From the Guidelines
Statistics for medical research should prioritize the application of sound statistical methodology to draw meaningful conclusions, as emphasized by the American Heart Association in their 2021 report 1. When conducting medical research, it is essential to focus on mastering key statistical concepts, including:
- Descriptive statistics (means, medians, standard deviations)
- Inferential statistics (p-values, confidence intervals)
- Study design principles Researchers should select appropriate statistical tests based on their data type and research questions, such as:
- T-tests and ANOVA for comparing groups
- Regression analyses for examining relationships between variables
- Survival analyses for time-to-event outcomes Sample size calculations are crucial to ensure adequate statistical power, typically aiming for 80-90% power to detect clinically meaningful differences, as recommended by the SISAQOL consortium in 2020 1. Additionally, researchers should understand concepts like statistical significance (typically p<0.05) versus clinical significance and be familiar with measures like number needed to treat (NNT) and absolute risk reduction (ARR) when evaluating interventions. Statistical software packages such as SPSS, R, or SAS are valuable tools for conducting these analyses. Presenting clinically relevant outcome measures, estimates with 95% confidence intervals, and matching the order of results to statistical analyses are also essential, as outlined in the guidelines for statistical reporting in the British Journal of Dermatology in 2015 1.
From the Research
Importance of Statistics in Medical Research
- Statistics plays a crucial role in medical research, starting from the planning stage of a clinical trial or laboratory experiment to establish the design and size of an experiment 2
- It is used to make inferences valid in a wider population, and advanced statistical methods are often needed to make relevant inferences from observed data 2, 3
Types of Statistical Methods Used
- Regression analysis is a standard method in medical research, including linear regression, logistic regression, Cox regression, and Poisson or negative binomial regression 4
- Descriptive and inferential methods, including regression modeling and propensity scores, are also frequently used in clinical research studies 3
- Other methods, such as N-of-1-pathways, can be used to unveil personal deregulated mechanisms from a single pair of RNA-Seq samples 5
Applications of Statistical Methods
- Statistical methods are used to quantify the relation between several variables and the outcome variable, and to control for potential confounders in observational studies 4
- They are also used to diagnose and treat diseases, such as pulmonary embolism, by evaluating the accuracy of different diagnostic tests 6
- Advanced statistical methods, such as systematic review and meta-analysis, can be used to pool estimates of sensitivity and specificity of different diagnostic tests 6
Collaboration and Interpretation
- Close collaboration between statisticians and other members of a research team is needed to ensure a seamless integration of the statistical elements into the reporting and discussion of research outcomes 2, 3
- Researchers need a basic understanding of statistical models to ensure that they are well defined and their findings are fully reported and correctly interpreted 4