Version 1.2 | Designed by Rob ter Heine
A freely available, open-source Shiny application for pharmacokinetic non-compartmental analysis (NCA), bioequivalence testing, study planning, and data visualization. Developed by the Radboud Applied Pharmacometrics research group at Radboudumc, Nijmegen, The Netherlands.
Built for pharmacokinetic scientists at the start of their career. Every setting has a plain-language explanation. No abbreviations without context.
The app offers six self-contained workflow paths accessible from a central hub:
1. Plan a Study — Calculate sample size or power for bioequivalence studies. Supports standard ABE, highly variable drug designs (ABEL/EMA, RSABE/FDA), narrow therapeutic index (NTID/FDA), dose-proportionality, and non-inferiority. Interactive power curves and sensitivity analyses for all design types.
2. Upload & Check Data — Import CSV or Excel files, auto-detect column names (including CDISC, WinNonlin export, and European naming conventions), set BLQ handling rules (6 rules including LLOQ/2 for drugs with absorption lag), and run 20+ automated data quality checks.
3. Visualize Data (new in v1.2) — Create publication-ready concentration-time plots directly from uploaded data. Individual profiles (spaghetti plot) with flexible colour grouping, and geometric mean ± geometric CV% summary curves with treatment overlays for crossover data. Export to PNG, PDF, or SVG at journal-submission resolution (up to 600 DPI). Includes an auto-generated figure legend ready to paste into a manuscript.
4. Analyze One Subject at a Time — Step through individual profiles with Previous/Next navigation. Includes manual data entry, interactive half-life inspector with point-by-point adjustment, and per-subject dose auto-fill from data.
5. Analyze All Subjects (Batch) — Run NCA on all subjects at once. Population summary statistics, individual profile grid, spaghetti and mean±SD plots, half-life review with recalculate, and steady-state display with AUCτ as the primary parameter.
6. Bioequivalence Testing — Complete BE workflow: NCA → ANOVA → 90% confidence intervals → forest plot → regulatory conclusion. Supports 2-period crossover, fixed-order crossover, 3-period crossover, replicate crossover, and parallel designs.
Plus: Statistical Methods page (publication-ready, expert-reviewed), Data Preparation Guide (10 scenario tabs), and About & Packages with version history.
Both the Batch NCA and Bioequivalence paths can generate a Complete Analysis Record — a self-contained zip file for regulatory submissions, publication supplements, or audit trails:
- results.xlsx — Individual NCA parameters, summary statistics, and (for BE) confidence intervals and ANOVA tables
- analysis_settings.json — Every setting that affects the analysis, with package versions, timestamp, and (if used in the same session) visualization settings
- reproduce_analysis.R — Standalone R script that reproduces the exact analysis without the app, with step-by-step comments
- data_integrity.txt — SHA-256 hash of the original data file
- analysis_summary.html — Self-contained summary with statistical methods, software environment, and instructions
- Original data file — Copy included so the package is self-contained
- R ≥ 4.1.0
- Required packages: NonCompart, PowerTOST, nlme, shiny, bslib, plotly, DT, readxl, dplyr, tidyr, ggplot2, openxlsx, jsonlite, digest
Clone the repository and install dependencies:
source("install_and_run.R")Or manually:
shiny::runApp()Available at robterheine.shinyapps.io/NCAassistant.
A consolidated validation package is available in validation/, following a risk-based approach consistent with ICH Q9 and GAMP 5 Category 5 principles.
Single validation script — run from the project root:
Rscript validation/validation.RThis executes automated tests and generates a results CSV with environment details, per-section results, and URS traceability.
Validation deliverables:
- User Requirement Specification (
validation/NCA_Assistant_URS.docx) — requirements across 8 categories (GEN, DAT, NCA, BE, PWR, EXP, UI, VIZ), with FMEA risk assessment, supplier assessment, ALCOA+ data integrity framework, and change control procedures - IQ/OQ/PQ Protocol (
validation/NCA_Assistant_IQOQPQ.docx) — every automated and manual test listed individually with method, expected result, URS cross-reference, and criticality classification - Consolidated test script (
validation/validation.R) — automated tests + manual test definitions, covering IQ, data handling, NCA accuracy, bioequivalence, power/sample size, export/reproducibility, usability, and visualization (URS-VIZ)
NCA accuracy validated against analytical ground truth (mono-exponential IV bolus) and R's built-in Theoph and Indometh datasets. Every test is classified CRITICAL (failure blocks qualification) or SUPPORTIVE (failure requires risk assessment). Visualization tests are classified SUPPORTIVE (non-critical: graphical output does not affect NCA parameters or regulatory conclusions).
See validation/README.md for detailed instructions on running the validation and adapting it for your organisation.
ter Heine R. NCA Assistant (v1.2). Radboud Applied Pharmacometrics, Radboudumc, Nijmegen, The Netherlands. https://github.com/robterheine/ncaassistant
Kim H, et al. NonCompart: Noncompartmental Analysis for Pharmacokinetic Data. Transl Clin Pharmacol. 2018;26(1):10-15.
GPL-3.
Radboud Applied Pharmacometrics — Radboudumc, Nijmegen