O retain the methodological strengths of prospective, randomized design, but use large numbers of patients, more flexible patient entry criteria and multiple study sites to generate effectiveness data and improve external validity fewer types of data may be collected for each patient easing the participation patients and also clinicians