Project Summary
Our client tasked us with designing an AI-enabled EHR (Electronics Health Record) integration that uses historical patient data to predict and flag patients who are at risk of a mental health crisis. Its main purpose is to alert healthcare professionals about impending mental health episodes before they become crises.
The project's core goals centered around establishing the tool's viability in the U.S. market, understanding its use within current clinical processes, and devising a prototype attuned to the American healthcare environment, particularly within the EPIC system (a widely used healthcare software platform) and prevailing clinical workflows.
My Role
As an Associate Creative Director of UX on a team of five, my role spanned the entire project lifecycle. Collaborating with a Director of Client Strategy, a Product Manager, a VP of Experience, and a Researcher, my duties encompassed industry-specific research, opportunity discovery, journey map creation, conceptual and detailed wireframe designs, client presentations, and prototyping for user testing sessions.
Challenges
No user research
Because of a limited client budget, and privacy concerns with regard to users in the healthcare profession, we were unable to conduct user research and instead had to rely on existing client research.
No access to the system we were designing for
We were unable to access demos or any relevant information about EPIC, the platform we were designing for, because of confidentiality agreements and the software’s proprietary nature. Instead, we had to research old screenshots and EPIC manuals online in order to make reasonable assumptions about the system’s capabilities, functionality, and flow so that we could use them for templates and as a design foundation.
Indirect user testing
Due to privacy concerns, we were unable to directly conduct user testing sessions nor could we observe them live or via recorded audio or video. Instead, we had to rely on written transcripts and summary documentation.
Discovering Key Opportunities
To identify key opportunities under the restriction of not being able to conduct user research, we relied on Koa's research documents, which included clinical best practices, existing clinical pathways, and EPIC usage in emergency room settings to determine user types and to create provider clinical journey maps.
Key Opportunities We Identified:
1. Additional integration points within EPIC
The team at KOA expressed some concerns regarding Foresight's efficacy for opportunities 2-4. However, they agreed that Foresight seemed optimally positioned for integrating within Epic, recognizing this as its primary value, and this would be the priority.
2. Passive data capture and alerts/notifications
The KOA team was cognizant of the necessity for a rigorous regulatory process for the implementation of a reliable alerts system and deemed it impractical for the initial launch. However, recognizing that such a system aligned with the core value of Foresight, they suggested including this opportunity in the mockups for potential future adaptation. They also emphasized the importance of integrating passive data capture and visualization, as these are essential features of Koa Health.
The KOA team was cognizant of the necessity for a rigorous regulatory process for the implementation of a reliable alerts system and deemed it impractical for the initial launch. However, recognizing that such a system aligned with the core value of Foresight, they suggested including this opportunity in the mockups for potential future adaptation. They also emphasized the importance of integrating passive data capture and visualization, as these are essential features of Koa Health.
3. Cross-platform integration with other EPRs
KOA's feedback suggested caution in treating Foresight as a mere instrument for enabling EHR integrations. They emphasized that the main aim should be to access and share clinician notes across platforms to facilitate continuity of care. They deemed this opportunity as not requiring a comprehensive mockup or landing page.
KOA's feedback suggested caution in treating Foresight as a mere instrument for enabling EHR integrations. They emphasized that the main aim should be to access and share clinician notes across platforms to facilitate continuity of care. They deemed this opportunity as not requiring a comprehensive mockup or landing page.
4. Actions and next steps for providers
The team at KOA viewed the actions and next steps more as extensions to EPIC's primary functionalities. They were uncertain if Foresight was the best choice to provide these extensions, taking into account the development costs. However, they did request the inclusion of a reminder setting feature in the mockup.
The team at KOA viewed the actions and next steps more as extensions to EPIC's primary functionalities. They were uncertain if Foresight was the best choice to provide these extensions, taking into account the development costs. However, they did request the inclusion of a reminder setting feature in the mockup.
Concept Validation
From the opportunities we identified in the journey maps, we proceeded to create initial concepts. However, because EPIC would not provide us with any product materials, including recent product images or demos, we found ourselves combing the internet in order to locate screenshots of the software and EPIC user guides, which were often out-of-date, but crucial in gaining insights into what its current features and functionalities might be.
For proof of concept validation, we presented users with three static screen designs during an initial round of usability testing sessions, facilitated by the client. Each design showcased different integration points of the tool within EPIC. Budget constraints necessitated a quick exposure of these designs to users, and bypassed a click-through prototype for this initial round. The focus of these screens centered around the following EPIC sections:
1. Patient List View
2. In-Basket (EPIC’s messaging system)
3. KOA Tool Patient Mental Health Details Overlay with customization options
2. In-Basket (EPIC’s messaging system)
3. KOA Tool Patient Mental Health Details Overlay with customization options
Addressing User Feedback
After the proof of concept was validated by users, we used the information gathered in that first round of testing sessions to improve the design by understanding the HCP’s current experience, the process, and the tools used to currently assess, manage, and treat mental health crises across different clinician types and practice settings, identify any unmet needs for mental health patient management, and to gather feedback on utility and where in the workflow the tool may be relevant.
Riskscore Modal
At the core of the user experience is the Riskscore modal, which can be accessed from any integration point. Our choice to utilize a modal pattern for this tool was influenced by its adaptability. Systems like EPIC are highly customizable and appear across a spectrum of screen sizes. Therefore, a flexible design pattern such as a modal can be advantageous as it can adapt to diverse customization and usage requirements of various environments, including doctors' offices, hospitals, psychiatry offices, and emergency rooms.
Integration Point: The List View
The list view is where healthcare professionals view upcoming appointments and have morning huddles. Clicking on a patient's Riskscore value displays a modal with details about that patient's mental health score.
Integration Point: Patient Details Page
Clicking on the banner displays a modal with details about that patient's mental health score.
Integration Point: In-Basket
In-Basket is EPIC's version of a messaging/email system. The message contains static information from the modal and a banner that when clicked, displays the full-featured modal.
Usability Testing
After the updates had been made to the designs, following the concept validation activity, I created prototypes representing two scenarios.
Scenario 1
Dr. Weber, a primary care physician, is logged into EPIC to host the morning huddle and is reviewing the patient list of the day’s appointments with his staff.
Dr. Weber, a primary care physician, is logged into EPIC to host the morning huddle and is reviewing the patient list of the day’s appointments with his staff.
Scenario 2
Dr. McQue, a psychiatrist, is logged into EPIC. While reviewing her patient list for the day, she receives a new In-Basket alert from Riskscore.
Dr. McQue, a psychiatrist, is logged into EPIC. While reviewing her patient list for the day, she receives a new In-Basket alert from Riskscore.
We spoke to a mix of clinical experts
We engaged in discussions with a diverse group of U.S.-based physicians, which included seven psychologists, four primary care physicians, two emergency room physicians, and two physician’s assistants.
- All of them either currently, or in the past, used EPIC as an EMR.
- All played an active clinical or operational role in the assessment, management, treatment, or prevention of mental health crises.
- All played an active clinical or operational role in the assessment, management, treatment, or prevention of mental health crises.
Key Findings Summary
The Riskscore tool indicated a positive outlook for feasibility driven by potential improvement in outcomes and utility in daily workflow.
Product Assessment
Other Findings
- All providers interviewed stated that this tool would be realistic to use in their day-to-day workflow.
- Nearly all providers cited a willingness to pay for this tool.
- People want Foresight to be the one-stop shop for Mental Health Evaluation.
- Nearly all providers cited a willingness to pay for this tool.
- People want Foresight to be the one-stop shop for Mental Health Evaluation.
Hurdles to Overcome to Encourage Adoption
Despite the significant approval, the tool had garnered among healthcare professionals, there were a few obstacles that needed to be navigated to ensure its widespread adoption:
Physicians are Predisposed Against New Tool Adoption
Physicians' resistance to learning new tools, hesitance to add time to their workflow, and "portal fatigue" from multiple tools pose challenges to adoption.
Physicians' resistance to learning new tools, hesitance to add time to their workflow, and "portal fatigue" from multiple tools pose challenges to adoption.
High Burden of Proof for Tool Adoption
The tool must compellingly demonstrate its ability to enhance patient outcomes and operational efficiency to overcome clinicians' skepticism.
The tool must compellingly demonstrate its ability to enhance patient outcomes and operational efficiency to overcome clinicians' skepticism.
Potential for Disregarding Risk Scores
Factors such as unreliable or scores that don't update over time, alert fatigue, and limited patient scores might lead providers to disregard the risk scores.
Factors such as unreliable or scores that don't update over time, alert fatigue, and limited patient scores might lead providers to disregard the risk scores.
Lack of Cross-Platform Integration Creates Information Silos
At present, one of the largest issues of integrating new technologies is the need for healthcare professionals to log in to multiple platforms that do not communicate with one another.
At present, one of the largest issues of integrating new technologies is the need for healthcare professionals to log in to multiple platforms that do not communicate with one another.
