الفهرس 
Assessment of statistical methodologies and pitfalls of dissertations carried at National Cancer Institute, Cairo University
ABSTRACTOnly 14 pages are availabe for public view
 |  | from | 174 |  |  |
| | | from | 174 | | |
Abstract
Objective: To identify statistical errors, flaws and pitfalls in dissertations done during the period from 2009 to 2013 as a part of the requirements for the Medical Doctorate (MD) degree in 3 departments in the National Cancer Institute (NCI), Cairo University (CU) to improve the quality of medical research required in the institute. Results: The study included 62 MD dissertations; 12 clinical trials and 50 observational studies. Statistical methods were mentioned explicitly in 53 studies (85.5%), but were complete in 28 (52.8%) and appropriate in 13 (24.5%). The most common descriptive measure used was frequency and percentage. The most common statistical tests used were chi-square test followed by log-rank test, than Mann-Whitney U test. Only four studies had a basis of sample size estimation and mentioned power of the study. Only 37.1% and 38.7% of dissertation results support aim and answer the research question, respectively. Most of results were misinterpreted (82.3%) with misuse of statistical words (77.4%). Tabular and graphical data display was independently informative in only 36 dissertations (58.1%) with accurate title and labels in only 17 (27.4%). Statistical material was reasonably presented in 37.1%. Statistical tests fulfilled the assumptions only in 29 studies; with evident misuse in 33 studies. Ten dissertations reported non-significance of their primary outcome measure. The power of the test of these studies ranged from 6% to 60% with a median of 35.5%. Title representing the aim (p = 0.025) and abstract format and proper writing increased significantly (p = 0.010) in recent dissertations. There was an improvement in summary and conclusion characteristics by time. There was no significant change in the characteristics of the results section and appropriate use of statistical analysis over the time periods. Collaborative work between NCI and other centers relatively increased in more recent years