Inside the black box of shared decision making: distinguishing between the process of involvement and who makes the decision

BACKGROUND: Shared decision making has practical implications for everyday health care. However, it stems from largely theoretical frameworks and is not widely implemented in routine practice. AIMS: We undertook an empirical study to inform understanding of shared decision making and how it can be operationalized more widely. METHOD: The study involved patients visiting UK general practitioners already well experienced in shared decision making. After these consultations, semi-structured telephone interviews were conducted and analysed using the constant comparative method of content analysis. RESULTS: All patients described at least some components of shared decision making but half appeared to perceive the decision as shared and half as 'patient-led'. However, patients exhibited some uncertainty about who had made the decision, reflecting different meanings of decision making from those described in the literature. A distinction is indicated between the process of involvement (option portrayal, exchange of information and exploring preferences for who makes the decision) and the actual decisional responsibility (who makes the decision). The process of involvement appeared to deliver benefits for patients, not the action of making the decision. Preferences for decisional responsibility varied during some consultations, generating unsatisfactory interactions when actual decisional responsibility did not align with patient preferences at that stage of a consultation. However, when conducted well, shared decision making enhanced reported satisfaction, understanding and confidence in the decisions. CONCLUSIONS: Practitioners can focus more on the process of involving patients in decision making rather than attaching importance to who actually makes the decision. They also need to be aware of the potential for changing patient preferences for decisional responsibility during a consultation and address non-alignment of patient preferences with the actual model of decision making if this occurs.