My Role


Literature Review, Interviews, Analysis


Paper Mockups, Wireframes, Prototypes


Poster Feedback Sessions, User Evaluations



  • Login feature for security and privacy
  • Social Sensing component that displays feedback from trusted others
  • In Vivo component that alerts the clinician as to whether a peer thinks the patient is fully engaged during their in vivo session
  • Messaging component that allows a clinician to message a trusted other with specific questions

Research & Analysis

Literature Review

Understand existing issues with PE therapy and PTSD treatment in general

Literature Findings

  • Current technological developments are patient-focused
  • Novice clinicians can be overwhelmed by manualized process
  • Clinicians may fear pushing the patients too far or fail to help them engage properly

Task Analysis

Understand the manualized therapy and gold-standards of its practice in outpatient PTSD treatment

Literature Findings

  • Identified 7 major opportunities for technological intervention

Clinician Interviews

Understand how therapy is practiced in the real-world and the problems that clinicians face

We conducted 12 semi structured interviews with PE clinicians with the goal being threefold:

  1. To understand the challenges clinicians face when delivering PE therapy
  2. To better comprehend how clinicians assess patient progress
  3. To determine what data clinicians are missing from their existing processes

We used Braun and Clarke’s Thematic Analysis method to analyze our data. Using an open coding method, we generated over 120 excerpts which were then coded into 28 themes. Three main themes emerged that informed our design process.

  1. First, clinicians wanted a more efficient workflow in order to focus their energy on the therapeutic tasks with the patient rather than on collecting or analyzing data.
  2. Second, while self-reports are an important part of monitoring patient progress, clinicians struggled to effectively monitor all of the self-report data due to the large amount of self report homework that needed to be processed and analyzed.
  3. Third, clinicians would like a better way to assess their patients' symptoms and behaviors outside of therapy and expressed a desire for more frequent assessments of symptoms in the patient's home, and information about the patient's daily life.



After our data analysis, we started ideating on different ways to represent solutions to the pain points that clinicians faced. Based on these themes, we came up with 2 main design directions, a social sensing system, and a clinician facing dashboard to help track homework and guide therapy

We proposed four use cases in which social sensing could be leveraged in PE therapy

  1. After group sessions (in which group members would provide data for the patient)
  2. After individual sessions (in which John would provide an assessment of how he thought the clinician visit went)
  3. After in vivo sessions with peers (in which peers would indicate how engaged John was during his in vivo session)
  4. A general social support use case (in which trusted others would inform the clinician as to how well/poorly John has been doing, as well as any concerns they had)


Poster Feedback

We conducted a poster feedback session with users to see which scenarios we should move forward with. We found that scenarios 3 and 4 were the most useful to clinicians. From our poster session, we developed wireframes which we then developed into a hi-fidelity prototype





Social Sensing

Social Sensing

In Vivo

Sent Message

Recieved Messages


After coming up with our basic screens, we came together again to talk about the color scheme for our app. In talking about the "personality", or feel, for our app (described above), we decided that we wanted to go with a mentor sort of feel – friendly and approachable, but also in a position of authority. We want people to be comfortable using the app, but we also want them to trust it. We kept this in mind as we picked the color scheme. Ultimately, we decided on purple and yellow, with variations on these shades of purple and yellow being determined by Google's material design palettes.

Hi-Fidelity Prototype

Login Screen

Login Hover

Sign In

Sign In Hover


Patients Hover


Jane Hover

Social Sensing

Social Sensing Hover

Partner Expand

In Vivo

In Vivo Graph


New Message

Recipient Search

Select Recipient

New Message to Recipient

Social Sensing Drop Down

Social Sensing Multiple

In Vivo Drop Down

In Vivo Multiple

Social Sensing Send

Message Sent

User Evaluations

Understand how clinicians would use our systems and gather feedback on the systems

We conducted 5 user evaluation sessions with PE clinicians to determine the usability of the Social Sensing System


Device Usage Survey

Importance of Components

All 5 clinicians ranked the importance of each tool in the following order:

  1. In Vivo
  2. Social Sensing
  3. Messaging

Task 1

Find the Social Sensing for patient Jane D. and determine if the patient is doing well or not.

  1. Is this what you expected to see? Why or why not?
  2. What do you like/dislike?
  3. Is there anything you would change?

Task 2

View the In Vivo information for Jane D. and determine based on what is available, how useful or not the information is.

  1. Is this what you expected to see? Why or why not?
  2. What do you like/dislike?
  3. Is there anything you would change?

Task 3

Send a message to the patient’s partner asking the following questions:

  1. Social Sensing: Has Jane been displaying avoidance behaviors?
  2. Social Sensing: Has Jane been going to work?
  3. In Vivo: Did Jane seem better after the session?
  4. Did Jane engage fully in the practice session?
  5. Which questions are not so helpful?
  6. Are there any other questions you would like to ask?
  7. Is this what you expected to see? Why or why not?
  8. What do you like/dislike?
  9. Is there anything you would change?

Post Task Questions

  1. What is the purpose of social sensing?
  2. Can you reflect on the pros and cons of such a system for you as a clinician and for the patient?
  3. Can you tell me of an example from your past patients/clients where this would have been useful?
  4. Can you tell me an example from one of your patients/clients where this system would not have worked?

Summary of Findings

  • Liked Social Sensing and think it would be a helpful way to get more information about patient and how they’re doing outside of therapy
  • Privacy and confidentiality issues, need to get consent to disclose information
  • Would want to discuss including trusted other and what their role is - Operationalize some of these terms to make sure trusted other understands what they mean
  • InVivo
  • Clinicians have to spend amount of time filling data in themselves and graphing, so having it there for them is really useful so they don’t have to do it themselves, saves time
  • Liked that message questions were pre populated
  • Sometimes, clinicians aren’t tech savvy - might not use this method
  • Patients might no have access to technology
  • Patients might not have a trusted other
  • Think that this could be useful for patients they’ve had in the past
  • Liked the ease of use of entire system

Improvements Based On Findings

Social Sensing

  • Emoticons a little confusing - might need a tutorial or something about those
  • Emoticons drawing more attention at first than the social alert

In Vivo

  • Should have description of what tasks are
  • Should describe if this is trusted other’s opinion or patient’s opinion
  • Either have different graphs showing pre, peak and post or using a line to connect the data points to make it easier to see a trend
  • How long patient is doing In Vivo for and if they’re staying in it long enough
  • Expected numbers instead of graph


  • System needs to be slightly more detailed and have more specific wording for some questions
  • Include Imaginal Exercise questions too
  • Have a way to write free form messages


User Centered Design

The approach we took when starting on this project was based on the user centered design process, where you first research the problem space, ideate and design, prototype, evaluate and test then iterate again and again. This method argues that this is always a formative process and it never reaches the summative process, because the solution can always be improved upon.


Future work would include the following:

Interview Trusted Others

  • Under what conditions they are willing and to provide information.
  • What type of information do they want to provide>

Interview PE Therapy Patients

  • Who should be a trusted other? What data should trusted others provide?
  • How do patients interpret self-report visualizations?
  • What features will encourage patients to track data?
  • What features will encourage patients to take ownership of their treatment?