OBJECTIVE--To test patient compliance for faecal occult blood testing in suburban and inner city general practice. DESIGN--Prospective opportunistic trial using the Haemoccult test kit. Tests were offered during routine surgery attendance. SETTING--Three group general practices in Birmingham. SUBJECTS--All patients aged 40 years or older on the start date who routinely attended surgery during two years. MAIN OUTCOME MEASURES--Numbers of patients approached for testing and the numbers refusing, accepting, and returning the test kits. RESULTS--Only 26.3% (1230/4677) of the potential target population had been screened within the two years, although 988 (39.3%) of the suburban practice target were screened. However, 55.7% (1230/2207) of patients actually offered a test returned completed kits, with only 6% (133) refusing the kit. 683 (61.6%) patients aged 50-69 returned kits, compared with 343 (54.3%) aged 70 or over and 204 (43.8%) aged 40-49. These differences were significant (p less than 0.001). Patients from the inner city practice were significantly less likely to be offered the test than those in suburban practice (242 (11.2%) v 988 (39.9%), p less than 0.001) and return the samples (242 (38.8%) v 988 (62.4%), p less than 0.001). Patients from the inner city practice were also more likely to refuse the test (78 (12.5%) v 55 (3.5%), p less than 0.001). CONCLUSIONS--Opportunistic testing for occult faecal blood in asymptomatic patients was reasonably acceptable to patients, especially those in a suburban practice. If the test is shown to reduce mortality from colorectal cancer then formal screening would probably achieve acceptable target rates, especially among patients aged 50-69, who represent the prime risk group.