if you don't have the staffing in house, actually work?”

Thank you,
Linda Wedemeyer

At the end of this paper I will share some of the insights that I gleaned from the answers to these questions.

I was able to locate answers to my first question, what is it? The GNU Project (The Free Software Definition, 2003) was launched in 1984 to develop a complete Unix-like operating system intended to be free software. The concept is nicely described on their website, which states that “'Free software' is a matter of liberty, not price. To understand the concept, you should think of “free” as in “free speech,” not as in 'free beer.'"  

A more precise definition is found on the website for the Open Source Initiative (Open Source, 2003).Open Source Initiative (OSI) is a non-profit corporation dedicated to managing and promoting the Open Source Definition for the good of the community, specifically through the OSI Certified Open Source Software certification mark and program.” This organization provides licenses for open source software, designed to protect rather than restrict the rights of those who own the software (The Open Source Definition, 2003). An OSI license requires much more than just access to source code. Very importantly, it requires free redistribution. This doesn’t mean that you can’t charge for providing the software to others. What you cannot do is charge them for the right to redistribute it to others. They are free to sell it or give it away, and they may do this in aggregate with other software if they like. They must provide the source code itself, or easy access to it, along with any distribution of open source software. There is a concept of derived works, which means that an OSI license must allow modification of open source software, followed by distribution to others under the same terms as the original product. The license may, however, provide a means for protecting the integrity of an author’s code. This is accomplished by requiring that redistribution be in the form of source plus patch, to clarify who was responsible for writing what. There can be no discrimination in an open source license, whether it is against persons, or groups, or fields of endeavor. Whenever open source software is distributed, the same rights apply as was the case with the original product. If an application is distributed as part of another application, the same rights still apply—the license is not specific to a product. An OSI license cannot restrict the distribution of other software in conjunction with the open source application. For example, it cannot require that all software distributed on the same CD be open source. For software to be licensed as open source it must be technology neutral—it cannot require certain functions, such as a popup dialog box for establishing an agreement. 

The open source license that I have seen mentioned most often is the GNU General Public License (The GNU General Public License (GPL), 2003). Introductory statements on the OSI Website begin: "Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed. 

Preamble
The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users. This General Public License applies to most of the Free Software Foundation's software and to any other program whose authors commit to using it. (Some other Free Software Foundation software is covered by the GNU Library General Public License instead.) You can apply it to your programs, too." 

My next question, what is it good for, finds lots of answers in internet sources and e-mail conversations at AMIA, but very little in peer-reviewed medical journals. One of the most frequently mentioned advantages in AMIA e-mails is the avoidance of dependence on unstable vendors of proprietary software (The Open Source Case for Customers, 2003). Health care institutions invest enormous sums in information systems, only to find that the vendor goes out of business. This leaves the institution with a system that they cannot upgrade or maintain because there is no access to the source code. Even if the vendor stays in business the software owner is dependent upon the vendor for needed upgrades and maintenance. The vendor may be unwilling to provide what is needed without additional expense, and the software owner is unable to go elsewhere (Carnall). Security and reliability are enhanced when the source code can be inspected for bugs (Carnall). Effort is directed toward innovation, rather than toward reinventing what others have already done (Carnall). Users of an open source application do not have to be concerned about the need to learn a new user interface when it becomes necessary to purchase a new proprietary product. Proponents of open source software point out its rapid evolution, as a result of its availability to the whole world for review and improvement. They claim that the result is better quality software.  

The European Union is placing a strong emphasis on development of open source software. (Carnall). There is an organization called the Open Source Health Care Alliance (OSHCA) (The Open Source Health Care Alliance, 2003), which defines itself as “A collaborative forum to promote and facilitate open source software in human and veterinary healthcare”. They discuss numerous benefits of the open source approach to healthcare IT on their website. They feel it is key that the approach is user driven, with the focus on usability of software products. They point out the unprecedented flexibility, and rapid innovation. They value the resource sharing, which allows collaborative development effort. Software licensing fees are minimal, allowing more room in healthcare budgets for support, training, and customization. (Carnall) Developing countries can access applications affordably. The software cannot be controlled by any organization, most notably governments and large software companies. It is claimed to support business models that are innovative, stable, and sustainable. It facilitates peer review--anyone can evaluate the code.  

The openEHR Foundation (openEHR, 2002) is an international non-profit group working to bring about comprehensive, interoperable electronic health records. Their published goals are to: This group supports open source software as a means to accomplish their goals: "openEHR recognises the value of open source as a means of fostering a community of developers, and of prototyping and evaluating components in a field of innovation such as the EHR. openEHR endeavours to produce reference implementations of the information architecture as soon as is practical, and to make these available through an open source licence, in order to foster a network of demonstrator sites contributing to the evaluation process and to a global community of EHR use. openEHR will also collaborate openly with commercial software developers seeking to adopt the openEHR specifications, and will foster inter-vendor collaboration."

The openEHR Foundation describes itself as committed to supporting legislative and industry standards. They work closely with standardization bodies, as well as with national and international project teams. 

The answer to my third question, is anybody actually using it, is clear in some areas and remarkably unclear in others. The effort began with the Internet and the World Wide Web, and is now becoming commercial (Products, Open Source, 2003). As regards specific products, let’s first look at operating systems. The Linux operating system, which is open source, is used on handhelds, PCs, and supercomputers. BSDs, Berkeley Systems Distribution of UNIX, are open source operating systems. There are numerous Internet applications. Apache runs 50 percent of the world's Web servers. BIND provides domain name service for the entire Internet. Send mail is the most ubiquitous e-mail transport program. Mozilla is a redesigned version of Netscape, which is open source, and which is regaining some of Netscape's market share. OpenSSL provides encryption over the Internet. The claim is that commercial software can't compete with the quality of these applications. There are numerous programming applications available. These include Perl, Zope, and PHP for live Web content. Also included are Python, Ruby, and Tcl/Tk, which are high-level programming languages. Compilers and tools include GCC, Make, Autoconf, and Automake. These are regarded as having outstanding quality. 

In reviewing lists of companies that deal in open source software (Open Source Companies, Open Source, 2003), there are some names that we all know very well, including IBM, Apple, HP, and Sun. Others on the list include SGI, Sharp, Cyclades, Red hat Software, ActiveState, Sleepycat Software, Inc., Covalent Techonologies, and more recently government and non-profit agencies. 

Since our primary interest in this class is in medical applications, I will list a few of the hundreds that I found. I found much more evidence of existing applications than evidence of any widespread use. Some of the applications that I found include the Good European Health Record (Good Electronic Health Record in The Spirit Project, 2001-2002), which is related to the openEHR Foundation. DHCP/VISTA (DHCP/VISTA in The Spirit Project, 2001-2002) is the excellent system in nationwide use at the VA. It is described as the largest collection of open source health care software worldwide. Medsphere Systems Corporation (Medsphere, 2003) distributes it as OpenVista under an open source license, but it is actually available unrestricted as public domain software. VISTA Hardhats (Hardhats, 2002) is an active developer community working on expanding and supporting VISTA. Open Paradigms, LLC (Open Paradigms, LLC) is a commercial company providing training and support for open source health care applications. SourceForge.net is an open source software development website (SourceForge.net, 2003). OSCAR (Open Source Clinical Application Resource) (OpenSourceClincalApplicationResource) is an open source, web-based electronic patient record system.  

Medsphere (Medsphere, 2003) develops, sells, and supports OpenVista. They claim that this is the first huge, affordable, open, integrated health information system. It uses a GT.M database on Linux. GT.M is an M database available to GNU/Linux-based users as open source freeware (Sanchez, 2000). Medsphere (Medsphere Signs Contract with Pacific Telehealth and Technology Hui, 2002) signed its first $300,000 contract in July of 2002. This is a joint effort by the VA and Department Of Defense.  

What I learned from the question that I posted to the AMIA list group is that open source for healthcare is a movement in its infancy (Shreeve, 2003). Products have been in the development stage for several years, and it is only recently that real world implementations are occurring. OSCAR (OpenSourceClincalApplicationResource), for example, reports that they have 20 implementations in place. It appears that the size of these implementations includes groups up to about 20 physicians. SQL Clinic (Good, 2003) has been in use at Saint Vincents Catholic Medical Centers of New York, Division of Residential Services, for the past three years. They have a few paying customers.  

Regarding the claim that open source software is more cost-effective than proprietary software, Ignacio Valdes (Valdes, 2003) at AMIA has the following to say: "No one that I know of has direct evidence for economic benefit of FOSS for medicine. There is much indirect evidence in other industries. I am in the planning stages of a study that will attempt to answer your question." An obvious questions to ask about anything that is free, is how can it survive in our economy? In fact companies can make money by charging for distribution, warranties, support, installation, and customization (Benson, 2001). Robert Young (2001) has written a detailed discussion of the economics of open source software. He challenges the notion that it is any more difficult to make money in the open source business than it is to make money selling proprietary software.  

I asked the AMIA workgroup to talk about support for open source software. I have mentioned above the various commercial companies providing support. The key to this approach is that it fosters competition among these companies in the provision of quality support. There is no such competition for the vendor who provides proprietary software-- only that company can support it. The comments that I got back suggested that it is easier to buy a proprietary package and pay for support (Good, March 2, 2003), but that the cost and quality of support and software very widely, and customization is expensive. At SVCMCNY (Saint Vincents Catholic Medical Centers of New York, Division of Residential Services) the choice was made to develop in-house because "clearly it is much more cost effective".

Dr. Good at SVCMCNY provides some recommendations for support of open source applications. His group recommends that each of their clients locate a database administrator, either a non-IT person or someone hired specifically for this purpose." This person then joins the SQL Clinic team and gradually learns enough to become a developer." This provides SVCMCNY with a group of individuals from different agencies, with different areas of expertise, who can collaborate on development. All members of the team and their representative employers benefit from the efforts of others in the group. Dr. Good believes that this cooperative approach works better than the competitive approach, although SVCMCNY is still very new to the process.

Dr. Shreeve (March 2, 2003) in the AMIA workgroup addresses my concern about hype vs. objective data very nicely: "as the growing OSS/FS Healthcare community gains additional experience with real implementations, we will begin to generate the OBJECTIVE evidence that you seek." Apparently I was unable to find it because it doesn't exist. 

I chose to discuss this topic because it is clear that a new approach is needed. Integrated clinical information systems are not widely distributed (van Ginneken, 2002). Attempts to computerize healthcare records have been in progress for many years, yet still very few hospitals in the United States have implemented them. Proponents of open source claim that their approach provides enormous benefits in cost efficiency. We know that we need independence from unstable vendors. Given the enormous cost of these systems, it does not seem reasonable to take the risk that a vendor will go out of business, leaving an institution with a product that can no longer be maintained or upgraded. One of the greatest difficulties that our information technology departments have is that the needed functions simply don't exist, and adequate usability is very difficult to provide (Ash, 2000). If open source really can provide rapid software evolution, it could be of great benefit to us. Seemingly viable options for support are available, including commercial efforts as well as combinations of in-house staff with the commercial efforts.
 

References 


Ash, J.S, Gorman, P.N., Lavelle, M., and Lyman, J. (2000). Multiple perspectives on physician order entry. Proceedings of the American Medical Informatics Association Symposium 27-31.

Benson T. Medical software's free future. All software developed at public's expense should be licensed as open source. BMJ 322(7290),863. 

Carnall D. Medical software's free future. BMJ 321, 976 

DHCP/VISTA in The Spirit Project (2001-2002) [Online]. Available: http://www.euspirit.org/sw_scheda.php?id_software=1001831024.9498&url_local=&url_originator=&swnlang_src=&sw_summary=&sw_classification=.

Good Electronic Health Record in The Spirit Project (2001-2002) [Online]. Available: http://www.euspirit.org/sw_scheda.php?id_software=1001631471.8968&url_local=&url_originator=&swnlang_src=&sw_summary=&sw_classification=.

Good, T. AMIA email (March 2, 2003) [Online]. Available: mailman.amia.org.

Hardhats (2002) [Online]. Available: http://www.hardhats.org/index.html. 

Medsphere (2003) [Online]. Available: http://www.medsphere.com/home.  

Medsphere Signs Contract with Pacific Telehealth and Technology Hui (2002) [Online]. Available: http://www.medsphere.com/media/press/20020712-Pacific_Teleheath_Hui.pdf.

openEHR (2002) [Online]. Available: http://www.openehr.org/openehr_home_aims.htm. 

Open Paradigms, LLC [Online]. Available: http://www.openparadigms.com. 

Open Source (2003) [Online]. Available: http://www.opensource.org. 

OpenSourceClincalApplicationResource [Online]. Available: http://67.69.12.117. 

Open Source Companies, Open Source (2003) [Online]. Available: http://www.opensource.org/docs/products.php.  

Products, Open Source (2003) [Online]. Available: http://www.opensource.org/docs/products.php.  

Sanchez Offers GT.M Database as Open Source Freeware to GNU/Linux Users, in Sanchez (2000) [Online]. Available: http://www.sanchez.com/news/archivedreleases/nr_001107.htm. 

Shreeve, S. AMIA email (March 2, 2003) [Online]. Available: mailman.amia.org.

SourceForge.net (2003) [Online]. Available: http://sourceforge.net. 

The Free Software Definition, GNU web page (2003) [Online]. Available: http://www.gnu.org/philosophy/free-sw.html.

The GNU General Public License (GPL) (2003) [Online]. Available: http://www.opensource.org/licenses/gpl-license.php.

The Open Source Case for Customers, Open Source web page (2003) [Online]. Available: http://opensource.org/advocacy/case_for_customers.php.

The Open Source Definition (2003) [Online]. Available: http://www.opensource.org/docs/definition.php.

The Open Source Health Care Alliance (2003) [Online]. Available: http://www.oshca.org/. 

Young,R. Open Sources: Voices from the Open Source Revolution. Giving It away. How Red Hat Software Stumbled Across a New Economic Model and Helped Improve an Industry (1999) [Online]. Available: http://www.oreilly.com/catalog/opensources/book/young.html. 

Valdes, I. AMIA email (March 2, 2003) [Online]. Available: mailman.amia.org.

van Ginneken, A.M. The computerized patient record: balancing effort and benefit. Int. J Med Inf. 65(2),97-119.

© Copyright 2003   The Informatics Review
Informatics-Review > Thoughts > The Open Source Movement