The challenge
Clinical, finance, and operations teams had built more than 4,000 Power BI reports on nightly extracts from the EHR reporting database. Numbers disagreed across departments, refreshes failed weekly, and nobody could say which length-of-stay metric was right.
What we did
Tulan landed EHR, claims, and workforce data in a HIPAA-governed lakehouse and used usage telemetry to find the 600 reports people actually opened. Those collapsed into 60 certified Direct Lake semantic models with a shared clinical metrics layer. Row-level security by facility and role was enforced through Purview and OneLake security.
The outcome
Executives, nurse leaders, and finance now use one definition for every core metric. Dashboards load in under two seconds, month-end close is 11 days faster, and capacity planning uses near real-time census data.
“We stopped arguing about whose number was right and started arguing about what to do about it.”
SVP and Chief Analytics Officer