• quality indicators
• mental health
• primary care

The involvement of patients in health care is now a matter of government policy, at least in the UK. Partly as a result of the scandal surrounding paediatric heart surgery in Bristol and the ensuing enquiry,¹ public and patient involvement is now firmly on the agenda. To researchers who have been advocating patient centredness and shared decision making for years, if not decades, this is a welcome initiative. However, neither policy makers nor researchers have made much progress in working out how patient involvement is to be measured. Involvement can occur at many levels—from citizens’ juries to patient participation groups—but perhaps the most important arena for many people is the face to face consultation with a clinician.

The value of the carefully conducted work by Elwyn et al² published in this issue of QSHC is, firstly, that it established—by means of a systematic review—that there was no existing measure of the extent to which clinicians involve patients in decisions within consultations and, secondly, that it has succeeded in developing such a measure. The resulting instrument, OPTION, will enable researchers to investigate the links between process in the consultation and subsequent outcomes. The literature is full of theoretical and abstract models which in many cases have not been turned into research tools.³ The OPTION instrument will allow the model of shared decision making to be evaluated in terms of its occurrence in actual practice and its consequences for both patient and professionally defined outcomes. It will allow researchers to explore whether or not shared decision making leads to better clinical decisions, the achievement of patients’ goals, higher levels of adherence, and better health outcomes. Another important question which can now be addressed is the extent to which patients from different social backgrounds have different expectations and/or are treated differently.⁴ The results of the study by Elwyn et al showed that the doctors in their sample rarely established patients’ preferred level of involvement in decision making. Thus, if doctors are treating less educated patients differently from those with more education, or if they are less likely to involve patients with fewer resources in decisions about their care, such differences may reflect doctors’ assumptions and not the actual preferences of these patients. Such differential involvement, if it exists, may help to explain the inequalities in access to and take up of health care. The instrument developed by Elwyn et al will allow these and other possibilities to be explored.

However, as the authors acknowledge, OPTION does not measure the patient’s contribution or the actual interaction between patient and professional. To that extent it is a professional or doctor centred instrument. The “skills framework” identified refers to the skills of the professional, not of the patient. It can be argued that the onus is on professionals to change their behaviour as they spend more time in consultations than patients usually do. In fact, in any interaction both (or all) participants are usually contributing in some form or other. The level of variation within practitioners in the study by Elwyn et al, which exceeded variation between practitioners, suggests that patients’ behaviour might explain at least some of what practitioners were doing. You cannot understand fully what clinicians are doing if you do not also look at what patients are doing, which is to say that health care consultations are co-constructed.⁵ As well as establishing patients’ preferences for involvement in decision making, researchers need to examine what patients of all backgrounds are actually doing in the consultation already, and the ways in which current patterns of involvement or non-involvement are co-produced by clinicians and patients.