OIRM: Too small trial? unable to detect clinically important effects. Too large? Lack of uniformity may show null effect when subset can benefit.