up a level
post article
search
admin
Contact
main
|
| Massively Expensive IT Failures in the 90's |
 |
 |
 |
 Posted by Saint on Monday August 14, 2000 @ 08:31 AM
from the closed-source-in-healthcare-still-doesn't-work dept.
Lest the previous article on San Mateo's $12 million+ failure be seen as an aberration www.jimintriglia.com again provides this link to a Health Informatics article detailing the massively expensive IT failures of the 90's. 'Nearly every healthcare delivery system in the country invested between $5 million and $50 million in CPR [Computer-based Patient Record] efforts between 1993 and 2000. Despite serious criticisms from some experts that almost all investment plans to implement the vision of a CPR were flawed at inception, the majority prevailed, creating tremendous industry momentum.' With numbers like these, open source medical software CAN'T be any worse. At the risk of a big 'I told you so!' read the original LinuxMedNews article: Medicine's Dirty Software Secret. Digg this article
<
|
>
|
|
The Fine Print: The following comments
are owned by whoever posted them.
( Reply )
|
Re: Massively Expensive IT Failures in the 90's
by Hemant Shah on Monday August 21, 2000 @ 03:03 PM
|
That the closed-source medical software does not work is no secret. Indeed, that is stating the obvious and no dispute about that.
However, the surmise that it does not work because it is closed source is at best tenuous. The fallacy in the argument, which open-source zealots make, is due to the fact that only software that did exist till now was closed source. The open source medical software is just a speck on the horizon now and as yet untested.
To my mind, the key reasons for failure of medical software are:
· Trying to naively apply the principles of other domains in medicine – when medicine is radically different from all others
· Not adequate effort in identifying what constitutes the clinical requirements and clinical information – which is central to healthcare
· Research organizations’ obsessive focus on esoteric areas rather than the needs at the point of care
· Lack of approaches where the healthcare professionals become participants in the information modeling and management
· Too many disparate tools or technologies – no significant effort in creating an integrating framework to accommodate these or encourage development of compatible tools
Amongst other things, these have led to alienation of the physician from the entire effort. The physician has become, for no fault of hers, the weak link in the chain of information.
The problems stem from flaws at analysis and design level not at the developmental level.
Much of the value of open source is from what it brings to the coding. Whereas the problems of healthcare software are due to lack of understanding (as even the referenced article mentions repeatedly). Much of this “understanding” is open source, anyway.
So, I am not sure we will make any significant progress, even if the code is open. It’s back to the drawing board situation, really.
I do believe that open source movement in medicine has a promise, but if we attribute everything that was wrong to source being closed, we will be repeating the same mistakes – may be at far greater costs.
Regards,
Hemant
Hemant Shah MD, M. Surg.
Research Fellow, Medical Informatics
National Library of Medicine
38A/Floor 10
8600 Rockville Pike
Bethesda MD 20894
Ph (301) 435 3262
h.shah@computer.org
shahh@nlm.nih.gov
|
[
Reply to this ]
|
Re: Massively Expensive IT Failures in the 90's
by Saint on Monday August 21, 2000 @ 08:14 PM
|
You have good Kung-Fu. However, open software isn't new. Unix has inspired generations of programmers since 1969 in a quasi-open source way. i.e. before it was called open source. The Free Software Foundation was begun in 1984. Open source software has only become a cultural phenomenon relatively recently. Many people forget that the existence of the Internet (and BTW, the software that runs this site) owes a great deal to many unpaid, unsung software engineers working over the years in Universities and other institutions in a quite open source way.
Your point about painting everything with the same brush is well taken. The most likely scenario is that clinical computing systems of the near future will be a hybrid of open and commercial software. Having said that, the sins of closed source software in medicine are huge. Had a shared, cooperative method been widespread like the development of Unix we would be in much less of a predicament than we are now: fragmented, lacking in many areas and hugely expensive. Analysis and designs abound, development is precisely the problem in medicine. The great thing about open source is that 'going back to the drawing board' means that someone has already written part of the design. An example is StarOffice 6.0 available as open source in October. These are essential applications that can be embedded into clinical computing platforms. One of many things you can't do that with closed source. On the other hand, I like closed source because it makes open source work harder.
-- Saint
|
[
Reply to this ]
|
|
The Fine Print: The following
comments are owned by whoever posted them.
( Reply )
|
|
