The objective of this study is to update a previous evidence‐based medicine (EBM) review on Parkinson's disease (PD) treatments, adding January 2001 to January 2004 information. The Movement Disorder Society (MDS) Task Force prepared an EBM review of PD treatments covering data up to January 2001. The authors reviewed Level I (randomized clinical trials) reports of pharmacological and surgical interventions for PD, published as full articles in English (January 2001–January 2004). Inclusion criteria and ranking followed the original program and adhered to EBM methodology. For Efficacy Conclusions, treatments were designated Efficacious, Likely Efficacious, Non‐Efficacious, or Insufficient Data. Four clinical indications were considered for each intervention: prevention of disease progression; treatment of Parkinsonism, as monotherapy and as adjuncts to levodopa where indicated; prevention of motor complications; treatment of motor complications. Twenty‐seven new studies qualified for efficacy review, and others covered new safety issues. Apomorphine, piribedil, unilateral pallidotomy, and subthalamic nucleus stimulation moved upward in efficacy ratings. Rasagiline, was newly rated as Efficacious monotherapy for control of Parkinsonism. New Level I data moved human fetal nigral transplants, as performed to date, from Insufficient Data to Non‐ efficacious for the treatment of Parkinsonism, motor fluctuations, and dyskinesias. Selegiline was reassigned as Non‐efficacious for the prevention of dyskinesias. Other designations did not change. In a field as active in clinical trials as PD, frequent updating of therapy‐based reviews is essential. We consider a 3‐year period a reasonable time frame for published updates and are working to establish a Web‐based mechanism to update the report in an ongoing manner. © 2005 Movement Disorder Society