retain the methodological strengths of prospective, randomized design, butuse large numbers of patients, more flexible patient entry criteria and multiple study sites to generateeffectiveness data and improve external validity.