Advancing the standards of clinical research: the urgent need for new methods and better data

Some of the weaknesses of the randomized controlled trial (RCT) are pointed out, among them unavoidable biases, the possibility of coming to erroneous conclusions by the 'luck of the draw' and the fact that most such experiments lack a doctor/patient relationship and are of little assistance in clinical decision making. An exaggerated propensity on the part of the medical profession to regard the RCT as the only 'scientific' method for research into therapeutics is noted. The potential advantages of post-marketing surveillance of new interventions in a physician-centred manner are outlined. Such a data base could ultimately be expanded into a data base that records all activity in a given practice. The PACE strategy, employed by haematologists in the north of England, and the sentinel practice primary care research networks used by family doctors contain elements of the system proposed here. However, each possesses disadvantages that limit its applicability to special cases. Random choice of practices for any particular observational study of a new intervention could lend it some of the features of an experiment.