preview

Diabetes Essay

Better Essays

Over 415 million people globally have either type I or II diabetes. The total is expected to reach 600 million by 2040. It is a heavy burden to the healthcare system, costing over $240 billion a year in the United States alone. It is a condition which largely needs to be managed by the individual, testing and managing glucose levels to maintain the appropriate amount of sugar in his or her blood. Over time, high blood sugar can lead to cardiac disease, lost limbs, renal failure and blindness.
Medtronic recognized the need to provide tools for diabetes to manage their diabetes more easily than regular finger sticks. They already market continuous glucose meters and insulin pumps to help patients manage their condition on a daily …show more content…

In 2013, IBM told Forbes that it would be tackling clinical trials and be in use with patients within months. After four years with M.D. Anderson, it had not produced any tool ready to go beyond pilot testing.
For all its complexities, Watson can not exist without being fed knowledge by humans. To learn how to make a cancer diagnosis, it must be programmed with thousands of cancer diagnoses. If there is an anomaly that has never been seen before, Watson will not necessarily have the capability to solve it. Missing or incompatible data could seriously skew a result. This is not a problem unique to Watson; all healthcare applications of machine learning suffer from this dilemma. https://www.technologyreview.com/s/607965/a-reality-check-for-ibms-ai-ambitions/
Another problem resulting from the information fed to Watson is bias. For example, all of the software being tested in 30 cancer centers worldwide was programmed by Memorial Sloan Kettering Cancer Center. MSK is an American company which manages highly complex cancer cases, primarily for the affluent. Although it is a premiere cancer center, the solutions they would implement may not be appropriate in a country with different types of insurance coverage, or different levels of disease prevalence.
The most obvious answer to reduce treatment bias is to enter more global information. However, that exposes yet another problem with Watson. Either the EHR system needs to be compatible with Watson’s

Get Access