Blüm Health
Insurance Member Portal
I led the end-to-end design of the major feature flows for a health insurance product, exploring how to better meet users' needs from their health insurance portals.

Role
Product Designer (End-to-End: Research, Strategy, UI/UX)
Timeline
6 Months
Tools
Figjam (ideation), Figma (design), UXtweak (research), Gemini (copy)
Constraints
Limited resources; self-directed domain research

by Miles Mitchell

Contents
Overview
High-Level Project Summary

Background
The Story
The Blüm healthcare project tackled simplifying health insurance by centralizing coverage, cost transparency, and health records into a single intuitive portal. It emerged from my own personal struggles navigating health insurance as a young adult and the countless usability stories I had heard regarding products in the space.
The Mission
Translating an inherently complex system — ridden with industry jargon and laden with confusing processes — into an accessible experience was a large undertaking. My goal was to reimagine an insurance product that actually met the needs of its users while still accomplishing business goals, such as decreasing the volume of support inquiries.


My Process
1
Discovery
  • Market research
  • Competitive analysis
  • Problem statement
  • User surveys and interviews
  • Affinity mapping
  • User personas
2
Ideation
  • Card sorting → Sitemap
  • User journey mapping
  • User flow diagramming
  • Sketching low-fidelity wireframes
3
Prototype
  • Mid-fidelity prototype
  • Design System
  • Hi-fidelity prototype
4
User Testing
  • Usability testing
  • Preference testing
  • Iteration and refinement


Major Insights
While this was a personal project, I utilized evaluative research to validate my design decisions, as well as project potential business outcomes and establish success metrics to measure them.
1
Improving Operational Efficiency
By enabling users to easily estimate costs and view coverage details, we can preemptively answer common questions. This reduces administrative burden of support inquiries

Success Metric: Decrease in support tickets
2
Mitigating Risk and Cost
Testing flagged a 50% failure rate in the records workflow. Redesigning this flow now prevents costly post-release engineering fixes later




Success Metric: Increase in Task Success Rate
3
Gaining User Trust
Research showed that users hesitated to book without explicit coverage confirmation. Adding clear "In-Network" and "Covered" badges removed uncertainty, increasing users confidence.



Success Metric: Increase in Confidence Score
Login Screen
Home Dashboard
Coverage detail screen
Manual entry of health records
The Challenge
Defining the Problem
Discovery
Equipped with my own negative personal experiences engaging with my health insurance, I sought to identify the business problem and hypothesize how exactly that affected users. This began with a combination of secondary research — market research and then competitive analysis.

Market Research
Conducting secondary market research, I identified the business case for a better digital insurance experience. The data revealed a critical gap: users are digitally ready, but the industry isn't delivering.
65%
Digital Portal Usage
65% of individuals accessed their patient portal in 2024 (more than double the rate in 2014). Mobile app-based access jumped to 57% in 2024, while web-only access dropped to 42%.
39%
Usability Gap
The #1 driver of member satisfaction is "Ease of Finding Information," yet current health plans fail to deliver this level of digital experience 39% of the time.
85%
Retention Impact
Digital experience directly affects loyalty. When satisfaction scores are high, 85% of members say they “definitely will” renew with their current plan.
Competitive Analysis
To understand why satisfaction is low, I conducted a deep-dive audit of major competitors like UnitedHealthcare and Blue Care Advisor (BCBS MN), and documented some findings that aligned with my market research.
A Fragmented Experience
While functional, the Blue Care Advisor app suffered from poor information hierarchy. Critical tasks like viewing benefits often redirect users to external websites or non-mobile PDFs, breaking the user journey and increasing frustration.

Features vs Reliability
Although UHC set a high bar with features like Digital Wallet integration and Rewards programs. App reviews for UHC and my observations using BCBS revealed that stability issues were common.
A link redirecting users outside of the app
A failure to load the web app in a browser

The Opportunity
What I found was that competitors were often:
I concluded Blüm could find success by focusing on a seamless and centralized experience — no PDFs, no external links, just clear answers in an app that is intuitive to navigate.

Problem Statement
With a business case in hand, I then had to hypothesize how this equated to a user need. I did so by establishing a problem statement.
The User
Blüm health plan members who feel disconnected from their health plan details.
The Problem
Users currently struggle to make informed healthcare decisions because critical information is fragmented and hard to interact with.
The Goal
A convenient, centralized location to access comprehensive information about their health insurance and available, covered services.
Defining the Strategy & Scope
To translate this problem into an actionable MVP, I established some guardrails before moving into primary user research.
Mobile-First Rationale
Primary use cases — like checking coverage at a reception desk or pulling up a digital ID — are inherently mobile. Perfecting the "on-the-go" experience offered the highest return on investment for launch, and its prioritization was supported by market research.
A Centralized Experience
Based on competitor failures, a core business requirement was that all information must live natively within the app—no redirecting users to external mobile browsers or PDFs.
The MVP Constraint
While a full health insurance portal requires dozens of features, my timeline required strict prioritization. I knew I needed to focus on the most critical user needs in members' journeys, but to find out exactly what those were, I needed to speak directly with the users.
Understanding the Users
Empathizing and Identifying Needs
Discovery
To move beyond assumptions I established in my secondary research, I gathered qualitative and quantitative data to understand the root causes of user frustration and core user needs.

Research Methods
User Surveys
20 participants (screened for active health insurance users)
User Interviews
3 sessions representing distinct user groups such as "Chronic Care" and "Basic Access" users.
Synthesis and Analysis
I utilized Affinity Mapping to organize raw survey and interview data into thematic clusters. This process revealed that while users have distinct goals, their friction points are almost identical: financial ambiguity, fragmented systems, and confusing terminology.

Key Research Insights

Cost Ambiguity
The single biggest pain point was financial ambiguity and the ability to forecast healthcare costs
The Insight: Transparency is the primary driver of trust. Users feel helpless when they cannot predict out-of-pocket expenses.
Cost transparency is super important to me. I don't want hidden fees."
77%
of users didn’t know how much was covered for a specific service or procedure before it was performed

Information Fragmentation
Viewing coverage details is the #1 priority for users, but this data — along with their health history — is scattered across portals, third-party sites, and physical mail.
The Insight: Because there is no single source of truth, users resort to manual workarounds (like saving PDFs to digital folders) just to piece together their coverage info and health history.

The Jargon Barrior
Users frequently abandon tasks because they don't understand terms like
"Co-insurance" or "Deductible"
The Insight: Users desire a quick, understandable explanation of their health insurance and what it covers, delivered in plain English.
70%
of users experienced difficulty understanding health insurance terminology
Defining the User
User Personas
Based on the research, I identified two distinct users to guide my design decisions.

Eric
The "Just Need the Basics" User
A 27-year-old leasing agent who only engages with insurance when absolutely necessary (e.g. urgent care).
"Just tell me in plain English what I'll owe for a regular doctor visit. I shouldn't need an insurance glossary to understand my benefits"
Behavior: Eric is a reactive user. He avoids the app because he finds terms like "co-insurance" confusing, often delaying care or calling support just to get a simple cost estimate.
Design Implication: The interface must prioritize "plain language" and speed. Information like Member ID and Urgent Care costs must be visible within moments of opening the app.
Sarah
The "Informed But Frustrated" User
A 38-year-old mother managing the healthcare logistics for her entire family.
"Why do I have to log into three different places just to figure out if a claim was processed correctly and paid from my FSA? It should all be connected."
Behavior: She currently self-organizes by maintaining manual spreadsheets and Google Drive folders to track her daughter’s vaccination history and family expenses because the portals fail to do it for her.
Design Implication: The platform must act as a "centralized hub," aggregating records from disparate providers into a single, searchable timeline.
User Journeys
I mapped the end-to-end experience for Sarah and Eric to pinpoint exactly where they abandon the digital product and resort to manual workarounds.

Journey 1
The Records Scavenger Hunt
User: Sarah (The Family Manager)
The Breakdown: Sarah’s journey revealed a chaotic manual process. To submit school forms, she had to physically sift through filing cabinets, scan paper documents at home, and merge files manually.
The Opportunity: Centralized Digital Records. A feature to upload, store, and verify physical records so users never have to "scavenge" again.

Journey 2
The Dead End Search
User: Eric (The "Just Basics" User)
The Breakdown: Eric tried to self-serve by searching "strep test cost," but the app failed to understand the query. He was forced to abandon the app and wait on hold with Member Services just to get a price.
The Opportunity: Smart Cost Estimation. A tool that understands natural language (e.g., "Strep Throat") and calculates out-of-pocket costs based on real-time deductible status.

Information Architecture
Structuring the Experience
Ideation
Equipped with user needs and the major tasks they were trying to accomplish, I was ready to define how users would move through the product. Before defining the site structure, I needed to identify exactly what screens and features were required to solve their core problems.

User Task Flows

The Cost Estimator Flow
Addressing the #1 User Pain Point: "Estimating Costs"
I focused on Eric’s journey to determine the cost of a Strep Test. This flow was critical because it required the user to interact with with the system via three separate data points:
  1. Service Type (Strep Test)
  1. Provider Status (In-Network vs. Out)
  1. Real-Time Benefit Status (Insurance phase – Deductible )
By mapping this first, I identified the need for specific screens like "Service Selection/Search", "Coverage Details", and “Cost Estimator” that would later need a home in the site architecture.

Other Key Workflows:
  • Vaccination Records: Converting physical papers to digital records (Sarah's Journey). (view here)
  • Provider Search: Filtering specialists by location and reviews and booking an appointment. (view here)
Card Sorting
Once I characterized key features via the User Flows, I needed to know where users would expect to find them. I conducted a Hybrid Card Sort with 15 participants, which resulted in structural insights such as:

"Get Care" vs. Search
Users found a single "Search" tab too vague for finding doctors. They grouped provider lookups and scheduling under action-oriented labels, leading me to create a distinct "Get Care" section.
Validating the Dashboard
High-frequency items like Digital ID Cards were scattered across categories. This validated the need for a Home Dashboard to bridge the gap and offer jumping points to frequent flows.
"Health Profile" over "Records"
Participants felt "Health Records" sounded too clinical. They preferred "Health Profile" as a softer, more personalized home for their history, Care Team, and digital documents.

Sitemap
Synthesizing the user flows (what we need) and the card sorting data (where it goes), I constructed the final Sitemap. I flattened the hierarchy to ensure no critical feature—like the Digital ID or Cost Estimates—was ever more than two taps away from the home screen.
Lo-fidelity Ideation
Sketches and Wireframes
Once the site structure was defined, I moved into low-fidelity ideation to solve interface challenges such as visualizing cost transparency.
Coverage Details and Cost Estimator
While I developed all core workflows in parallel, I want to highlight the evolution of Eric’s Journey (Checking Coverage), as it presented a tough design challenge: simplifying the math of "Deductibles" and "Co-insurance" into a visual that any user could understand in seconds.
Sketch
My initial concept used a "Deductible Meter" and a binary toggle for "In-Network vs. Out-of-Network" to let users simulate different scenarios.View
Wireframe
Moving to digital low-fidelity allowed me to refine the hierarchy, prioritizing the "Estimated Cost" as the primary CTA element while tucking other cost details and requirements into secondary info on the page.

Visual Design &
The Design System
The Interactive and Visual Experience
prototype
With the core architecture mapped, I shifted focus to the visual and interactive experience. I validated layouts with realistic healthcare copy in mid-fidelity, established a scalable design system, and built a high-fidelity V1 prototype to test with real users.

Mid-Fidelity
Refining the Structure
Before applying brand identity to the UI, I moved my low-fidelity wireframes to mid-fidelity prototypes. This step was critical for making refinements to layout and copy that would define interaction.
  • The Goal: Validate that the information hierarchy worked when real text was applied, rather than just placeholder shapes.
  • Outcome: Introducing copy for complex components like the "insurance phase” component or tags, ensured that when I moved to high-fidelity, they would still fit in their elements and sit in the hierarchy as expected
Coverage categories offer quick starting points for the flow while a search bar offers specific queries.
Visualizations of user data improves scannability and quick reference of user's cost responsibility.
Cost estimation calculator as a module to maintain context about what the user is estimating.
Allows the user to easily calculate an estimate and quickly jump back to coverage details.
Design System and Guide
To bridge the gap between functionality and brand, I first established a Style Guide to serve as the visual North Star for stakeholders. This defined the product's identity—calming greens and clean typography—which I then scaled into a comprehensive Design System of reusable components.
Style Guide
The Guide (Identity)
Defined the "Calm & Trustworthy" aesthetic to lower user anxiety.
Design System
The System (Mechanics)
Standardized elements to ensure consistency across the app.
From Concept to Prototype
Bringing the Design to Life
With the initial design system applied to my mid-fidelity wireframes, I built the complete high-fidelity MVP prototype.

The “Cost Estimation” Evolution
This 4-step progression illustrates how the "Coverage" screen matured from a rough idea into a trustworthy financial tool.
Sketch
Defined core concept
Lo-Fidelity
Established heirachy of elements
Mid-Fidelity
Refined layout and copy
High-Fidelity
Applied brand and emphasized interaction and status
Validation and Iteration
Testing and refining the solutions
Testing
Before finalizing the design system or building the final screens, I subjected the V1 prototype to evaluation through multiple methodologies. By synthesizing data from moderated usability testing, expert peer critiques, and WCAG accessibility audits, I identified and resolved critical friction points to ensure a seamless, inclusive user experience.

Usability Testing
To validate the high-fidelity flows, I conducted remote, moderated usability testing with 6 participants. I evaluated the design's effectiveness using three core metrics: Task Success Rate, User Error Severity, and the Single Ease Question (SEQ) .

Task Success Rates
100%
Find a Provider
92%
Check Cost & Coverage
100%
Find a Health Record
50%
Add a Health Record
The data and my observations revealed a critical failure in the 'Add Record' workflow.
Also, while the "Find a Provider" had a 100% task success rate, after Affinity Mapping, I revealed there was still notable friction in the flow.
An affinity map from the synthesis of my usability testing results
The Fixes
I prioritized these major friction points for immediate redesign based on data from my usability testing results — specifically severity and frequency.

Fixing the "Add Record" Workflow

Improving Provider Search Efficiency

Maintaining a Transparent System
Problem: Even when the app successfully calculated a cost, 67% of users hesitated because the screen lacked an explicit signal confirming the service was actually covered by their plan.
The Fix: I added a high-visibility "Covered" Badge with a green checkmark. This simple visual cue provided immediate psychological safety, confirming the user's financial status without forcing them to guess.
Clarifying Navigation
Problem: Users confused the "Profile" tab when looking for health records, assuming it only contained account settings (like passwords).
Fix: I renamed the tab to "My Health". This simple semantic change immediately clarified the navigation's purpose and matched user mental models.
Further Refinements
Polishing the Design

Peer Feedback
Hierarchy & Scannability
To simulate a collaborative team environment, I engaged in design critiques with senior designers and peers. Their feedback focused heavily on removing micro-friction and improving visual hierarchy.
  • The Feedback: The product felt dense. Peers noted that without ample spacing and stronger visual separation, the cognitive load was too high for a quick-reference tool
  • The Iteration: I introduced more white space using standardized spacing variables, grouped important sections into distinct "cards" with clear headings, and removed redundant iconography to drastically improve scanning.
Before
After
Accessibility Audit
Designing for Inclusion
Before locking in the final UI, I conducted a full WCAG 2.1 accessibility audit.
  • The Fix: I adjusted color contrast on various UI elements—such as "Status Badges," tertiary buttons, input field borders, and utility helper text—to ensure all components passed strict AA standards for low-vision users .
Before
After

Final Solution
An Intuitive Centralized Hub for Plan Members
The final product consolidates a fragmented user journey—where users previously managed health data across 3+ disparate platforms—into a single, centralized dashboard. By reducing cognitive load and surfacing critical data upfront, the app transforms "insurance management" from a source of anxiety into a source of clarity.

Provider Search & Booking
Key User Flows
A streamlined experience that reduces friction by automating location data and matching user mental models. Coupled with prominent "In-Network" signaling, it drives booking confidence without forcing users to leave the app.





Transparent Cost Estimation
Key User Flows
An intuitive flow for identifying covered services, combined with a financial transparency tool that visualizes insurance phase status and member responsibility—removing the guesswork from healthcare costs.




The "My Health" Records Hub
Key User Flows
A centralized repository for health history that unifies fragmented data. It leverages a guided, step-by-step process that matches the user's mental model for uploading physical documents.






Retrospective and Next Steps
Looking Ahead
As outlined in the project overview, the ultimate success of this redesign relies on improving operational efficiency, mitigating risk, and building user trust. With an MVP successfully validated by users, the next phase of this product would involve engineering handoff and expanding the feature set. Here is what I learned during the process and what I would prioritize next.

Key Learnings
1
Design cannot always fix a broken system, but it can explain it.
My biggest challenge was reconciling user expectations ("Just tell me the exact price") with the reality of the immense complexity of insurance estimates. Instead of oversimplifying the math, I focused on transparency. I learned that users tolerate complexity if you are honest about it — labeling a price as an "Estimate" builds more trust than a false promise of accuracy.
2
The Danger of the "Happy Path"
I initially designed the "Provider Booking" flow assuming full API integration with every doctor. In the real world, many providers lack digital scheduling. Future iterations of this product must account for these edge cases with "fallback flows" (e.g. Click to Call or Link to External Site).
3
You are the liaison, not just the designer.
Navigating the labyrinth of "CPT Codes" and "Pre-authorizations" taught me that complex systems require subject-matter experts (SMEs). My role isn't going to be to know everything; it's to translate the SME's technical constraints into the user's practical language.
The Backlog
The Future Roadmap
Contextual Education
Implement "Smart Tooltips" or in-text glossary link functionality for complex terms (like Co-insurance or Pre-authorization) that appear only when relevant, supporting users without cluttering the UI.
Financial Consolidation Hub
Create a "Claims & Billing" center that aggregates claims history, premium billing, and out-of-pocket maximum spending, giving users a holistic picture of their finances.
Responsive scaling for larger screens
While the mobile app excels for "Quick Access" scenarios (like Urgent Care visits), I would scale the design system to desktop to support complex, deep-dive administrative tasks.
Thanks for Looking 👀

👋 Thanks for looking! Let's Connect
Made in a lot of places, but most recently Denver, CO 🌄
Made with