Adverse drug reactions as cause of admission to hospital

Editor—Pirmohamed et al, in assessing the burden of adverse drug reactions, excluded from their analysis any patients with either deliberate or unintentional overdose.¹ This seems inappropriate, as overdose is obviously among the risks of prescribing. The same error is inherent in the “gold standard” randomised clinical trial: it is unrealistic to expect that drugs will be used in all cases, or even most cases, precisely as intended. It is the real world experience, including misuse and abuse, which should guide our assessment of benefits and risks.

Hence, for all practical purposes, Pirhomamed et al underestimate the burden of adverse drug reactions. It would be of interest to see the results if overdoses are included. It is less clear that relapse due to non-compliance, which Pirmohamed et al also exclude, should be classified as an adverse drug reaction. However, the probability of non-compliance should be included in weighing one intervention against another, or against doing nothing.