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:
"promote and publish the formal specification of
requirements for representing and communicating electronic health
record information, based on implementation experience, and evolving
over time as health care and medical knowledge develop;
promote and publish EHR information
architectures, models and data dictionaries, tested in
implementations, which meet these requirements;
manage the sequential validation of the EHR
architectures through comprehensive implementation and clinical
evaluation;
maintain open source "reference"
implementations, available under licence, to enhance the pool of
available tools to support clinical systems; and
collaborate with other groups working towards
high quality, requirements-based and interoperable health information
systems, in related fields of health informatics."
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. BMJ322(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.
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.
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.