I used to be involved medical software. And, yes, you do go to federal pound me in the a$$ prison if MEDICAL software you create has bugs in it. The whole FDA approval thing cuts two ways, and it is a VERY sharp blade. That's why most software engineers won't write such software. How long could they stay out of prison using PHP or Python? Imagine a trauma center's server going down. Yep! Off you go to prison son. A missed tumor because your imaging code did not bring out enough contrast. That's good for 7 - 12 years. The lack of guarantees in the floating point code alone is probably enough to get a lot of these guys 5 to 10.
And yes, it has happened . . . several times!
These sorts of laws would make you code like your life depends on it . . . because it does.
Be careful what you wish for. That's all I can say.
Punishments are generally so harsh because it is considered fraud. The defendants were selling software that didn't work. They either knew it did not work, in which case it is pretty much straightforward fraud, and God help them if someone dies. Alternatively, they did not know that the software did not work. In which case they committed fraud when they represented themselves to the health care organization that purchased the device as being qualified to make a determination as to the efficacy of the software or device. Again, God help them if someone dies.
But all of that is really beside the point.
The fact is that in order to get your software FDA approved you signed off on literally thousands of legally binding sheets of paper. The short version of these sheets of paper is that you guarantee, for instance, that the software ran an 8bit lut through a 10bit colorspace so that the full range of that 10bit colorspace would be visible in the 8bit window at one time or another. That is actually a fairly standard guarantee that medical software makers have to give. If your software does not give the appropriate contrast at a given setting, then the FDA knows that you and whoever else signed that sheet of paper lied. You could not have tested it for all values. This is fraud.
Or let's say your software flips images left for right. You guessed it, there is a slip of paper you and a lot of other people sign indicating that your software will not do that. You also gave that slip of paper to the FDA to get FDA approval. Later some poor nurse somewhere is prepping the left leg for amputation instead of the right one. The doctor cross checks with the FDA approved software, and verifies that the problem is in the left leg. And . . . do you see where this is going? This too is fraud. And it is criminal in the United States. You WILL be prosecuted.
Just a couple off the top of my head, many, many more just Lex/Nex it. Or probably better, if you are thinking about getting into this area, PLEASE, talk to your lawyer.
Punishments are generally so harsh because it is considered fraud. The defendants were selling software that didn't work. They either knew it did not work, in which case it is pretty much straightforward fraud, and God help them if someone dies. Alternatively, they did not know that the software did not work. In which case they committed fraud when they represented themselves to the health care organization that purchased the device as being qualified to make a determination as to the efficacy of the software or device. Again, God help them if someone dies.
Comments
Actually,
I used to be involved medical software. And, yes, you do go to federal pound me in the a$$ prison if MEDICAL software you create has bugs in it. The whole FDA approval thing cuts two ways, and it is a VERY sharp blade. That's why most software engineers won't write such software. How long could they stay out of prison using PHP or Python? Imagine a trauma center's server going down. Yep! Off you go to prison son. A missed tumor because your imaging code did not bring out enough contrast. That's good for 7 - 12 years. The lack of guarantees in the floating point code alone is probably enough to get a lot of these guys 5 to 10.
And yes, it has happened . . . several times!
These sorts of laws would make you code like your life depends on it . . . because it does.
Be careful what you wish for. That's all I can say.
wow, that's crazy. If a doctor makes a mistake like that they don't go to jail! how does that work?? it doesn't even make sense!
As I wrote above:
Punishments are generally so harsh because it is considered fraud. The defendants were selling software that didn't work. They either knew it did not work, in which case it is pretty much straightforward fraud, and God help them if someone dies. Alternatively, they did not know that the software did not work. In which case they committed fraud when they represented themselves to the health care organization that purchased the device as being qualified to make a determination as to the efficacy of the software or device. Again, God help them if someone dies.
But all of that is really beside the point.
The fact is that in order to get your software FDA approved you signed off on literally thousands of legally binding sheets of paper. The short version of these sheets of paper is that you guarantee, for instance, that the software ran an 8bit lut through a 10bit colorspace so that the full range of that 10bit colorspace would be visible in the 8bit window at one time or another. That is actually a fairly standard guarantee that medical software makers have to give. If your software does not give the appropriate contrast at a given setting, then the FDA knows that you and whoever else signed that sheet of paper lied. You could not have tested it for all values. This is fraud.
Or let's say your software flips images left for right. You guessed it, there is a slip of paper you and a lot of other people sign indicating that your software will not do that. You also gave that slip of paper to the FDA to get FDA approval. Later some poor nurse somewhere is prepping the left leg for amputation instead of the right one. The doctor cross checks with the FDA approved software, and verifies that the problem is in the left leg. And . . . do you see where this is going? This too is fraud. And it is criminal in the United States. You WILL be prosecuted.
> And yes, it has happened . . . several times!
Citation?
Example of what happens when your medical device just plain doesn't work:
http://www.highbeam.com/doc/1N1-118387E2698885C8.html
Example of what happens when you give 'deal sweeteners' to get your device into a health care facility, or IDN (Integrated Delivery Network):
http://physicianlaw.foxrothschild.com/2009/03/articles/fraud...
Just a couple off the top of my head, many, many more just Lex/Nex it. Or probably better, if you are thinking about getting into this area, PLEASE, talk to your lawyer.
Punishments are generally so harsh because it is considered fraud. The defendants were selling software that didn't work. They either knew it did not work, in which case it is pretty much straightforward fraud, and God help them if someone dies. Alternatively, they did not know that the software did not work. In which case they committed fraud when they represented themselves to the health care organization that purchased the device as being qualified to make a determination as to the efficacy of the software or device. Again, God help them if someone dies.