Bio_Bulletin_Board Discussion List Opens

The BIO Bulletin Board is a newly created, Mailman-based listserver designed to facilitate discussions among the Open Source
bio-informatics community
. The mailing list has been implemented in response to numerous requests for an e-mail based alternative to our ‘Ask The Open Lab’ web-forum, which lacks this facility. It is intended to provide a convenient resource and archive for members of the bioinformatics community. Free / Open Source software products and service announcements are encouraged; however, advertising commercial software products and services, or spamming of any kind is NOT permitted. With this post, LMN is also introducing its ‘BioInformatics’ subject area.

Latest Future-Proof Information Systems Paper

Brainiac Thomas Beale at the Good Electronic Health Record (GEHR) has his latest (HTML draft here, PDF draft here) of his paper on ‘Future Proof Information Systems’. With an Introduction here: ‘…the usual legacy IS systems…have a limited lifespan and are expensive to modify and extend in order to accommodate changing needs…The approach proposed here is a rigorous knowledge-modelling one, and is founded on a basic tenet: the separation of domain and technical concerns in information systems. In real terms, this translates to:

  • The removal of domain concepts from concrete software and database models, into independently managed, standardised vocabularies and domain-specific models.
  • Re-engineering the software and database as a knowledge representation (KR) system architecture, designed to process information by using externally supplied domain definitions…’

Fresno III: Commercial Model Has Failed

Fresno III was recently held to discuss free and open source medical software. The full press release follows. Some of the highlights of the meeting where as follows:

  • Agreement reached that commercial model of software development in family medicine/general practice has failed.
  • Continued collaboration and partnership is advantageous.
  • A model for patient involvement with the EHR was agreed.
  • A framework for further international collaboration has been achieved, although funding sources remain to be secured.
  • Specific aspects of primary care computerization needs are better met in the UK than the USA and vice versa, though neither fully reflects the concepts identified in GEHR (Good Electronic Health Record)

Summary Statement and Press Release

Fresno III OS-IHI Working Group

Feb 2-4th 2001

In summary; its Open Source, Output vs. Input and Object oriented programming that were the “kernels” of this meeting. Here are the summary points as framed by Dr. Ellis and Dr. Pepper. All addendums and comments appreciated.

In Attendance:

Dr Alex Caldwell Family Practitioner. Tulare CA. USA (

Tim Cook President FreePM. Miami, Florida. USA (

Dr Nikki Ellis Research Fellow. University of Manchester. UK (

Dr Chason Hayes Orthopaedic Surgeon. Charlotte N Carolina. USA (

Peter Hudson Nurse Practitioner. Fresno USA (

Dr Adrian Midgley General Practitioner. Homefield Surgery. Exeter (

MJ Midgley Nurse Practitioner. Homefield Surgery. Exeter UK (

Dr David Pepper Assoc.Prof UCSF@Fresno. CEO-MAP Masters (

Dr Robert Shepard Medical Director. Helena Healthcare. Montana USA (

In addition to this core group, an additional 20 people attended an evening session that further explored the issues being raised and discussed by the core.


The third in a series of Open Source International Health Informatics (OS-IHI) retreats was held in Fresno, California during the 2nd to the 4th February 2001 to explore ways to improve electronic health care record keeping for outpatient primary care practitioners.

  • A review of six primary care systems, and associated decision support software, was undertaken.
  • Progress to date on the Open Source (OS) projects represented by those present was discussed.
  • A demonstration of the TkFP GUI Front end was held.

Key Points

  • Agreement reached that commercial model of software development in family medicine/general practice has failed.
  • Specific aspects of primary care computerization needs are better met in the UK than the USA and vice versa, though neither fully reflects the concepts identified in
    GEHR (Good Electronic Health Record)

  • Continued collaboration and partnership is advantageous.
  • Agreement that an Object Orientated approach to systems development will be explored within an Open Source environment as a likely emerging and powerful technology.
  • A framework for further international collaboration has been achieved, although funding sources remain to be secured.
  • A model for patient involvement with the EHR was agreed.
  • Agreement that the unique needs of health informatics dictate specific requirements for Output versus Input modes for viewing data.


Submission for Theatre Style demonstration to AMIA 2001 (Fall Congress)

Submission for Knowledge Sharing House Panel to Medinfo2001: Commercial research aspects of OS-IHI

Two Papers in the OS&IHI series have been drafted and will be completed and submitted for publication (First paper in series: Open Source & International Health Informatics: Placebo or Panacea? Ellis et al submitted to medinfo2001)

  • Open Source & International Health Informatics: Consideration Of Object Orientated Design
  • Open Source & International Health Informatics: The Business Case
  • Application of MAP-Masters to the NIST ( for funding

Areas Requiring Further Work

In keeping with the OS Development model, anybody who can offer assistance in these areas would be welcomed to either participate in this work virtually or to attend future retreats.

  • An Open Access/Target international drug preparation database
  • An Open Access/Target clinical coding system or thesaurus

Further Information

Please contact anybody listed as in attendance at the head of this statement for further information on any aspect outlined above.

Support from B-M-Squibb is graciously appreciated for the speakers expenses.

Merger of ResMedicinae with OpenEMed

After many long discussions, the members of agreed some days ago,
to bring their development power into the
project from now on.

Both projects follow quite similar aims in that
they want to develop many CORBAmed modules,
mainly using Java, to create a prototype useable
in practice within one to two years.
Also, the GEHR kernel as another standard will
possibly be implemented.

Another aim is to strongly work together with
related projects such as OIO/FreePM/LittleFish
who want to provide CORBAmed interfaces as well.

I have a dream: Slashdot Readers on Medical Information Systems

You can always count on Slashdot readers to speak their minds and quickly generate a pool of diverse opinions. In this AskSlashdot piece today titled “High Tech Medical Clinics“, more than 300 comments (from the point of view of patients) cover many issues central to the future of medical information systems. If the customer is always right, who is the customer? Who is going to listen? Most importantly, how well can the OIO (and other open source projects) support these requirements?

NSA Develops Secure Linux

The National Security Agency (NSA) has a paper on a ‘security enhanced Linux’ it has developed which vividly shows the free and open source software process in action. Could it be useful in medicine? From the paper: ‘…This version of Linux has a strong, flexible mandatory access control architecture incorporated into the major subsystems of the kernel. The system provides a mechanism to enforce the separation of information based on confidentiality and integrity requirements. This allows threats of tampering and bypassing of application security mechanisms to be addressed and enables the confinement of damage that can be caused by malicious or flawed applications…the integration of these security research results into Linux may encourage additional operating system security research that may lead to additional improvement in system security…

Interview: OIO’s Andrew Ho

Andrew Ho, MD has the unlikely combination of being an assistant clinical professor in the department of Psychiatry, Harbor-UCLA Medical
Center and the leader of one of the many open source medical projects currently in
existence. The Open Infrastructure
for Outcomes
(OIO) project is his brain child. Ho’s comments and guidance on topics as diverse as meta-data, forms libraries and project convergence can be frequently found on many of open source medical computing’s discussion lists. LinuxMedNews caught up with Dr. Ho to learn more about him and his views on medical open source, the OIO, his competition, HIPAA and what he does for fun.

LMN: Tell us about yourself. What is your work background, where do you live, schooling, hobbies?

Ho: I am originally from Taiwan. I came to the U.S. at 11 years old and lived in New Jersey on a farm until I was 16, learning to be a farmer. Then I moved to Southern California to Orange County. In 1983, I went to college at UC Irvine and was introduced to the world of Apple II hackers and academic software engineering. I worked for 2 years as a programmer at Alfred Bork Educational Technology Center where his team was developing interactive software to teach Physics. That’s where I learned user-interface design and was introduced to the IBM PC, email, and ftp. We also did some pretty advanced stuff like separation of text content and program logic so that it is easy to run the software in another language. We developed graphics and windowing libraries on the UCSD p-system that offers write-once, run-everywhere through basically a virtual
machine layer, pretty much the same as Java. It is sad to see that in almost 20 years we haven’t gone that much further. Went to Medical school at UC San Diego with the fantasy that I will change the practice of medicine through better information tools.

LMN: When did OIO start and what prompted you to begin the OIO?

Ho: As an undergraduate I began looking at improving the practice of medicine through improved tools mostly from artificial intelligence. It has been a long term interest as to how these tools can be applied. My interest in these tools led me to brain imaging research which led me into Psychiatry and sleep research… However, during residency training, I realized how difficult it is to collect a sufficient set of treatment and outcomes data for input to artificial intelligence tools. Between 1996 and 1997, I was studying treatment outcomes at the VA in West LA and it took a year to organize and analyze the data. I realized that data management
was a big task and there was no tools for it. That’s when I started planning the OIO. Nobody was or is writing software for this type of application…I moved to Harbor-UCLA in 1998 in order to build a treatment program that incorporates these new tools.

LMN: Who is working on it with you?

Ho: A team which is using the OIO at Harbor-UCLA: John Tsuang another Psychiatry
Professor at UCLA who has been working with me since 1995. Together we’ve seen the hard way of doing a treatment outcomes study in a real treatment setting and now we are doing it the easy
way. Other participants are two social workers (David Haponski, Christi Finazzo), a nurse (Carol Giannini), secretary (Perla Rosales)as well as a part-time resident (Tim Fong).

LMN How about Funding?

Ho: Good question. We’re 100% clinically funded, there is no research funding for this project. It is an important issue because since the beginning we wanted to develop a software system that could be used without reliance on research funding. That’s why we decided to do the outcomes study part without funding. We would also have problems open-sourcing the software if we used grant funding because of University intellectual property rights. But now that we’ve open-sourced it, we can get funding to further develop it and to teach others to use it.

LMN: What is OIO’s target audience?

Ho: Clinical settings that want to have an outcome management tool that they could potentially use as an EMR. Alternatively, a
paper-based practice could also collect data in order to analyze outcomes. For the research setting, it can be used as a research data
management system, for managing study work flow and running clinical trials. Research applications are basically wide open since the OIO can host arbitrary data collection forms.

LMN: What makes OIO better than other projects like FreePM,
FreeMed, TK Family Practice and
the rest of the current open source medical projects?

Ho: I think the goal is actually a bit different. The aim of OIO is not necessarily a product that is tailor-made to a specific application. It is an infrastructure that applications can be built on. It isn’t necessarily the final product to install. For example, you might want to have certain screens and forms for your family medicine clinic. It is primarily about data-interchange but comes with some basic !’bundled!( applications. The best analogy is that it’s Windows, rather than Microsoft Word. Windows comes with Notepad and a calculator, but these ‘bundled’ applications are not necessarily
the strength of Windows. Windows is useful because you can built more complex applications on top. OIO is at a level below applications even though we have sample applications. We’d rather collaborate with other groups that want to develop full applications such as a Neurology practice management package, or depression treatment
algorithms. The selling points are #1: OIO will save application development time, and #2: it
will allow applications to interchange data with other applications that are also based on OIO. That’s why the focus of OIO development is not on creating forms that are ideal for a certain application. Of course, if nobody wants to write Microsoft Word, I may have to do that too.

LMN: When do you hope to have OIO complete?

Ho: Complete? Never complete! (laughs) We want to get certain features in by version 1.0. We released our first public version as 0.9.0 in September of last year and we’ve had six updates since then. It went from 0.9.0 to 0.9.6. We’re a little bit past the halfway point to version 1.0. We have added all the XML import and export capabilities and advanced forms features such as
Java applet integration since 0.9.0. We expect to reach version 1.0 by May of 2001. There are two to three major features we hope to have by then: #1: A modular report-generation system, so it would be plug and play just like the forms are. #2: a ‘packet’ of forms, for example during this office visit, you need to fill out these three forms. We hope to have the logical branching and other more complex processing of forms in place by then. This involves building a level of abstraction above the current forms metaphor. #3: we want to have a version that works with Palm OS so that you can fill out the same forms on the Palm Pilot and then transport the data back to the web-based data management system.

LMN: Why do you think OIO will succeed?

Ho: So far I think it is unique in what it aims to achieve. Other systems want to do it all. The OIO is more focused. It is in a spot that is not as obvious or attractive to other projects.
We want to provide an underlying data interchange service. Using the Windows analogy, instead of
having multiple metadata import and export drivers we want to provide a single metadata creation and management facility that other projects can take advantage of. To my knowledge other projects don’t provide this service. It boils down to making it easier to produce medical applications that are interoperable and standard compliant.

LMN: How does this differ from HL7?

Ho: HL7 is a data interchange standard. It doesn’t provide the ease of programming that OIO offers. With the OIO you can use its software
services such as the metadata editor, exchange, and rendering engine that are already built into the OIO. It can be seen as the same distinction between an operating system that provides print API and the data format that some printers expect.
OIO makes it easier for applications to generate and use standard compliant messaging and semantic constructs.

LMN: What prompted you to go open source and not proprietary? If OIO is so good, why not try to profit from it?

Ho: It is a good question that I had to answer early on: I think a big part of its value comes from the fact that it is open source. This is generally true of infrastructure tools. It would be much more valuable if more people choose to use it. This is a necessary step to data-interchange and interoperability which is an endemic problem facing medical data systems. The current licensing makes it an incentive not a dis-incentive to adopt OIO as part of another system. Besides, it is just the beginning of many applications that can be built on top of it. If it
where proprietary I think it would be a much, much more shaky foundation. By open-sourcing it, I can entice others to not only use it but improve upon it. The money thing? It would be nice, but who is to say that in the long run it won!&t be more lucrative to open-source the OIO? Also, the way to measure return is not just in money. I derive a lot of pleasure from knowing that it is useful.

LMN: Where do you see OIO in 2 years, 5 years, 10 years?

Ho: In 2 years we should be able to have a few more sites and research projects that use it. If we get the funding from NIMH, we will have an active program for teaching people how to use it.
We also want a hosted application service alternative for researchers and clinicians
who do not have the time or expertise to install OIO at their own sites. Hopefully a larger selection of forms in specific domains like Neurology or Orthopedic Surgery will also be available.

LMN: In 5 years?

Ho: Wow, 5 years is a long time. By then we should have some good data on the limitations of our architecture. It will be a 7 year-old design by then, so we will know a lot more as to what we need to do for the next generation system. We will be able to see if this idea of massive interchange of data has panned out. It will be either chaos or the Nirvana of data interchange. By five years there will have large datasets collected using the OIO so we will be able to see how useful it really is. Also, by that time Microsoft’s .NET strategy would have time to mature so we can see what they can do that we can’t do and vise versa.

LMN: 10 years?

Ho: By then hopefully all software systems will have incorporated all the capabilities of OIO. I will be working on a different project. Perhaps something even more exciting.

LMN: What problems are you having with OIO? What are its flaws?

Ho: I think a weakness is in the dependence on ZOPE so far. I would like to see it implemented based on other technologies. If there is a security flaw in ZOPE then the OIO will be directly affected. If there is an alternative implementation then this might
not be as big a problem. That’s why we want to do it on Palm using Java which will also give us a version of OIO on Java for non-Palm environments.
The second thing is, we don’t have a large set of applications available yet. I suppose I mean OIO forms. It would be nice if there could be an adequate library of forms that you can install on OIO. We have the facility to import and export these forms, but we don’t have these forms yet. While it isn’t difficult to create these forms,
we only have forms that we use in our current projects. As more and more people use OIO, then there should be a wider choice of forms. Finally, we have a very aggressive development schedule, so when I implement certain features that are not immediately useful in our own research project, it is hard to get exhaustive testing of these features before code release. So when we release enhancements, they could contain bugs. Hopefully we will be able to grow more users and get better feedback, especially for the new features. For example, the patient record import/export feature that we released in 0.9.6 is not used in our research project. So I had to release that feature without extensive testing.

LMN: Not specific to OIO, where is this all going?

Ho: It boils down to helping users do things quicker, easier and more accurately. That of course, is sometimes limited by technology, but there is also this idea that you are fitting the tool to the use. Where it is going is general-purpose tools, that we can call infrastructure, that have applicability. The users don’t care about all this, they care about getting their work done. They want a custom application that works for them in that particular moment. You either have to do this infrastructure thing real well or this highly customized application thing real well. It’s going to be impossible to do both in a single product. We will have a niche for either an infrastructure product or a application product. If you look at the history of computer
tools, it is pretty clear that this is what happens as a domain matures. For example: Windows and its applications. For this general purpose set of tools we call an operating system, as users begin to be able to use a wide range of customized applications that run on Windows, that’s when it becomes truly useful to the users. So more customized medical applications and specialized infrastructure tools that make it easier to develop these applications is where this is all going.

LMN: Any thoughts on HIPAA?

Ho: I think it is a step in the right direction. At the same time it is a big mess because of the lack of technology. One can try to solve these problems by law which is the best that we can do now. When we have the technologies to safeguard data in ways that we can’t provide now, I think the rules will be simplified. Writing the law is only one of the steps in influencing behavior. How the rules are enforced is another crucial part. I’m actually very excited about HIPAA because there is another product that we are proposing that aims to safeguard the confidentiality of patient records. That is a separate project with a focus on data security.

LMN: What do you do for fun?

Ho: We have three children and it is always fun to take them out to the park or show them the neatest toys on the web.

LMN: Like what?

Ho: (laughs) They like the Disney site.

LMN: Any last thoughts?

Ho: Thank you for doing this piece, I look forward to seeing it on your page.

Computerworld: IBM/MDS Proteomics Build Linux Medical Supercomputer

Computerworld has an article on IBM’s announcement today that it will be developing a Linux and AIX based medical research supercomputer with Canadian firm MDS Proteomics. ‘…The deal between IBM and Toronto-based MDS Proteomics Inc. is aimed at developing a supercomputer that can handle 700 billion floating-point operations per second and that will let scientists at the research firm conduct complex analyses of proteomics – the study of proteins and how they function. The two companies said they also plan to create a free online protein analysis database that will be accessible to scientists and other users from outside of MDS Proteomics…’

Wired: Microsoft Exec Slams Linux

Wired has an article quoting Microsoft group software manager Doug Miller as saying: ‘…that Microsoft believes that “in the rush to get on the enterprise bandwagon,” the new Linux kernel lacks some of the key elements required for enterprise use.”Based on the warnings from the developers and confusing messages from the distributors, it is clear the long-heralded 2.4 Linux kernel is a long way from being ready for business use,” said Miller. “The kernel is just the beginning, still raw technology.”…’ The article has a number of rebuttals to Miller’s statements. has a piece that ‘quotes’ Larry Augustin, CEO of VA Linux: ‘…Augustin denounced as “absurd” allegations that Microsoft might be utilizing its infamous FUD tactics to spread Fear, Uncertainty, and Doubt about an opponent in an effort to steal market share. “That would be deceitful,” he replied, “and Microsoft has stated repeatedly that it does not lie or cheat or mislead.”