Shared Decision Making Between Patients and Doctors Has Promise, But Obstacles Remain

“Currently, the health care system works against providers who want to make patients full partners in the medical decision-making process. Giving patients truly personalized care will require a long-term commitment to redesigning the system.”

— Mark W. Friedberg

doctor listening

In order to make health care more patient-centered, Medicare has encouraged providers to implement shared decision making. In shared decision making, providers help patients understand medical evidence about the decisions they face, while patients help providers understand their preferences and values. In theory, this leads to personalized decisions rather than “one-size-fits-all” health care.

In primary care, where there is currently little financial incentive to guide patients toward particular treatment choices, shared decision making may improve quality while avoiding unwanted and costly interventions. It may also increase patient engagement.

However, shared decision making is still uncommon; U.S. providers regularly make treatment recommendations and deliver care without adequately explaining options or soliciting input from patients.

To learn more about implementing shared decision making, RAND researchers interviewed representatives from eight primary care sites participating in a three-year demonstration.

What are the key steps for successful shared decision making?

There are four main steps for executing shared decision making between a provider and patient:

  1. Patient and provider recognize the opportunity to make a health care decision. For example, a practice might send a booklet about colorectal cancer screening options to patients who turned 50 in the prior year. Booklets and videos that give patients unbiased, expert-reviewed information about testing or treatment options are called “decision aids.”
  2. Patient uses decision aids. For example, the patient might read the booklet (or watch a video) on colorectal cancer screening, perhaps with instruction from a provider.
  3. Provider and patient have a conversation. Patients and providers engage in a “post-decision aid” discussion during which they reach a shared decision. For example, the provider might check whether the patient understands that major colorectal cancer screening options have roughly equal effectiveness but that some are more invasive and others are more frequent. The patient might then verify that the provider understands the patient's preference for more frequent over more invasive testing.
  4. Patient receives care consistent with the shared decision. If the patient receives care that is inconsistent with the shared decision—perhaps because the decision was not communicated to other providers—this process will have failed to deliver patient-centered care.

What obstacles might prevent providers from implementing these steps?

Researchers identified three main barriers to implementing shared decision making:

  1. Overworked Physicians

    Since primary care physicians typically address multiple issues during each visit, shared decision making that began with physician-driven distribution of decision aids was unreliable. Site leaders noted that reminding physicians of opportunities on a patient-by-patient basis proved unsustainable and estimated that only 10-30 percent of patients facing shared decision opportunities received decision aids.
  2. Insufficient Provider Training

    Some site leaders reported having to convince physicians that they were not already performing shared decision making, even before introduction of decision aids. Training providers in shared decision making may be essential to its successful implementation.
  3. Problems with Information Systems

    Nearly all sites' records were unable to flag patients as candidates for decision aids or indicate which patients had received them. Systems lacked the capability to track patients through the key steps of shared decision making (e.g., whether a post-decision aid conversation had occurred). Finally, systems could not integrate with decision aids. For example, even when patients completed questionnaires about their values and preferences, electronic records lacked mechanisms for making these responses available to physicians conducting post-decision aid conversations.

What could help improve implementation of shared decision making?

Among the sites participating in the demonstration, two factors emerged that could help implementation of shared decision making:

  1. Automatic Triggers

    Recognizing decision opportunities and using decision aids occurred more reliably when these steps were “automated” than when providers had to remember to take them. For instance, providers could distribute decision aids automatically based on:
    • patient demographic characteristics (e.g., age and gender) – for decision aids about screening
    • specialist referrals – for decision aids about surgical procedures
    Relative to primary care physicians, specialists typically cover a more limited range of issues. This comparatively narrow focus improved the chances of performing post-decision aid conversations in their offices.
  2. Engaging providers other than physicians

    Some sites reported that distribution of decision aids occurred more reliably when it was handled by providers other than physicians, like nurses and health educators.

What implications are there for future policy and practice?

Implementing shared decision making will be challenging for provider organizations. Those that use automatic triggers for decision aids, engage providers other than physicians in shared decision making, and develop new information systems capabilities may have the best chances of success.

This study suggests that in order to measure whether accountable care organizations (ACOs) have engaged meaningfully in shared decision making, it will be important to measure all steps in the process. All-or-none measures that give providers credit only when all key steps of shared decision making are performed for a given patient may be appropriate.

Finally, the barriers identified in this study likely stem from a common source: a payment system that prioritizes the volume of visit-based care delivered by physicians over the reliability of longitudinal care delivered by teams. Payment reforms that tilt the balance of incentives away from volume and toward greater care coordination may be a prerequisite for successful shared decision making.

For more information, please see:

More on Shared Decision Making