Clinic operations, from reception to dispensary, on one system

The challenge

The organisation was delivering care at low cost in places with little of it, and running the whole operation on paper. Every stage of a patient's visit was a form, and every form was a place for the record to go missing or fall out of step with the one before it.

The constraint that shaped everything: consistency. Care had to follow the same protocols at every site, and on paper there was no way to enforce that or to know whether it was happening.

The context

A visit passes through several hands — reception, initial screening, the consulting clinician, the laboratory, the dispensary — and each needs a different view of the same patient and needs to hand something on to the next.

Protocols and standard operating procedures also change. On paper, a change means retraining every site and hoping.

The approach

I modelled the system around the patient's journey rather than around departments, because the journey is what was actually breaking. Each role gets the view it needs of one shared record, and handoffs are transitions in that record rather than a piece of paper crossing a room.

Protocols and medicine assignments were made centrally managed data rather than application logic — so a clinical change is a configuration change, not a release.

The solution

A mobile application covers clinic operations across the connected roles, backed by a web application for administration and resource management.

Protocols, standard operating procedures and medicine assignments are defined once, centrally, and applied everywhere. They can be added and edited without a code change, which is what allows clinical practice to be updated across the whole network at once.

AI & automation

None. This was a process and integration problem, and the answer was a well-modelled workflow.

It is worth being direct about that: an organisation delivering care on a tight budget needed its records to be consistent and its protocols to be followed. Nothing about that called for a model, and adding one would have consumed budget that was better spent on the parts that mattered.

Engineering

The dominant constraint was operating conditions. Clinics could not depend on continuous connectivity, so the system had to keep working during an outage and reconcile cleanly afterwards without losing or duplicating a patient record.

Role-based access had to be strict — clinical data with several roles touching one record is exactly where authorisation has to be right — while staying quick enough to use in a busy clinic. Making protocols configurable rather than hard-coded meant designing a data model that could express clinical rules without becoming a programming language nobody could safely edit.

Outcome

The system was deployed across more than twenty clinics, replacing the paper process with one connected record per patient and one central definition of how care should be delivered.

Staff stopped re-recording the same information at each stage of a visit.