AIM: To summarise the magnitude and mechanisms of surgeon-related variability in the outcome after radical resection for rectal cancer and to present a solution and targets. METHODS: A review of the literature, consultation of the "Guidelines for the management of colorectal cancer" published by the Association of Coloproctology of Great-Britain and Ireland, and analysis of data from the database of the Belgian Ministry of Health, RIZIV-INAMI, on radical resection for rectal cancer in Belgium during the years 1995-1997. RESULTS: The proportion of abdominoperineal excision of the rectum (APER) varies between 23-58% in specialised centres and 43-57% general practice. In Belgium the APER rate for rectal cancer located between 4 and 16 cm above the anal verge is 50% with an overall in-hospital mortality of 3.5%; both APER rate and postoperative mortality are lower in university than in community hospitals. Most studies observe an effect of specialisation, reducing mortality with a factor of 2.5-3. The magnitude of surgeon-related variability in the oncological outcome has been well documented indicating that the impact of the surgeon-factor is considerably larger than that of adjuvant therapy. When comparing subspecialised with general surgeons, relative risk factors of 0.3-0.8 are reported for local recurrence rate, and 0.7-0.8 for disease free survival. CONCLUSION: Inter-surgeon variability is to be related with surgical skill and adequate implementation of recent diagnostic and therapeutic methods. Guidelines and centralisation are appropriate concepts, but do not guarantee improved quality of care. The targets are an APER rate of < 40% with an operative risk of < 2%, a local recurrence rate of < 10% and a disease free survival of > 70%. External audit is essential, but subspecialty training is a prerequisite and the surgeon is not the only factor to be audited.