We assessed the effect of two different methods of contacting the target population on the rate of participation in a colorectal cancer screening programme. All individuals aged between 50 and 74 years enlisted in one primary health care centre in Barcelona (Spain) were included in a prospective randomized controlled trial. An invitation letter signed by a doctor together with two containers for faecal sample collection were sent by post to subjects in the 'standard' group (n = 1060), while subjects in the 'study' group (direct contact, n = 965) were visited by a trained non-health professional who supplied them with the same documentation as the standard group. The screening test consisted of an immunological method for the detection of faecal blood which does not require any prior specific dietary measures. Specimens were collected on two successive days. A significantly higher participation was observed in the study group (557/965, 57.7%) compared with the standard group (388/1060, 36.5%, P < 0.005). Specimen collection correctness was also higher in the study group (419/557, 75.1%) compared with the standard group (262/388, 67.5%, P < 0.014). There were no differences in terms of either age group or sex for the participation, nor for degree of correctness of specimen collection. Participation and specimen collection can be raised in colorectal cancer screening programmes by means of an invitation made through direct contact by a suitably trained non-health professional.