Correct staging of colon cancer is decisive regarding further oncological treatment, surveillance and prediction of long-term survival. This study investigated the variability in accuracy of pathology reports with focus on differences between pathology departments and their compliance to regional guidelines. Data from the colon cancer register (1997-2002) of the Uppsala/Orebro, Sweden, health care region were analysed and the seven pathology departments in this region were compared. Included were 3735 patients who had undergone resection of a colon cancer. Cumulative 5-year survival was the main end-point. For 64% (n = 2390) of the cases, the number of lymph nodes examined was given (median 8). Survival in stage II was lower when fewer than 12 nodes were examined or when the number of nodes sampled was not given (P = 0.001, log-rank test). In stage III, those with at the most 3 nodes positive (N1) had a better survival than those with 4 or more nodes positive (N2) (P < 0.001, log-rank test). An index of metastases (IM), derived from the number of nodes with metastases divided by the number of nodes examined, was calculated for stage III tumours. Examination of 12 nodes is necessary to assure stage III cases with the median IM (0.32), whereas 20 nodes are necessary to assure 90% of cases with the lower quartile of IM (0.16). Irrespective of the number of nodes investigated, overall survival was better among patients with IM < 0.33 vs. IM > or = 33 (P < 0.001, log-rank test). The prognostic information of the IM was higher than that of the N-stage. Quality of a pathology department, measured by the median number of lymph nodes investigated and by the proportion of reports where the number is given, was determined to indicate correct staging and management of the patient. An index of metastases (IM) is a possible basis for guidance in the choice of adjuvant treatments that appears superior to that of N-stage.