The objective of this paper is to compare the hospital costs of treating patients with colorectal cancers detected as a result of a faecal occult blood screening programme with those of patients whose cancers present symptomatically (control group). Patient-specific cost estimates are made, using case records and hospital accounts, for 360 patients over 3 years. Mean treatment costs for the group offered screening and for the control group are calculated to be 3,179 pounds and 2,966 pounds respectively, although the difference between these means is insignificant. Low treatment costs in the case of screen-detected cancers are largely accounted for by polypectomy with no subsequent readmission; in the control group case, they tend to be accounted for by early patient death. For the sample as a whole, the costs of treating very early-, and very late-, stage cancer are significantly lower than those of treating cancers in the intermediate stages. On the basis of trial evidence, the introduction of mass screening for colorectal cancer is unlikely to give rise to substantial economies in the costs of treatment.