Product: AI Scribe
Role: UX Designer, Visuals, Interaction
Duration: June 2023 - August 2023
ModMed Ambient Listening is a virtual scribe that creates AI structured notes in EMA right after the provider finishes recording the conversation with the patient, this aids the provider in their daily workflow to deliver better notes, increase focus on the patient, and provide a faster and better visit experience with the patient without a human scribe—striving to optimize the patient-doctor relationship and minimize interference the way a real or virtual human scribe would while bringing the value of a professional medical assistant.
​
Create the best user experience for providers, prioritize efficiency and accuracy so that the provider can focus on the patient while making faster and better notes, save on time and interface clicks, while acknowledging the crucial rising inflation and labor shortage Modmed customers are facing
The Goal
Let's Begin
A strong foundation was built through previous research findings and establishing an infrastructure for this AI Scribe to provide high-quality training and testing data for initial AI model training and continuous training to enhance model performance. Developing a mobile version for AI scribe was prioritized to collect recording data and produce transcriptions. This stage was also used to consider how the user would interact with the recording functionality and determine the appropriate auto-stops and recording length for each speciality.
The user interface for the AI Scribe mobile application is designed to optimize usability and efficiency in a professional healthcare setting. This layout strategically places paramount focus on recording functionality and the management of patient information while streamlining user interactions and mitigating the complexity associated with an excessive information interface.
In this feature, the auto-stop functionality represents a vital safety feature designed to mitigate potential operational lapses by ensuring the timely cessation of recording processes. This sophisticated mechanism serves as a protective buffer against the inadvertent and unchecked continuation of recording activities, thereby safeguarding both the integrity of recorded content and the efficient allocation of resources.
Project Vision
Pain Points
Excessive Clicks
Time
Traditional EMR systems often require healthcare professionals to navigate through a maze of menus, forms, and fields, resulting in a laborious and time-consuming process that involves an overabundance of mouse clicks or touchscreen taps.
This relentless demand for clicks not only consumes valuable time but also introduces the potential for errors and distractions. Providers are forced to divert their attention away from patients, spending more time on data entry and documentation than on direct patient care. This not only impacts the quality of the patient experience but also increases the risk of burnout among healthcare professionals.
​
Healthcare professionals face immense pressure to manage an extensive caseload efficiently while adhering to strict schedules. The time-intensive nature of tasks such as documentation, note-taking, and administrative duties leaves healthcare providers with limited time for direct patient engagement.
Inefficient Assitance
Inefficient assistance, particularly when utilizing medical scribes who lack familiarity with medical knowledge, poses a significant challenge in the healthcare setting. Medical scribes are often employed to assist healthcare providers with documentation, data entry, and administrative tasks. However, when scribes lack adequate medical knowledge and understanding, they can inadvertently introduce errors, misunderstand clinical context, or fail to capture critical patient information accurately.
Wireframes
In the initial stages of conceptualizing the ModMed AI Scribe, a crucial step involved the creation of wireframes of the mobile application for the recording functionality. These hand-drawn sketches served as the foundational blueprints for the user interface and interaction design of the application.
This low-fidelity approach provided a valuable opportunity to iterate and refine the user interface before committing to digital design, facilitating collaborative discussions among the project team.
As the project progressed, this design was reiterated for iPad but still served as a reference point for the development of the digital interface in the next stage.
AI Scribe on iPad
The iPad, with its touch-sensitive interface and powerful design tools, became the main device for this feature and the canvas upon which the initial ideas were transformed into a functional prototype. This transition allowed for the translation of paper-based sketches into dynamic, interactive wireframes and prototypes, bridging the gap between concept and execution.
This phase not only facilitated the visualization of the application's user interface in a more refined and detailed manner but also paved the way for rigorous usability testing and user feedback incorporation.
The recording functionality underwent a redesign to enhance its compatibility with iPad usage and seamlessly integrate with the EMA software's user interface.
The UI was meticulously optimized with a translucent design, ensuring it doesn't obstruct or hinder the visibility of patient notes underneath. This deliberate choice empowers users to seamlessly utilize the recording function while maintaining a clear view of the content it overlays. Furthermore, users have the flexibility to reposition the recording function anywhere on the screen, allowing for a personalized experience.
To maintain consistency with the iPhone version, these user-friendly features were retained within the UI. Additionally, recognizing the importance of user control, a convenient cancel option was introduced, providing users with an efficient means to opt out of the recording function when needed.
The reconciliation process for the transcript data, derived from the recorded interactions, was accorded the utmost priority in the UX journey. It underwent rigorous refinement through multiple iterative phases, involving intensive work sessions with a diverse team comprising healthcare providers, engineers, mobile architects, and AI machine learning specialists
Leveraging an in-depth comprehension of the functionalities inherent in the preexisting capabilities of the EMA software, our team embarked on the strategic design of a dedicated reconciliation page. This page represents a crucial junction in the user experience journey, where the user has the opportunity to evaluate the AI-generated suggestions in a manner that aligns with the priorities inherent in the patient visit note.
In our design approach, we meticulously prioritize certain aspects that are not only reflected in the patient visit note but also deemed highly pertinent for the provider's seamless workflow. The reconciliation page serves as a dynamic bridge between the AI-driven transcript and the final patient visit note, ensuring that the content aligns with the specific needs and preferences of the healthcare provider.
One noteworthy feature introduced to augment the user's experience is an additional recording option. This feature offers a supplementary layer of flexibility, allowing the provider to capture additional context or insights that may enhance the overall quality of the patient visit note. Recognizing the importance of accessibility and user-friendliness, we also implemented a streamlined and readily accessible transcript display.
This thoughtfully designed interface element ensures that the user can effortlessly review the transcribed content while making necessary adjustments or additions. It simplifies the process of fine-tuning the note and guarantees that the AI-driven suggestions seamlessly integrate with the provider's unique insights and expertise.
​
Usability Study
These studies involved a diverse panel of medical professionals hailing from various specialties, ensuring that the application's design and functionality catered comprehensively to the unique needs and expectations of doctors in different clinical domains.
Our commitment to delivering a user-centric solution prompted us to engage doctors as key stakeholders and users in these studies. By doing so, we aimed to attain a deep understanding of their workflow, preferences, and pain points, enabling us to fine-tune the application to align seamlessly with their daily practices.
In this endeavor, we collaborated closely with these medical experts, inviting them to actively participate in usability sessions, provide valuable insights, and offer constructive feedback.
High Fidelity Prototype
This prototype represents the culmination of our efforts to create a user-centric AI solution tailored for healthcare providers. It features an intuitive user interface designed to streamline providers' tasks, from recording patient interactions to refining patient notes. Based on insights from our collaboration with healthcare experts, different stages of iterations were done while considering additional features that would enhance the experience. This prototype embodies our vision for a transformative tool that empowers and supports healthcare professionals in their daily work.
Key Takeaways
As I reflect on the journey of designing the ModMed AI Scribe feature, several key moments emerged from our usability study, the creative design process, and extensive work sessions with our medical doctors and legal team. While working closely with impeccable team members of seniority level such as the President of Engineering, Senior AI Machine Learning Engineer, Senior Architects, Chief Medical Officer, and several experienced developers, it was great to gain insight on the best ways to collaborate to work towards our common goal and produce impactful work. It's important to note that, due to a signed NDA and the non-disclosure nature of the feature, specific visual representations and in-depth details about the feature cannot be publicly disclosed until after the Momentum event in October 2023.
Nevertheless, our collaborations with medical professionals across specialties have been instrumental in shaping an application that truly prioritizes the needs and preferences of healthcare providers. The usability studies reaffirmed the critical importance of a user-centric design, and the high-fidelity prototype serves as a testament to our commitment to delivering an intuitive and efficient tool for providers.
​
Overall, I'm proud to have been a part of such an impressive team and excited to see how this feature grows and aids our healthcare providers.