Medication Adherence : A Solution for tracking medication adherence during Clinical Trials
Clinical trials suffer from unreliable medication adherence data, particularly for eyedrop medications. Current manual tracking methods are inefficient and error-prone, creating incomplete trial data and adding significant administrative burden to Clinical Research Coordinators (CRCs). The goal of the project was to design a connected eyedrop tracking system for an elderly user base (65+) who are not tech-savvy, while meeting the stringent data requirements of clinical trial protocols.
Project Summary
Context
10XBETA
Timeline
8 weeks (Research & Early Design Phase)
Collaborators
Product Manager, Software Engineer, Creative Technologist
User research, stakeholder interviews, information architecture, early wireframes
Responsibilities
Contributions
I joined the project during the foundational research phase to:
Conduct user research with clinical stakeholders and patients
Map the clinical trial ecosystem and identify key pain points
Define information architecture for patient app and CRC dashboard
Create early wireframes establishing product structure
The company used these deliverables as the foundation for their MVP hardware device.
50-60% of patients fail to adhere correctly to eyedrop regimens
Manual data entry consumes 50% of CRC administrative time
23% of trial data is incomplete or inaccurate
Overview
Research Methodologies
Subject Matter Expert Interviews
2 in-depth interviews with clinical research professionals to understand the ecosystem and validate assumptions:
Clinical Research Coordinator (CRC): Explored daily workflows, manual data entry processes, patient management challenges across multiple concurrent trials, and observations of common patient adherence issues
Principal Investigator (PI): Discussed trial integrity requirements, data quality standards, regulatory compliance needs (HIPAA, FDA 21 CFR Part 11), and decision-making criteria for trial technology adoption
These SME interviews provided critical insights into:
Current manual tracking methods and their inefficiencies
CRC administrative burden (estimated 2-3 hours daily on data entry)
Regulatory and compliance requirements
Inconsistent Electronic Data Capture (EDC) system integration needs
User Research & Patient Understanding
Due to regulatory constraints preventing direct patient contact during early research phase, I employed alternative methods to build patient understanding:
Clinical studies on medication adherence rates in elderly populations (65+)
Usability research on medical devices and IoT products for seniors
Accessibility guidelines and age-related capability changes (motor skills, vision, cognition)
Statistics on smartphone ownership and internet access in 65+ demographic
SME-Informed Assumptions:
Based on the CRC interview and demographic research, developed validated assumptions about the patient population:
Primary demographic: 65+ with varying technology literacy
Physical constraints: Mild arthritis affecting bottle handling, vision impairment
Cognitive challenges: Short-term memory uncertainty about dose timing
Technology access: Limited smartphone ownership (~18%), unreliable WiFi (~35%)
Motivations: Strong desire to complete trial successfully, fear of "letting down" research team
Research Synthesis & Key Insights
Memory is the primary barrier - Memory uncertainty drives missed doses and inaccurate self-reporting
Technology must be passive - 70%+ of patients express anxiety about "complicated" devices
Connectivity cannot be assumed - 35% unreliable WiFi, 18% no smartphone ownership
Real-time data is essential - Monthly check-ins too infrequent for timely intervention
Manual data entry is the bottleneck - 2-3 hours daily impacts CRC capacity
Design Recommendations
Based on these insights, I developed recommendations that directly addressed each pain point:
For Patients:
Automated tracking eliminating need for manual logging
Multi-modal reminders (device + app notifications)
Simple visual feedback confirming dose recorded
Offline-capable device with local data storage
For Clinical Research Coordinators:
Real-time adherence dashboard for all trial participants
Alert system for missed doses requiring intervention
Automated data export to EDC systems
Device health monitoring for proactive troubleshooting
Technical Requirements:
Dual connectivity: Primary WiFi + cellular data backup
Connection monitoring with patient reminders when offline
Local buffering to store dose events when disconnected
Automatic sync when connection restored
Design Principles
From research insights, I established core principles to guide interface design:
Invisible Technology - System requires zero active management from users
Graceful Degradation - Must work even when connectivity fails
Accessibility First - Design for limited dexterity and vision
Device-First Design - Core functionality without requiring app
Trust Through Transparency - Clear feedback about what's being tracked
Impact & Outcomes
Research Impact:
Validated product-market fit with clinical stakeholders
Identified critical design constraints (elderly UX, offline-first architecture)
De-risked major assumptions about technology adoption and connectivity
Service Design Impact:
Service blueprint became critical alignment artifact for engineering team
Defined technical requirements for connectivity, reminders, and data sync
Provided clear vision for how frontstage/backstage/support systems integrate
IA & Design Impact:
Created scalable structure supporting future features
Established clear separation of patient vs. CRC needs
Defined MVP scope based on IA complexity analysis
Wireframes guided development team and enabled early stakeholder buy-in
The company moved forward with hardware development and higher-fidelity UI design based on these deliverables. Later patient testing validated the assumption-based personas and design decisions.
Key Learnings
What Worked Well
1.Resourceful research within constraints - Using SME expertise and video ethnography as patient proxies yielded actionable insights despite access limitations
2. Service blueprint as alignment tool - Mapping frontstage/backstage/support systems helped engineering understand UX requirements and made connectivity decisions more user-centered
3. Alert-first IA - Structure emerged directly from understanding CRC exception-driven workflow, validated through testing
4. Progressive disclosure strategy - Assumption-based persona development successfully guided interface complexity decisions
What I'd Do Differently
1. Direct patient research earlier - While SME insights were valuable, direct patient validation would have reduced assumptions and uncovered edge cases sooner. I would advocate strongly for at least 5-6 patient interviews even in constrained environments.
2. More diverse SME perspectives - 2 interviews provided solid foundation, but adding a data manager and regulatory specialist would have enriched understanding of backend requirements
3. Prototype testing with proxies - Could have recruited elderly users outside clinical trials (friends/family of team members) for early wireframe feedback to validate assumptions before company testing phase
Want to Learn More?
This summary covers the highlights of a complex, multi-phase design process. Reach out to discuss the full case study, including detailed research findings, design iterations, and lessons learned from bringing hospital-grade care to the road.