Advanced Clinical Decision Support System for Diagnosing Parkinson’s Disease
University of Michigan investigators lead by Drs. Dinov, Dauer and Hampstead developed a new method for diagnosing, predicting, and tracking neurodegenerative disease (PMID:27494614, http://www.ncbi.nlm.nih.gov/pubmed/27494614). This clinical decision support system predicts Parkinson’s disease with 96% consistency, accuracy, sensitivity, and specificity, in line with the rate of diagnosis by physicians. This machine-learning-based approach relies on multi-source Big Data including clinical, demographic, genetics information, and derived neuroimaging biomarkers. An added benefit of this predictive analytics and forecasting approach is that when excluding the longitudinal clinical information used in the clinical diagnostic procedure, the accuracy of the decision support system still exceeds 80%. The same methods, software, and protocols developed to study Parkinson’s disease are now being employed to examine other neurodegenerative disorders like Alzheimer’s, Huntington’s, and Amyotrophic Lateral Sclerosis.