Verified Fair Models for Predicting Progression in Alzheimer's Disease

Alzheimer's disease is a major public health problem that affects millions of people and causes gradual loss of memory, thinking, and independence. Because the disease develops slowly over many years, it is important to identify people at high risk as early as possible, before symptoms become severe. Researchers now collect many types of information that may help predict disease progression, including brain scans, memory testing, biological markers from spinal fluid and blood and genetic data. Together, these different types of information can give a fuller picture of the disease. However, in long-term studies, some of this data is often missing because people miss visits, drop out of studies, or do not complete every test. This can make it harder to build accurate prediction tools that adequately account for health differences between different groups of people. One way to overcome these challenges is by utilizing the ever-increasing capacities of artificial intelligence, or AI, to perform unbiased analyses of these data points to build prediction models that may more accurately determine who is at greatest risk for Alzheimer's.