Modernizing a Pharma Giant's Shiny App Framework with Open-Source Tools for Scalable Clinical Insights

Project overview

‍

The client launched a clinical submission automation initiative to standardize how teams produce regulator-ready outputs across Phase I to III studies. Appsilon supported the initiative by bringing clinical submissions expertise (submission standards and validation expectations) and turning it into a repeatable workflow teams could run at scale. This work sits directly in the regulatory submission path and supports teams producing ADaM datasets, define.xml, and other regulator-facing artifacts.

‍

‍

What was happening

‍

Submission dataset generation was slowed by legacy processes and manual checks. The bigger issue was key-person risk: progress depended on a small number of specialists who understood submission standards, what “regulator-ready” means in practice, and how to validate outputs. When exports or define.xml generation failed, the legacy tooling provided unclear errors, forcing teams into trial-and-error reruns and pushing fixes late in the cycle. That made changes hard to roll out consistently and timelines hard to predict. 

‍

‍

What we did

‍

We worked directly with clinical programming and standards users to capture submission requirements and validation expectations, then built them into standardized workflows. Outputs were validated in existing compliance tooling (P21). In addition, we improved transparency with clear, user-tested logging and validation messages so teams could see exactly what failed and fix issues faster, reducing rework and delays. 

‍

In plain terms, we helped turn expert judgement into repeatable rules, consistent outputs, and dependable checks, while also improving performance for large exports and setting up a fast feedback loop for continuous refinement. Appsilon also supported the client’s move toward open-source, helping them adopt R-based workflows without turning biometrics teams into an IT function.

‍

‍

Impact

‍

• We built a dedicated R package for submission exports and define.xml generation that replaces the legacy “black box” workflow with fast exports and clear, user-tested messages that explain what failed and what to fix
• 40GB submission dataset export improved from overnight to about 20 minutes. Predictable turnaround for high-volume dataset exports
• Reduced “key-person” dependency by embedding submission know-how into the process, not just in people’s heads
• Faster adoption across studies by making changes easier to roll out consistently
• Increased flexibility - previous solution only allowed for sas7bdat datasets, now the client can use any input format that is readable as a data.frame

‍

‍

‍

Who this helps

‍

• Clinical programmers and data standards teams producing submission datasets and outputs
• Regulatory submission teams relying on consistent, audit-ready artifacts
• Study teams and leaders who need quicker, more predictable turnaround

‍

“ If you lead biometrics, clinical programming, or submissions, we help you make it faster, repeatable, and easier to scale across studies.”

‍

‍

Ready to Modernize Your Clinical Systems?

‍

If you're facing similar challenges with aging internal tools, tight resources, or complex regulatory environments, we can help. Appsilon specializes in helping pharma teams transition to open-source in a compliant, practical way, so you gain speed and flexibility without increasing operational burden. Let's talk.

‍

‍