This is pretty much a no brainer in concept as many of the new devices and what is being approved revolves around software, thus they might think about 
understanding the basic concepts of what software (algorithms) does. Actually this could be a good move as more would perhaps jump into the consumer digital awareness arena. <Algo Man = Algorithms>
The FDA can be very transparent, but will they understand what is going on is the big question. Software unfortunately is not perfect and we have more devices that report and record data today and depend on it, so a general summary for transparency that they want won’t cut it. BD
A bipartisan group of U.S. House representatives send a letter to the Food & Drug Administration asking for more transparency and greater detail from the watchdog agency as it implements changes to the 510(k) medical device clearance process.
Six Democrats and six Republicans from the U.S. House Energy and Commerce and Health, Education, Labor and Pensions committees signed the letter, writing that at least five of the FDA's proposed changes are "controversial."
As a PhD in computer science who is on CPAP, I thought I might chime in here.
ReplyDeleteAn algorithm is simply a way to do something; for example, recipes are algorithms. On a computer, an algorithm is a way for a computer to do something -- perhaps display something on a screen or compute a square root. It may also be a way for an expert system to use information about medical conditions and symptoms to determine a likely medical diagnosis.
Software embodies algorithms; a software program is not an algorithm in and of itself.
As far as transparency is concerned, I'm more concerned about transparency in Congress -- the process for writing the Obamacare bill was about as opaque as possible.
I do have concerns about algorithms used in medical devices. Doctors do not understand the process of taking an idea or a prototype and doing the engineering (software and other) and developing it into something that can work in all the situations that it would be applied. For example if you were developing CPAP headgear, you would include headgear that worked with patients who had hair as well as those who did not -- as a CPAP user, I can tell you that headgear in the mid-90s and for not a few years later did not work well if you were not balding. This type of mistake is not made in something that is well engineered. I made my own headgear which I still use.
I agree the FDA needs an "Algo man", preferably someone who has written software since that is the focus, and someone who truly understands algorithms which is apparently rarer than I thought.