Parkinson’s disease, the chronic, long-term degenerative disorder of the central nervous system, affects many people worldwide. Its symptoms, including tremor, rigidness, slowness of movement, etc., are easily recognizable, however, once the person develops these symptoms, the disease is already in its full swing.
To help Parkinson’s patients get the help they need as early as possible, researchers at the Washington University School of Medicine in St. Louis have developed a new algorithm that can predict the condition early on. The algorithm looks for signs of Parkinson’s in a patient’s medical history – signs that usually get overlooked.
The researchers started developing the algorithm by using demographic factors already associated with the development of the disease. So things like age, gender, race, tobacco smoking history, etc., were added to a simplified medical history. The team started with 26,468 codes for diagnoses or medical procedures the patients had received in the last five years, and then they identified which of those codes were predictive of Parkinson’s. They ended up with 536 codes.
The algorithm was 73% accurate in predicting who would develop the Parkinson’s, and 83% accurate in predicting who would not develop the disease.
Source:
Digital Trends (https://www.digitaltrends.com/cool-tech/parkinsons-algorithm-washington/)
The Wonderful World of Algorithms 