Services

PK/PD, CDISC, HEOR & automation-packaged as real deliverables.

Deoyin PX provides project-ready service packages. Each engagement is scoped with clear inputs, outputs, and timelines, using reproducible code and validated workflows.

Service Lines

PK/PD Analytics & Reporting

End-to-end PK workflows: from raw concentration-time data through QC, NCA-ready tables, and submission-friendly appendices.

Cmax, Tmax, AUC NCA datasets PK listings & figures

CDISC & Regulatory Programming

ADaM programming (ADSL, ADAE, ADTTE, and study-specific ADaMs), TLF production, and Pinnacle 21 validation with issue tracking.

ADaM packages TLF libraries P21 reports

HEOR & ICER Modeling

ICER models in R and Excel, sensitivity analyses (SA), probabilistic sensitivity analyses (PSA), and visual decision dashboards for market access.

Cost-effectiveness PSA Budget impact

Reporting Automation & Data Engineering

Automated generation of Word/Excel reports, ETL pipelines for clinical and PK data, and “one-click” workflows for repeat analyses.

R / Python / VBA ETL Report templates

AI-Assisted Medical Analytics

AI-supported drafting of PK narratives, CSR sections, and evidence summaries that remain anchored to validated analytical outputs.

Narrative drafts Summaries Evidence review

Workflow Apps & APIs

Shiny apps and Plumber APIs that package PK, CDISC, or HEOR logic into reusable internal tools with auditability and version control.

Shiny apps APIs Internal tools

Example Engagements

  • PK package for Phase I trial – NCA-ready datasets, PK tables/listings, and key exposure figures.
  • ADaM + TLF bundle – Core ADaM datasets, reusable TLF programs, and a validated P21 package.
  • ICER model build – Base cost-effectiveness model in R, Excel front-end, and PSA routines.
  • Reporting automation prototype – Turn a repetitive Word/Excel report into a parameterized script.
  • Shiny dashboard – Interactive PK or HEOR explorer for internal teams.

How scoping works

Each engagement is scoped around four questions:

  • What data will be available, and in what structure?
  • Which outputs are essential vs. “nice-to-have”?
  • What are the regulatory or internal standards to follow?
  • Where could automation meaningfully reduce future workload?

Discuss a project