|
After my recent posts on the future of open source biotechnology, I received an email from Ray Van De Walker to provide a contrary viewpoint:
Briefly, my credentials:
I have seven years of experience developing biomedical software. Several products have achieved FDA and CE approvals, and are in mass-produced use by patients.
Here's the problems:
1. The market itself does not tolerate design defects. The established standard is less than one unanticipated risk per billion patient hours, and the evidence trail has to hold up in court or sellers will be successfully sued by patients for whom the therapy failed.
2. In the U.S., to advertise or sell a therapy, the FDA requires evidence of safety and effectiveness. The European union just requires evidence of safety; effectiveness is left to doctors and patients.
a) Therefore the therapy has to have traceable continuity in testing or safety engineering, including maintenance of test results at an agency or person that can be held liable by the FDA or (in the case of the EU) a government's regulatory agency. Most organizations and sane people will not accept liability without an income stream to pay for the liability insurance, which of course is extremely expensive for medical therapies.
b) To meet effectiveness requirements in the U.S. there must be patient trials to prove the therapy. The patient trials are extremely expensive, because the doctors, hospital time, etc. are expensive (Most do not work pro-bono). I.e. close to quarter billion dollars for a major drug using a new metabolic pathway to be certified. Medical devices (much cheaper) only cost 25 to a 100 million if the therapy they perform is not experimental.
Contrast that with open-source software, which basically says that if there's a defect, then it's up to the user to fix it or live with it.
Production would not be a problem, I think. The market selling open source therapies could be much like that for generic drugs. The producers would have to have certified quality organizations, but generic producers already do that.
For my part, I have to say that agree with these two central points:
- Centralized regulation kills open source approaches to development for exactly the same reasons it damages the market - imagine the consequences of centralized regulation of server software. We'd still be paying $50,000 for 200MHz machines and hand-testing factory RAM for errors.
- Many medical applications require a level of reliability that traditional-model open source groups do not approach. The same is true of many software applications.
I do think that we can fairly quickly point to a large swathe of medicine and medical science that could benefit from improvements in lower-reliability methodologies. The practice of neurosurgery will probably not be a part of this swathe - but I'm willing to bet that most in vitro and simulational medical research certainly will be.
A larger range of medicine would also benefit from lower-reliability methods in the absence of centralized, monolithic, risk-averse regulation - say, with distributed, competitive, accountable review instead, allowing individuals to adopt risk levels as they feel appropriate, just as in all other walks of life. Most applications of medicine and most basic medical research does not need the fault tolerance expected of air traffic control or rocketry systems ... or neurosurgery for that matter.
|
|
Posted by Reason
|
|
|
|
