Essential statistics for the clinician

This article provides a working introduction to statistics for the core trainee, specialty trainee and keen medical student to aid their daily practice, interaction at journal clubs and scientific meetings and for examination preparation with a focus on surgical research. Data can be described as metric (quantitative) or categorical (qualitative). The mode, median or mean can be used as measures of location and the variation ratio, range, inter-quartile range and standard deviation as measures of spread. Cohort studies can be used to define risk reduction of a given intervention; case-control studies rely on odds ratios to approximate risk. Hypothesis tests can be used to establish statistical significance, which is distinct from clinical significance. Confounding can be controlled with multi-variable analyses. Correlation coefficients can be used to establish the relationship between two sets of continuous data, but is distinct from causation. Power calculations can be used to estimate the sample size required to generate a clinically and statistically significant result from a trial. It is important for clinicians to be able to understand these concepts and translate them into simple and understandable format for patients, colleagues and commissioners to maintain high quality care within the NHS.