Medication Decision Aid: a shared decision-making tool for use by clinicians and patients at the Mayo Clinic in the prescription of antidepressant medications




The purpose of this tool is to facilitate conversation between the clinician and patient during the in-office process of choosing an antidepressant medication. Importantly, the goal is not to optimize decision-making or automate the role of either interactor. Instead, the goal is to draw out the expertise of each person by presenting information and interactions that facilitate conversation. A foundational concept of Mayo Clinic’s Shared Decision Making Philosophy is that “while clinicians know information about the disease, tests and treatments, the patient knows information about their body, their circumstances, their goals for life and healthcare.“ In this model, both sides are equally important.




We observed video documentation of in-office visits that featured the use of a paper card version of this tool. These observations led to a rhetoric-driven design that considers the affordances of in-office tools to create a digital version of the currently used paper aid. Digitization is valuable in this situation because it offers an encyclopedic affordance that allows for updating the tool with the latest information as medical research advances. We began with an existing prototype and developed it with these design notions in mind.


Rhetorical analysis: How do conversations unfold between clinicians and patients? How can these conversations drive design?  

After watching video documentation of conversations between clinicians and patients at the Mayo Clinic during office visits, we noticed that:

  1. Patients tend to guide their decision-making with a primary concern. For example, a patient may make a final decision after considering only one issue of primary concern. Patients tend to know what issue matters to them most, and that issue remains central.

  2. When secondary issues inform the decision-making process, they tend to be discussed  in comparison with the primary issue, and the primary issue remains central. For example, if a patient is considering sleep issues primarily and she also considers sexual issues, the primary issue takes priority over the secondary one even though it is seriously considered.

  3. When a third issue informs the decision-making process, it also tends to be discussed in comparison with, and secondary to, the primary issue of concern.

The metaphor that we kept central to our design process was the notion of turning. When two people collaborate in the decision-making process, the input of one “turns with” the other’s input. Each piece builds on the other, yet logical progression returns to prior premises in order to come to a conclusion.

Instead of a linear pattern, shared decision-making takes a more circular one, revolving until a conclusion is found. This circular pattern emphasizes participation and shared agency, while a linear pattern can be seen in traditional top-down, clinician-controlled rhetoric and reinforces an imbalanced experience of agency.

Importantly, when a patient and clinician’s conversation takes this circular pattern, what is situated at the center is the patient’s issues. We think this model fits the goal of the Mayo Clinic “to help patients make well-informed decisions that reflect their values and goals with their clinician.”

Affordances: How will the tool be used? How might the opportunities and limitations of the medium support that use? 

How will the tool be used by people?

  • physically shared by two interactors
  • unobtrusively -- it doesn't draw attention to itself

How will the tool be used in the environment?

  • in a clinic or doctor's office that cannot be assumed to be equipped with latest technologies


By making a wireframe, we learned that garnering authority for a digital tool has primarily to do with branding. In order to present a trustworthy and confident tool, we would have to modify the branding images. We role-played using the wireframe multiple times and came to the realization that the repetition of medication names across cards was distracting and unhelpful. We also learned that switching among selected secondary issues was confusing. It was also easy to make navigational mistakes. 


Iteration 1

In this iteration, we primarily fleshed out the interaction using Axure. Interactors drag and drop the primary issues into place on the screen, and they can select secondary issues by clicking or tapping. Primary issues could be changed through drag and drop as well. The secondary issues could be swiped through in a carousel.

Through role-play and in-class critique sessions, we learned that: 

  • we needed to create authority for the tool through branding
  • the secondary issue carousel was difficult to navigate and caused confusion
  • the combination of drag and drop with clicking for issue selection was confusing

Our goal for the second iteration became to develop the application's voice -- authoritative yet friendly. 

Iteration 2

In this iteration, we addressed the issue with content on the Information Page by structuring it like an online conversation with "voice" represented by left and right alternating justification. We also integrated branding by including the Mayo Clinic logo. We revised the selection events to drag and drop for both primary and secondary concerns to maintain the same interaction vocabulary. Instead of changing navigation for the "stack" of secondary issue cards, we decided to omit it altogether and limit the selection to two cards at a time. 

Through role-play and in-class critique sessions, we learned that:

  • The formatting of questions on the Information Page wasn't clear. If we wanted to use the conversational idea, we needed more visual indication.
  • Instead of creating authority, the logo seemed out of place and diminished authority.
  • The language on the Information Page was still uninviting, particularly the title and button.
  • The limitation to two cards wasn't overly restrictive because it is rarer for a patient to choose three cards. When an additional secondary issue is selected, it replaces the current secondary issue card while not affecting the primary issue card. 
  • The absence of leave-taking from the interaction made it feel incomplete.

Our goal for the final iteration was to address: authority through branding, language, and leave-taking.

Final Iteration