Statistics is the science and arts of extracting meaningful information and hidden underlying trends from apparently randomized data. There are many parameters, procedures and tests to achieve this. Central tendency of a data can be measured by calculating mean, median or mode; similarly, dispersion of the data can be measured by calculating variance or standard deviation; underlying trend can be measured by regression or correlation analysis. Test of a hypothesis can be done by one sample t-test or paired sample t-test, F-test, c2-test etc. When a sample is large enough and representative then one can even extrapolate the findings. However, care should be taken in inferring from the statistical parameters and the explanations should be used in right context.
Coming to this article by Parker-Pope T.; this article has tried to examine different claims made about the usefulness of the cholesterol-lowering statin drugs. The author himself has not conducted any statistical experiment or analysis; rather he has made use of the statistical experiments and analysis made by different research groups making these claims. In subsequent paragraphs, different statistical aspects used in this article will be briefly discussed.
Let us talk about the statement made by Dr. Valentine Fuster that Statins have many meaningful biological effects. In this statement the word ‘meaningful’ is important from the point of view of statistics. For this word to be used there has to be an underlying statistical experiment an analysis to support this statement. Such an experiment can be carried out by taking two groups of persons (subjects) of similar medical conditions say for example cholesterol level, age, inflammation etc. The one group should be given the Statin treatment and the other group should be given no treatment or should be given non-medicated syrup to subtract the placebo effect. After medication is given for the required time period then cholesterol level of the subjects from both the groups should be measured and recorded. The statistical analysis should be performed to determine whether there was significant reduction in the cholesterol level and also whether there is significant difference in the mean cholesterol level of the subjects in the two groups. This analysis will require calculation of mean, standard deviation and hypothesis testing by one sample t-test and paired t-test.
Similarly, the author has talked about necessity and utility of C-Reactive Protein or CRP test during blood test for determining risk level for heart attack. This is also based on an statistical analysis of an study. This study has shown that people with high CRP level are more prone to heart attack even with normal cholesterol level. This conclusion can also be drawn on the basis of a statistical analysis only. When an attempt was made to treat people with high CRP level by statins it was found that from a sample of as many as 90000 subjects only about 20000 were found suitable for treatment by statins. Because of large sample size, this finding can be extrapolated and it can be said that the number of people who can benefit from statin usage are not many.
Statistics can be used for misrepresentation is also pointed in this article. The author says that use of statins reduced the risk of serious heart problems from 1.8% to 0.9% leading to a reduction of nearly 1% and this is what is of importance from clinical consideration. But it was reported as 50% improvement (from 1.8% to 0.9%) and this 50% has hardly any clinical importance. Thus it can be seen as how statistics can be used to make exaggerated claims as well.
It can be concluded that the author has used many relevant statistical concepts and findings in this article to stress his claim that statins are no magic pills and not as useful as being claimed and advertised by many.