Statistical concepts for research in emergency medical services

There is a critical need for high-quality, rigorous out-of-hospital research to drive further developments in out-of-hospital medicine and to improve the quality of care and outcomes for patients served by EMS. For this research to be valid and meaningful, it must be built on a foundation of appropriate statistical design and data interpretation. In this chapter, we address classical hypothesis testing, type I and II error, power analysis, commonly used statistical tests and their underlying assumptions, confidence intervals, multiple comparisons, subgroup analysis, intention-to-treat, interim data analysis in clinical trials, multivariable analysis, clustered data, missing data, and strategies for using statistical consultants.