Objective: Multicenter randomized trials are required for pragmatic evaluations of health care interventions, but recruitment is difficult. Systematic reviews failed to identify robust strategies to improve recruitment. We developed and evaluated a complex intervention to increase levels of randomization and informed consent. Study Design and Setting: The ProtecT (Prostate testing for cancer and Treatment) trial compares radical surgery, radical conformal radiotherapy, and active monitoring for men aged 50-69 years with localized prostate cancer. The intervention was developed using qualitative research methods (content, thematic and conversation analysis). Rates of randomization and immediate acceptance of allocation were measured every 6 months to evaluate the impact of the intervention. Results: The complex intervention comprised reviews of centers falling below study targets, training programmes, documents and individually tailored feedback. Over 65% of eligible participants consented to randomization. Trial participants became increasingly well informed as immediate acceptance of allocation rose from 65% to 81% between 2001 and 2005. Conclusion: This complex intervention resulted in high levels of randomization and informed consent in a difficult trial. The generic aspects of the intervention could be applied to other trials to maximize randomization and informed consent, and allow the mounting of trials previously considered too difficult.