What Is Statistical Analysis in Clinical Trials?
Most of us have a relationship with statistics. It is used as a tool throughout the community and in all disciplines, from economics to physics, from aquaculture research to knitting. As a researcher, you have probably taken at least one course in statistics at the university, and have used statistics in your clinical research in connection with publication.
Statistics is one of the tools that can be employed to make use of observations from the past to say something about the future. It is often divided into prediction analysis and causal cause-effect analysis. Prediction describes what will happen without any manipulation or influence, while causality deals with what happens when you actively go in and change something.
When conducting a clinical trial, the clinical decision maker is interested in causality and informing about the effects of choice. What happens to my patient if I use treatment A versus treatment B? A lot can happen during treatment, but I want my choice to have the highest likelihood of a good outcome for the patient. One of the most common ways to get answers to such clinical questions is to plan and conduct a randomized controlled clinical trial, which generally involves randomly assigning patients to one or the other treatment, and then seeing how the two groups are doing. However, there are many factors that can affect the patient’s condition beyond just the one clinical choice. Therefore, it is important to consider how much of the group difference is due to chance and how much is due to the choice of treatment. This is the goal of statistical analysis.
Since statistics are so central to the analysis and interpretation of the clinical trial, it is very important that you plan the analyses and interpretations before you start recruiting patients for the study. This is usually done in a trial protocol which describes the entire study. In most clinical trial protocols, statistics is a key factor. An important part of the planning is to decide how many patients are needed, which usually involves knowing which statistical analysis method that will be used.
In addition to the trial protocol, it is common to write a statistical analysis plan, where the statistical analyses are described in more detail than in the protocol. Such analysis plans are increasingly requested by medical journals. The analysis plan should be finalized before starting any analysis, so that the choice of method of analysis is not affected by the results of the study.