Assessing clinical trials--rash adventures.
Assessing clinical trials--rash adventures.
复制标题
评估临床试验——鲁莽的冒险。
作者：
S. Gore
Eagerness to begin analysis may compromise the careful planning and execution of the trial if the important step of checking the data before analysis is omitted. No amount of statistical ingenuity can make bad data good. What is needed is first to prepare an account of the numbers of patients lost to follow-up or withdrawn from the trial, giving the reasons for withdrawal and evaluating potential bias; secondly, to check that the observations are sensible-within range, for exampleorderly, and consistent for a given patient; and thirdly, to compare baseline characteristics between randomised groups to verify that the patients in these groups are initially similar or to be aware of chance differences that should be taken into account at analysis.
