Intended for healthcare professionals


Decision aids, empowerment, and shared decision making

BMJ 2014; 349 doi: (Published 25 September 2014) Cite this as: BMJ 2014;349:g5811
  1. Ian Hargraves, lead designer,
  2. Victor M Montori, professor of medicine
  1. 1Knowledge and Evaluation Research Unit, Mayo Clinic, Rochester, MN 55905, USA
  1. Correspondence to: V M Montori Montori.victor{at}

Each works or fails to work in patient-clinician conversations

Finding unexpected results when testing our ideas is the basis of learning, experimentation, and scientific method. In a linked paper (doi:10.1136/bmj.g5651), Denig and colleagues report on a randomized trial of a decision aid for patients with type 2 diabetes.1 After much well executed work in the development and testing of this decision aid, the authors concluded that it had a trivial, if any, effect on patient empowerment. What might we learn from this?

The trial tackles a real problem: the difficulties faced by patients and doctors in collaborating to target and control risk factors for adverse diabetes outcomes.2 Type 2 diabetes is both a medical disease—rarely the only chronic disease affecting patients—and a patient life issue. Targeting risk factors for diabetes requires commitment by both clinicians and patients, as it involves drug selection as well as the instigation of changes to patients’ lifestyles. This care can be overwhelming,3 and patients and clinicians should collaborate to prioritize the focus of treatment. Collaboration is, however, threatened by knowledge and power differences in the clinical encounter.

The hypothesis of the trial was that use of a decision aid would lead to patient empowerment, which would lead to patient involvement, which in turn would lead to greater adherence to treatment. This causal chain of empowerment is a reasonable idea. Indeed, it lies squarely within current thinking about shared decision making.4 The negative results of this trial might lead us to conclude that there was some flaw in the design or the use of the intervention. They might also lead us to question the widely accepted assumption of a causal chain linking the introduction of shared decision making to patient empowerment.

Both lines of questioning focus our attention on what is happening in the clinical encounter. We do not know if Denig and colleagues’ complex intervention succeeded in supporting the collaborative deliberation necessary for shared decision making. The trial set up the clinical encounter as a black box. The decision aid served as the input to the box, patient empowerment as the output of the box, and the researchers found little association between input and output. Inside the black box was the work of patients and clinicians coming together in conversation to make decisions about diabetes care. Without paying close attention to this interaction of patients and clinicians in the encounter it is impossible to fully appreciate the extent to which the decision aid has limitations in its design or use.

The authors report, for example, that several healthcare providers thought some of the information contained in the decision aid was unnecessary. Digging deeper into this suggestion requires looking at how the questionable information functions or fails to function in conversation—does it usefully contribute to patients and doctors talking through the problem at hand? With this focus, the decision aid and its constituents may be evaluated as factors operating internally, within the conversation.

Why was the encounter left unexamined? While this study focuses on the problems of patient empowerment and involvement, its view of patient empowerment assumes that a patient’s power to make decisions in clinical encounters is contingent on factors outside the clinical encounter. These factors external to the encounter include the content of the decision aid in both a paper based and an electronic format along with the training of doctors in motivational interviewing and risk communication. It follows from this assumption that the clinical encounter can be left unexamined.

We propose that the work of shared decision making—collaborative deliberation in the face of uncertainty5—happens within the encounter between patients and doctors. Decision aids function or fail to function in this environment. Successful decision aids enliven conversation and are at the same time enlivened by conversation, just as patients and doctors themselves may enliven conversation and be enlivened, or empowered, by conversation.6 A decision aid, patient power, medical skill, and scientific evidence do not simply result in good decisions by being in a room together. Each may potentially contribute, but we think that their potential is drawn out and realized in conversation.

Following this line of thinking means paying closer attention to conversations between clinicians and patients when developing and evaluating decision aids. Researchers should deploy observational methods that try to identify those factors internal to the encounter that cause the decision aid to function as a decision aid—that is, to support interpersonal deliberation grounded in evidence, the real and particular problem being discussed, and the patient’s context, preferences, and values.7

Denig and colleagues have advanced our understanding of how we should plan and conduct studies on shared decision making, decision aids, and other forms of patient centered care. Their meticulous work invites us to focus our attention on what is not seen. Their results advance the notion that testing shared decision making (and possibly patient empowerment) requires studying the logic of conversations between patients and their doctors.

This analysis should persuade researchers to partner with patients and doctors to enter and illuminate the black box of clinical encounters. In there, we shall find the conversations that shape health care.


Cite this as: BMJ 2014;349:g5811


  • Research, doi:10.1136/bmj.g5651
  • Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following: IH and VMM work on designing and testing shared decision making tools, but these tools are freely available ( without income to them, their research group, or their institution.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract