Governance
Demo control
Where this run stands, who you are acting as, and the demo script.
Sample values. Synthetic study, not real patients.
Demo role: Viewer: reads only. Switch. Roles are a demo switch, not sign-in.
This run
| Synthetic study | Type 2 diabetes, seed 7 |
| Primary endpoint | change in HbA1c from baseline to week 52 (continuous, classified by keywords; a person confirms it) |
| Signed export from the hospital zone | exp-71acbbf231 |
| Profile version | 1 |
| Clinician review | 0 of 28 values reviewed below 80% |
| Design report | Not signed. Nothing exported yet. Sign-off |
| Code version | 0.1.0+g089e218.src.52797a48 |
Demo role
Clinician review needs the clinician reviewer role; designs and sign-off need the analyst role; a fresh run needs the data steward.
Run the demo
Follow the acts in order. Each line is what to say on that screen.
| Act | Screen | Say |
|---|---|---|
| 1 Ingest | index.html | The package is in. The checklist shows what was found; patient-level files stayed in the hospital zone. |
| 1 Ingest | index.html | The protocol as structured data. Each field opens the paragraph it came from, in Word or PDF. |
| 1 Ingest | upload | Drop in any protocol. Rules read it now; a model reads it through the gateway when a key is set. |
| 2 Understand | retrospective.html | Planned against actual: enrolment took longer, and the outcome varied more than assumed. |
| 2 Understand | feedback.html | What ethics committees, monitors and investigators said, by theme, each with its source. |
| 3 Learn | profile.html | Local behaviour by layer. Each value says learned, elicited or assumed, with an interval for a new study. |
| 3 Learn | review | An investigator confirms values, corrects one (provincial attendance at harvest time) and rejects one. |
| 4 Apply | simulate.html | The same brief, sized on global defaults, loses power under local behaviour. Design B fixes it. |
| 4 Apply | design-report.en.html | The report a sponsor would buy. Every number has its label, interval and source. |
| 4 Apply | sign | A named biostatistician signs. Only then can the report be exported. |
| 4 Apply | design | Start from a published trial, say FLAURA, and see how it would run in Vietnam. |
| 5 Trust | trust.html | Replay: fitted on the first two-thirds, the last third predicted, then revealed. Calibration on synthetic studies. |
| 5 Trust | trust.html | What one study can and cannot show. |
Start again
Generates a new synthetic study for the chosen disease (Word, PDF and Markdown protocols in English and Vietnamese, patient-level tables, letters and reports), de-identifies it in the hospital zone and prepares the export. All reviews start pending.