A regional health authority came to us with a familiar problem: critical reporting bed occupancy, patient throughput, resource allocation took up to three days to compile from disconnected facility systems. Decisions that needed same-day data were being made on numbers that were already stale.
01 The challenge
Data lived across a dozen facility-level systems, each with its own format, refresh schedule, and data quality quirks. Manual consolidation was the bottleneck not the analysis itself.
02 The approach
- Deployed lightweight, standardised ingestion connectors at each facility feeding a central lakehouse in near real time.
- Built a shared data model so 'bed occupancy' and 'patient throughput' meant the same thing across every facility.
- Automated data quality checks flagged anomalies at the source before they ever reached a report.
- Delivered role-based dashboards so ward managers and regional leadership saw the same numbers, at the granularity each needed.
03 The result
Reporting time dropped from three days to under two hours. More importantly, the health authority could make resourcing decisions the same day an issue emerged, rather than reacting to a problem that was already three days old.
For the first time, we're managing today's numbers, not last week's.
04 What made it work
The technical architecture mattered, but the bigger unlock was standardising the data model before automating anything without a shared definition of each metric, faster pipelines would have just delivered inconsistent numbers faster.
Salcon Team
Dec 2024