work
about
cv

How might we...

design CGM apps for real life, not just clinical data?

How I redesigned the Freestyle Libre app to prioritise human context over medical metrics

Want to see it live? See it here

Foreword: this is an old case study, with dated UI. So enjoy the UX thinking, disregard the 2010's UI.

TLDR

Freestyle Libre app is designed like clinical tool - tracking data for doctors, not insights for patients. It didn't prioritise everyday usability or human context.

Personal project: Research-driven redesign of Freestyle Libre app

  • Type 1 diabetics commonly develop vision issues - yet app ignores accessibility

  • App doesn't capture context (food, stress, exercise) that actually explains blood sugar changes

  • Redesigned to prioritise everyday context over raw data

THE PROBLEM

App treats diabetes like a data problem. Doesn't capture context, no accessibility.

THE JOURNEY

Diabetic patients using CGM testing and logging blood glucose.

THE APPROACH

Redesign around real-life context: logging, long-term patterns, night mode.

ROLE

Personal project. Research, strategy, design, validation. Drawing on lived experience as Type 1 diabetic.

IMPACT

Validated design through testing, demonstrated human-centred approach to digital therapeutics.

Recently went live, still tracking metrics. So far: Reduction in drop-off post order, reduction in patient care messages, >80% guide completion rates.

GETTING STARTED

A brief introduction to type 1 diabetes

Patients don't need more information

If you’re a little unsure what diabetes is — I recommend having a quick skim of this NHS site. Even if you don’t read the rest of this case study, you might meet an attractive diabetic one day and need a more appropriate response than— ‘Oh yeah…I think my cat had that. We had to put him down.” 

Living with diabetes means tracking everything

I’m a Type 1 diabetic, and I use a CGM (Continuous Glucose Monitor) called Freestyle Libre - a small sensor stuck to my arm that tracks my blood sugar 24/7 without finger pricks. It’s incredible technology. Genuinely life-changing. I have a tiny sensor continuously monitoring my glucose levels and sending data to my phone. We’re living in the future. Except the app designed to make sense of all this futuristic data looks like it was built by someone who’s never actually had diabetes.

Managing diabetes means my blood sugar doesn’t exist in a vacuum. It’s affected by everything I do, everything I eat, everything that happens to me. When I see my blood sugar spike at 3pm, I need to know: Did I eat something? Am I stressed? Is it just being dramatic for no reason? (It does that sometimes.)


Managing diabetes means my blood sugar doesn’t exist in a vacuum. It’s affected by everything I do, everything I eat, everything that happens to me. When I see my blood sugar spike at 3pm, I need to know: Did I eat something? Am I stressed? Is it just being dramatic for no reason? (It does that sometimes.)


Food. Exercise. Stress. Sleep. Illness. Hormones. Medication timing. Alcohol. Temperature. Whether Mercury is in retrograde. Literally everything affects blood sugar. And yet, the app has no way to track any of it. Just numbers on a graph. Data without context, patterns or history is just... data. Seeing "8.2 mmol/L" at 3pm means absolutely nothing without knowing why it's 8.2 and what to do about it.

CGMs like Freestyle Libre are incredible technology. They track blood sugar 24/7 without finger pricks. Life-changing for diabetics. I used to have to pull out scary machinery in a cafe while mothers shielded their children from me. But the app is not designed for what I, using it more than 20 times a day, need it for.

Food. Exercise. Stress. Sleep. Illness. Hormones. Medication timing. Alcohol. Temperature. Whether Mercury is in retrograde. Literally everything affects blood sugar. And yet, the app has no way to track any of it. Just numbers on a graph. Data without context, patterns or history is just... data. Seeing "8.2 mmol/L" at 3pm means absolutely nothing without knowing why it's 8.2 and what to do about it.

CGMs like Freestyle Libre are incredible technology. They track blood sugar 24/7 without finger pricks. Life-changing for diabetics. I used to have to pull out scary machinery in a cafe while mothers shielded their children from me. But the app is not designed for what I, using it more than 20 times a day, need it for.

The app is designed like a clinical tool. It prioritises data collection over patient insight. It's built for what doctors need to see in a 15-minute appointment, not what patients need to manage their condition daily.

PROBLEM

Doesn't capture context

No way to log food, exercise, stress, medication changes, illness - the things that actually affect blood sugar.

PROBLEM

Single readings, not patterns

Hard to see long-term trends or scroll back through history. Each day is isolated.

PROBLEM

Effort to log anything

Clunky interface makes logging feel like work, not helpful.

PROBLEM

Ignores accessibility:

Bright interface when diabetics check blood sugar at night. No consideration for vision issues

This isn't just frustrating. It's dangerous. When patients can't understand patterns, they can't make informed decisions about insulin, food, exercise. The data is useless without context.

RESEARCH INSIGHTS

The accessibility crisis nobody's talking about

What's crazy is that diabetics are far, far more likely than most users to have vision issues. So it's shocking that a medical app like this has essentially ignored basic accessibility guidelines.

50%

of Type 1 Diabetics have some form of retinopathy

30

people a week have eyesight seriously affected by diabetes

99%

of people who have had diabetes for 20 years have eyesight issues

These aren't edge cases. This is the majority of long-term diabetics. But the app is designed with tiny text, low contrast, bright white backgrounds. Diabetics frequently check their blood sugar at night - waking up to blinding white screens. Night mode was one of the most consistent requests in my research. Not a nice-to-have. A genuine accessibility need.

These aren't edge cases. This is the majority of long-term diabetics. But the app is designed with tiny text, low contrast, bright white backgrounds. Diabetics frequently check their blood sugar at night - waking up to blinding white screens. Night mode was one of the most consistent requests in my research. Not a nice-to-have. A genuine accessibility need.


I did research with other diabetics to validate

  • Surveys to identify common pain points

  • 5 in-depth interviews about daily usage patterns

  • App store review analysis (quantified recurring complaints)

  • Usability testing with 4 diabetics on existing app

  • Surveys to identify common pain points

  • 5 in-depth interviews about daily usage patterns

  • App store review analysis (quantified recurring complaints)

  • Usability testing with 4 diabetics on existing app


What I learned

App doesn't align with diabetic mental model

People already track food, exercise, stress in their heads or in separate apps. They want to connect those dots to their blood sugar, not just see a graph.

No broad overview or long-term patterns

Scrolling back through days of data is painful. But that's where you actually spot what's happening - "my blood sugar is high on hot days" or "I crash after swimming."

Too much effort for basic tasks

Every additional tap to log information or view history adds friction. When you're checking your blood sugar 10+ times a day, friction compounds.

Night mode isn't optional

Multiple diabetics mentioned checking blood sugar at 2am and being blinded by white screens. For people with vision issues, this is even worse.

Data without context is just numbers. "My blood sugar is 8.2 mmol/L" - okay, why? What do I do about it? The app doesn't help answer those questions.

APPROACH

What I redesigned

Context, everyday usability, and pattern management

  • Reduced taps for common actions. Made logging feel fast and helpful, not like homework you're avoiding.

  • Monthly and weekly overview screens showing trends, not just individual readings. Scrolling back through weeks of data to spot patterns manually is unrealistic - this surfaces them visually.

Graph view for context, patterns, and ease

Diabetes isn't managed day-to-day. It's managed over weeks and months. You need to see: "My blood sugar is consistently high on Mondays" or "I always crash after certain types of exercise."

  • Monthly and weekly overview screens showing trends, not just individual readings. Scrolling back through weeks of data to spot patterns manually is unrealistic - this surfaces them visually.

Contextual logging

In interviews, every single diabetic mentioned trying to figure out what caused blood sugar changes. They were already doing this mental work. The app just wasn't supporting it.

One participant said: "I know I need to track what I eat, but opening a separate app to log food and then checking Libre and trying to remember what I ate 2 hours ago... I just don't do it."

So I made context logging part of the blood sugar check flow. One action, all the information captured together.

Night Mode

Dark interface option for nighttime checks. Diabetics check their blood sugar at night constantly. Bright white screens at 2am are genuinely painful, especially for people with vision issues. Multiple participants mentioned this unprompted. Diabetics check blood sugar at night - during sleep, when they wake up, before bed. A blinding white screen at 2am is genuinely painful, especially for people with vision issues.

This should have been a default feature, not an afterthought.

TESTING

Validating through testing

I tested the redesign with 4 diabetics who regularly use Freestyle Libre

Key learnings:

From dose guide to platform-wide support system

This was never just for weight loss. It was a framework for any scenario needing nuanced, personalised support. Nutrition advice, side effect management, contraceptive selection, treatment switching, eventually integration into initial questionnaires. We also built support infrastructure: redesigned site architecture, central support hub, better wayfinding, reformatted dense PDFs into videos and infographics.

Context logging

All 4 said they'd actually use this feature. It aligned with how they already think about their diabetes management.

One said: "This is what I've been trying to do in my head for years."

Pattern view

Unanimous positive response. Participants immediately spotted patterns in the demo data they wouldn't have seen otherwise.

"I didn't realise how much my stress affects my blood sugar until I saw it like this."

Overall usability

Every additional tap to log information or view history adds friction. When you're checking your blood sugar 10+ times a day, friction compounds.

Night mode

"FINALLY!".

Enough said. There was a reason that it was the most requested feature on the app store.

…

Why this matters for digital therapeutics

This project demonstrates a core principle for healthcare apps: design for human context, not just clinical data.

Medical devices and apps often prioritise precision and data collection over usability and insight. This makes sense for clinical tools used by healthcare professionals. But patient-facing apps need different priorities.

Patients don't need more data. They're drowning in data. They need:

  • Context to understand what the data means in their actual life

  • Patterns to make informed decisions over time

  • Accessibility because chronic conditions often come with complications

  • Low friction because they're using these tools constantly, not occasionally

The research validated what I knew from lived experience: diabetics are already doing complex pattern recognition in their heads. They're already tracking context in separate apps or notebooks. They're already trying to make sense of their data.

The app should support that work, not ignore it.


Ready to see more?

Click through to the next case study for more high-octane UX fun

See next case study: ZavaMed BMI verification