Designer | Four members | April 2024 - Present | Figma
A digital tool designed to aid professors in generating patient cases.
RxPert is a tool meant for professors to use for patient case generating. A usual patient case takes between one to two hours to create but with the help of RxPert, professors are able to put in a few key pieces of information and have their case be generated for them.
Outcome
Streamlined patient case creation through an AI patient case generator, shortening the time spent on case creation by a sixth of the time.
Problem
Current pharmacy professors are spending hours on creating cases or simply reusing old ones because they don't have the time.
Originally, I was introduced to the idea of a patient simulator by three students who were looking for a UX designer to design the tool. This would've been meant for students who wanted practice with patient interaction. I jumped on the idea because I believed in the purpose of the tool, as well as getting the opportunity to create something new. After doing some informal surveying and talking to professors, we found that there was more of a market for a patient case generator, thus switching courses with our product.
Goals for Case Generator
01
Case Generation
Professors will be able to generate on-paper cases based on their selected topic, learning objectives, and level.
02
Editable Display
Intuitive flow to edit the case, whether it be through chat or the case editor for specific parts.
03
Integration
For the future, we are interested in making cases interactable where students can talk with the patient in the case.
User Personas
Our primary users are within pharmaceutical academia, specifically professors and curriculum planners.
User Quotes
"Cases take one to two hours to create"
"Uncertainty with relying on AI for racial/cultural competency"
"Generated questions for the case are basic"
User Flow
It was important for us to understand how a professor would navigate through our product and creating a user flow allowed us to connect the different interfaces together smoothly.
Lo-Fi and System Design
Sketching out my ideas on paper let me experiment more in less time, helping me narrow down what designs were feasible versus just a fun idea. When sketching ideas, it was important for me to reference UI designs that already exist so that the user has some familiarity going into the product.
For the case generator, I was inspired by ChatGPT's document view, that creates a new side panel where users are able to access the document in a separate window rather than in the chat. I liked this design because it allowed the user to focus on the document rather than getting lost in the chat.
Access everything you need from one screen.
Whether you want to generate, edit, or view a case, the dashboard provides quick and easy access for all the desired tasks.
Fill out the form and generate your case.
Input the case information and have it be available to use within a few seconds.
Share your cases and interact with others.
Upload your cases for feedback and have access to other’s cases to use in your curriculum.
* Received feedback that professors would not want to use their time to leave feedback/interact with cases - this feature was put on hold
Edit your case with the help of AI.
Click the icon in the corner after the case is generated and edit it till you are happy with the result.
User Testing
When testing our design of the AI chat bot in the corner of the screen, we ran into the problem of different generations having different exposure with corner chat bots.
Professors from the older generation had trouble navigating how to edit the case, missing the edit button on the corner of the screen.
Younger professors were able to use the edit button to interact with the chat bot and edit their case.
We realized that, to ensure all professors were able to navigate our tool, we needed to guide them through the generation, editing, and refining process.
Iteration
Step One: Generating the Case
We included a progress bar at the top to inform professors what step they are on.
Iteration
Step Two: Edit the Case With the Chat Bot
Chat with AI on the right and see your changes in real time on the left. Approve the changes you like and decline those you don't.
Iteration
Step Three: Manually Edit Your Case
Type directly in the case to fix any minor details that were not fixed during editing with the AI.
Potential Future Design: Chain of Thought
Professors were unsure as to whether they would be able to trust AI to develop accurate patient cases.
By providing them with an alternative UI design, they would be able to verify that the case information is accurate based on the chain of thought process.
Future with RxPert
We are looking forward to release RxPert out to the public within the next couple months.
Currently, we are refining the user experience of creating and editing the case. We want to ensure that it is intuitive for professors.
Users/User Testing
We conducted a meeting with one of the pharmacy professors who creates exams and they shared with us that out of the 50 question exams, a third of them are mini cases. They also mentioned that it takes them around three days to create one exam, needing to make three exams in total.
We generated a mini case for them and after reading through it, they were impressed by the content and asked for us to send them the case.
Optimization
It is always important to keep testing your design and updating it for better usability. As new interfaces will continue to roll out, we will continue to test them and work the feedback into our design.
Personal Reflection
Taking time to understand the user, not only about how they navigate the platform, but who they are as a person.
One of our biggest struggles for this project has been user testing. The professors’ priority is their class and they often have little time to give elsewhere. From the limited user testing that has been conducted, we often find ourselves needing to make decisions on the user experience by way of extrapolating based on what was said by the professor. By getting to know the professors outside of their classroom selves, I believe that that would help us gain more insight on their thought processes since introspection can often be flawed.
made @ home