AI patient profile dashboard

Overview

This is the member profile a Twin Health coach works from: one screen with a member's glucose, metabolic score, activity and body composition, and — on top — the coach's own work for that member today. Coaches, providers and nurses all use it; it's designed around the coach, who opens it most.

The problem: Coaches look after a lot of members, and the data about each one was spread across several apps. Spotting a trend, deciding who to call first and having a conversation backed by data all meant hunting first.

What we built: A profile that reads top to bottom in the order a coach thinks. A task bar says what to do and why, with Chat, Mark as done and Next task. A header says who this is and who else is looking after them. Four cards answer how are they now, and over time? And the task decides which card is lit — a review lights glucose, a missed walk lights activity, a milestone lights the score — so the eye lands where the conversation will start.

Who it's for:

• Primarily: health coaches, care providers, nutritionists and nurses.
• Also: members, who see their own data in a different app.

01The profile — The coach's task on top, the member underneath.

Five questions, top to bottom

In the prototype above, step through the tasks with Next task and watch which card the page lights. Scrub the glucose chart to read an hour.

The questions a coach needs answered at a glance:

• What do I need to do next? The task bar.
• Who am I looking at? The member header.
• How are they doing right now? The cards.
• How have they been doing over time? Averages and a score that moves slowly.
• How do I dig deeper? Cards that open, so nothing needs a new tab.

02The task bar — The current task, why it exists, and Chat, Done and Next.

The task bar

It turns the profile from a place to look into a queue to work through: Review · 1 day ago, a line on why the task exists, and three actions. Chat opens a message already drafted from the task, which the coach edits rather than writes.

The member header

Age, gender and member ID; the conditions (Diabetes) and programme (Research MRT v5) that change how the numbers should be read; and the care team — coach, provider and nurse — by name.

03The header — Who this is, and who else is looking after them.

Four cards

• Blood sugar: the day against the target band, with today's and the five-day average. It's the biggest card because, for a member with diabetes, it's where most conversations start.
• Metabolic health: one score — 90, up 30 since joining — that opens into what it's made of. It moves slowly on purpose, so it rewards habits, not single good days.
• Activity: active minutes, calories and steps, each against a goal.
• Body composition: weight, BMI, muscle and body fat, with change since joining (weight down 7.3).

04The cards — Glucose, score, activity and body, one question each.

Visual design

Blue for actions, green for progress, warm tones for activity, so the kinds of number are told apart before they're read — and labels throughout, so none of it depends on colour. Clear type hierarchy separates titles, labels and values; generous space and defined card edges keep a dense screen calm.

05Hierarchy — Space and card edges carrying a dense screen.

What it's for

• Coaches: less time hunting, quicker to see who needs attention, and conversations that start from the member's own data.
• Members: more personal support, which is what keeps people on the program.
• The program: more members per coach without longer days.

Conclusion

The profile holds a lot of data, but it doesn't ask the coach to read all of it. It asks one question at a time, in the order a coach would, and lets the task of the moment decide where to look first.