Abstract
Medical statistics need to be properly understood and used in order to assess the significance of results obtained in clinical practice. To clarify whether the medical staff in our hospital appropriately understands medical statistics, this study conducted a questionnaire survey with an objective assessment of basic medical statistics. Of 1498 hospital staff, 464 responded to the questionnaire on medical statistics, which included the following items considered important for the evaluation and use of hypothesis testing: interpretation of p-values, reasons for using different tests (t-test and Mann-Whitney U-tests), presentation of test results for the Mann-Whitney's U-test, and selection of a statistical method to use as a measure of independence. The percentage of correct answers was 20.5%, 16.2%, 6.3%, and 15.1% for the interpretation of p-values, reasons for using different tests, presentation of test results, and choice of statistical method, respectively. The percentage of correct answers for each question was significantly higher for those with research and writing experience than for those without such experience. The number of correct answers was also significantly higher for those with research and writing experience than for those without such experience. Additionally, 20 of the 464 respondents were able to correctly answer all four questions. Multiple regression analysis showed that the number of correct answers was associated with experience in research, experience in writing papers, and job title (multiple regression analysis: R2=0.429758). Our medical staff’s understanding of medical statistics was low, and we were able to identify some of the factors that influence this understanding. These results suggest the importance of learning from clinical research and writing experience in order to improve understanding of medical statistics in the future.
