personal medical devices

One of the more interesting topics that came up at lunch today was the work being done by Ben Shneiderman's students on interfaces for "personal medical devices"--like the monitors used by diabetics to record blood sugar, for example. Coincidentally, one of the faculty members at the lunch was a diabetic, and she was wearing an automated insulin pump--which spurred some interesting dialogue.

Shneiderman told us about a physician he's been working with at Johns Hopkins who wants to work on how these devices record and report data, so that they (a) better match patient needs for record-keeping (think about all the ways Cory Doctorow criticizes human metadata assignment, and then extend to the even more critical data that medical patients are often expected to record about themselves...the potential for accidental or intentional error is enormous), and (b) better match physician needs for analysis. Even when a patient properly records blood sugar 4x/day in a 30-day log, for example, that information isn't generally in a form that's useful to the medical practicioner.

But, says Shneiderman, NIH is unwilling to fund research into interfaces to devices, and visualization of data--because medicine has traditionally not focused on these "peripheral" areas. In Leonardo's Laptop, Shneiderman uses some other disturbing examples of medical interfaces gone wrong--with disastrous results--and points to Peter Neumann's document Illustrative Risks to the Public in the Use of Computer Systems and Related Technology. (Neumann is the moderator of comp.risks on Usenet, as well as chair of the ACM Committee on Computers and Public Policy.)

This is one of the great frustrations for me (and many others) in academia--the pigeonholing of research, and the difficulty in obtaining support and funding for research that crosses over disciplinary boundaries, no matter how important it may be. My hope that weblogs are a start towards breaking down those boundaries may be naive, but I cling to it anyways...