Thinklytics

Mid-Size Pharmaceutical Company · Life Sciences · Boston, MA · 16 weeks

Clinical trial data governance framework deployed

A mid-size pharmaceutical company struggled with a 14-week FDA submission process due to clinical trial data scattered across six CROs and three internal systems in inconsistent formats. We created a data governance framework and built a CDISC-compliant data pipeline that cut the preparation time to three weeks.

Challenge

The company managed clinical trials through six contract research organizations, each using unique data formats and variable names. Preparing a CDISC-compliant submission meant manually converting data from all nine sources. This process took 14 weeks of biostatistician effort and created the main delay in meeting regulatory deadlines.

Approach

We created a data pipeline that met CDISC SDTM and ADaM standards, pulling data from six CROs and three internal sources. We applied consistent transformations and checked the results against CDISC rules. We also set up data governance to enforce standards on all future CRO contracts, stopping similar data issues before they start.

Outcome

We cut FDA submission prep time from 14 weeks to 3 by streamlining data workflows. Biostatisticians spent about 11 fewer weeks per submission on data transformation thanks to automated processes. CDISC compliance on first submission jumped from 84 to more than 99 of every 100 after tightening data standards. The governance framework we built is now the default for all new CRO contracts.

How We Got Our CRO Data Speaking CDISC for Cleaner, Smoother Reports

Alright, here’s the scoop. Every CRO we worked with had their own way of collecting and labeling data. And to make things trickier, they all used different file formats. Total chaos. So, we built this transformation layer that cleaned everything up and got it all into the CDISC SDTM format. Since their processes kept changing, we kept things flexible by maintaining CRO-specific mapping tables and updating them as needed. Plus, we set up automated checks to catch any data that didn’t meet the standards before it could cause problems down the line.

How We Tackled Managing Those New CRO Contracts

CROs were all over the map with their data submissions because there wasn’t a clear standard. So, we got together with regulatory affairs and clinical ops and hammered out a data standards appendix for every new CRO contract. This appendix laid out exactly which variable names to use, the formats, and how the data should be delivered. Problem solved, no more weird or off-the-wall data slipping through.

Results

  • 14 to 3 weeks FDA submission preparation time
  • 11 weeks Biostatistician time saved per submission
  • 84 to 99 of 100 CDISC conformance rate on first submission
  • 6 CROs Standardized under unified data governance framework

We used to spend 14 weeks assembling FDA submission packages because our CRO data was scattered. Thinklytics built a pipeline that cuts that down to 3 weeks and catches conformance errors early before they become bigger problems. It made our regulatory process a lot faster.

VP of Regulatory Affairs, Mid-Size Pharmaceutical Company

In the client's terms

The question

Which number is right, and who gets to decide?

The situation

Trials ran through six CROs, each with its own formats and variable names. Producing a CDISC-compliant submission meant converting nine sources by hand, fourteen weeks of biostatistician time, and it was the critical path on every regulatory deadline.

What we found

The fourteen weeks recurred every submission because nothing upstream had changed. Each new CRO contract added another set of conventions, so the conversion burden was growing with the pipeline.

What it produced

Submission preparation fell from 14 weeks to 3, returning about 11 weeks of biostatistician time per submission. First-submission CDISC conformance rose from 84 in 100 to over 99. The framework is now the default for all new CRO contracts.

The roadmap

A pipeline conforming to CDISC SDTM and ADaM standards ingesting all six CROs and three internal sources, with consistent transformations validated against CDISC rules rather than checked afterwards. Data standards were then written into the governance framework so future CRO contracts carry them, which is what stops the problem regenerating.

The work behind this result

Related case studies

Frequently asked questions

How long does CDISC submission preparation take?

Three weeks at this mid-size pharmaceutical company, down from fourteen, across six CROs. That returns about 11 weeks of biostatistician time per submission, and first-submission conformance rose from 84 in 100 to over 99.

How do you standardise data across multiple CROs?

Normalise on the way in and fix the contracts. A transformation layer maps all six CROs into CDISC SDTM with per-CRO mapping tables, and a data standards appendix specifying variable names, formats and delivery now goes into every new CRO contract.

Why does clinical submission preparation stay slow?

Because nothing upstream changes. The fourteen weeks recurred every submission here, and each new CRO contract added another set of conventions, so the conversion burden grew with the pipeline until the standards moved into the contracts.