It's called an artificial pancreas. It's basically a collection of sensors, a pump, and a tiny computer no bigger than today's insulin pumps.
The idea is that the sensors would continually test blood sugar, the computer would decide when insulin or glucagon is needed, and the pump would deliver the goods automatically.
The software is the key to the whole thing. And now the software works.
Steven Russell and his team at Harvard report they have tested their algorithm on 11 patients, collecting data and delivering medicine manually and running the calculations on a laptop.
Their description of the system in Science Translational Medicine triggered a response from MIT in the same issue, approving of the idea.
The goal is to have a device ready in 2015 for Type I diabetics, people who currently must give themselves daily injections of insulin based on regular checks of their blood sugar levels. The proposed Harvard device would include manual overrides.
The hope is that the automated testing, and inclusion of glucagon, will provide finer control over blood sugar levels, minimizing side-effects of the disease, allowing for both a more normal lifestyle and completely normal lifespan.
Amy Tenderich, community manager at Diabetic Connect and manager of her own site, Diabetes Mine, wrote recently that "usability wars" in diabetes pumps are just starting, with Roche's purchase of Medingo, makers of the Solo Patch pump. A completely automated system would turn this industry on its ear.
The Juvenile Diabetes Research Foundation, chaired by Mary Tyler Moore, has been pushing the artificial pancreas idea for years and its research director recently told HealthDay News it is jazzed about the whole thing, believing the coming device will "transform the management of diabetes." (The illustration is from the JDRF.)
The group recently announced it was working with Johnson & Johnson to bring artificial pancreas technology to market. The group has yet to produce a news release on the Harvard research. Harvard is not part of the JDRF's research consortium.
This post was originally published on Smartplanet.com