Implementing Shared Decision Making In The Real World

You read the Disease Management Care Blog and use its literature-based links.  As a result, you know a lot about shared decision making (SDM).   Based on the randomized control clinical trial data that are regularly linked by the DMCB, you're ready to implement it in your network of primary care clinics.  You call your physicians together for a meeting, explain SDM is the best thing since the invention of the Woods lamp and ask them to refer patients to the program.

Except for the part about the DMCB, that's what basically happened at the eight highly regarded medical institutions: Dartmouth-Hitchcock Medical Center, MaineHealth, Massachusetts General Hospital, Mercy Clinics in Iowa, the Oregon Rural Practice–Based Research Network, the Palo Alto Medical Foundation, the Stillwater Medical Group in Minnesota and the University of North Carolina at Chapel Hill.  The decision aids covered approximately 50 common conditions such as knee arthritis and prostate enlargement

And it didn't work out.

To investigate what happened, investigators at RAND interviewed 23 "key informants" from the 34 primary care sites that had implemented the Informed Medical Decision Foundation's shared decision making program. The interviews were conducted between December 2010 and March 2011.  Their study is published in the February issue of Health Affairs.

Three barriers were identified:

Overworked physicians: given all the other health care needs of their patients, it was difficult to count on physicians to distribute the decision aids.  As a result, only 10%-30% of eligible patients participated.

Overconfident physicians: many doctors felt that they were already providing sufficient patient education or that their patients would not benefit from the decision aids.  Just telling the docs that this worked was not enough.

Underperforming EHRs: the state-of-the-art information technology could not flag potential patents, remember which individuals had been exposed to the decision aids or record the patients' preferences.

The answers?

Well, if the someone from the eight institutions had simply called the DMCB and asked, it would have told them about the many times it explained stuff to a roomful of physicians and had zero impact.  There also good peer-reviewed literature on why it's so difficult to change physician behavior. 

The DMCB ultimately found the answer is to take physicians out of the work flow.

That's what the RAND researchers found.  Based on their interviews, they recommend that if you're going to implement SDM in your clinics:

 Automate the process as much as possible and remove human decision-making from the process the triggered the decision aid.  That could be done on the basis of pre-existing clinical criteria or when a specialist referral had been arranged.

 If automation was not feasible, rely on non-physicians to trigger the decision-aid.  For example, office assistants could offer SDM to patients in the course of check-out.

Image from Wikipedia