A case study about who gets found in timeFellowship research · 24 weeks · 2023

Oral cancer is easyto see, and stillfound too late.

OncareaOral cancer screening for rural outreach camps, built at iHub Divya Sampark, IIT Roorkee. I was the Interaction Designer and Research Lead.

A mouth can be checked in a minute with a light and a trained eye. Yet most oral cancers in India are found at an advanced stage, and fewer than 2 in 100 rural adults have ever been screened. This is the story of why rural patients are left behind, and how Oncarea puts a doctor, a guided camera and a careful AI into the one day a camp is in town.

1.3%of rural men ever screened17
70%found at an advanced stage2
Doctor app: habit and symptom questions for a patient Guided oral scan with a mouth overlay over the live camera Doctor app home with the next session and patient totals
Where screening breaks4 links
01 · Question one

Six steps, each backed by a published source. The person is Rohan, our research persona.

    Every screening is a race against the stage it is found at.

    The context · IndiaArrows open the source
    02 · Question two

    Each step answers one of the breaks above, and links to the design decision behind it.

      Project at a glance

      A 24 week fellowship project. I led research, system design and the final screens across app, web and clinic dashboard.

      IIT Roorkee29.8649° N 77.8966° E
      03 · Research

      We asked doctors about liability, and patients about fear.

      Semi structured interviews let us follow what people cared about instead of a script. These were the questions that opened the most.

      What the research told us to build: a doctor first system, with limited patient autonomy and human in the loop AI.

      04 · People

      One wants context. The other wants calm.

      Two personas and empathy maps from the interviews. They pull in opposite directions: the doctor wants more information, the patient wants less. The system has to serve both at once.

      05 · What changes

      From an app that diagnoses to a system that supports.

      We leave nothing to chance

      Design focus: a clinically realistic, ethically constrained system for on site screening and outreach programmes.

      06 · The system

      Clinical architecture before any screen.

      I mapped the clinical journey as five layers held together by three principles. The AI has a place in it, but only where a doctor is watching.

      ActionRuleWho
      07 · Blueprint

      Three surfaces and one flow of data. Open any section for the full detail.

      Behind the screens: the full blueprint.

      Deliverables4
      Information architectureDOCTOR APP

      Four tabs, matching the four things a doctor does in a day. Screening lives one tap from Home, because it is the reason the app exists.

      Feature map8 GROUPS
      User flows4 FLOWS
      Concept and wireframesLOW FIDELITY

      We explored every possible flow through scenario role play, then turned them into low fidelity pencil prototypes for early testing.

      Pencil wireframes of the doctor home, patient details, patient report and appointment screens
      The original feature flow diagram
      08 · Design decisions

      For each one: what triggered the call, the evidence behind it, the options we weighed, and what it cost.

      Nine decisions, each backed by evidence.

      09 · Final screens

      Snapshots of every deliverable.

      Through several rounds we simplified patient interactions, added scan readiness checks, improved confirmation feedback and reduced cognitive load. For patients, most screens follow the same pattern as the doctor's screening flow.

      10 · Style guide

      One primary ramp, a ten step neutral scale, four semantic colours and Poppins in six weights. Built with library components so doctor, patient and web stay consistent.

      One purple for action, everything else steps back.

      Primary
      Neutral
      Semantic
      Type · Poppins
      Components, live · tap them
      Buttons
      Clinical question
      Does the patient have pain opening the mouth?
      Habits
      Does the patient use any of these?
      Appointment status
      UpcomingPendingCancelledCompleted
      Tuesday
      Wednesday
      Wallet
      Your balance1340.56Coins
      Rewards will be available after successful onboarding!
      Nudge
      Set your availability to start attending patients!

      Nudges sit where the problem is, say what to do, and use coral only when something blocks the next step.

      Rules I held the screens to

      2023
      fellowship

      Completed and certified by iHUB DivyaSampark, IIT Roorkee. The work ended at design and doctor feedback, so here is what I would measure next.

      11 · Outcome
      Pilot metricWhat it tells usTests
      12 · Sources

      Where the numbers come from

      Every outside number on this page links here. Project facts come from our own research.

        Next case study01 · 2026

        Why people fail at languages with apps built for retention, and what I changed.

        120+ learnersCHI 2027 under review13 weeks
        Read the case study
        mygermanly.com
        Back to portfolio All case studies
        Chanakya UG Fellowship certificate · iHUB DivyaSampark, IIT RoorkeePDF
        Chanakya UG Fellowship certificate: Abhishek Jha completed the fellowship on Cancer Detection and Cancer Care Ecosystem under Prof. Partha Pratim Roy, 2 Nov 2022 to 30 Apr 2023