
DocBot
A mobile healthcare application with a chatbot companion

The challenge
According to the National Library of Medicine, nine out of ten adults struggle with health literacy.
Personal health literacy can be defined as the ability to find, understand, and utilize information to make informed decisions related to health and wellness.
How might we help adults feel more confident in their ability to find and utilize health information?
The idea
DocBot is a conceptual mobile application featuring an embedded healthcare chatbot companion. Its inception can be traced back to a literature review I conducted, which explored the growing adoption of chatbots in the service sector, particularly in customer service applications.
As a designer with a background in health, I wanted to brainstorm ways that we could leverage chatbot technology to help users increase their personal health literacy.
Timeline
August 2022 - December 2022
Role
UX Designer and UX Researcher

User requirement gathering
Initial survey responses (n=21) revealed that the primary goals users sought to achieve through the application were scheduling appointments and accessing medical information.
Furthermore, the respondents expressed a notable concern regarding the accuracy of information provided by both the chatbot and the application. This sentiment aligned with my previous literary research findings:
Users exhibit a general openness to engaging with chatbots, yet often express a lack of familiarity and hesitation towards artificial intelligence technology.
Striking the right balance in chatbot design is crucial, as both an excessively human-like or insufficiently human-like appearance can lead to a negative user experience, evoking the uncanny valley phenomenon.
Users' perception of chatbots is influenced by visual presentation, identity cues, conversational style, and gender stereotypes.

Ideation, design, prototype
There were two major areas of focus for the design of DocBot: the chatbot and the application itself.
My goal for the chatbot was to strike a balance in the messaging tone and pacing to avoid the pitfalls of becoming excessively human-like and triggering the unsettling sensation known as the uncanny valley.
Due to the text-heavy nature of the information repository, I thought through ways to break down the information into easily digestible chunks. I opted to incorporate descriptive illustrations and dropdown text sections to enhance the user experience and ensure that the content was easily navigable and manageable.

Usability testing and analysis
Through usability testing (n = 13), I identified three friction areas for users.
#1. Despite being informed that the messaging system utilized a chatbot, a few users were inadvertently misled by the stock image avatar, mistakenly believing they were engaging in a conversation with a real person.
#2. While searching for information about a particular medical condition, users anticipated finding a button or link within the "When to see a doctor" section that would enable them to schedule an appointment with their healthcare provider.
#3. Users encountered difficulties when attempting to update crucial medical information, such as allergies and prescribed medications. Although they searched within the settings tab, they overlooked the "Profile" section as they assumed it only contained basic identifying details like name and contact information
Design iteration
To enhance the user experience, I implemented a series of minor changes in three key areas of friction:
#1. Chatbot avatar
To reinforce the understanding that users are interacting with a chatbot and not a human, I replaced the stock photo avatar with an engaging illustration. This visual change helps users remember the nature of the interaction.
#2. Streamlined appointment creation
Under the "When to see a doctor" section, I incorporated a link button that allows users to directly schedule appointments. This eliminates the need for users to navigate to the appointments screen or request an appointment through the chatbot, making the process more efficient.
#3. Refined UX copy
To provide users with clearer context regarding the information stored, I renamed the 'Profile' settings tab to 'Medical Information.' This modification offers users a better understanding of the type of data contained in that section.
In addition to these improvements, I introduced a comprehensive onboarding tutorial for first-time users.
This tutorial serves as a guide, familiarizing users with the application's layout and features. It can be accessed at any time through the settings menu, ensuring users can refer back to it whenever needed.

Reflection
Combining my background in healthcare with my passion for technology and design, this project provided an exciting opportunity to merge my interests. However, like any other project, it was not without its constraints.
Constraint #1: Participant demographic
The usability testing participants consisted solely of college students in their early 20s. This skewed the gathered data, as younger participants tend to be more comfortable with mobile applications and more receptive to adopting new technology. To ensure a more representative user base, a diverse range of participants would have been ideal, providing more accurate insights.
Constraint #2: Engineering feasibility
During this capstone project, the design possibilities were limitless within the confines of Figma. However, in a real-world application development scenario, engineering feasibility would play a significant role in both the overall application design and the implementation of the chatbot.
Constraint #3: Legal considerations
In an ideal DocBot application, all doctors would be part of the DocBot network, enabling users to easily make appointments and manage their medical records. However, the real-world implementation would require adherence to legal expectations for protected health information to ensure compliance with privacy and security guidelines.
Constraint #4: Time
If I were to continue working on this project, I would conduct a second phase of usability testing to validate the updates made to the original design. While the changes were minor, obtaining concrete data showcasing improvements in task completion time and success rates would be valuable.
Nonetheless, I thoroughly enjoyed the opportunity to showcase my design skills, which I have honed over the past two years. This project allowed me to combine research and design, reaffirming my passion for the field. As a designer, I strive to be well-rounded and welcome any chance to apply user research methods to my work.
If you are interested in reading my full report, you can access it here. Send me a email or a message on LinkedIn if you would like to discuss my work in more detail! =)