Dissertation > Excellent graduate degree dissertation topics show

Statistical Comparison and Research on Simple Randomization, Stratified Blocked Randomization, Minimization and Dynamic Balanced Randomization in Clinical Trial

Author: WuChunFang
Tutor: HeJia
School: Second Military Medical University
Course: Epidemiology and Biostatistics,
Keywords: Dynamic balanced randomization Balanced Random Ⅰ error Test performance Central random
CLC: R195
Type: Master's thesis
Year: 2011
Downloads: 149
Quote: 0
Read: Download Dissertation

Abstract


Clinical trials is the systematic study of any trials in humans (patients or healthy volunteers) medicines and other medical interventions, in order to confirm or reveal the role of the medical intervention, adverse reactions, test drug absorption, distribution, metabolism and excretion determine the efficacy and safety of experimental drugs. The test results of people's lives and health are closely related, and the conduct of the test subjects may cause invasive So whether it is domestic or international, all kinds of drugs, medical devices reported, the review is extremely prudent. This prompted the trial sponsors, and principal investigator of the reference to the views of the parties of the statistical experts and clinical experts, to ensure that the scientific nature of the trial process in the entire period of the test. Randomized program development and implementation, and plays a vital role in the beginning stages of the entire test, decided to test whether the scientific. Randomized clinical trials in a variety of ways, including completely randomized block randomization, stratified random and dynamic randomization. With the rapid development of new drug development in China and abroad, new drug research and clinical trials base gradually transferred to the domestic, the form and content of the domestic clinical trials, more diversified, traditional randomization methods include completely randomized been unable to meet existing demand, these methods there are some defects in the face of a particular type of test. Through years of development, statistical experts in clinical trials and gradually introduce the method of dynamic random resolve the traditional randomized method does not resolve the problem. Designed by computer simulation in a balanced distribution of prognostic factors and prognostic factors unevenly distributed both cases, the dynamic equilibrium random method and the method of completely random, stratified block randomization method as well as the past two years to discuss more minimum packet randomization method, to explore the method of grouping balanced and randomness; calculation of the method has been applied packet data by simulating Ⅰ error and test performance using the traditional method of hypothesis testing, to explore the use of this method whether a packet class I error will increase the statistical analysis of the test data and class II error. The balanced randomized comparison of several methods, the balanced distribution of prognostic factors crowd, the simulation results show that the dynamic balanced randomization method not only can guarantee each treatment group into the group of the same number of cases, at the same time to ensure the distribution of prognostic factors between the treatment groups is balanced, better than the other three methods; prognostic factors for uneven distribution of population, under the condition of most of the sample size, the dynamic equilibrium random method to ensure that the number of cases treated balanced balanced distribution of prognostic factors and minimum the results of comparable, are better than the other two methods, in the case of the small sample size and the level of prognostic factors more dynamic balanced randomization method provides better results; randomness, regardless of prognostic factors in try a balanced population distribution or the number of randomly assigned dynamically balanced randomization method is essentially twice the minimum number of randomized, randomness greater than the minimum. Ⅰ error using the t-test method to explore the dynamic equilibrium random method Ⅱ Type I error, the simulation results Ⅰ error is smaller than the other two methods, test performance and Stratified the randomized method results close, in the grasp greater than 0.5, the dynamic equilibrium randomized to grasp the degree is greater than the completely random method, that is to use the data dynamically balanced randomization method did not increase the statistical test Ⅰ error and class II error. By Monte Carlo simulation, the paper argues that from either the grouped balanced, randomness or statistical analysis Ⅰ, Class Ⅱ error affected the dynamic equilibrium random method is a the ideal random method, applied to sample size than small, there are stringent requirements of the test balanced test. Researchers should focus on the characteristics of the clinical trials, a reasonable selection of random method, give full play to the role of the random process.

Related Dissertations

  1. On Research about Yunnan Ordinary Higher Education Region Balanced Development,G649.2
  2. Study of Synchronization and Adaptive Frequency in Frequency-Hopping System,TN914.41
  3. Design and Processing of Hybrid Modulated HF Radar Waveform,TN958.93
  4. Research and Design of High Layer Protocol in DMR Trunking System,TN929.52
  5. Research on Temporal Information Recognition and Normalization,TP391.1
  6. Development of Pxibus High Performance Digital I/O Module,TP274
  7. The Research on the Writing and Directing Strategies of the Science-Educational TV Prigrams Based on Constructivism Learning Theory,G222.3
  8. Based on statistical methods of magnetic resonance imaging of the human brain image segmentation and three-dimensional analysis of the data,R445.2
  9. An Approach for Identifying a Plant Resistance Gene Based on the Random Forest,Q943
  10. Effects of Low-protein Diet on Production Performance and Feeding Environment of Piglet,S828.5
  11. Construction of G Bank Yunnan Branch Office of Performance Management System,F272.92
  12. Performance Evaluation of University Academic Team Based on Balanced Scorecard,G647
  13. Studies on Construction of Specific Gland Expressional Vector of HLF and Modified Fibroblast Cell Line,Q78
  14. The Balanced Scorecard China’s Public Department Performance Management Application,F272
  15. Xi'an Lianhu District Compulsory sharing of resources exploration,G522.3
  16. Garment Co., Ltd. Anshan SF Marketing Strategy,F426.86
  17. Cao Mine Ventilation Optimization of multi-regional,TD724
  18. Increments with generalized negative dependency and the asymptotic behavior of random,O211.5
  19. Research on How to Regularly Develop Basic Education of Gansu Provinc,G527
  20. Calibration and Application of Inertial Measurement Unit Based on MEMS Gyro,V241.5
  21. Research on Opinion Target Extraction,TP391.1

CLC: > Medicine, health > Preventive Medicine,Health > Health care organizations and career ( Health Management ) > Health Survey and Statistics
© 2012 www.DissertationTopic.Net  Mobile