Bayesian two-stage designs for phase II clinical trials.
Explore the source record for details and available documents.
Bayes Theorem↗
Biomedical subjects
Say-Beng Tan
Publications and source records attributed to Say-Beng Tan.
Phase II trial designs in the presence of stratification.
Explore the source record for details and available documents.
Bayes Theorem↗
A Bayesian dose finding design for dual endpoint phase I trials.
We propose a dose-finding weighted design for an early clinical trial which aims to determine the optimal dose, selected on the basis of both efficacy and toxicity, to be used in patients entering subsequent studies in a drug development process. The goal is to identify the optimal dose, while using a minimal number of subjects. For each dose under test, a decision table is defined with a utility value attached to each possible decision. The relationship between the utility and the target probability for each outcome is shown. A Dirichlet prior is used and we illustrate the process of maximizing the expected utility under the resulting posterior distribution to find the optimal decision at each stage of the trial. We show how this affects the eventual choice of optimal dose in various scenarios. Properties of our design are discussed and compared with a current standard design.
Anti-HIV Agents↗
A cohort study of incident myopia in Singaporean children.
PURPOSE: To determine the risk factors of incident myopia in a school-based cohort study in Singaporean children. METHODS: A 3-year prospective cohort study was conducted in Singaporean school children aged 7 to 9 years in three schools at entry. Chinese children without myopia at baseline (n = 994) were included in the analysis. The main outcome was incident myopia, defined as spherical equivalent (SE) at least -0.75 D based on cycloplegic autorefraction. Other definitions of incident myopia, at least -0.5 D and at least -1.0 D, were also assessed. RESULTS: After controlling for school, age, gender, income, reading in books per week and intelligence quotient (IQ) test scores, we found the relative risk (RR) of incident myopia defined as -0.75 D to be 1.55 (95% confidence interval [CI] 1.18-2.04) for two versus no myopic parents. The multivariate RR of myopia for IQ in the third versus first tertile was 1.50 (95% CI, 1.19-1.89). However, the RR of incident myopia was 1.01 (95% CI, 0.97-1.05) for every unit increase in books read per week. Similar results were obtained with definitions of -0.5 and -1.0 D for incident myopia. CONCLUSIONS: These data provide new prospective evidence of essential links between parental myopia, IQ scores and subsequent myopia development. However, reading in books per week was not associated with incident myopia.
Child↗
Bayesian designs with frequentist and Bayesian error rate considerations.
So far, most Phase II trials have been designed and analysed under a frequentist framework. Under this framework, a trial is designed so that the overall Type I and Type II errors of the trial are controlled at some desired levels. Recently, a number of articles have advocated the use of Bayesian designs in practice. Under a Bayesian framework, a trial is designed so that the trial stops when the posterior probability of treatment is within certain prespecified thresholds. In this article, we argue that trials under a Bayesian framework can also be designed to control frequentist error rates. We introduce a Bayesian version of Simon's well-known two-stage design to achieve this goal. We also consider two other errors, which are called Bayesian errors in this article because of their similarities to posterior probabilities. We show that our method can also control these Bayesian-type errors. We compare our method with other recent Bayesian designs in a numerical study and discuss implications of different designs on error rates. An example of a clinical trial for patients with nasopharyngeal carcinoma is used to illustrate differences of the different designs.
Bayes Theorem↗
Where's the utility in Bayesian data-monitoring of clinical trials?
BACKGROUND: Data monitoring is now an established part of good practice in clinical trials. Bayesian procedures for data-monitoring of treatment trials have been proposed and used, but sometimes without explicit consideration of utilities. A natural statistical framework for evidence-based medicine is a Bayesian approach to decision-making that incorporates an integrated summary of the available evidence and associated uncertainty with assessment of utilities. METHODS: We explore this approach to data monitoring, explicitly addressing separately the individual, scientific and public health perspectives. The Data Monitoring Committee's decision can then be thought of as a weighted combination of these perspectives. These ideas are illustrated with a trial of treatments for oesophageal cancer. RESULTS: For a Bayesian approach without explicit utilities we show that a utility structure is, in fact, implicit, and that it may be viewed as a weighted sum of the individual and scientific utilities. CONCLUSIONS: We argue that explicit consideration of utilities leads to decision-making that is more transparent, and lays foundations for data monitoring of more complex trials.
Algorithms↗
IQ and the association with myopia in children.
PURPOSE: To evaluate the association between intelligence and myopia in children. METHODS: Cycloplegic refraction and ocular biometry parameters, including axial length, vitreous chamber depth, lens thickness, anterior chamber depth, and corneal curvature were obtained in 1204 Chinese school children aged 10 to 12 years from three schools who were participants in the Singapore Cohort Study Of the Risk Factors for Myopia (SCORM). Intelligence quotient (IQ) was assessed using the nonverbal Raven Standard Progressive Matrix test. RESULTS: After controlling for age, gender, school, parental myopia, father's education, and books read per week, myopia (spherical equivalent [SE]) of at least -0.5 D was associated with high nonverbal IQ (highest quartile) versus low IQ (lowest quartile) (odds ratio = 2.4; 95% confidence interval, 1.7-3.4). Controlling for the same factors, children with higher nonverbal IQ scores had significantly more myopic refractions (-1.86 D for children with nonverbal IQ in the highest quartile compared with -1.24 D for children with nonverbal IQ in the lowest quartile; P = 0.002) and longer axial lengths (24.06 mm versus 23.80 mm; P = 0.022). Nonverbal IQ accounted for a greater proportion of the variance in refraction compared with books read per week. CONCLUSIONS: Nonverbal IQ may be an independent risk factor of myopia, and this relationship may not be explained merely by increased reading (books per week) among myopes. An interesting observation is that nonverbal IQ may be a stronger risk factor for myopia compared with books read per week. The complexity of the relationships between nonverbal IQ, reading, and myopia warrant additional studies to clarify any cause-effect relationship.
Books↗
Strategy for randomised clinical trials in rare cancers.
Explore the source record for details and available documents.
Data Interpretation, Statistical↗
Elicitation of prior distributions for a phase III randomized controlled trial of adjuvant therapy with surgery for hepatocellular carcinoma.
A randomized, controlled clinical trial of radioactive iodine tagged with lipiodol in patients with resected hepatocellular carcinoma was criticized for its early stopping and resulting small sample size. To clarify its results, a new, larger multicenter trial was therefore proposed. This paper describes the elicitation of the pretrial opinions of the investigators involved in the new trial and the construction of appropriate clinical and skeptical prior distributions based on their responses. The prior distributions provide a useful tool in assessing the state of equipoise before the start of the trial. They can also be used in Bayesian analyses, both at the interim stage(s) as well as at the end of the trial. We illustrate these analyses, assuming that the data resulting from the new trial was the same as that obtained in the earlier trial when it was stopped.
Bayes Theorem↗
Quick-FLIC--a short questionnaire for assessing quality of life of cancer patients.
Health-related quality of life instruments tend to include a great many items. This imposes a burden on the respondents as well as undermining response rate and data quality. In this study we developed a shortened version of the Functional Living Index-Cancer (FLIC), now called Quick-FLIC, and examined its measurement properties. A questionnaire package, self-administered by 140 patients, included the FLIC and the Functional Assessment of Cancer Therapy-General. A factor analysis and clinical judgement were used to shorten the FLIC, which included 22 items. Each subscale of FLIC was shortened to include two or three items only. The Quick-FLIC included a total of only 11 items. Nevertheless, the measurement properties of the Quick-FLIC and its subscales were comparable to those of the original FLIC. It is concluded that the shortening of established health-related quality of life instruments is viable in oncology research.
Adult↗
Bayesian two-stage designs for phase II clinical trials.
Many different statistical designs have been used in phase II clinical trials. The majority of these are based on frequentist statistical approaches. Bayesian methods provide a good alternative to frequentist approaches as they allow for the incorporation of relevant prior information and the presentation of the trial results in a manner which, some feel, is more intuitive and helpful. In this paper, we propose two new Bayesian designs for phase II clinical trials. These designs have been developed specifically to make them as user friendly and as familiar as possible to those who have had experience working with two-stage frequentist phase II designs. Thus, unlike many of the Bayesian designs already proposed in the literature, our designs do not require a distribution for the response rate of the currently used drug or the explicit specification of utility or loss functions. We study the properties of our designs and compare them with the Simon two-stage optimal and minimax designs. We also apply them to an example of two recently concluded phase II trials conducted at the National Cancer Centre in Singapore. Sample size tables for the designs are given.
Antimetabolites, Antineoplastic↗