Computer Algorithm Allows for Diabetes Testing Online

An ophthalmologist at the University of Tennessee has unveiled a new imaging technique used to diagnose diabetes over the Internet. A specialized camera is used on a potentially diabetic patient to analyze the inner surface of their eye. The photo is then uploaded to a database of thousands of other images (a mix of healthy and diseased pictures) for comparison.

Algorithms then assign a disease level to the uploaded image by taking into account similar factors such as damaged blood vessels and discoloration. (Source: whytewolf.us)

While the creator of the system, Andrew Chaum, still overlooks the computer-generated matches for accuracy, he believes that the system will be able to diagnose patients with very little oversight in about three months.

Chaum has reason to be confident. The system is said to identify between 90 and 98 percent of diabetic patients, while issuing a disease level ranging from healthy to severe cases of diabetes.

Diabetes a Growing Concern in the U.S.

More than 25 million Americans suffer from diabetes, which, if left untreated, can cause a number of health-related complications, including blindness. There is also a great disproportion between the number of ophthalmologists and the number of people requiring optical attention.

It is estimated that only half of those people needing to be treated for diabetes are able to get tested. With the new system, more people can be diagnosed using less time, effort, and resources.

While the system is set to debut in the United States (a first-world nation) the technology has been implemented in mostly rural areas where access to doctors is limited. One of the initial testing areas, the Mississippi Delta region, has diabetes affecting up to 20 percent of the population. (Source: wired.com)

No Substitute for Human Interaction

While many will be interested to see how well the system works, some analysts are critical of the entire process. Their main concern rests with the potential of a computer misdiagnosis, whereas the presence of a doctor reduces this risk.

Another concern: computer programs, software and databases become static and outdated, as opposed to the active problem-solving and decision-making skills of a physician.

Although many may be reluctant to abandon their face-to-face visits, it is important to remember that the system is designed out of necessity more than luxury. The bottom line is that systems such as this could lead to quick and (hopefully) accurate diagnoses for serious medical concerns.