System Walkthrough

Chipatala Field

How data moves from a village HSA post all the way to the Minister of Health — and back down again as decisions.

HSA→ Clinic→ Doctor→ DHO→ Minister

The Big Idea

One shared record book

Every single thing that happens — a patient registered, a referral made, medicine arriving or given out, a patient admitted — is written down as one entry in a shared, permanent log. Nothing is ever erased or overwritten, only added to. Every dashboard is just a different view over that same log, filtered to what that person is allowed to see. This one idea is what makes offline work, what makes the audit trail trustworthy, and what makes disease-outbreak alerts possible.

Minister & DHO

Account creation — who can create whom

There's no public "sign up" page. Nobody creates their own account.
The Minister or a DHO creates every account — an HSA, a clinic nurse, a doctor, another DHO — setting a username, a temporary password, a role, and which facility they belong to.
The very first account (the Minister's own) is created directly against the database once, before anyone exists to create it through the app.
Every staff member changes their own password on first login. Only the person themselves can change it — and only by typing their current one first.
If someone leaves, their account is deactivated, not deleted — they can't log in, but their name stays attached to every record they touched, for the audit trail.

HSA Field Worker

Where a patient's story begins

An HSA opens the app in the field, on a basic phone, often with no signal.
They register a patient — name, age, gender, village, a quick vitals check, and a working diagnosis picked from a list (malaria, cholera, TB, measles...). This field is what later powers outbreak detection, so it has to be a proper pick, not free text.
If the case looks serious, the HSA creates a referral to the linked clinic, with a reason.
All of this saves to the phone instantly — it never waits for signal.

Clinic Nurse & Doctor

What happens to a referral

When the HSA's phone gets signal — even a few seconds — it quietly pushes everything queued up.
The clinic nurse sees the referral in their inbox, taps Accept, and can walk the patient through to hospital if needed.
The doctor sees it in their own queue, opens the patient's full history, and can confirm the diagnosis — a doctor's confirmed diagnosis always overrides the HSA's initial guess for anything downstream.
If the patient needs to stay, the doctor admits them — a ward, a bed number. They show as "currently admitted" at every level above, and are later discharged with an outcome. This is what makes bed occupancy real, not guessed.

Doctor · DHO · Minister

Medicine — tracked, not typed in

There's no manual "current stock" number anyone types in. Every delivery and every dose given out is logged as its own event — stock in, stock out.
The actual stock level at any moment is just: everything that came in, minus everything that went out — it can't silently drift from reality the way a spreadsheet can.
A hospital doctor sees their own pharmacy's stock, low items flagged (under 30 units).
A DHO sees that same picture rolled up across their whole district — which facilities risk running out.
The Minister sees it rolled up nationally, by district.

Every Device

Offline vs. online — the actual mechanism

Every device keeps its own local copy of whatever it's created, whether or not it has signal.
No signal → new records pile up in a small "waiting to sync" queue on the device. Nothing is lost, nothing blocks the worker from continuing.
Signal returns → that queue pushes up to the shared server. A record sent twice by accident is simply ignored as a duplicate, never double-counted.
Clinic, doctor, DHO, and Minister are assumed to have decent connectivity — their views refresh automatically every few seconds, with a visible "last updated" time so nobody mistakes it for instant.

Doctor · DHO · Minister

How an outbreak actually gets flagged

Every disease has its own trigger number for a single facility, over a rolling 7 days — tuned to how dangerous a handful of cases of that disease actually is.
A second, stricter trigger exists specifically for children under 5 — the group most at risk for diseases like measles and pneumonia — so a small children's cluster can raise an alarm even if the total count alone wouldn't.
Crossing either number doesn't wait for someone to notice a chart — a red alert banner appears automatically for the doctor at that hospital, the DHO over that district, and the Minister nationally, all at once.
Every alert breaks down by age group and gender, so it's immediately clear whether it's a children's outbreak or spread across all ages — which changes what action makes sense.

Cholera / acute watery diarrhoea2 cases

Measles3 cases (2 if under-5)

Tuberculosis (TB)5 cases

Malnutrition (under-5)6 cases

Pneumonia10 cases (5 if under-5)

Malaria15 cases (8 if under-5)

The shape of it, top to bottom

Nobody re-enters anything twice — every level above is looking at the same original record, aggregated one step further.

HSA

enters raw data in the field

→

Clinic / Doctor

treats, confirms diagnoses, admits patients

→

DHO

watches the whole district's load, stock, outbreaks

→

Minister

watches the whole country, one level further out