Parkinson's disease is a common movement disorder and affects many older adults. The use of deep brain stimulation has been shown to have good results in symptom reduction, but quantitative methods for the adjustment of deep brain stimulator parameters are required. In this paper, we show that features derived from wearable sensors (accelerometers) are able to characterize changes in the severity of bradykinesia observed when turning the stimulator off and on as well as changes while the stimulator is off for a period of time. We also demonstrate results derived by means of predictors that accurately estimated the clinical scores associated with the motor activities performed during the experiments. These preliminary results are very encouraging and show the potential for the developed methodology to provide clinicians with assistance in adjusting deep brain stimulator parameters.