Biostatistics and statistical analysis for clinical research
Statistical strategy should follow the scientific question—not the other way round.
Data analysis is not a detached activity that begins once a database is locked. Decisions about objectives, primary outcomes, comparisons, data structure and uncertainty are part of study design. Bringing biostatistical thinking into the process early makes the eventual findings more transparent, interpretable and reproducible.
Heorfy provides specialist biostatistics consulting for clinical trials, observational research, epidemiology and health technology assessment projects. Support can extend across the analytical lifecycle or focus on a defined need within a research team, CRO or healthcare organisation.
How Heorfy can contribute
Specific scientific capabilities, tailored to each study design and project objective.
Statistical analysis plans (SAPs)
Analysis populations, estimands where relevant, methods and sensitivity analyses.
Trials and observational studies
Analysis strategies that reflect design, clinical objectives and uncertainty.
Advanced statistical modelling
Regression, survival analysis, mixed models, longitudinal and correlated data.
Inference and simulation
Assumption checking, robustness, resampling and alternative scenarios.
Reproducible analysis
Documented programming, analytical outputs, tables, figures and reports.
Interpretation and reporting
Scientific appraisal and support for manuscripts and presentations.
An analysis plan that anticipates the important choices
A statistical analysis plan (SAP) links the study questions to the methods used to answer them. Where appropriate, it specifies analysis populations, variables and estimands, primary methods, sensitivity analyses, missing-data strategies and the presentation of findings.
Prespecification reduces decisions driven by observed results and helps distinguish planned analyses from exploratory investigations. A well-designed SAP is not a list of statistical tests. It explains why each method is appropriate for the study design, the available data and the clinical objectives.
Methods matched to the actual complexity of the data
Research may call for regression models, survival analysis, repeated-measures methods, mixed-effects models, approaches for correlated data or recurrent-event analysis. Observational studies often require careful treatment of confounding, selection, heterogeneity and identification assumptions. Other projects benefit from simulation, resampling or quantitative evidence synthesis.
Methods should be chosen because their assumptions are defensible and their results are useful, not because they appear sophisticated. Where relevant, sensitivity analyses and alternative scenarios form an integral part of the work, allowing the robustness of key conclusions to be evaluated.
Reproducible findings with meaningful scientific interpretation
Useful analysis requires checked data, documented procedures and reporting that distinguishes statistical estimates, uncertainty, clinical relevance and limitations. Project outputs may include data-structure assessment, statistical programming and analysis, tables and figures, technical reports and support with scientific manuscripts or conference communications.
The result should be more than a collection of p-values. Effect sizes, confidence intervals, assumptions and consistency across analyses are essential to understand what the evidence does—and does not—show.
Flexible collaboration with methodological continuity
Heorfy can contribute to study design, prepare an SAP, deliver defined analyses or provide an independent statistical review. This approach supports both end-to-end projects and teams requiring specialist capacity for a particular stage, while maintaining close communication with clinical and scientific leads.
Biostatistics is equally important to the clinical validation of health technologies and to health economic evaluation, where analytical decisions determine the quality of evidence used downstream.