The purpose of The Michael J. Fox Foundation’s Parkinson’s Data Challenge was to develop a way to help patients and clinicians using objective, passively collected data points from smartphones. The goal was to use the data to distinguish Parkinson’s disease patients (PD) from control subjects and/or to quantify PD symptoms in a way that could enable the measurement of disease progression.
Through the Challenge, we presented and validated that a machine learning approach is superior to conventional statistical methods for the detection, monitoring and management of PD via passively collected smartphone data. In spite of the very sparse data of the Challenge’s subset, we were able to predict incidence and monitor progression of the disease with 100% accuracy. In addition, a machine learning approach paves the way for disruptive innovation in the monitoring and management of the disease as an aid to clinicians and patients alike.
Relevance to Diagnosis/Treatment of Parkinson’s Disease:
With the proliferation of smartphones among the global population, passively collected data from Parkinson’s patients offers great promise to helping clinicians and patients themselves better understand disease progression in between visits with their doctor.
Lionsolver’s winning submission to the Challenge demonstrated the value of using smartphones as an aid to supporting clinical and self-management of PD. Further research will investigate how to make the data more useful and engaging to doctors and their patients, and to understand the best way to deliver the data through a web-based service.