Biostatistics
OXON provides high-quality and accurate data analysis for your observational study
Our expert biostatisticians identify the best methods for assembling, analysing and reporting on your data in compliance with regulatory guidelines.
STATISTICAL ANALYSIS
Real-world observational data analysis addresses different study designs and thus requires expert statistical skills that are different to the analysis of clinical trials. OXON bridges the worlds of RWD and trials with analysis of pragmatic unblinded trials and meta-analysis of trials.
Our epidemiological statisticians work closely with our SAS programmers, data managers, outcomes researchers and medical writers to generate high-quality statistical analyses and reporting from a wide range of study types, including descriptive and comparative prospective, retrospective, case control, cross-sectional, surveys and meta-analysis of epidemiological data as well as trials.
Our team ensures that the methods for reporting and analysis are consistent with the needs of the sponsor and appropriate for the study.
We offer the following services:
• Statistical considerations in study design
• Sample size calculation
• Randomisation
• Statistical Analysis Plans (SAP)
• Programming of analysis datasets, tables, listings and figures
• Data transfers
• Statistical reports
• Statistical analysis is performed with SAS® software.
• Statistical analysis is run in accordance with relevant SOPs, Study Protocol, Statistical Analysis Plan and results in a Statistical Report.
STATISTICAL MODELLING
We have strong ties with the Medical Statistics Unit at the London School of Hygiene & Tropical Medicine and tailor the analysis to the needs of the data and design.
OXON performs non-standard advanced statistical techniques for observational data, including but not limited to the following:
• Propensity scores
• Disease risk scores
• Instrumental variable analysis
• Markov models
• Discrete event simulation
• Time series
• Trend analysis
• Analysis of trajectories for treatments and disease states
• Self-controlled studies
• Multiple imputation
• Competing risks (novel technique published in Lancet)
• Risk-benefit analysis (novel technique such as the ‘Win ratio’)
• Meta-analysis of different study types
• Bayesian statistics
• Simulation studies
• Machine learning