Smarter Hospital System Design for Early Disease Prevention
Transforming healthcare with AI to reveal hidden risks for enhanced patient care
Industry
HealthTech
Timeframe
1 years
with Implement
Team
Business, Product, Innovation and Developer
Role
Research, IA,
Wireframing and UI Design
INTRODUCTION
Thailand is facing a crisis
Millions of lives in Thailand are silently suffering from preventable diseases like diabetes and hypertension.Doctors are overwhelmed. Patients are unaware until it’s too late.
​
I designed an AI-powered system to catch the warning signs early, helping doctors act faster and giving patients the power to change their future before it's lost.
40K+
Thai people die each year from NCDs.
That’s around 1,000 people every day. Preventable conditions like diabetes, heart disease, and stroke are the leading causes of death, and they come with a massive social and economic cost.
Reference: National plan for the prevention and control of
non communicalble disease (2023-2027), Thailand
1.66 Trillion THB
Thailand’s economic loss from NCDs.
That’s 9.7% of GDP, with 91% of the loss hidden in absenteeism, underemployment, early retirement, and productivity lost to
premature death.
Reference: National plan for the prevention and control of
non communicalble disease (2023-2027), Thailand
Process Overview: User-centered solutions for better healthcare efficiency
Design thinking ensures we create solutions that address real user needs, improving healthcare efficiency and outcomes.
Empathize
Define
Ideate
Prototype
Test
Research needs of doctors, nurses, patients,
business team
Clarify barriers to early disease detection and intervention
Brainstorm solutions
with stakeholders
Design wireframes, IA, and UI
Validate with real user feedback and iterate
EMPATHIZE: UNDERSTAND THE PROBLEMS
Gain real insight with all stakeholders
We conducted 10+ user interviews with key stakeholders, including founders, doctors, nurses, patients, and teams from business, technology, and innovation, to deeply understand their challenges, painpoint and their goals.
UX Research Canvas
UX Research Canvas Process
1
2
3
4
5
Define the Goal
Clarify the business objective and
user impact.
Frame the Inquiry
Create open-ended questions and list assumptions to validate.
Plan the Research
Select users who reflect real contexts and
Choose suitable methods.
Conduct & Synthesize
Run sessions, engage stakeholders, and gather insights.
Deliver & Communicate
Summarize key findings in personas, reports, and testing results.
Research Method: User Interviews
Doctor
Goal
-
Understand the doctor’s workflow in managing chronic diseases and identify barriers to efficiency.
Research Question
-
How do you currently track and follow up with patients?
-
How do you currently track and follow up with patients?
-
How do you explain complex health info to patients?
Nurse
Goal
-
Explore the nurse's role in assessments and patient communication to improve workflow efficiency.
Research Question
-
How do you communicate health
status to patients?
-
How do you currently monitor patient progress?
-
What tasks are most time-consuming during patient assessments?
Patient
Goal
-
Understand the patient experience and identify barriers to understanding and following treatment plans.
Research Question
-
What challenges do you face when you see the doctor or nurse?
-
How do you receive and understand health information?
-
How do you prefer to receive health information about your condition and treatment options?
Business Team
Goal
-
Understand the business team's goals for improving hospital efficiency and adopting new technologies.
Research Question
-
What are the main business goals for implementing this hospital system?
-
What barriers do you face in adopting new technologies within the hospital?
Research Method: Affinity Diagram
Dr.Kasem, 44 years old
Doctor
An experienced internal medicine specialist managing chronic diseases, he sees 40–50 patients daily. He values prevention but struggles with time constraints and outdated systems.
Tech Proficiency
"I want patients to understand their risks and how to adjust their behavior, but there’s too much information, And they often don’t get it"
Goal
Inform the patient about the factors contributing to risk and methods for behavior modification to reduce it.
Frustrations (Pain Points)
There is too much information to convey, and patients might
not understand it.
Motivation
Passionate about preventive care, using AI-driven insights for better decisions, and reducing hospital overload through early intervention.
Expectation
Real-time reports on patient outcomes and hospital efficiency.
Ms.Somporn, 48 years old
Head Nurse
An experienced internal medicine specialist managing chronic diseases, he sees 40–50 patients daily. He values prevention but struggles with time constraints and outdated systems.
Tech Proficiency
"Another problem is Patients often ignore medical advice, leading to worsened conditions and unnecessary repeat treatments."
Goal
Enable faster risk screening, improve patient compliance, enhance doctor coordination, and streamline documentation for better efficiency.
Frustrations (Pain Points)
overwhelmed with paperwork and patient load, struggles with communication gaps, and finds it frustrating when patients don’t follow medical advice.
Motivation
Advocates for patients, values AI for reducing errors, and seeks a system that enhances teamwork and efficiency.
Expectation
The system should be simple, mobile-friendly, and visually intuitive, providing clear risk scores and automated patient education.
MS.Sukpimon, 53 years old
Office Worker
She is an office administrator with a sedentary routine. she dismisses symptoms and rarely gets check-ups, believing that feeling fine means being healthy. she remains unaware of her personal risk.
Tech Proficiency
I need a simple, clear system to track my blood sugar and blood pressure, with easy-to-understand advice for daily life. and avoid complex medical terms"
Goal
Enable faster risk screening, improve patient compliance, enhance doctor coordination, and streamline documentation for better efficiency.
Frustrations (Pain Points)
overwhelmed by confusing health data like blood sugar and blood pressure numbers, and struggle with too much online advice.
Motivation
The system should provide clear, personalized health risks and simple daily actions, with easy smartphone integration and reassurance to keep her on track without overwhelming medical terms.
Expectation
Stay healthy as she ages, preferring simple, natural lifestyle changes over medication, and values clear, easy-to-understand health advice.
Company
Business Team
She is a business development at an AI healthcare startup, working with diverse teams to align their solutions with hospital needs, tackling business, regulatory, and adoption challenges.
Tech Proficiency
"We aim to partner with hospitals and insurers to offer AI-driven risk prevention,
driving meaningful change in Thailand's healthcare."
Goal
Motivation
To scale the project, need to partner with hospitals, clinics, and insurance companies to offer AI-driven risk prevention services.
Frustrations (Pain Points)
Struggles with complex data, doctor resistance, fragmented IT systems, slow approvals, and convincing hospitals to invest.
Using technology and data-driven innovation to improve healthcare, to make a meaningful impact on Thailand's healthcare system.
Expectation
The AI system should be intuitive, clinically validated, cost-effective, and supported by positive feedback from early adopters.
!
Painpoint
Doctor
Overloaded with complex explanations
Nurse
Struggle with patient compliance and
admin load
Patient
Confused by medical data, prefer simple
daily advice.
Business
Need a system that's easy to scale
and adopt.
Goal
Doctor
Advise patients on risk factors and guide them in behavior changes to reduce their health risks
Nurse
Streamline screening, boost compliance, enhance coordination, and improve efficiency
Patient
Improve efficiency through faster risk screening, better patient compliance, streamlined doctor coordination, and simplified documentation.
Business
To scale, the project aims to partner with hospitals, clinics, and insurers to deliver AI-powered risk prevention services
DEFINE: PROBLEM STATEMENT
Problem Statement
Many Thai patients at risk of NCDs lack early detection and guidance, leading to worsening health conditions and costly treatments. How might we create an intuitive hospital system that enables early intervention, risk assessment, and proactive health monitoring
Vision
To prevent NCDs from using AI-driven risk analysis, early warnings should be provided to improve health outcomes
before diseases develop.
IDEATE: EXPLORING SOLUTION
Process Overview: A User-Centered Design Journey
An agile 3-step process: brainstorm, sketch & wireframe and test & Feedback designed to turn insights into intuitive user experiences through continuous feedback and iteration.
Test &
Feedback
Sketch &
Wirefrane
Brainstrom
Explore ideas and define solutions through collaboration and creativity.
Turn concepts into low-fidelity visuals to map out structure and flow.
Validate designs with users and refine based on real insights.
Brainstorm: From Brainstorm to Breakthrough
We explored ideas like Hologram Health, Gauge Charts, and Ranking Leaderboards, welcoming feedback from all stakeholders, including the Hospital Team, patients, and the Internal Team. This collaborative approach led to intuitive data visualizations, comprehensive analysis reports, and an efficient back-office system, tailored to real hospital needs.
WHY
Healthcare has diverse users with different goals. Brainstorming uncovered key pain points, aligned priorities, and bridged gaps—turning insights into real solutions.
Uncover hidden operational bottlenecks
that were not visible through interviews alone
Generate cross-functional ideas
balanced medical accuracy with user-friendly design
Prioritize features
based on shared pain points and high-impact opportunities
WHAT
The goal was to co-create meaningful, practical, and scalable ideas. Key topics included:
How can we predict and prevent disease before it becomes critical?
How can we make complex health data
understandable to patients?
Which features would save time without sacrificing care quality?
WHO
We brought together doctors, patients, managers, and cross-functional teams to ensure every perspective shaped the outcome.
Design for all stakeholders' needs.
Balance user experience, service efficiency, and business growth.
Discovery 3 keys oppotunity.
Our research revealed urgent needs in hospitals
Predictive Health Analytics
Forecasting disease risk is complex, requiring diverse patient data and AI models to detect early signs of conditions like diabetes and hypertension.
Improved Efficiency in Healthcare​
Balancing AI-driven health interventions with personalized care is challenging, as it must account for lifestyle, genetics, and environment while maintaining a human touch.​
Early Disease Detection and Warning
Deliver timely, trustworthy AI-driven health warnings to encourage preventative action.
Final Concept: A Preventive Health Dashboard
Lifestyle Modification
Personalized health recommendations based on individual habits, aiming to improve long-term wellness and prevent chronic diseases.
Digagnos Analysis Report
A comprehensive report that consolidates health data, offering insights and detailed analysis to support accurate diagnoses and treatment plans
AI to detect early disease risk
Leveraging artificial intelligence to analyze patient data and identify early warning signs of diseases, enabling proactive care and timely interventions.
Validating Product-Market Fit Early
We validated strong product-market fit by addressing Thailand’s rising NCD rates, aiming to help doctors detect risks early and reduce healthcare costs.
Product-Market Fit Process
Define Core Value
.png){kind=small}
.png)
Identify Target Audience
Uncover
Underserved Needs
Align Value
with User Needs
Translate into
feature set and design
Prioritize the feature,
Then turn it on to wireframe
UX
How do we turn this fit into a usable product?
Feature set
How does our value solve what they’re missing?
Core Valuve
What meaningful value are we offering?
Underserved Needs
What pain points are not being addressed today?
Target Audience
Who will benefit most from our solution?
Sketch & Wireframe: From Concept Ideas to Design the Structure
After brainstorming solutions, I translated many early idea sketches into wireframes, establishing a solid foundation for
a user-focused design.
Design decisions guided by human behavior
Grounded in UX laws from the start, this project ensures a user-centered experience by simplifying complex data, accelerating actions and clarifying decisions
Hick's Law and Von Restorff effect
To adhere to Hick's Law and reduce cognitive overload,we prioritized the display of the most critical risk indicators at the top of the AI Health Risk Analysis section. This allows users to quickly focus on patients needing immediate attention, rather than sifting through a large amount of data.
Faster decision making
Reduce cognitive load
Better prioritize of care
Jakob's Law
we enhance usability by mirroring familiar UI patterns in our Patient Information layout, navigation bar, and search box. This minimizes learning curves and maximizes efficiency for healthcare professionals.
Reduce learning curve​
Reduce cognitive load
Increased Efficiency of task completion
Test & Feedback: From Wireframes to Insights Feedback
I gathered feedback and made necessary adjustments to enhance the design, ensuring it aligned with user needs and improved functionality.
Doctor
​"There is concern about accuracy
and insufficient information,
which may lead to errors in initial patient diagnosis."
Business Team
"Is it easy to use?
If not, it could limit sales growth."
​
"If more than four diseases
how is it displayed?
Change for the future
If we ignoring feedback from both doctors (accuracy & usability concerns) and business stakeholders (ease of use & scalability) can lead to severe consequences for both UX and business success.
Impact of Not Making These Changes
1
Low Adoption & Loss of Trust from Doctors
70% of digital transformation projects in healthcare fail due to low adoption from medical professionals.
2
Increased Errors & Patient Safety Risks
Poorly designed information displays can lead to misdiagnosis, treatment delays, or incorrect prescriptions,
3
Business Loss: Decline in Sales & Market Competitiveness
If the system is hard to use, hospitals and clinics may reject it, leading to financial losses and loss of market share.
PROTOTYPE: BRINGING DESIGN TO LIFE
AI-Powered Patient Risk Analysis for Doctors and Nurses
Designed a patient analysis dashboard for a healthcare platform to help doctors and nurses analyze at-risk patients using AI, recommend personalized health modifications, and advise on further diagnostic tests
Front System
Simplified, Clear, and Initiative
Based on research insights, I designed a simplified patient dashboard with easy-to-understand risk scores, using color-coded indicators for quick decision-making.
Front System
Make more Trust
Enhance user trust and confidence that the software is properly connected to the HIS system.
Problem
Data Accuracy: Inconsistent or delayed updates between the
software and HIS reduce trust in patient data.
Solution
Add a "Sync with HIS" button in the dashboard for connect
with HIS system.
*HIS = Hospital Information System
Easy to Control
Prioritize diseases by urgency for doctors and nurses to quickly identify critical cases, ensuring smooth usability.
Problem
Awareness: Users are unaware of which diseases to be concerned about or the key symptoms to monitor, often overlooking important signs of illness.
Solution
-
Add a "Next" button to improve desktop navigation.
-
Display the "Total number of diseases in the system" to help users understand the scope of diseases in the system.
-
Prioritize the most concerning disease by positioning it first (to the left of the disease bar) for user's immediate attention.
Front System
Front System
Ensure accurate data
Improve data accuracy and Information depth to prevent diagnosis errors.
Problem
Doctors and nurses face challenges with incomplete or inaccurate patient data, leading to diagnostic errors and difficulty in confidently answering patient questions with personalized advice.
Solution
-
Adding more blood test data, X-ray results, and personal lifestyle information will improve the accuracy of patient analysis.
-
Add a "Show only abnormal" filter to help users quickly focus on high-risk lab test results.
Users' Feedback after Redesign
Dr.Kasem, 44 years old
Doctor
"With the improved accuracy and comprehensive information, the initial patient diagnosis is now more reliable and precise."
Ms.Somporn 48 years old
Head Nurse
"With accurate data in hand, we can make informed decisions, ensuring reliable diagnoses and building stronger trust with our patients.
Ms.Sukanda, 34 years old
Business Develoment
"Is it easy to use? Absolutely, and its simplicity will accelerate sales growth."
A Dashboard Built for CEO, Admins and Marketing Teams
Designed a centralized admin dashboard (back system) for a healthcare platform to help internal teams manage users, track patient examination statuses, and visualize marketing data
Back System
Data-Driven Insights
This visual dashboard empowers the management team with clear, real-time patient data to support strategic decisions and improve hospital direction.
Back System
Simplify User Oversight Across Hospital Networks
All from one centralized view, designed for better control and faster decisions.
Back System
Manage Hospital Teams with Clarity and Control
View and control user roles and access within each hospital for streamlined operations.
Developer Collaboration
Clear communication is key. Define your desired display and functionality precisely and collaborate to establish realistic
development timelines.
Search Box
This search box displays the results of user searches by Name and Surname, ID card and Hospital Number​
Search Box
Use a conditional check to determine if no results are returned from the search query. If no results are found, display a friendly reminder suggesting the user verify spelling or adjust search keywords for better accuracy​
AI Health Risk Analysis
The Box displays a risk analysis by comparing the health average of each disease, indicating the risk status as High, Normal, or Low. Diseases with urgent risk levels are prioritized and shown at the top
Detail of Disease
The disease Box displays a health data analysis of patients, including lab tests, X-rays, and lifestyle information, with data synced from the hospital's HIS system
Patient Information
This Patient Information Box displays real-time patient data, including Date of Birth, Weight, Height, Blood Pressure, Heart Rate, and BMI. The data is synced from the hospital's HIS (Hospital Information System)
Design System
Create a design system to scale efficiently and stay consistent, splitting it into frontend and backend to serve both user-facing and internal needs effectively.
Front System
TEST : VALIDATING THE SOLUTION
Expected Outcome
As a UX designer, my focus went beyond visual design. On this project, I focused on how our work improved workflow efficiency and delivered measurable business value in the healthcare sector.
Satisfaction rate
User Expreience Metric
60-70%
Goal
understand how satisfied users are with the system's usability, features, and overall experience.
How to Messure
The satisfaction rate likely came from initial feedback. I would have pushed for ongoing surveys to track trends and improve.
Dr.Karun 52 years old
Doctor
"AI Prediction Disease had 60-70% satisfaction. Improving accuracy, usability, and clarity will boost confidence and adoption."
Time Save
User Expreience Metric
25%
Goal
To assess how the system reduces assessment time, improving workflow, and patient care.​
How to Messure
The satisfaction rate likely came from initial feedback. I would have pushed for ongoing surveys to track trends and improve.
Ms.Panisara, 51 years old
Head Nurse
"Initial use showed a 25% reduction in assessment time, but a larger study is needed to confirm the impact."
Adoption Rate
10+
Professional User
Business Metric
Goal
Track active healthcare users to gauge adoption, improve workflows, and guide system scaling.
How to Messure
Dashboards show 10+ active users
An early sign of adoption.
Mr. Pattanadech, 44 years old
CEO
"The AI prediction system has over 10 active users, showing good adoption and improving workflow efficiency. User data allows for continuous performance tracking and optimization."
Scalability and Market Reach
5+
Clinic
Business Metric
Goal
Use partnership and user growth as key indicators of market demand and scalability beyond pilot clinics.
How to Messure
Tracking the number of partnerships and user growth would be key metrics for assessing scalability.
Ms.Sukanda, 34 years old
Business Develoment
"The AI prediction results in the first 5 clinics met the expected criteria, improving treatment efficiency. For scalability, we are tracking the growth of partnerships and users to assess future expansion potential."
KEY TAKEAWAY
Learning through hands-on experience.
This project taught me valuable lessons about being resourceful, putting the user first, and working effectively with others when
developing a product
Resource are critical factor
Initially planned for a holographic human body, but strategically refocused on core pain points due to budget, personnel, and time constraints
Design for users, Not for ourselves
We design based on assumptions, but user feedback often reveals our designs aren't liked. To avoid wasted effort, create quick prototypes and gather user feedback early and often.
Take action on what truly matter
Gather, analyze, and negotiate stakeholder feedback to assess its impact on the design and overall outcome.
More Project
SafeGuard
Real-Time Fraud Detection App Experience for Trust and Safety
Breathe Easy
Air Quality App That Turns Data into Actionable
Health Decisions
