Tu, Li-Chuan
(2013)
The performance of biased-coin minimization in multicenter randomized clinical trials.
Doctoral Dissertation, University of Pittsburgh.
(Unpublished)
This is the latest version of this item.
Abstract
Randomized clinical trials (RCTs) are widely used as the gold standard for comparative medical studies. Using randomization to determine treatment assignment assures that all patients have the same chance of being assigned to each treatment group and that the treatment groups are comparable in terms of the distributions of prognostic factors. When treatment groups are not comparable, the power of statistical test will be decreased. Moreover, the problem of imbalance becomes more notable when it occurs in the important prognostic factors because it could result in a significant bias when assessing differences by treatment group.
The most intuitive and simple form of randomization is complete randomization. However, with complete randomization there is still a chance for an imbalance on prognostic factors. In order to overcome the problem of imbalance when using complete randomization, restricted randomization procedures were proposed. However, some have argued that an unintended consequence of the restrictions placed on randomization is that they could create patterns that allow for the prediction of future treatment allocation. Furthermore, some have questioned the accuracy of model-based statistical inference using conventional asymptotic test under restrictions placed on the treatment allocation.
This dissertation is concerned with an assessment of the performance of biased-coin minimization. The assessment is twofold. The first aspect is to determine in terms of balancing properties and also in terms of the probability of predicting treatment assignment when using biased-coin minimization. The second aspect is to compare the results from the classical statistical test, log-rank test, based on population model and the randomization test from the randomization model while biased-coin minimization is applied.
Randomized clinical trials are the gold standard of research for demonstrating the efficacy of therapies used to treat patients in the general community. Allocation methods that promote balance in key prognostic factors between treatment groups are important to assure the accuracy and validity of results from clinical trials. It is important to assess the properties of dynamic allocation methods to demonstrate the validity of these methods as they are applied in research that is designed to develop treatments that are used to enhance the public health.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
27 June 2013 |
| Date Type: |
Publication |
| Defense Date: |
15 February 2013 |
| Approval Date: |
27 June 2013 |
| Submission Date: |
2 April 2013 |
| Access Restriction: |
1 year -- Restrict access to University of Pittsburgh for a period of 1 year. |
| Number of Pages: |
99 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
School of Public Health > Biostatistics |
| Degree: |
DrPH - Doctor of Public Health |
| Thesis Type: |
Doctoral Dissertation |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
minimization, randomization, multicenter |
| Date Deposited: |
27 Jun 2013 23:03 |
| Last Modified: |
15 Nov 2016 14:12 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/18660 |
Available Versions of this Item
-
The performance of biased-coin minimization in multicenter randomized clinical trials. (deposited 27 Jun 2013 23:03)
[Currently Displayed]
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[feed]](/images/feed-icon-32x32.png){kind=icon}