What is medical statistics?

Sensitivity is the probability of not misdiagnosing (false negative) when diagnosing diseases. To put it bluntly, specificity is the probability of not being misdiagnosed (positive) when the index value is diagnosed. If the independent index value improves the sensitivity of diagnosis, it will definitely reduce the specificity of diagnosis, that is to say, reducing misdiagnosis will definitely improve misdiagnosis, and vice versa.

The sensitivity and specificity of an index, sensitivity is the probability that people who are already sick are identified and diagnosed as sick by this index. Specificity is the probability that people who are not sick can be identified by this index and diagnosed as people who are not sick.

Medical statistics is a subject that studies the collection, analysis and inference of digital data by using the principles and methods of probability theory and mathematical statistics combined with medical practice. The object of medical research is mainly the human body and various factors related to human health.

Characteristics of medical statistics

The same is true for statistical analysis of medical data. If an indicator is missing, the missing data will be filled with corresponding methods, and the differences of statistical results before and after filling will be compared, and this change will be discussed and analyzed. This can be called sensitivity analysis.

For example, in clinical trials, multiple statistical analysis data sets (such as PPS, FAS and SS) are often defined and formed. Of course, each data set has a specific function and application rules, but the comparison of the results of multiple data sets can also be called sensitivity analysis to some extent. Similarly, whether the statistical analysis and thinking of multi-center clinical research need to add the central effect term is also the category of sensitivity analysis.

The above contents refer to Baidu Encyclopedia-Sensitivity, Baidu Encyclopedia-Specificity and Baidu Encyclopedia-Medical Statistics.