Kitchener 2004.

Study characteristics
Methods	2‐arm RCT with Zelen randomisation
Participants	712 women pre‐randomised, with mild dyskaryosis for the first time or recurrent borderline change on routine cervical screening in primary care. Aged 20 to 60, not pregnant, no abnormal vaginal bleeding Of 712 pre‐randomised women, 476 decided to participate in the study. Study arms: No choice arm (n = 243): surveillance for 12 months Choice arm (n = 233) Inclusion criteria: Recurrent borderline or first mild dyskaryotic smear Aged 20 to 60 Exclusion criteria: Pregnancy Abnormal vaginal bleeding
Interventions	Group 1. No‐choice arm (n = 243) Cytological surveillance at 6 months & 12 months, with colposcopy +/‐ treatment (LLETZ) if either follow‐up smear abnormal General health questionnaire at baseline, 6 months, and 12 months Group 2. Choice arm (n = 233) Women were given the opportunity to choose between a) and b): choice a) immediate colposcopy (n = 130) immediate colposcopy with biopsies +/‐ treatment (LLETZ) General Health Questionnaire at baseline, at immediate colposcopy, 6 months, and 12 months choice b) cytological surveillance (n = 103) Cytological surveillance at 6 months and 12 months, with colposcopy +/‐ treatment (LLETZ) if either follow‐up smear abnormal General Health Questionnaire at baseline, 6 months, and 12 months
Outcomes	Default rates reported at 6 months and at 12 months defined as women lost to follow‐up at respective time points Cumulative histology at the end of 12‐month surveillance period Based on biopsy or LLETZ‐specimen Psychological morbidity and anxiety levels (General Health Questionnaire scores) GHQ‐caseness, defined as 4 or more points from the questionnaire
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation into two study arms prior to consent using computer‐generated random numbers
Allocation concealment (selection bias)	Low risk	Central allocation used. In one of the study arms the participants themselves chose the intervention group.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and investigators were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcome assessment was not blinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	712 pre‐randomised, 476 participated. Reasons for non‐participation reported, no clear differences between randomisation groups 172/476 were lost to follow‐up, no specific reasons stated. Adequate sample size might not have been achieved for all outcomes.
Selective reporting (reporting bias)	Low risk	Pre‐trial protocol was not available. Results still reported comprehensively
Other bias	High risk	Project funding reported; conflicts of interest not stated. Local ethical approval and informed consent from participants obtained. Missing information still considered not to introduce bias. Used Zelen randomisation, where some participants were able to choose whether to have immediate colposcopy or cytological surveillance, which might well bias the results included here.