An MD&DI December 1997 Column
When I was growing up in the 1960s, the year 2000 seemed impossibly far away. It was the province of Arthur C. Clarke and the Jetsons, and other fantastic visions of the future. Now that that distant time is almost here, I have discovered that lots of older and wiser people 35 years ago were just as convinced as I that the year 2000 was merely a fiction to read about, not a reality to plan for. The result of their failure to think past 1999 is the year 2000 problem, or millennium bug, about which we have all heard so much in the last year.
Put simply, the problem results from the efforts of computer programmers, beginning in the 1960s, to save precious machine memory by truncating dates to the last two digitsi.e., to use 65 to mean 1965. At the time, no doubt, it seemed like an efficient practice. In the year 2000, however, it could mean chaosor something less severeas computer programs try to figure out whether 00 means 1900 or 2000.
Given the extent to which computers are used nowadays, the year 2000 problem could affect virtually every aspect of business for a medical device manufacturerincluding software for manufacturing and information systems as well as software embedded in devices. One critical area that may be affected is clinical trials, as author Sunil Kumar Gupta discusses on page 64 of this issue. As Gupta notes, this is a potentially serious problem, but it can be solvedprovided that companies start dealing with it now.
To my knowledge, no one has thoroughly canvassed the medical device industry to determine the extent of the year 2000 problemindeed, many device companies may not yet know the full impact. But in light of the potential risks, FDA has suggested a few actions in a letter from device center director Bruce Burlington (available on the Internet at http://www.fda.gov/cdrh/yr2000.html).
For future medical device premarket submissions, Burlington says, FDA will need evidence that products will not be adversely affected by the year 2000 problem. Such products would not seem to be in much danger except as they rely on off-the-shelf hardware, which may need to be revalidated with respect to the year 2000.
For current devices, FDA urges companies to conduct safety and hazard analyses to determine the extent of the problem. If a problem is uncovered, the companies should then correct it in devices in production and contact customers that have purchased the affected products to work out a solution.
Finally, for any design, production, or QC process controlled by computers, manufacturers should ensure that any problems are resolved before January 1, 2000.
With such advice from FDA, it would be extremely unwise to minimize the perils of the year 2000 problem. But neither does it do any good to overestimate them. The popular press has, at least until recently, tended to overdramatize the situation. Now, however, the backlash is starting, with articles suggesting that while the problem is real, it does not foretell Armageddon. Many companies are already addressing the problem, and many software programs, especially those embedded in chips, do not rely on two-digit date codes anyway.
The year 2000 problem is no doubt tied up with the millennial madness that Western society will experience in the next couple of years. Against this backdrop, device companies should carefully assess their exposure to this very real problem, but they should also do their best to ignore the hysteria that goes with it.