We have been taking foot selfies for years now. The idea is almost embarrassingly simple: a person living with diabetes points a phone at the sole of their foot, sends us the picture, and we catch trouble while it is still small and reversible. We hold weekly foot selfie rounds at SALSA to do exactly that. For the back story, see our original home system, our early work on which imaging approaches actually work, and how the foot selfie carried our patients through the pandemic.
Here is the uncomfortable truth, though. Most of the elegant remote-monitoring tools we build in Los Angeles or London or Manchester are designed for places that look like Los Angeles, London, or Manchester. The vast majority of the world’s diabetic feet are not in those places. They are in low- and middle-income countries, where a beautiful app built in a high-resource lab can land like a spaceship in a cornfield: technically impressive, practically useless.
So this new paper, just out in BMC Health Services Research, is one I am proud to be a small part of. Led by Zafra-Tanaka JH, Almeida-Huanca G, Jauregui J, Yudith Quispe ,and a wonderful interdisciplinary team across Peru, Geneva, London, Sydney, Boston, Chicago and USC, it describes how we co-created a Foot Selfie intervention for Peru. Not for Peru, and not about Peru, but with Peru.
What they actually did
Between December 2024 and April 2025 the team ran eight workshops, four in Lima (urban) and four in Piura (semi-rural), with more than fifty people living with diabetes, their caregivers, and the doctors and nurses who care for them. They followed the NIHR principles of co-production: power sharing, including all perspectives, respect, reciprocity, and building relationships. In plain terms, everyone in the room was treated as an expert in something. They mapped the patient journey, then handed people a working beta of a smartphone app, one with an AI model trained to spot pre-ulcerative lesions in foot photographs, and asked them to doubt it, break it, and improve it.
What the patients taught the engineers
This is the part I love. The end-users did not ask for more technology. They asked the app to meet them where they actually live.
People who cannot bend down to press a button needed a voice-activated shutter and on-screen visual guides to take the picture. So those got built.
People were frightened by scary results, so the team sat with patients and asked what each message made them feel and do, then rewrote the wording to prompt action (go see your doctor) without provoking panic.
People wanted to learn and to be nudged, so a new My Care module appeared, with educational resources validated by people with lived experience, plus an alarm and a calendar.
People wanted to find each other and find services, so a More module arrived with directories of diabetes organizations and clinics with foot units. And the Home screen grew instructional videos, an FAQ, and a help chatbot, because in every single group people asked the same question: how do I actually use this thing?
My favorite detail in the entire paper has nothing to do with code. When the clinical term callo hemorrágico (hemorrhagic callus) sailed over people’s heads, the team swapped in the local term sangre molida, roughly mixed blood. Suddenly everyone understood. That is co-creation in a single phrase. You cannot prevent an ulcer you cannot name.
What they said no to, and why that matters
Co-creation is not a suggestion box where every idea wins. Patients wanted WhatsApp peer groups; the team declined, honestly, citing cost and the need for constant moderation, and offered directories instead. Clinicians wanted a provider dashboard; the team set it aside because this tool was built for patients first. They also had to gently manage big expectations about chatbots that were simply not safe to ship without proper testing. Naming those limits out loud, while still honoring the input, is the hard and honest part of this work.
Why this sits squarely in the remission framework
Regular readers know I think about the healed diabetic foot the way oncology thinks about cancer: not cured, but in remission, requiring lifelong surveillance for ulcer-free, hospital-free, activity-rich days. A foot selfie is surveillance you can run from your own living room. The potential advance in this paper is the insistence that the surveillance itself be designed in the living rooms where it will happen, in Tambo Grande as carefully as in Toluca Lake. It belongs to the same family of work as our friends and partners in Manchester building consistent foot imaging apps and AI-assisted DFU detection, with the added insistence that the human design the machine, not the other way around– see the diabetic foot grand challenge underwritten by NVIDIA when they were only a tiny multi hundred billion dollar gaming company.
The team is candid about the limits. They did not directly measure the impact of participation, and a pilot study is still to come. But the products themselves, the modules, the wording, the voice shutter, are the fingerprints of the people who will use them. That pilot will tell us about usability, adherence, and early signals of impact.
This work was supported by the UK Medical Research Council (MR/X004163/1).
Read it, and watch this space. The foot selfie keeps finding new feet.
Almeida-Huanca G, Jauregui J, Bustos D, et al. The co-creation of Foot Selfie, a patient-centered mobile intervention to prevent diabetic foot ulcers. BMC Health Serv Res. 2026. doi:10.1186/s12913-026-14718-5
#FootSelfie #ActAgainstAmputation #DiabeticFoot #RemoteMonitoring #mHealth #DiabeticFootInRemission #Peru #GlobalHealth #LMIC #ToeFlowandGo
Leave a Reply