Clinical trial biostatistics
Biostatistics is the cornerstone of a clinical trial’s design, execution and interpretation. We provide it for clinical development, with a focus on oncology and haematology and a specialism in survival and longitudinal analysis: from endpoints, estimands and sample size, through the statistical analysis plan and any interim analyses, to the final analysis and reporting.
Michael Crowther, who directs our statistical methodology, has co-authored published methods on treatment switching, trial sample size, and joint models for trial data.
Four kinds of question
The statistics should follow the question the trial has to answer. These are the questions we help with.
Is the trial the right size?
Sample size and power for the endpoint you will actually analyse, including time-to-event outcomes, and adaptive or Bayesian designs where they fit.
What will the analysis estimate?
Primary, secondary and exploratory endpoints defined as estimands in the protocol and the analysis plan. A hazard ratio, a difference in restricted mean survival time and a survival curve answer different questions.
What bends the result?
Treatment switching and dropout can change what a trial shows. Michael Crowther has co-authored published methods for both.
Who watches the trial, and what is reported?
Interim analyses for efficacy, futility or safety, independent statistical reporting to a data monitoring committee, and the tables, listings and figures for the clinical study report.
Statistical services across the life of a trial
Services can be tailored to your needs, by project or as an allocation of resource. Our clinical development work is focused on haematology and oncology; our haematology page describes our experience with haematological malignancies. These are the services we offer; the published methods and analyses below show the statistical work in use.
Design and sample size
The right endpoints and statistical methods, sample size and power to detect a meaningful difference, estimands, and adaptive, Bayesian or innovative designs where appropriate.
Randomisation and blinding
Randomisation schemes that guard against bias, with proper stratification and allocation concealment, and support for the blinding processes that protect the trial’s integrity.
The statistical analysis plan
A detailed plan that predefines how data are handled, analysed and reported, spells out the estimands for each endpoint, and sets the interim analyses and stopping rules.
Interim analyses and monitoring
Planned interim analyses to assess efficacy, futility or safety, and independent statistical reporting that supports the data monitoring committee.
Final analysis and reporting
The full statistical analysis of the trial outcomes, tables, listings and figures for regulatory submissions, and support for clinical study reports and scientific publications.
Advanced analytics and modelling
Advanced survival analysis, longitudinal and mixed-effects modelling, real-world data and external controls where applicable, and simulation to optimise trial outcomes.
Trial methods co-authored by Michael Crowther
Nine peer-reviewed papers on the design and analysis of clinical trials, each with Michael Crowther as an author, published between 2012 and 2025. They cover treatment switching, sample size and the case for a further trial, dropout, outcomes such as length of stay, and trial-based cost-effectiveness. Most were published before Red Door Analytics was founded in 2021.
- 2025
- 2021Multivariate Generalized Linear Mixed-Effects Models for the Analysis of Clinical Trial–Based Cost-Effectiveness Data. Medical Decision Making 2021;41(6):667-684. View →
- 2020Assessing the Course of Organ Dysfunction Using Joint Longitudinal and Time-to-Event Modeling in the Vasopressin and Septic Shock Trial. Crit Care Expl 2020; 2:e0104. View →
- 2019Measuring and Analyzing Length of Stay in Critical Care Trials. Medical Care 2019;57(9):e53–e59. View →
- 2018Assessing methods for dealing with treatment switching in clinical trials: a follow-up simulation study. Statistical Methods in Medical Research 2018;27(3):765-784. View →
- 2017Adjusting for treatment switching in randomised controlled trials – a simulation study and simplified two-stage method. Statistical Methods in Medical Research 2017;26(2):724-751. View →
- 2014Adjusting survival time estimates to account for treatment switching in randomised controlled trials – an economic evaluation context: Methods, limitations and recommendations. Medical Decision Making 2014;34(3):387-402. View →
- 2013Simulation-based sample size calculations for designing new clinical trials and diagnostic test accuracy studies to update an existing meta-analysis. The Stata Journal 2013;13(3):451-473. View →
- 2012
Published analyses of two phase 3 trials
Two 2026 papers in European Urology Oncology use joint models to analyse data from the ARASENS and ARANOTE trials in metastatic hormone-sensitive prostate cancer: one of PSA trajectories and survival, one of quality of life and progression. Sara Ekberg and Michael Crowther are co-authors of both.
- 2026Associations Between Quality of Life and Disease Progression in Metastatic Hormone-sensitive Prostate Cancer: Insights from ARASENS and ARANOTE Trials. European Urology Oncology 2026. View →
- 2026Association of Deep and Durable Prostate-specific Antigen Responses with Outcomes in Metastatic Hormone-sensitive Prostate Cancer: Insights from ARASENS and ARANOTE Trials. European Urology Oncology 2026;9(2):251-258. View →
The people you would work with
The expertise behind this work is published and open to inspection.
The sample size and power calculator
Sample size and power for two-sample means, proportions, log-rank, Cox and flexible parametric survival, free in your browser.
merlin
Our framework for survival, joint, multi-state and competing-risks models. In Stata it fits all of them; the R version covers survival models so far.
Courses
Short courses on survival analysis and related methods, open to anyone or taught in-house for your team.
Co-author of published methods on treatment switching, trial sample size and joint models for trial data, and the author of the merlin framework for survival, joint and multi-state models.
View profile →
Research focused on cancer survival and survivorship, and a co-author of the joint-model analyses of the ARASENS and ARANOTE trials (European Urology Oncology, 2026).
View profile →Engagements that fit the problem
Expert advisory
Hours to days. A focused conversation or short review: a second opinion on a design, a sanity check on an analysis plan, or direction before you commit to an approach.
Project engagement
Weeks to months. Scoped to a defined deliverable, such as the design and analysis plan for a trial, with a named expert team, an agreed scope and milestones, and reproducible code and documentation.
Ongoing support
Continuous access across several studies a year, where the relationship compounds.
The methods, in the open
Primers and tools on the questions above. Nothing needs an account.
Sample size for survival analysis
Why survival studies are sized in events, the Schoenfeld formula with a worked example, how events become patients, and what changes the number.
Sample size and power calculator
Sample size and power for two-sample means, proportions, log-rank, Cox and flexible parametric survival.
RMST or a hazard ratio?
What restricted mean survival time is, how to estimate it and choose the horizon, how it compares with a hazard ratio and the median, and what to report.
When the hazard ratio misleads
See how one Cox hazard ratio hides a delayed, waning or crossing effect, and how RMST, milestone survival and the hazard ratio over time tell the real story.
What is non-collapsibility?
Why adjusting for a prognostic factor changes an odds ratio or hazard ratio without confounding, which measures are collapsible, and what to report instead.
What is censoring?
What censoring means in survival analysis: right, left and interval censoring, independent censoring, and how Kaplan–Meier and other methods handle it.