Why We Should be Wary of Single-Center Trials

Rinaldo Bellomo, MD, FRACP, FJFICM; Stephen J. Warrillow, MBBS, FRACP, FJFICM; Michael C. Reade, MBBS, MPH, DPhil, FANZCA, FJFICM


Crit Care Med. 2009;37(12):3114-9. 

In This Article


Single-center studies are valuable hypothesis-generating investigations and remain an essential part of our craft. However, the evidence shows that, so far in the field of critical care, they have typically not been confirmed by larger multicenter studies, leading a recent review to comment that "the history of this field argues that two beneficial RCTs [randomized controlled trials] are necessary with at least one being a confirmatory trial".[49] Accordingly, we should be very careful in taking single-center trials as a demonstration of a biological truth and should avoid giving them undue weight or value. This is particularly true when such studies are unblinded. Especially in the light of the finding by NICE-SUGAR that a previously recommended intervention was not only ineffective, but actively harmful, changing practice or issuing strong recommendations on the basis of single-center trials seems premature and ill advised. Such steps should only follow after confirmation in suitably powered, robust, independently supervised and monitored multicenter, randomized, controlled trials. Fortunately, this problem is increasingly recognized,[50,51] with multicenter effectiveness trials of many recommended interventions currently under way. In general, we consider that it would be wiser for clinicians to wait for such trials to be completed and reported before making decisions about the application of a therapy only found effective in a single center.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: