Pharmacometrics is our deepest specialism. It asks the questions a standard clinical report does not: how does exposure vary between patients, what drives that variation, how does exposure relate to benefit and risk, and which dose should we take forward?

We use mathematical and statistical models to answer those questions with data, from preclinical translation through to late-phase planning. We work in Phoenix WinNonlin for pharmacokinetic analysis and NONMEM for population PK and PK/PD modelling.

Questions we help answer

  • How does drug exposure vary across patients?
  • Which patient characteristics, such as weight, age or organ function, influence pharmacokinetics?
  • How does exposure relate to efficacy or safety?
  • What dose or regimen should we test next?
  • How would alternative clinical scenarios play out?
What should the design of the next study take from this one?
  • Noncompartmental analysis. We assess key PK parameters, including Cmax, Tmax, AUC and half-life, as well as clearance and volume estimates where supported by the available data. This provides an exposure baseline for comparisons and modelling.
  • Compartmental PK modelling. Models that describe absorption, distribution and elimination, chosen to fit the data and the question, and used to simulate other regimens.
  • Population PK modelling. Nonlinear mixed-effects models that capture typical behaviour and the variability between individuals, with covariate analysis to see what explains it.
  • Pharmacodynamic modelling. Emax, sigmoidal Emax, indirect response and other models linking exposure to effect, chosen according to the mechanism and the data.
  • Integrated PK/PD modelling. One framework linking concentration, biomarkers and clinical response.
  • Exposure-response analysis. Relationships between exposure and efficacy, safety and biomarkers, to support dose selection and benefit-risk discussions.
  • Allometric scaling and preclinical-to-clinical translation. Cross-species scaling, human exposure projections and first-in-human dose planning, alongside the pharmacology and toxicology evidence.
  • Dose optimisation and simulation. Model-based comparison of doses, intervals and scenarios, including expected exposure, variability and target attainment.
  • IVIVC. Where data allow, relationships between in vitro dissolution and in vivo performance.
  • Model-informed drug development. Pulling PK, PD and other quantitative evidence together to inform trial design, dose and development strategy.
  • PK/PD data programming. Model-ready datasets, data checks, diagnostics and reproducible reports.

How we approach a modelling project

Define the objective. What decision is this analysis meant to inform?
  • Assess the data. Concentrations, dosing records, covariates and endpoints are reviewed before any model is fitted.
  • Build the model. Methods are chosen for the question and the data. Assumptions are written down.
  • Evaluate it. Goodness of fit, diagnostics, uncertainty, predictive performance and sensitivity to key assumptions.
  • Simulate and interpret. Scenarios are run and results are translated into plain findings.
  • Report it. Methods, assumptions, results and limitations are documented so another scientist can review and reproduce the work.

Why it matters beyond the model

Because pharmacometrics sits next to data management, programming and biostatistics here, model datasets are built from cleaned, standardised data and the results can be read against the statistical findings of the same study.

If you have PK data and a dose question, send us both.