As event rates fall, if mortality and disability are to improve further there is increasing need for large, well-designed trials. These should enroll more patients, more rapidly and at lower cost, with better representation of infants at highest risk and greater integration with routine care. This may require simpler datasets, linkage with routinely collected data, and international collaboration. It may be helpful to draw attention to recent evidence that participation in Phase III randomized controlled trials (RCTs) is at least as safe as receiving established care. Nationally coordinated clinical research networks employing local research staff may be the single most effective strategy to integrate clinical trials into routine practice. Other goals are: international standardization of outcomes; consensus on composite endpoints, biomarkers, surrogates and measures of disability; greater efficiency through randomized factorial designs and cluster or cross-over cluster RCTs; and equipping parents as partners in all aspects of the conduct of RCTs and in implementing their results.