Using prescribing indicators to measure the quality of prescribing to lderly medical in-patients

SIR—Batty et al. have recently reported on the quality of prescribing using a wide range of performance indicators [1]. Their inclusion of indicators that measure under-prescribing as well is welcome. However, as with all performance indicators, one should be cautious in assuming that these indicators are not without their own deWciencies. I illustrate just two problems with reference to the use of anti-thrombotic stroke prophylaxis in atrial Wbrillation, for which the prescription of either warfarin or aspirin 300 mg was regarded as appropriate. First, the group studied, elderly medical in-patients, must have included a signiWcant proportion of patients who were in the last few months of life. I am not aware that for this group, there is proven evidence of beneWt. Their criteria of ‘death of patient imminent’ [2] is too restrictive. An interesting recent approach to health promotion activities has suggested that older people could be classiWed into four groups: robust elderly life expectancy≥5 years; frail, life expectancy >5 years; moderately demented, life expectancy 2–5 years and end of life, life expectancy <2 years [3]. This classiWcation has the advantage of relative simplicity and may be a useful way of dividing older patients into groups for which prescribing an antithrombotic for atrial Wbrillation may or may not be appropriate. Secondly, these indicators do not take any account of the patient’s view of whether the beneWt is real or not and therefore whether in their judgement it is worthwhile taking the medication. We have all had experience of patients, who knowing all the information, choose not to take medical advice and refuse prescriptions. For such patients prescribing is clearly inappropriate. However, detecting that this is the case through audit will require examination of the medical records rather than just the prescription chart. It is hoped that further reWnements of the prescribing indicators may take these points into account.