Suspended judgment ethics, evidence, and uncertainty

Abstract

An article in the August 7, 1989, issue of The Boston Globe carried the headline A Harvard study on newborns draws fire: Doctors faulted for limiting lifesaving [1]. The story was yet another installment in the controversy concerning extracorporeal membrane oxygenation (ECMO), a recent innovation in the field of neonatal intensive care. This time, however, the results of the clinical trial and the ensuing debate were reported by the mass media 4 months before the study was published in a peer-reviewed medical journal [2]. ECMO has been proposed as a lifesaving treatment for infants with respiratory failure who have expected neonatal mortality of at least 80% with conventional ventilator support. It is a highly invasive and expensive procedure. It involves cardiopulmonary bypass, to provide adequate time and conditions for recovery of the impaired lungs. Pioneered by Robert Bartlett and associates, ECMO was tested by this team in a randomized clinical trial (RCT) [3]. The outcome was strongly in favor of ECMO, but the study design had been of the randomized play-the-winner type and provoked controversy [4]. It so happened that of 12 patients in the study, only I had been assigned to conventional treatment; this patient died, and all 11 ECMO patients survived. The Harvard study, using a different adaptive design, has been the second RCT undertaken for the assessment of neonatal ECMO. The results are again in favor of ECMO, but problems persist, involving ethical as well as medical considerations. The issues are well known: When is it appropriate to carry out an RCT? When should patient entry cease? What is the absolute mandate for informed consent? For ECMO there is the added problem of patient selection; other investigators are reporting equally good results without ECMO [5J and some are critical of the treatment given the control group in the trial. In the meantime, the technology is in a phase of rapid diffusion.