Three areas of intense investigation in Parkinson's disease clinical trials include symptomatic treatment of Parkinsonism, disease-modifying therapy (or neuroprotection), and the prevention and treatment of motor complications of dopaminergic therapy. Difficulty interpreting the results of many studies in recent years has been attributed to problems with the chosen outcome measures. This article reviews the most common outcome measures used, assesses their positive attributes and proposes needs for future research. The Unified Parkinson's disease Rating Scale has been extensively validated and is by far the most common outcome measure used in trials of symptomatic therapy. Ambiguities in the response scale descriptors, poor inter-rater reliability of some items and a lack of items addressing nonmotor features of the disease are being addressed in a revision of the scale. Quality of life outcomes are being used in the minority of clinical trials, and no single generic or disease-specific quality of life measure is being used most frequently. Additional work validating several of the disease-specific instruments is needed. When a generic measure is used, its validity for use in Parkinson's disease must be critically assessed despite its previously established validity in other diseases. With respect to measuring motor complications, significant unmet needs include a consensus as to the best way to define the first motor complication and validating time to the first occurrence of motor complications as a surrogate of future disability and quality of life. Measuring the effectiveness of a potential neuroprotective agent presents unique challenges, particularly since symptomatic effects of the experimental agent or concomitant treatment can obscure any neuroprotective effects. Study designs and biomarkers are being developed that may overcome this problem. Currently, neuroimaging techniques that reflect function of the dopaminergic system are the most promising biomarkers but still require additional validation.