Methods: Retrospective review of randomly selected case notes to identify drug treatments started and the problem(s) for which they were prescribed, followed by literature searches.
Results: A total of 132 treatment-problem pairs were found, comprising 85 unique treatment-problem pairs. An evidence base was found in support of 78 of the treatments started (59.1%). A further 41 treatment-problem pairs could be argued to be reasonable practice (sometimes included in guidelines), even though no published trial data support them. Ninety per cent of drug treatments started on the medical admissions unit have either an evidence base or are accepted practice.
Conclusions: Regular audit of this nature could be carried out on units admitting acute medical patients. Similar audits in internal medicine have delivered consistent results (50%–60%); there is a baseline level against which units can compare themselves. Clinical audit is an integral feature of clinical governance; all wards admitting acute medial patients could conduct similar audits on a random sample of patients.