The aim of the study was to compare standards for the process of care and 2-year survival between two cohorts of patients with head and neck cancer in the south and west of England. A total of 566 and 727 patients presented in 1996-97 and 1999-2000, respectively. The median number of cases treated per surgeon was 4 (1997, range 1-26) and 4 (2000, 1-23) and per radiotherapist was 10 (1-51) and 19 (1-70). For all 'nontemporal' standards, the overall standard increased, without reaching minimum high targets, while most 'waiting times' increased. Overall 2-year survival was 64.1% in 1997 and 65.1% in 2000. There was no difference in survival between networks (range 56-68, 1997, log-rank test 4.1, P=0.4; 62-69, 2000, log-rank test 1.26, P=0.69). Patients assessed by a multidisciplinary clinic exhibited improved survival (1997: P=0.1; 2000: hazard ratio 0.7, P=0.02), as did those with a pretreatment chest X-ray (hazard ratio 0.7, P=0.03). Despite an increased incidence, standards for the process of care for patients with head and neck cancer improved between 1996 and 2000, while waiting times increased and 2-year survival rates remained unaltered. Two out of five networks demonstrated centralisation of services between audits. Being seen in a multidisciplinary clinic correlated strongly with patient survival.