Skip to content

Latest commit

 

History

History
39 lines (24 loc) · 2.53 KB

File metadata and controls

39 lines (24 loc) · 2.53 KB

Product

Register

product

Users

Jaga serves community health workers conducting guided, phone-based research capture with symptomatic adults aged 18 or older. Participants need plain-language results that avoid false reassurance. Demo operators need a complete, resettable synthetic-data flow. Programme and research reviewers need access to model scope, calibration, evidence boundaries, and limitations.

Product Purpose

Jaga is an investigational tuberculosis triage research prototype. It combines guided cough audio with supported clinical inputs to produce a calibrated, model-estimated relative risk band that can help study follow-up prioritization. It does not diagnose or rule out tuberculosis. Every participant is directed to confirmatory evaluation regardless of the estimate.

Success means the capture and result flow is understandable on a phone, honest about uncertainty, transient with patient inputs, bilingual in English and Bahasa Indonesia, and clear about the next clinical step.

Brand Personality

Calm, transparent, and evidence-disciplined. The interface should feel trustworthy without implying clinical authority, and technically capable without using medical-scan theatrics.

Anti-references

  • Diagnostic interfaces that imply certainty, clearance, or treatment advice.
  • Celebratory green states, checkmarks, or reassuring motion for lower model-estimated risk.
  • Futuristic medical-scan aesthetics, neon visualization, decorative glass effects, or unexplained technical imagery.
  • Dense research dashboards that make community health workers interpret model internals before finding the next step.
  • Marketing-led health interfaces that hide uncertainty, limitations, or cohort boundaries.

Design Principles

  1. Make the mandatory next step impossible to miss.
  2. Expose uncertainty without making users interpret research metadata.
  3. Keep evidence, model inspection, and clinical action visibly distinct.
  4. Use progressive disclosure for reviewer detail while keeping the participant-facing result concise.
  5. Design for guided field use first, with every critical state usable on low-end Android devices and intermittent networks.

Accessibility & Inclusion

Meet WCAG 2.2 AA. Preserve critical copy and controls at 320 CSS pixels and 200% zoom. Use touch targets of at least 44 by 44 CSS pixels, keyboard-operable controls, visible focus, accessible names and text alternatives, reduced-motion behavior, and redundant non-color status cues. English and Bahasa Indonesia must remain equivalent, human-reviewed, and layout-safe.