Patient empowerment is a high priority for healthcare policy makers in many countries. By increasing the role of patients, health care providers should become more responsive to patients' needs and preferences and deliver better quality care. Patients can participate in health care in many ways. These include communicating directly to healthcare professionals in patient centred consultations; contributing to routine practice outside the consultation through quality of life questionnaires and patient satisfaction surveys; and, increasingly, by using information on health and health care through internet channels.

Many questionnaires on patient satisfaction have been developed in the past decade but they do not all meet the needs of either the consumers or the healthcare professionals.¹ Before being used on a wide scale, new approaches to capturing patients' views need to be assessed in well designed studies with a similar stringency to that which is applied to clinical interventions. In this issue of Quality in Health Care Grogan et al describe the validation of an instrument for measuring patient satisfaction with general practice.²

Users of these patient satisfaction instruments should have several concerns. Firstly, they need to be assured that the instrument does provide valid and reliable data. While there is no gold standard, Grogan and colleagues have shown that their instrument has good content and construct validity with internal consistency. Getting patients to contribute to the choice of aspects of care included in the questionnaire is crucial to making sure that it is their priorities on health care—and their priorities as seen by healthcare professionals—that the questionnaire reflects. For instance, patients in general practice value having sufficient time in a consultation, seeing a healthcare professional with good information giving skills, receiving effective treatment; and the availability of a practitioner in case of emergencies.³ Organisational aspects of care such as waiting times tend to be less important to them. If the instrument is used to assess practitioners or practices, it should also provide reliable figures at that level of aggregation. This type of reliability requires a sufficiently large number of respondents per practitioner, probably at least 60.⁴

Secondly, it should be feasible to use the instrument in routine clinical practice. It is important that its use should not absorb a disproportionate amount of resources—either time, material, or effort. Many practitioners may require external support, particularly for analysing patient surveys and translating results into a usable format. It is also important to make sure that a patient satisfaction instrument is acceptable to both staff and patients. The purpose of using such instruments needs to be clearly understood by everyone concerned. Some practitioners may be reluctant to perform patient satisfaction surveys but may be encouraged to do so through a focus on learning and quality improvement rather than research or accountability. Acceptability for patients is also important. The low response rate in the study by Grogan et al is worrying and may indicate a lack of motivation by patients—a factor which reduces the generalisability of the findings. Response rates in general practice can reach 70% or higher if the questionnaire is carefully designed, is handed out by practitioners rather than by assistants, and if reminders are sent to patients after a few weeks.⁵

Thirdly, the instrument should have proven effectiveness as a tool for quality improvement. Thousands of studies on patient satisfaction have been performed but very few were designed to assess the effects of feedback on patient satisfaction with either process or outcomes of health care. One approach to using patient satisfaction questionnaires to stimulate quality improvement is to publicise the results.¹ Of course, case mix adjustments are needed. For instance, practices with a large number of older patients may expect more positive evaluations of care than those with a younger age group. Research on patients' attitudes and behaviour suggests that, when needing health care, few individuals behave as “rational consumers” and compare and choose actively between different care providers,⁶ but this may change in the next decade. The reality is that patients often simply lack real choice because of a shortage of care providers.

For patient satisfaction to become a tool for improving the quality of care, not only do the instruments themselves need to be well validated and tested but the results need to be fed back to practitioners and co-workers in a way that enables them to learn and change. If this is not done, the views of the patients may be collected but will become redundant if no action is taken. An educational approach is probably the best option and should be integrated within a programme of continuing professional development and quality improvement.

Patients' views should not be treated separately from the delivery of clinical care. Empowering patients and understanding their views should be central to activities such as the implemention of evidence-based practice and development and dissemination of clinical guidelines. Evidence-based medical care makes no sense if patients' preferences are ignored. Most patients want their practitioners both to respect their preferences and to provide adequate information about effective treatment options. We look for evidence of effectiveness of clinical interventions. We need to be just as critical about methodologies that purport to promote patients' views.