How it works
TPAS asks the questions your protocol asks, weighs what the health worker observes, and recommends only what the evidence supports. When it is unsure, it refers.
The reasoning
A score can be high and still not count. Until the health worker has examined the child, answered the safety checks and taken the decisive measurement, TPAS holds its confidence down.
Every answer records where it came from: the health worker's observation, a vital sign, a rapid test, or only the complaint. A complaint scores nothing until someone has checked it.
A diagnosis must pass its threshold and be confirmed by its decisive measurement. A high score without the measurement keeps asking.
A danger finding ends the questions, raises the severe diagnosis and fires the referral. Home treatment is not offered, and a pre-referral dose is given where the protocol says so.
Weights, thresholds and dose bands are set by the clinician who wrote the protocol. The evidence behind each score is kept with the visit.
A child with fever and cough is assessed for malaria and pneumonia together. Shared questions are asked once, and one plan is checked for drug interactions.
A corrected answer replaces the old one and keeps a link to it. Answers that depended on it are asked again, and the referral fires or clears.
A health worker can skip a question they cannot ask, and must say why. The reason stays with the visit for the supervisor to read.
In the field
Once a visit is open, questions, scoring, vitals, rapid tests, medicines given, referrals and escalations are recorded on the phone. Each kind waits in its own queue and is sent in order when the phone reconnects. Unsent work is never cleared.
A 60-second breath-count timer, age-aware ranges for eight vital signs with critical flags, and the dose for this child's weight, with a warning when no weight was taken.
A high-contrast mode one tap from the home screen. Screen-reader labels throughout, with urgent results announced as they appear.
The handoff note follows the protocol's template and can be printed or shared from the phone. Leaving a visit without referring asks why, and the supervisor sees the answer.
Each treatment comes with the protocol's own counseling, written for the caregiver: how to give the medicine, when to come back, and what should bring them back sooner.
Escalations carry a deadline. When it passes without an outcome, the supervisor is alerted. The outcome is recorded when it comes.
One engine
The decision engine is one program, built once and run in two places: on the health worker's phone and on the server. The signed protocol is the only input that changes what it does.
Clinical content
The demonstration protocols follow WHO's Integrated Management of Childhood Illness and community case management guidance, as adopted in Liberia's community health assistant algorithm.
Asked first for every child under five. Any one sign refers, whatever else is wrong.
Cough or difficult breathing, 2 to 59 months. Timed breath count, chest indrawing, amoxicillin by weight.
2 to 59 months. Dehydration classified, dysentery recognized, ORS and zinc.
Uncomplicated and severe. Treated only after a positive rapid test. Severe signs refer with pre-referral treatment.
Next in line: 26 diagnoses are already defined, including hypertension, diabetes, asthma, COPD and pre-eclampsia, and 32 WHO reference protocols sit in the library as drafts for a medical director to complete.
How we test it
Each serious defect found on a handset is fixed at its cause and given a lasting guard: a server rule that refuses the unsafe state, or a test that fails if it ever returns. We also attack our own approval rules on purpose, trying to get around signatures, roles and the order of steps.
Request a demonstration
In one hour, with demonstration data:
info@diggysolutions.com · +231 775 711 477
We reply within one business day. Oldest Congo Town, Monrovia, Liberia.