Statistics in medicine
-
Teaching statistics to non-specialists is a challenge for which most statisticians are unprepared by their own training within an academic mathematics department. Most statistics courses for medical undergraduates still focus on research statistics, whereas it would be more appropriate to concentrate on statistics relevant to clinical decision-making about an individual patient. Teaching statistics to Master of Public Health students presents further challenges because of the wide variety of their backgrounds and the greater demands from mature postgraduates. Whatever the audience, however, the same principles apply: medical statistics should be taught as non-mathematically as possible, only introducing formulae when absolutely necessary and explaining their components; plenty of practical applications should be given; there should be ample opportunity for practice to gain hands-on experience using both calculators and computers (preferably with MINITAB); and tutorials should be streamed according to perceived mathematical ability, with remedial mathematics teaching available to those who need it.
-
Statistics in medicine · Dec 1994
Publications from multicentre clinical trials: statistical techniques and accessibility to the reader.
Articles from multicentre randomized clinical trials were analysed by methods adapted from Emerson and Colditz and Juzych et al. to compare the frequency with which different statistical methods are used by clinical trials investigators with the frequency used by other researchers, and to determine how much statistical knowledge is required to interpret the statistical treatment of data from clinical trials. We observed differences between the frequency of usage of statistical methods and the accessibility of the clinical trials publications and those of all medical research articles published by specific journals. Clinical trials publications are less accessible than others in medical journals to the reader who knows only descriptive statistics, t-tests, contingency tables, power calculations, and life table methods. Many more statistical methods must be known by a reader to understand fully publications regarding treatment group comparisons for the primary outcomes of interest from clinical trials.
-
Group sequential methods are becoming increasingly popular for monitoring and analysing large controlled trials, especially clinical trials. They not only allow trialists to monitor the data as it accumulates, but also reduce the expected sample size. Such methods are traditionally based on preserving the overall type I error by increasing the conservatism of the hypothesis tests performed at any single analysis. ⋯ These procedures have good expected and maximum sample sizes, and lead to natural point and interval estimates of the treatment difference. Hypothesis tests, point estimates and interval estimates calculated using this procedure are consistent with each other, and tests and estimates made at the end of the trial are consistent with interim tests and estimates. This class of sequential tests can be considered in both a traditional group sequential manner or as a Bayesian solution to the problem.
-
We propose and discuss several methods of monitoring multi-armed trials comparing means or survival. These methods combine multiple comparison procedures such as Fisher's LSD, Newman-Keuls and Tukey's with monitoring boundaries such as those of O'Brien and Fleming and Lan and DeMets. Tables of boundaries are provided for the equal variance or equal censoring distribution case.
-
Statistics in medicine · Jul 1994
The relationship between clinical trials and clinical practice: the risks of underestimating its complexity.
Two main points are addressed in the following remarks. The first is that the survival of randomized clinical trials (RCTs) as the gold standard by which to assess the effectiveness of medical technology is being seriously challenged. ⋯ Contrary to what health services research has repeatedly indicated, trialists still, by and large, seem to believe in a simplistic model that assumes a one-way linear and rapid relationship between scientific knowledge and clinical practice. In the concluding section the paper pinpoints some issues to be further discussed from the point of view of monitoring and conducting trials.