Fair-triage routing for peri-urban clinics in Colombo

This illustrative profile follows a fictional 2025 pilot in three Colombo peri-urban clinics. A routing model suggested who should be seen next, using arrival time, stated symptoms and a crude geographic risk score. Clinic staff noticed that patients travelling from underserved settlements were repeatedly placed later in the queue even when their symptoms matched earlier arrivals.

Sundara Health Lab convened patient advocates and the illustrative Office of Digital Public Infrastructure to compare predicted wait times against observed outcomes. The team published a model card, removed the geographic risk feature, and added a weekly disparity report that clinic managers can read without a data-science background.

The case is included because it shows a modest, reversible public-service experiment: the model remains assistive, a human nurse can override every suggestion, and the disparity table is posted in the waiting area. It is not a claim about any live Sri Lankan system.