IBM Watson being trialled for diagnosing rare and hard-to-diagnose diseases (DE)

The Centre for Rare and Undiagnosed Diseases in Marburg, Germany is known internationally as a healthcare resource for medical centers and physicians who suspect their patients have an unconventional and statistically rare medical condition. Often patient histories are inch-thick and replete with unstructured data such as X-rays, lab results, doctors’ notes and scans, reviewed by multiple medical professionals before reaching Marburg. How to review this data more promptly and better? The first step is to input this anonymized data into IBM Watson. The Marburg Centre will first ask patients to answer a 1,200-question digital questionnaire. This information will also include factors that may impact the diagnosis, from symptoms and family history to the environment they live in and their jobs. The data is also translated from German to English so that it can be cross-referenced to English-language medical information. For the 12-month pilot, Watson will provide Marburg’s staff with a differential diagnosis — a list of potential illnesses that fit the patient’s symptoms — which they then review. Watson can provide a percentage of possibility, which Marburg may skip as these are, after all, rare diseases. The Watson decision tool is being tested retrospectively on already-diagnosed patients. ZDNet

Categories: Latest News.