Hubbeling and Chang [1] are right to draw attention to issues concerning measuring patients' reactions to the care they receive from their psychiatrists in training. We accept the letter writers' point that patient feedback on medical practice can be related to factors other than their experience of the professionalism of the physicians that they have consulted. There are more specific issues with psychiatry patients because patients with serious mental illness and dementias may perceive situations in a different manner and/or they may comment on structures rather than processes.

The most important reason for measuring patient satisfaction is that there is a link between patient satisfaction with their physicians and patient outcomes. In the UK, the General Medical Council, which is the statutory regulator of the medical profession, has drawn attention to the fact that there appears to be a correlation between the number of complaints that they receive about physicians and measures of the quality of patient experiences in the hospitals where the physicians work [2]. Basically, safer hospitals have physicians that trigger fewer complaints from patients. While acknowledging that the direction of causation is uncertain, it would seem sensible to measure patient satisfaction.

A major European review of patient feedback about physicians [3] concluded that the development of good-quality instruments is a task that requires expertise and resources to achieve our findings regarding the Patient Satisfaction Questionnaire should be seen as a contribution to this work. We agree entirely that more work is required to perhaps differentiate the quality of therapeutic interaction from the physical structures within which patients are seen.