Humanly: Closing the say-do gap in respiratory healthcare innovation

The Challenge

Healthcare innovation has a problem that doesn't show up in clinical trials: people only stick with things they actually want to use.

Clinical trial settings are highly controlled. People are asked to adhere to an intervention so the trial can find out whether it's clinically effective. That data matters, but it's not the same as knowing whether someone, alone in their kitchen on a Tuesday evening, will pick up an app or a device and choose to use it. Without that, even a clinically effective intervention can fall flat in the real world.

What people say is often really different from what people do.

For Humanly, a design team focused on social impact projects across healthcare, charities and government, testing healthcare innovations meant going further than clinical trial data, further than focus groups, and further than even shadowing in person. Interviews surface what people predict they'd do. Reflective debriefs surface end-of-experience impressions, often missing the highs and lows in the middle. Shadowing surfaces real behaviour, but only briefly. As Jenni put it, it would be a bit weird to move in with somebody for a month and just watch them.

The team needed a way to observe over time, in real homes, around real lives, including the days where someone's health declined and they couldn't engage. Especially for lung conditions, where improvement only shows after weeks of consistent behaviour change, longitudinal context wasn't a nice-to-have. It was the whole point.

‍

The Approach

Humanly partnered with Asthma + Lung UK and a lived-experience co-design team to test two respiratory health innovations across an eight-week mobile ethnography study with 31 participants. Crucially, the research questions, the criteria by which products would be evaluated, and the mobile ethnography tasks themselves were co-designed with people living with lung conditions, not just designed for them.

Co-designed evaluation criteria. Before testing began, the co-design team developed product evaluation criteria combining lived experience, design and respiratory research perspectives, so each innovation was judged through the right lens, not a generic usability checklist.

Mobile ethnography over eight weeks. 31 people living with lung conditions tested the two innovations through Indeemo, sharing daily video diaries, screen recordings, photos and text updates as they used the products in real life.

Weekly task probes. Each week the team set a more focused task on a specific feature, often asking participants to use a feature while screen-recording and narrating their thoughts out loud, producing granular usability and content feedback.

60 semi-structured interviews and usage data analysis. The mobile ethnography sat alongside structured interviews and usage data, triangulating what people said with what they actually did.

A flexible field window. Eight weeks was deliberate. With lung conditions, improvements only really show after a few weeks of behaviour change, and participants frequently moved in and out of hospital, so the window had to accommodate real life, not control it out.

We can say to people, just use it as you would use it, do whatever works for you, and report back honestly to us. That's probably the closest we can get to finding out what people actually do in the real world.

‍

The Results

The richest output was a real-time view of how people actually used the products, not what they said about them in a debrief. That changed both what Humanly recommended and what the product teams could do mid-study.

A mid-study product update. Indeemo's live dashboard surfaced a significant usability issue with one of the products early on. Humanly flagged it to the product team, who released an update halfway through the study. The before-and-after evidence, from the same participants on the same product, showed a clear shift in user experience post-update.

Three usage segments emerged from the data. Humanly identified Adopters who followed the program as intended, Adapters who took what worked and tweaked it to fit their lives, and Rejectors who gave up early because the threshold for behaviour change was too high relative to their motivation. The segmentation gave the product teams a much sharper view of who their early users would be and what features might keep more people engaged.

Both products won additional funding and support. The impact data Humanly gathered through the study fed directly into the product teams' next funding rounds.

Peer-reviewed publication. The study was written up and published as Co-designing the future of respiratory healthcare in the journal CoDesign in December 2025, external validation of the rigour of both the co-design approach and the mobile ethnography methodology.

Stakeholders got close to the research, not just the headlines. Multimedia evidence including videos, screen recordings and photos meant product teams could see the experience for themselves rather than relying on a researcher's filtered summary.

The story moments mattered as much as the numbers. One participant who used to dread a hill on her walk home sent a picture of the sunset over that hill at the end of the eight weeks, with a note saying she now felt great when she got to the top. That kind of arc, visible only because the study spanned weeks of real life, is exactly what falls out of shorter or more controlled methods.