Case study · Series-B biotech · Multi-trial
HIPAA-aligned clinical data warehouse on GCP — in 8 weeks
A Series-B biotech needed a single source of truth across three active trials, two CROs and a real-world-data partner. We built it on Google Cloud, on credits, with a path to AI-ready datasets from day one.
The situation
Trial data lived in two EDCs, two CRO SFTPs and a Snowflake instance owned by RWD partner. Biostats was stitching exports by hand for every readout. Leadership wanted one place to ask "how is the program doing?" and a story for AI on top.
What we did
- • Set up a HIPAA landing zone on GCP: Organization, folders, VPC-SC perimeter, CMEK, audit-log sink, Security Command Center, BAA executed.
- • Helped the client qualify for GCP credits covering 12 months of platform cost.
- • Built ingest pipelines: EDC SDTM exports, CRO SFTP, Snowflake share, all into a BigQuery raw zone.
- • Modeled the warehouse with dbt: subject, visit, AE, lab, dosing, with study-level conformance.
- • Delivered Looker Studio dashboards for ops, medical monitor and leadership.
- • Set up a de-identified zone using DLP and OMOP-flattened tables for downstream AI work.
The result
8 wks
Zero to first dashboard
12 mo
Funded by GCP credits
3
Trials in one warehouse
OMOP
AI-ready downstream
Leadership runs program reviews off live dashboards instead of email. Biostats spends time on analysis, not stitching CSVs. The de-identified zone is now the starting point for the team's first internal AI cohort builder.
What the client said
"We had GCP credits sitting on a shelf. Elintex turned them into a working clinical data platform — and the architecture is one we can actually maintain."— VP, Clinical Operations