Objective To determine whether choice of colposcopy or six month cytological surveillance would be beneficial to women with mildly abnormal smears when compared with the national policy of six months surveillance in terms of psychological morbidity.

Design A randomised trial based on the Zelen design.

Setting A hospital-based research clinic.

Population Four hundred and seventy-six women who had had a recurrent borderline or mildly dyskaryotic smear on routine cervical screening in primary care.

Methods Women were randomised either to six months cytological surveillance or to make a choice between that or colposcopy and were followed up for 1 year.

Main outcome measures The primary outcome measure was caseness (score ≥4) on the General Health Questionnaire at 12 months follow up. Other measures were the Spielberger State and Trait scores, default rates and cytology/colposcopy outcomes.

Results There was no significant difference between the arms for General Health Questionnaire (GHQ) scores and Spielberger State and Trait at 12 months. There was a significant reduction in psychometric morbidity between baseline and 12 months in both arms. Overall rates of default from the protocol were the same in both arms, but default that led to uncertain ascertainment of cervical pathology was greater in the no-choice arm.

Conclusions This trial indicates that having choice did not impact favourably or harmfully on anxiety or feelings of wellbeing. If a patient is anxious, allowing the patient to choose immediate colposcopy may be preferable because it will improve ascertainment of underlying disease in a group who are more likely to default.