HURDLE REGRESSION MODELING: PARKINSON'S DISEASE DATA, SPECIFICALLY CLINICAL PARTICIPANTS OF MOVEMENT DISORDERS SOCIETY UNIFIED PARKINSON'S DISEASE RATING SCALE

Abstract

Parkinson's disease is a disease that attacks the motor parts of the body so that it can reduce the quality of life of sufferers. Treatment is limited to dopaminergic and physical therapy. This study aims to determine the factors suspected of influencing motor complications, especially in clinical participants of the Movement Disorder Society's Unified Parkinson's Disease Rating Scale (MDS-UPDRS) who are diagnosed with Parkinson's disease. MDS-UPDRS is an assessment instrument used to measure the severity of motor complication symptoms. Motor complication data obtained from MDS-UPDRS shows excessive distribution characteristics at excess zero values (zero inflation), which indicates overdispersion due to excess zero. Therefore, Hurdle regression is needed to overcome this problem. The predictor variables used are part of the measurement of MDS-UPDRS, namely motor aspects, non-motor aspects in daily life, and motor examinations by medical professionals. The results shows that Hurdle negative binomial regression model was better than the Hurdle Poisson and Hurdle Conway Maxwell Poisson when applied to motor complication data. Based on the Hurdle negative binomial regression model, it is known that clinical participants who do not experience motor complications are significantly influenced by non-motor aspect variables in daily life. In addition, in the count model, each increase in the score on the non-motor aspect variable tends to increase the average motor complication score by 1.0217.

Access to Document

Other files and links

• Link to publication in Scopus
• Open Access Version Available

Fingerprint