New algorithm uses medical history to predict Parkinson’s disease
NIEHS-funded researchers developed an algorithm that identifies people with a high probability of being diagnosed with Parkinson’s disease (PD), which is a debilitating movement disorder characterized by tremors, movement slowness, and difficulty with balance and coordination. The predictive model relies on demographic data, as well as medical tests and diagnoses available in Medicare claims data.
Researchers analyzed the Medicare claims of more than 200,000 people to look for associations among a diagnosis of PD, medical conditions, and demographic factors. Using this information, they developed a model that correctly predicted 73 percent of people who would be diagnosed with the disease in 2009, and 83 percent of the people who would not.
Much of the Medicare claims information that helped predict the disease referred to problems known to be associated with PD, such as tremors, posture abnormalities, psychiatric or cognitive dysfunctions, gastrointestinal problems, fatigue, and trauma. Other factors associated with the disease included weight loss and chronic kidney disease. Factors associated with a lower probability of PD included cardiovascular disease, obesity, cancer, gout, and history of tobacco smoking.
According to the authors, this algorithm could provide an early indicator to physicians that patients may need evaluation for the movement disorder. This method may also be used to identify other factors, such as environmental hazards, that may be associated with the disease or have a protective effect.
Citation: Searles Nielsen S, Warden MN, Camacho-Soto A, Willis AW, Wright BA, Racette BA. 2017. A predictive model to identify Parkinson disease from administrative claims data. Neurology 89(14):1448–1456.