Off-label use of medicines is bad medicine flying the flag of clinical freedom

Abstract

Let us start by recognising that, in daily practice, most clinicians do, from time to time, use licensed medicines in unlicensed indications. This doesn’t mean that the debate is a lost cause. Rather, I propose to limit the discussion to what I would label as ‘‘silly’’ or ‘‘non-evidenced’’ unlicensed use, specifically in relation to obstetrics. As clinicians we have to be a little humble and aware of the dangers of relying on ‘‘in my opinion. . .’’, ‘‘in my experience. . .’’ and in particular ‘‘I have done X of these and I have never seen a complication.’’ Proponents of the latter should be tactfully referred to the excellent statistical note pointing out that in a series of cases without complication, the upper limit of the 95% confidence interval of the probability of such an event is 1:1/3n. Thus a thousand uneventful uses of remifentanil for obstetric patient-controlled analgesia (PCA) do not exclude the possibility that catastrophic adverse events might occur with a frequency as high as 1/333. Before we embark on adventures in off-label use you might consider that the major problem for quality of clinical practice is not lack of evidence, rather it is heterogeneous practice despite evidence, in other words failure to implement the available evidence. Maynard reminds us that ‘‘translating evidence into practice requires radical changes in the behaviour in the producers of evidence’’ for example ‘‘academics,’’ so it is certainly no use just blaming the clinicians. However he goes on to comment ‘‘practitioners lack the skill and incentive to access and apply evidence. This reflects their poor training.’’ Clinicians have an ambiguous relationship with evidence-based medicine (EBM). When it suits us we are quick to cite it (for example the Cochrane Review of epidural versus non-epidural or no analgesia in labour and