Will Drinking too Much Coffee Result in a Heart Attack? Risk Factors and Health
Many medical studies involve observation rather than experimentation in circumstances where the latter is not possible. The area of science that attempts to assess risk factors associated with a person, their diet, their environment, etc., which may resul
- PDF / 327,032 Bytes
- 20 Pages / 439.37 x 666.142 pts Page_size
- 4 Downloads / 169 Views
One of the first things taught in introductory statistics textbooks is that correlation is not causation. It is also one of the first things forgotten. Thomas Sowell. You’ve got to keep fighting-you’ve got to risk your life every six months to stay alive. Elia Kazan.
3.1
Introduction
In a well-designed and well-performed randomized clinical trial, an observed treatment effect can, with some confidence, be said to be caused by the different treatments patients received. Suppose, however, interest lies in investigating whether smoking cigarettes causes high blood pressure. In a clinical trial, people would be allocated randomly to smoking and non-smoking groups and then. . . There is no need to go any further. Such a trial would be completely unethical for obvious reasons. An alternative approach is to take say 100 people who are smokers, and an equal number who are non-smokers, and measure their blood pressures. Whilst such a study is possible (see Sect. 3.3 for an account of the classical study of the effects of smoking on health), the observation that there are significantly more people in the smoking group with high blood pressure than amongst the non-smokers does not lead to an unambiguous conclusion about whether smoking cigarettes causes an increase © Springer International Publishing Switzerland 2016 B.S. Everitt, Health and Lifestyle, DOI 10.1007/978-3-319-42565-8_3
47
48
Health and Lifestyle
in blood pressure. Any observed association (correlation) between smoking and high blood pressure has three possible explanations: • Smoking causes an increase in blood pressure. • People with high blood pressure tend to smoke more. • Some unidentified factor affects both the tendency to smoke and the tendency to high blood pressure. We are now in the area of epidemiology, the study of disease in the widest sense, including non-communicable and infectious diseases and their risk factors; the latter are circumstances associated with a person, their diet, their environment, etc., which may result in an increased risk of developing some disease or other. Risk factors may not always be causally related to the disease, as explained in the smoking example in the previous paragraph. (Interest may, of course, also centre on identifying factors that reduce the risk of suffering from a particular disease.) Epidemiological studies are characterized by observation rather than intervention and experiment which are the quintessential components of clinical trials. Sinclair (1951) made the rather unkind comment that ‘the use of the experimental method has brilliant discoveries to its credit, whereas the method of observation has achieved little.’ This is not only unkind but wrong, as will be seen later in this chapter. However, in many cases when dealing with human subjects, experimental studies are simply not possible and an observational approach is necessary. The researcher undertaking this type of study should perhaps keep in mind the following wise words of Austin Bradford Hill made in the 1960s: The observer may well have to b
Data Loading...
