Objective ‐ To examine the prevalence of dysphagia in idiopathic Parkinson's disease (IPD) in the outpatient setting and to determine what assessment criteria to use to select patients with dysphagia for referral to speech and language therapists. Material and methods ‐ Sixty‐four patients with IPD and 80 age‐matched controls were interviewed in clinic about their swallowing history and an objective swallowing test administered. All patients were assessed over the next few weeks by an experienced speech and language therapist and dysphagia rated according to a modified Rehabilitation Institute of Chicago Dysphagia Rating Scale and a novel global rating scale. The ability of various clinic criteria to predict patients with severe dysphagia were examined. Results ‐ Dysphagia for food was found in 30% of patients, significantly more than in controls. Swallowing speed and bolus volume were significantly lower in patients compared with controls and were correlated with declining Hoehn and Yahr score. Swallowing speed fell significantly on withdrawal of medication. The therapist's global rating score and Chicago score declined with Hoehn and Yahr score and duration of disease. However, only 10% of patients required dietary advice and none needed gastrostomy or tracheostomy. Discriminant analysis showed that various combinations of clinic selection criteria were no better than the presence of dysphagia for food at predicting which patients had significant dysphagia requiring advice from a therapist. Conclusions ‐ Patients with idiopathic Parkinson's disease should be questioned about dysphagia for food on a regular basis and, if present, should be referred to a speech and language therapist for further assessment and treatment. The outcome of this protocol should be tested prospectively.