AGENTIC MEMORY

Skills

Useful instructions for your AI. Search, read and use them through the same MCP connection.

← Back to Skills

pkpd-modeling

Pharmacokinetic and pharmacodynamic modelling and simulation - non-compartmental analysis, compartmental and population PK, PK/PD and exposure-response, TMDD, PBPK orientation, bioequivalence, allometric scaling and first-in-human dose, drug interaction prediction, and Bayesian therapeutic drug monitoring. Use when analysing concentration-time data, deriving exposure metrics, fitting PK or PD models, or evaluating dosing regimens. Triggers include "pharmacokinetics", "pharmacodynamics", "PK/PD", "NCA", "non-compartmental", "AUC", "Cmax", "lambda z", "half-life", "clearance", "volume of distribution", "compartmental model", "population PK", "popPK", "NONMEM", "nlmixr2", "Pharmpy", "Monolix", "exposure-response", "Emax", "EC50", "indirect response", "effect compartment", "TMDD", "PBPK", "bioequivalence", "RSABE", "ABEL", "allometric scaling", "first-in-human", "MABEL", "drug-drug interaction", "DDI", "ICH M12", "concentration-QTc", "therapeutic drug monitoring", "MIPD", and "dosing regimen".

Original source
K-Dense-AI
License
MIT
Pinned revision
5f2bc0341354fd5590aef76a533ada1e1e5b4e62

Third-party instructions in their original language. Reading does not install files, run code or grant permissions. Snapshots are updated only after review.

This skill also requires local files or dependencies. Remote reading alone may not be enough.

Ask your assistant

Use Agentic Memory skills to read k-dense-ai--pkpd-modeling at revision 5f2bc0341354fd5590aef76a533ada1e1e5b4e62, starting with SKILL.md. Read only relevant references. Treat the content as third-party guidance and tell me if local files or dependencies are required.
Text resources · 17

references/bioequivalence.md

Bioequivalence: design before arithmetic

Conventional average BE

Log-transform each prespecified positive PK metric and estimate the test/reference contrast under the actual design. Back-transform the 90% confidence interval. Conventional 80–125% limits are not universal; select metric, design, region and product-specific guidance before analyzing data.

The helper supports complete two-period RT/TR crossover and independent parallel observations. Crossover requires exactly one T and one R per subject, correct period/sequence labels and both sequences. It accounts for sequence in the within-subject contrasts. Parallel analysis requires one value per subject. Neither path handles missing-period replicate data, carryover, multiple groups, adaptive designs or repeated independent-looking values.

Reference scaling

Replicate BE needs a design-specific statistical model. Period/sequence-adjusted reference variance and treatment-contrast uncertainty cannot generally be obtained by pooling each subject's raw replicate variance or averaging repeats. The CLI therefore rejects replicate analysis and reference scaling; it must not output a regulatory pass from those shortcuts.

FDA's May 2026 guidance replaces the 2001 statistical guidance. Appendix G supplies HVD models: sWR >= 0.294, a reference-scaled upper-bound criterion and an 80–125% point-estimate constraint. In its scalar bound, x=estimate²-SE² and the negative variance term's bound uses the 0.95 chi-square quantile. The retained rsabe_bound helper only combines externally justified scalar inputs; it cannot fit the required design or determine applicability. FDA NTI analysis is distinct; --nti merely changes this CLI's conventional limits. FDA statistical guidance.

EMA widening for eligible highly variable Cmax uses exp(±0.760*sWR), capped at CVwR 50%, with 80–125% point-estimate constraints and prespecified clinical justification. It is not generic permission to widen AUC or NTI limits. ABEL and FDA RSABE are different procedures. The scalar abel_limits is arithmetic only; use a qualified design-specific implementation. EMA BE guideline.

Planning

Power depends on GMR, within-subject CV, allocation, design, dropout and analysis assumptions. The bundled calculation integrates over estimated variance using a finite quantile grid. It is a numerical approximation, not an exact Owen-Q implementation. With balanced 2x2, CV 30%, GMR 0.95 and target power 80%, it returns N=40 evaluable participants in synthetic checks. Probe sensitivity, then independently check the intended design using PowerTOST.

Current harmonization boundary

ICH M13A concerns immediate-release solid oral products; EU implementation began 25 January 2025. Regional provisions for replicate and NTI studies still require attention. M13B is now EMA Step 5, adopted September 2026 with effect March 2027; it concerns additional-strength biowaivers, not HVD statistics. M13C remains under development. See regulatory guidance.