Biostatistics in Science

Biostatisticians study the application of statistical theory and methods to analyze data collected by Scientists, such as geneticists, epidemiologists, medical doctors, dentists, etc. Such data are collected to further each of the above field study. The book <Basic Biostatistics for Geneticists and Epidemiologists> mentions two good examples to explain the statistical application in the field study.

For example, a genetic study to investigate whether one or more genes might predispose people to an increased risk of developing a specific disease would require an application of statistics to reach a valid conclusion.Another example, an application of statistics is required to reach a valid conclusion when a clinical study is conducted for the purpose of investigating which of the two pharmaceutical treatments is preferred in managing patients with a specific disease.

In the application of statistical theory, it is often but not limit to the responsibility of Biostatisticians to educate and communicate with Scientists about how the statistical methods need to adjust according to the complex study design and/or specific aim of the field study. The application often required some basic knowledge about the science review and background scientific literature review.

We will need to differentiate the two types of reasoning methods. In general, we use deductive reasoning to prove thins with certainty. In scientific method, we use inductive inference and can never prove anything with absolute certainty. 

Deduction Reasoning: General principles are applied to specific situation at hand in order to reach the best decision possible for a particular patient. This is from general to specific.  Most of the basic medical training centers around deductive reasoning, which is based on general scientific laws and what we can deduce from them.

Inductive Reasoning: We conduct experiments and comparative studies to focus on questions that arise in our work. We study a few patients, and from what we observe, we try to make rational inferences about what happens in general. This from specific subjects at hand to general. The Scientific method has the following basic steps:

1. Making observations - that is, collecting data.
2. Generating a hypothesis- the underlying law and order suggested by the data
3. Deciding how to test hypothesis - what critical data are required?
4. Experimenting - this leads to an inference that either rejects or affirms the hypothesis

If the hypothesis is rejected, then we go back to step 2.  If it is affirmed, this does not necessarily mean it is true, only that in light of current knowledge and methods it appears to be so.There are some clinical decisions must be made with variability in mind, such as,

• whether an observation on a patient should be considered normal or abnormal?
• Is a particular observations more typical of a person with disease or of a person without disease?
• Is the observation outside the range typically found in a healthy person?
• If the patient were examined tomorrow, would one obtain essentially the same observations?
• If more observations obtained, would the results be very close to the observations already had?

Inductive inference is a much riskier procedure than deductive inference. Like in mathematics, start with a set of axioms.Refereed to the following book for details,

Elston RC, Johnson W. Basic Biostatistics for Geneticists and Epidemiologists: A Practical Approach. 1st ed. Wiley; 2008.