Slashdot Mirror
Introducing Open Source to the Doctors
TCook writes "Dr. Daniel Johnson introduced the morals of open source to the American Medical Informatics Association. Read his
presented paper.
" Medical technologies are one of those areas that I think should be open-sourced for obvious reasons. The notion of bugs and flaws that no one can get at in medical technology gives me the willies. As well, if by applying the "thousands of eyes" we can save lives, I think that's good for all.
I agree with you 100%, you are one of the rare few. But you are wasting your breath on slashdot, particularly on something so broad as to capitalism. Most slashdot-commentors do not know their history very well, and lack even a basic understanding of what makes capitalism work (and other systems fail).
I am not a GNU/Linux or free software / open source zealot; I simply recognize its genuine strengths and enormous potential.
It's a bad sign that this sort of disclaimer has to go on top of articles about open source; I've done it myself on a couple of occasions.
I'm not going to make the usual plea to the zealots to be polite when talking about OSS; I know by now that it's not going to work. What's needed is for those of us capable of pressing the point without being an asshole to do so when the opportunity presents itself.
OSS' greatest enemy isn't Microsoft; it's its own zealots.
If a member of my family dies because somebody put = where they meant ==, I want to know who they are, and I want to sue the pants off them. Most open-source software has disclaimers that absolve the developers of all blame. I guess the solution is for companies like Redhat to distribute the software with a guarantee. Any other ideas?
grep -ri 'should work'
Their business is just software, network and servers, not medical equipment, that's the Equipment Department's business.
The software was both in-house development, and analysis and ordering of external solutions for information exchange software (secure mail system for the hospital) and software such as electronic medical-records, care planning tools and even intranet solutions.
The one thing that struck me most was that commercial or closed-source software was always the only thing thought of, even if most of the developer's headaches seemed to be chasing external software consulting companies who had failed to deliver the product wanted or failed to deliver it on time, and even more often with a significantly larger price than expected. There were also external contracted companies who had, without warning, stopped developing products entirely that were essential to the hospital. All these problems ended up at the IT department. I'm not an software development expert but I can tell you that there was easily more people involved in keeping track of external vendors and external projects and patching this software up so that it would work as expected than would have been needed developing all this software themselves.
(As a side-note, add to this the fact that the hospital had an ancient policy not just to focus on commercial software, but only Microsoft software everywhere, i.e. Windows9x, NT, IIS, SQL Server, ASP, MS Exchange, MS Proxy, Internet Explorer, MS Office, and the various MS development tools... Such a single-minded policy isn't helping productivity in my opinion)
Due to the problems with external vendors it seemed that doing more stuff on their own in the future was the new policy. Therefore the IT department had doubled the number of employees the last year. And the policy seemed to be that almost all of the development should be made in-house, with as little input from the outside as possible. And from what I heard this was also the situation at most other hospitals that could afford a large IT department. And therefore cooperation between hospitals or even the thought of sharing software between hospitals seemed to be out of the question. I spoke to a lot of the developers there during my work, and even if the historical explaination of the situation varied, this seemed to be the situation right now.
So from what I've learned, hospital IT departments have a lot to learn from open source, co-development and code-sharing. I was truly happy when I saw this article.
GNU/Linux. The Freshmaker.
I think the guy has a point, and I tend to agree with that point. However, I don't think anything like this will be accepted in the near future for one reason: people outside the OSS comunity just dont get it.
Admittedly this is anecdotal, but I'll repeat it here because I found it illuminating:
I hang out on undernet's #linuxhelp quite a bit. A 15 year old was on asking questions. Someone suggested he buy a copy of RedHat from cheapbytes, but he pointed out that cheapbytes wont let a 15 year old purchase from them. Of course, someone told him to get his parents to buy it for him. His response? "my parents dont want me to get linux, because they think I will become a cracker"
Draw your own conclusions.
I work in the medical software field, so I'm qualified to comment:
"Even highly scientific stuff needs spiffy GUIs and the like, stuff that anyone can write."
Bullshit! The user interface needs to be simple, plain, unadorned, with all necessary information instantly accessible. When the patient is pried wide open during open heart surgery, a "spiffy" interface will kill him!
"Got the fastest FFT out there, and only five people in the universe can understand how it does it? I can write a GIMP plugin frontend for it."
If you don't understand FFT, and you program medical software that uses it, you're a idiot. This isn't a GIMP plugin, it's a person's life!
"The point is morality."
I'm surprised. You got that one right! It's immoral to subject someone's very life and death to a particular political view. A truly moral physician would use the most accurate, bug-free, and technically superior software. It doesn't matter that if it's open or closed source. You may believe, and even be right, that the open source solution fits the bill, but if it doesn't, then what? Let the patient die?
"It's WRONG to sell the right to use software. If you do it, you're a software hoarder, and you're evil."
I will tell you what true evil is. True evil is disallowing a physician to save a person's life just because the software he uses doesn't suit your personal fancy. Well fuck you! If that's what GNU is, I don't want anything to do with it!
"RMS is way more eloquent than I am."
And apparently way more intelligent as well.
A Government Is a Body of People, Usually Notably Ungoverned
"as for linux, in 2.4 SMP kernel, it is about 3000% more efficient than sun solaris and scales better."
I thought so. You really don't know what you're talking about, do you? Somebody scrubbed too hard while washing your brain.
A Government Is a Body of People, Usually Notably Ungoverned
Why exactly do you need to sue the pants off them? Is it to get revench (the economic equivalent of a pointless punch in the face) or because you have just won on the American wheel of litigation and want the chance cash your reward?
If a company guarantees that something works and it doesn't, then they have broken there promise and should possibly be sued (such legislation being there so people have to stand up for their promises). What does that have to do with whether the software was developed as open or closed source? The guy who forgot the second = was not doing it because negligence, he made a mistake. It is whoever promised you that the system would work that was negligent, and people can promise that for regardless of how it was developed (closed or open source, they better have tested it well).
Personally (but, of course I'm not American) I would prefer if there was no suing going on at all, as long as less people are dying. This is what the thousand eyes reference was about if you missed that.
-
We cannot reason ourselves out of our basic irrationality. All we can do is learn the art of being irrational in a reasonable way.
1979: This will never amount to anything.
1989: Ok, this can be used to make developement tools, but you'll never make a whole OS.
1999: Ok, you could make a whole OS, but you'll never innovate on it.
Wanna bet?
-
We cannot reason ourselves out of our basic irrationality. All we can do is learn the art of being irrational in a reasonable way.
Consider:
(Although, with relative velocity vectors of a goodly number of Km/s, a little Sun 3/60 could probably do a good number...)
I have not the Real Time skills to deal with that; the absolutism of the comments were what offended me.
Consider that:
RT is not the only issue; free software has considerable things to offer in the non-hard-RT areas.
If you're not part of the solution, you're part of the precipitate.
Sounds like a challenge for a distributed-processing project... :-)
You don't seem to grasp how quickly Linux Grew. It came from a 1 man operation to a better operating system that you can't beat the price/performance ratio. As Windows devolves and regresses, linux explode's with growth.
It seems to me to be very difficult to decide what Linux's growth rate is. First you have to decide what you mean by Linux. Do you mean the kernel? GNU/Linux? Open Source that is compatible with the Linux kernel? How many Linux articles there are in ZDNews? Then you have to decide if you want to talk about the user base, or include commercial applications too.
Whichever you choose will determine the perception of growth rate. If you chose the body of open source that Linux is compatible with, well the history goes back a LONG way. Progress has in fact been quite slow over most of that time. Empires have risen and fallen during that time. If you chose GNU/Linux, well, GNU started in 1984. Look at the growth of MS since 1984.
Certainly Linux has increased it's growth rate due to the increase in participation in the internet. But there are a lot of questions with the open source movement yet. Will the application base come to be? Right now the kernel development part of the picture looks good. But where is an applications catalog that includes offsets to most of what is avaiable for Windows going to come from? Heck, it isn't even clear if we are going to get a good browser and a financial planner.
Don't get me wrong, I love Linux. I run it wherever I can. I go back a long way in computers, and I *HATE* what Microsoft has done to the industry. But Linux is really only starting to scratch the sand on the beach next to the Microsoft Ocean. We have a long way to go before we cross that ocean. It is difficult to extrapolate from where we are today to the other side of that ocean.
You've just proved why free software can be economically viable. You surely don't think Microsoft (for example) will provide a level of support that effective? Their sheer size forces them to offload the support burden onto just the sort of hostile drones you lambast.
You're absolutely right that it's the support. However, there's no reason this and the rigorous checking required can't be _combined_ with non-proprietary code. Everyone can benefit from sorts of code subjected to _severely_ ruthless testing and quality control (oh look, this routine causes a race condition/wrong answer/segfault one ten billionth of the time. *tweet* outta the pool!) and the resulting code could offer lessons to all OSS projects, whether or not they must be comparably reliable and safe. And again, it all comes down to who's willing to go the farthest for the customer, who can often be quite offbase. It's not possible to monopolise on this partly for simple reasons of efficiency, but in the normal flow of business, it's quite reasonable to stake out a really _solid_ niche based on such a level of support (think the Nordstrom's department stores on the West Coast). The emphasis in recent years of monopolise, cash out, quantity over quality business is a distortion of how the markets naturally work- normally trying to stake out a service niche is not only feasible but a really winning proposition. Only in situations of extreme competitive pressure from a monopolist dumping crapware and cutting off distribution, does it become unreasonable to try to establish a quality specialty product that doesn't attempt to seize the whole of the market as cheaply as possible.
You're right in that capitalism in how it is used today is much different than what academics teach and thus what most people beleive, if they have many thoughts on it at all, other than it's the best possible choice, if not the only one.
/.'ers who dislikes capitalism. I'd say today's version of capitalism is a bit fairer than it was years ago as you suggest. However this issue is incredibly complex and I don't have enough personal knowledge to really say much more than what I've stated below.
I am one of those
Capitalism suffers from what any system with a structured heirarchy has: a lack of accountability and thus a lot of corruption. Every modern political/economic system is structured like this. Thus I don't see capitalism as being much better than communism (don't take this the wrong way). I dislike communism as well for similar reasons.
Unfortunatly a political/economic system that is controlled by many tends to be inefficient. Perhaps sometime in the future this can be helped by technolgy and strong education for everyone (my personal belief).
The bottom line of capitalism is money. This takes away a sense of reality from what really matters. Scare resources have to be allocated somehow, but capitalism has utterly failed at this. It also has a lack of ability to account for the fact that the earth can only give us so much before we cause irreversable damage.
The interesting thing about the linux community is that it was able to create a very good, complex product completely outside of the context of a capitalist business model. Here we have a product that was designed to be a good OS instead of something that will make money through the artificial means of marketing, or brainwashing if you're cynical like me.
People who develop for linux don't do it because they're on some sort of jihad. There are a few zealots, sure. Most people agree, I think, that people develop for linux for personal reasons, to obtain respect in the community. So here a good product can be created that requires a large number of people working together yet they are able to maintain a sense of individuality. I don't know of any companies that can give you that.
A few points to concede:
Linux could be considered to be a "totalitarian" setup since it's code base is controlled by one person. Developing for linux doesn't pay the bills either.
I define progress by a few things. One is the amount of control the average individual has over his/her own life. Two is technologies (including the proper implementation) that allow people more time to pursue higher order functions: specifically education and personal enlightenment.
Linux doesn't really fit these per say, but I think it's a step in the right direction because linux isn't particularly tainted by the need for money. Companies like redhat do need to make money, but we do live in a capitalist society so there is no getting around that anytime soon.
"Money is the root of all evil". One of my favorite sayings. Technology is supposed to improve our lives. By my ideas of progress, it hasn't. It is used primarily to feed our materialistic urges, directly related to money. You can factor microsoft right into this. Bill gates limits our ability to control our lives. Lets face it: he has a lot of control over the future of the internet. His power will fade someday, a point even he concedes. People might change, but the instituions do not. As long as we have a capitalist society there will always be someone there to limit our control. I'm open to attack here and could ramble on for hours, but it's late.
I have no amazing solutions, of course. I just see linux as a Good Thing, a form of progress for society. We're generally off topic here but I think this is what the doctor was getting at, on a level he didn't really discuss.
Medical technologies are one of those areas that I think should be open-sourced for obvious reasons. The notion of bugs and flaws that no one can get at in medical technology gives me the willies. As well, if by applying the "thousands of eyes" we can save lives, I think that's good for all.
I would agree full heartidly, but most of the general public I have talked with about this, and many papers I have read, have the idea that the opensource model is "peicemeal" and that with code from many different coders you some organization, security, and stability. I fear that unless the public is far better educated about opensource, and more popular things come from opensource, that these fears will keep opensource from entering such a sensitive field as the medical field.
by the FDA before that piece of equipment may be placed in production. The code review process that is used for medical equipment makes the code review process used by almost any other company look like a peacemeal process.
The long and the short of it is that it seems to me that using open source techniques on medical equipment won't significantly improve the quality of the code--but it may reduce the amount of time it takes before the code quality demanded by an FDA review can be met. Further, code reuse of pieces of code that has already been verified as correct by the FDA may help reduce the development time. And that would be a Great Thing...
______/._____/.___/.__/._/._/././././
*bzzzt!* Stupid slashdot effect....
--
Has anyone even thought through the implications of Open Source medical software? Look at what a typical OSS project is like: release early and often; if it doesn't work today, don't worry there's a new release tomorrow; the users are the testers; etc., etc.
Aaaargh! Don't any of you even dare releasing any medical software until it has been 101% tested by experts in the field. A thousand eyes may see all bugs, but I don't want those eyes being medically illiterate hackers. And neither do I want it under a license that has a warranty disclaimer. If the developers don't trust it enough to warranty it, then neither should my physician. After it's finished, tested, beta tested and FDA approved, then, and only then, can you release it to the public.
If you think I'm off my rocker for saying this, keep in mind that this is what I do for a living. I'm a QA engineer for medical software. When the developers have done their own unit testing, integration testing, received FDA approval, and signed off on it, then I get it. And what usually happens is that I find a literal life-or-death bug in the first day of testing.
A Government Is a Body of People, Usually Notably Ungoverned
The medical software company I used to work for always had an open source policy. The code is not GPL'd etc, its on a restricted license, but the principle of open source is there - release the source code to the customer; don't hide it.
This dates back to 1979 when the company was setup.
When the system was installed all the source was left on the machine too. This meant that we could easily debug on-line without the messiness of tapes and the resulting delay.
Our customers developed and maintained the software.
I'm now working in a hospital, and I develop and maintain the same software as a customer.
It's useful - I can determine exactly where a problem occurred and report it. The fault is fixed within minutes.
Most of the competition had a turn around time of 24 hours in getting a bug fix to us.
This is where the open source in medicine should be.
Our laboratory information system is open source as is our patient database.
In 1996 if you even brought up the subject of computers in medicine at an interview they would have drawn and quartered you and used your remains to teach gross anatomy (personal experience). Now that they've opened that up, their next brick wall is allowing anyone but the most highly qualified MIT grads touch the source code used in medical applications. There's a reason why most of these medical technology companies are in the northeast. If you want to work on medical software professionally you need serious formal EE or CS degrees to the cieling. Managers in that area are more anal retentive about the formalities than Bill Gates is about using Windows. So maybe in 5 to 10 years if the medical profession becomes really really strapped for cash you'll be able to get an open source project running a patient information system but today it's more likely used as a web server, a mail transport agent, and the same drill.
I agree, mostly. I happen to work at a medtech company that is the process of developing a device (can't name it here) that is in many ways just as critical as a pacemaker, I can attest to the strength of the FDA review process. Futhermore, these companies are forced to review their own code simply to protect themselves from liability, if nothing else. There is a world of difference between doing extensive field and lab testing (which have obviously failed in the past), and verifying that the code does what it is supposed to do (relatively easy to verify).
I don't believe that Open Source would make any significant contributions in terms of development. In fact, I think it would be a really bad idea to seek out snippets of code from others. There is simply no substitute for a truely excellent and experienced programmer.
Open review might be worthwhile as a final test, though I think few would really provide review that even approaches the FDA--it is very labor intensive and requires certain detailed knowledge of the product. To really properly review the code (at a company like mine), it requires detailed knowledge of the product (as in mechanical, electrical, and optical engineering). Damn few people can even approach it...