The British Geriatrics Society has described 16 training objectives for senior house officers (SHOs) in geriatric medicine. We have developed these into an audit standard, and in a controlled trial we consider the effectiveness of audit as a means of monitoring and improving SHO training. SHOs from three hospitals were sent a questionnaire asking about the adequacy of formal and informal training for each objective; there was a response rate of 87%. Results were presented at an audit meeting in one hospital, inadequacies of training identified and the hospital's programme of teaching modified appropriately. Two control hospitals remained unaware of the audit findings and were therefore unable to modify their training programmes. Re-audit at six months completed the audit cycle; the control hospitals showing no change between the two audits. In the intervention hospital the adequacy of training over the 16 objectives improved from 59% at first audit to 73% at re-audit. The initial audit had highlighted three target objectives as needing special attention; here the improvement from 38% to 69% was even more impressive, and statistically significant at p < 0.05 on chi 2 test. Thus, although a questionnaire approach is inevitably subjective and affected by SHO satisfaction, the discipline of audit appeared effective in identifying and correcting deficiencies in an SHO training programme.