Healthcare
Software for medical research and patient care programs.
We work with university research teams and the companies running care programs. Patients, clinicians and caregivers use the same products, and nobody is obliged to keep using them.
Clients include
What we’ve built for research teams and health companies
Four projects, each with a different mix of the same team: research, interface, mobile engineering and the data researchers see.
Survey participation that doubled instead of declining
A Stanford Medicine-led research team was losing response rates to the same drift affecting medical research everywhere. We designed and built HIPAA-compliant iOS and Android apps for participants, and two dashboards that let researchers watch response rates, drop-off points and recruitment performance while collection was still running. Participation doubled compared to the methods they had measured before.
A chemotherapy assessment patients could complete on their own
The Clinical Pharmacy Department at the University of Michigan needed to monitor chemotherapy-induced peripheral neuropathy through gait, balance and finger-coordination assessments. The assessments came from Apple ResearchKit, which brought validated clinical frameworks and fixed flows with them. The work was making those flows usable by someone mid-treatment.
A product tested with physicians before any of it was built
mEMR is physician-led and had a clear vision that had never met a real user. We ran the discovery and the UX research across both sides of the product: how physicians decide, and how patients would hand over their own clinical data. That turned into high-fidelity prototypes people could actually use. The concept was validated without writing code.
mEMR. UX research and interactive prototyping, before any development commitment.
Care programs that reach patients by voice and text
EmpowerHealth runs multiple care programs in parallel, which meant outreach that works for patients who will not download anything. We built a HIPAA-compliant platform combining automated voice and SMS, conversational AI and real-time analytics, with iOS and Android alongside it.
What changes when participation is voluntary
Most software can assume its users have to be there. In research and in care programs nobody does, and three things follow from that.
Someone can leave and never say so
A participant who stops answering does not file a complaint, and a study with attrition does not fail loudly. It produces weaker findings a year later. The interface decisions that look cosmetic here are the ones that decide whether the data set is complete enough to publish.
The researcher needs to see it while it’s happening
Knowing at the end which recruitment channel worked is knowing too late. Response rates, drop-off points and channel performance have to be visible during collection, because that is the only window in which anything can still be changed.
The validated framework decides what you can change
When an assessment is clinically validated, its flow is not yours to redesign. The work is making a fixed sequence usable by someone who is unwell, without touching the part that makes the measurement mean something. That constraint is the job, not an obstacle to it.
Questions we get before the first call
Do you work with hospitals and health systems?
No. Our health work has been with university research teams and with companies running patient programs. Health-system IT is a different practice with its own specialists: EHR implementation, provider workflows, the integration surface inside a hospital. We don’t claim it.Do you build medical devices or software as a medical device?
No. We’ve built patient-facing software on top of validated clinical assessments, which is a different thing. We have not taken a product through an FDA submission and we don’t hold a medical-device quality system, so a Class II build needs a partner who does.How do you handle HIPAA?
As a design constraint from the first week, not a layer added at the end. On the Stanford work it shaped data handling, storage, access control and the participant flows themselves; on EmpowerHealth it shaped what the platform may say, store and send across every care program running on it. HIPAA has no certification a company can hold; what it has is trained people, and the developers, designers and project managers who work on the projects that require it are HIPAA certified before the work starts and again while it runs. Separately, Onetree holds ISO/IEC 27001:2022, issued by UNIT and valid through March 2029.Can you work inside a study that’s already running?
Usually yes, and it’s the common case. The Stanford apps went into a research effort that was already collecting and already losing response rates, so the work was fitting into an active study rather than starting one, which changes what you’re allowed to move.What does the team actually look like?
Multidisciplinary from the first week. On the Stanford project that meant working alongside epidemiologists, statisticians, healthcare professionals and survey experts, and designing to what they needed to measure. There is no phase where the work changes hands and has to be explained again.
Tell us who has to keep coming back.
In a 45-minute working session we’ll go through who your product depends on, what would make them stop, and what you’d need to see while it’s happening rather than afterwards. Bring the number that’s been drifting.