Chronic cough is common in the community and can cause significant morbidity. It is not clear how closely treatment guidelines are used in general practice, or how often specialist referral is indicated. We aimed to assess the management of chronic cough in primary care before referral to a cough clinic, and to assess the outcome of managing chronic cough with an approach of simple investigation and empirical treatment trials. Data were extracted from the records of all patients attending a district general hospital respiratory clinic over a two-year period with isolated chronic cough lasting ⩾8 weeks. The clinic assessed symptoms with a cough-severity visual analogue scale and the Leicester Cough Questionnaire. Among 266 patients, the most frequent diagnoses were asthma (29%), gastro-oesophageal reflux (22%) and angiotensin-converting enzyme inhibitor use (14%). In all, 12% had unexplained chronic cough. Common diagnoses had often not been excluded in primary care: only 21% had undergone spirometry, 86% had undergone chest radiography and attempts to exclude asthma with corticosteroids had been made only in 39%. In the clinic few investigations were conducted that were not available in primary care. Substantial improvements in symptoms occurred with a median (interquartile range) total of 2 (2-3) clinic visits. We estimated that 87% of patients could have been managed solely in primary care; we did not identify distinguishing characteristics among this group. Most cases of chronic cough referred to secondary care could be managed with a simple and systematic approach, which is potentially transferrable to a community setting.