Statistics in a Nutshell

Chapter 18. Medical and Epidemiological Statistics

Many of the statistics used in medicine and epidemiology are common to other fields: examples include the t-test (covered in Chapter 8), correlation (covered in Chapter 9), and the various types of regression and ANOVA (covered in Chapters 12–15). But some other statistics have been developed specifically to meet the needs of medical and epidemiological research (such as the odds ratio), and others, while common to many fields, are used so frequently in medicine and epidemiology that they are covered in this chapter (for instance, standardized rates).

Measures of Disease Frequency

Before getting into specific measures of disease frequency, it is worthwhile to discuss the meanings of several terms in common usage that are often confused. We can always report disease frequency in terms of the number of cases: there were 256 cases of tuberculosis (TB) in city A and 471 in city B last year, for instance. Raw numbers are useful for people who allocate current resources and plan future monetary and space allocation, because they need to know how many cases of TB and how many hip fractures to expect in the coming year so they can allocate resources accordingly. However, for research and planning at the national and international level, disease occurrence is more usefully described in terms of relative rather than absolute occurrence, because we often want to look at trends over time or across different geographical areas with different population ...