AI-Enhaced Telehealth Monitoring Interface
Specialised Product Design
UX/UI
Human-Centered AI
Experience prototyping
Project Information: 6 weeks, autumn 2024
Designing a complex product with a focus on
specialised use.
Partner: Philips Healthcare
Role: UX Research, Ideation, GUI Design
Team: Sander Randoja, Zeynep Emiroğlu,
Shelley Xiao
Challenge
With the rapid growth of Telehealth services, nurses face increasing pressure to manage complex patient cases remotely. Limited availability of experienced staff makes it difficult to monitor patients effectively, respond to critical events promptly, and maintain high-quality care.
Outcome
The result is a reimagined telehealth system featuring AI-powered decision tools. The system synthesizes a patient's entire history, which includes medical updates, alarms, and video, into an instant overview for immediate context. Going a step further, it proactively predicts potential future alarms before they occur. This combination of historical synthesis and predictive insight empowers nurses to respond faster in emergencies, enhances team collaboration, and fundamentally improves the quality of patient care.
AI-driven
Design
Research with medical experts revealed opportunities to integrate AI for automating tasks, highlighting critical patient insights, and supporting faster, data-driven decisions, reducing nurses’ workload and improving care quality.
Key Challenges Tackled
Inadequate Context
Telehealth professionals often lack a continuous, time-based view of patient history, making it harder to assess current or future needs.
Lack of Time-based Insight
Conventional tools often fail to provide a clear, evolving picture of patient health over time, preventing informed, timely decisions in care.
Increased RIsk Of Errors
Without time-based insight, the risk of misdiagnoses, delayed treatment, and inadequate care increases.
Strain on Tele-health Nurses
Missing key context hampers timely decision-making, increasing stress and inefficiency for telehealth nurses.
Identifying Key Contextual Cues in tele-ICU Care
Our research showed that telehealth nurses depend on four main event types for context:
Comments
Medical Events
Alarms
Media
Organizing these events along a timeline helps tele-ICU nurses see how they connect, making patient histories faster to interpret and easier to share across teams.
Hover to see more details
Monitoring View for the Tele-Nurse
The interface is designed around a dual-screen setup: the left shows an overview of all monitored patients, while the right provides a detailed timeline of a selected patient.
This structure enables telehealth nurses to shift seamlessly between broad monitoring and in-depth review, helping them spot trends, trace patient histories, and act quickly on critical events.
The Individual Patient Vitals Window Reimagined
Traditional vital windows showed vitals and alarms in isolation, offering limited context. The reimagined design adds a timeline graph to reveal trends over time and an AI prediction panel to flag risks like hypoxia before they occur.
These features give telehealth nurses clearer context, earlier warnings, and greater confidence in responding quickly to critical events.
Patient Context Overall View
Patient Overall View
Detailed 4-hour timeline showing alarms, medical events, and comments.
Intelligent hourly summary highlighting key occurrences within each time block.
Vital sign graphs providing a clear overview of patient trends.
Scrollable history to review
past patient context when needed.
Patient information tab with recent movements, diagnosis, treatment summary, and upcoming care plans.
Visual movement cues including images to give nurses richer context.
The Patient Overall View brings together critical context in one interface. A 4-hour timeline organizes alarms, events, and comments, paired with summaries and vital sign graphs to reveal trends.
Nurses can scroll back for history when needed, while a side tab shows key patient details, recent movements, and care plans. Visual cues like movement images further support quick interpretation.
Contextual Events View
When a prediction or alarm is triggered, nurses can quickly trace the related events highlighted on the adjacent screen. This contextual view helps them understand what led to the notification and navigate the situation with greater clarity and confidence, while keeping clinical judgment in their hands.
Snapshots of Our Process
Research Group Field Visit to Thorax ICU at Norrlands Universitetssjukhus.
Clustering and Synthesizing Research Findings.
Proposing Tele-ICU Nurse Workflows Based on Common Themes.
Roleplaying to Test Proposed Workflows.
Ideating Ways to Provide Nurses with Richer Patient Context.
Takeaways
1. Design for Augmentation, Not Automation
My primary takeaway was that in expert domains like healthcare, AI's greatest value lies in amplifying human intuition, not replacing it. Our research with clinicians at Norrlands Hospital made it clear that the goal was never to automate the final decision. Instead, my focus shifted to reducing the cognitive load that precedes it. I learned that the most effective AI doesn't just provide answers; it provides clarity, freeing the nurse to focus on what truly matters: delivering expert patient care.
2. The Currency of AI is Trust, Not Just Accuracy
Through prototyping and feedback sessions with the Philips Healthcare UX team, my understanding of trust in AI deepened significantly. I learned that for a nurse to rely on a predictive tool in a high-pressure moment, accuracy is only the starting point. Trust is built on transparency and explainability. This insight was pivotal, guiding us to design a system that always shows the "why" behind its suggestions, clearly linking predictions to the patient's data history.
3. Prototyping as a Tool for Inquiry Under Constraint
Faced with a tight timeline and limited user access, I learned to leverage prototyping and roleplaying as our primary tools for inquiry, not just validation. By making abstract AI concepts tangible and interactive, we could elicit deep expert feedback and communicate our vision effectively, proving that clear methodology can turn constraints into a catalyst for focused innovation.
Working on the mobile view,
please have a look on desktop