Re: Status of uw-prism packaging for Debian

["Steve M. Robbins" <steve@sumost.ca>]

Hi,

I'm tuning in late, but regarding Ira Kalet's question:

> 3. Finally, there is still the issue of what the US FDA might say
> about distribution within the US, 

We have to worry about more than just the US.  Almost every country regulates 
the marketing of medical devices.


> as software products that do what
> Prism does are considered medical devices and cannot be distributed
> without FDA 510K premarket approval, an onerous process to be sure.
> It does not matter that no money is involved.  What do you do about
> other debian-med packages?  Might any of them be considered a
> medical device?

In my opinion: yes, almost certainly.  PACS software, for example, is 
regulated as a medical device and Debian has some DICOM viewers if not full 
PACS implementations. 


Ira states that approval by FDA is required even if "no money is involved".  
In my superficial search, I didn't find any reference from the FDA stating this, 
but it wouldn't surprise me since I did find such a statement [1] from Health 
Canada:

	Note that the Regulations apply to Free and Open-Source (FOSS) device
	software, since the definition of "sell" in the Food and Drugs Act,
	includes transactions without compensation.

[1] http://www.hc-sc.gc.ca/dhp-mps/md-im/activit/announce-annonce/md_notice_software_im_avis_logicels-eng.php


I don't know whether this issue has been considered within Debian before.  But 
I'd like to point out that there is a long history of free/open source PACS 
and DICOM viewers [2] including the very well known Osirix [3] that can be 
downloaded for free (download is described as "not cleared for clinical use") 
or purchased as a fully approved medical device.

[2] http://www.medfloss.org/taxonomy/term/84
[3] http://www.osirix-viewer.com/


My guess is that the legality of distribution hinges on how the software is 
represented.  For example, [1] defines "device" as:

	any article, instrument, apparatus or contrivance, including any
	component, part or accessory thereof, manufactured, sold or represented
	for use in

	a.    the diagnosis, treatment, mitigation or prevention of a disease,
	       disorder or abnormal physical sate, or its symptoms, in human
               beings or animals,

	... etc.

If the software makes no claim about diagnosis or treatment, perhaps it's 
still OK, as in the Osirix case.


Note that I'm not a lawyer so these opinions are worth what you paid for them.  
I do, however, write medical device software in my professional life that has 
been cleared by regulatory bodies in US, Europe, Canada, and elsewhere.

Regards,
-Steve