Self-designing clinical trials

I present a method of sequential analysis for randomized clinical trials that allows use of all prior data in a trial to determine the use and weighting of subsequent observations. One continues to assign subjects until one has 'used up' all the variance of the test statistic. There are many strategies to determine the weights including Bayesian methods (though the proposal is a frequentist design). I explore further the self-designing aspect of the randomized trial to note that in some cases it makes good sense (i) to change the weighting on components of a multivariate endpoint, (ii) to add or drop treatment arms (especially in a parallel group dose ranging/efficacy/safety trial), (iii) to select sites to use as the trial goes on, (iv) to change the test statistic and (v) even to rethink the whole drug development paradigm to shorten drug development time while keeping current standards for the level of evidence necessary for approval.     

Increasing efficiency from censored survival data by using random effects to model longitudinal covariates

When estimating a survival time distribution, the loss of information due to right censoring results in a loss of efficiency in the estimator. In many circumstances, however, repeated measurements on a longitudinal process which is associated with survival time are made throughout the observation time, and these measurements may be used to recover information lost to censoring. For example, patients in an AIDS clinical trial may be measured at regular intervals on CD4 count and viral load. We describe a model for the joint distribution of a survival time and a repeated measures process. The joint distribution is specified by linking the survival time to subject-specific random effects characterizing the repeated measures, and is similar in form to the pattern mixture model for multivariate data with nonignorable nonresponse. We also describe an estimator of survival derived from this model. We apply the methods to a long-term AIDS clinical trial, and study properties of the survival estimator. Monte Carlo simulation is used to estimate gains in efficiency when the survival time is related to the location and scale of the random effects distribution. Under relatively light censoring (20%), the methods yield a modest gain in efficiency for estimating three-year survival in the AIDS clinical trial. Our simulation study, which mimics characteristics of the clinical trial, indicates that much larger gains in efficiency can be realized under heavier censoring or with studies designed for long term follow up on survival.     

A partially grouped logrank test

We propose a modified version of the standard logrank test for survival data in which the first contribution to the statistic is based on grouping of data before a pre-selected grouping time. The grouping is accomplished by artificially constructing the first table based on the product limit estimates of the proportions surviving at the grouping time. Remaining contributions to the statistic are identical to that of the standard logrank statistic. The approach has the advantage of being uninfluenced by non-proportional hazards differences prior to the grouping time while being almost as efficient as the usual logrank test for proportional-hazards alternatives. The statistic is particularly useful for interim monitoring in situations where early difference between treatments are unimportant and/or lagged treatment effects are anticipated. We report simulation studies to verify and investigate the test's properties.     

Ordinal properties of perceived average duration: Simultaneous and sequential presentations.

When an S evaluates the duration of 2 simultaneous events, the perceived durations are combined in a nonlinear manner, whereas comparable evaluations of successive events yield linear combinations. In the present study, 54 undergraduates rated average durations of pairs of time intervals ranging from 0.5 to 10.0 sec. The intervals making up a pair were presented successively and simultaneously on separate occasions within the same presentation sequence. The order of data was consistent with models previously proposed for perceived average duration. A nonmetric analysis of the combined ordinal data from both conditions yielded a single set of measures of perceived duration with ratio-scale properties. The common scale for the 2 conditions is consistent with differences in results from simultaneous and successive monitoring of time intervals, stemming from differences in the way duration information is combined rather than how it is perceived. The derived scale is related to physical time by a power function with a 1.06 exponent. (19 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Environmental influences in hand preference: an African point of view

In the analysis of survival data from clinical trials and other studies, the censoring generally precludes estimation of the mean survival time. To accommodate censoring, Irwin (1949) proposed, as an alternative, estimation of the mean lifetime restricted to a suitably chosen time T. In this article we consider the use of Irwin's restricted mean as an index for comparing survival in different groups, using as an example published data from a randomized clinical trial in patients with primary biliary cirrhosis. Irwin's method, originally based on the actuarial survival estimator, is extended to incorporate covariates into the analysis through the use of piecewise exponential models. For comparing two survival curves, the logrank test is known to be optimal under proportional hazards alternatives. However, comparison of restricted means may outperform the logrank test in situations involving nonproportional hazard rates. We examine the size and power of these two procedures under various proportional and nonproportional hazards alternatives, with and without covariate adjustment. For survival curves that separate early in time the censored data generalization of the Wilcoxon test is known to exhibit high power, and we examine how the comparison of restricted means performs relative to this procedure also.     

Conditional power calculations for early acceptance of H0 embedded in sequential tests

For ethical and efficiency concerns one often wishes to design a clinical trial to stop early if there is a strong treatment effect or if there is strong evidence of no treatment effect. There is a large literature to address the design of sequential trials for detecting treatment differences. There has been less attention paid to the design of trials for detecting lack of a treatment difference and most of the designs proposed have been ad hoc modifications of the traditional designs. In the context of fixed sample tests, various authors have proposed basing the decision to stop in favour of the null hypothesis, H0, on conditional power calculations for the end of the trial given the current data. Here I extend this procedure to the popular sequential designs: the O'Brien-Fleming test and the repeated significance test. I derive explicit boundaries for monitoring the test statistic useful for visualizing the impact of the parameters on the operating characteristics of the tests and thus for the design of the tests. Also, they facilitate the use of boundary crossing methods for approximations of power. I derive appropriate boundaries retrospectively for two clinical trials: one that concluded with no treatment difference (AIDS Clinical Trials Group protocol 118) and one that stopped early for positive effect (Beta-Blocker Heart Attack Trial). Finally, I compare the procedures based on the different upper boundaries and assess the impact of allowing for early stopping in favour of H0 in numerical examples.     

Degeneracy and discrete receiver operating characteristic rating data

RATIONALE AND OBJECTIVES: Observer performance studies sometimes use too few cases for estimating diagnostic accuracy from binormal receiver operating characteristic (ROC) curves. One important problem is degenerate data sets. We compared a new algorithm, RSCORE4, with the exact-solution approach to degeneracy in ROCFIT and with the Wilcoxon statistic. METHODS: Degenerate ROC solutions result from empty cells in the data matrix. We addressed this problem by adding a small constant to empty cells in a maximum-likelihood program, RSCORE4. When this method failed, the program branched to a pattern-search algorithm. We tested the program in a series of Monte Carlo studies. RESULTS: RSCORE4 converged to nondegenerate solutions in every case and gave results closer to population values than ROCFIT or Wilcoxon. ROCFIT converged to exact-fit degenerate solutions, those with zero or infinite parameter values, in more than 40% of the samples. The Wilcoxon statistic was biased. CONCLUSION: RSCORE4 seems to outperform other currently recommended methods for dealing with degeneracy.     

Group sequential monitoring of distribution-free analyses of repeated measures

In many clinical trials the principal analysis consists of a 1 degree of freedom test based on an aggregate summary statistic for a set of repeated measures. Various methods have been proposed for the marginal analysis of such repeated measures that entail estimates of a measure of treatment group difference (the treatment effect) at each of K repeated measures and a consistent estimate of the covariance matrix, where asymptotically these estimates are normally distributed. One can then obtain an overall large sample 1-d.f. test of group differences, such as by taking the average of these K estimates. These methods include the Wei-Lachin family of multivariate rank tests and a corresponding multivariate analysis using the Mann-Whitney difference estimator as a measure of treatment group differences. Other methods, such as O'Brien's non-parametric test, are based on a single summary score for each patient, such as the within-patient mean value. These, and other such methods, allow for some observations to be missing at random. Herein I employ sequential data augmentation to conduct group sequential analyses using a 1 degree of freedom test from a multivariate Mann-Whitney analysis and for the O'Brien rank test. Su and Lachin used this method to perform group sequential analyses of a vector of Hodges-Lehmann estimators. By augmentating the data from the sequential looks in a single analysis, one obtains an estimate of the covariance of the estimates at each look, from which one obtains an estimate of the correlations among the sequential 1-d.f. test statistics. I describe a simple secant algorithm to determine the group sequential boundaries based on recursive integration of the standard multivariate normal distribution with the estimated correlation matrix. Although the boundary obtains readily using the method of Slud and Wei, the more flexible method of Lan and DeMets may be preferred. The true information fraction at each look, needed to apply the spending function method of Lan and DeMets, however, is unknown. Thus, I also describe the use of a surrogate measure of information.     

Inhibition of Ras prenylation: a signaling target for novel anti-cancer drug design

This paper describes a measure of explained variation (MEV) of survival times for a given regression model used in survival analysis. It quantifies the predictive power of a set of prognostic factors in the model, and therefore provides useful information for more precise prediction of patient prognosis, and for designing randomized clinical trials with the capability of determining treatment effects. The MEV defined in this article is asymptotically derived from the squared product-moment correlation; it can be interpreted as an adaptation of the multiple correlation coefficient for the normal linear model to the survival time regression model. Monte-Carlo simulations are performed to investigate the statistical behavior of the proposed MEV. The MEV is applied to estimate the predictive power of several sets of prognostic factors for gastric cancer in Japan using data from a large clinical trial.     

Cultural effect on postoperative pain

OBJECTIVE: To investigate whether the S + G2/M fraction (proliferative index) is a prognostic determinant in breast cancers classified as Auer IV. STUDY DESIGN: Prognostic evaluation of Auer IV DNA histograms with respect to the high versus low S + G2/M fraction, obtained by image cytometry on consecutive breast cancer imprint preparations. RESULTS: When studying recurrence-free survival (n = 136), the prognostic value of S + G2/M was found to vary with time: it was negligible before the median time to relapse (1.5 years) but thereafter statistically significant, in both univariate and multivariate analysis. The same pattern was found when overall survival was used as the end point; the effect was delayed to about the median time until death (three years). Tumors with a low S + G2/M fraction were smaller and more often estrogen receptor- and progesterone receptor-positive than those with a high S + G2/M fraction. CONCLUSION: According to ICM-DNA values corresponding to the S + G2/M region, patients with breast cancers classified as Auer IV can be divided into subgroups with different tumor characteristics and prognoses.     

Analysis of prognostic factors and patterns of failure in dogs with malignant oral tumors treated with megavoltage irradiation

OBJECTIVE: To determine quality and duration of progression-free survival (PFS) time in dogs with malignant oral tumors after definitive megavoltage irradiation, to analyze prognostic factors for PFS time and patterns of failure, and to analyze the influence of tumor recurrence and development of metastasis on survival. DESIGN: Prospective clinical trial. ANIMALS: 105 dogs with squamous cell carcinoma, fibrosarcoma, or malignant melanoma of the oral cavity without evidence of metastasis. PROCEDURE: Dogs were treated with 48 Gy over 4 weeks on an alternate-day schedule of 4 Gy/fraction. Multivariate analysis was done by use of Cox's regression model to determine significant prognostic factors and by use of a competing risk model to determine the differential effects of prognostic factors on type of, and time to, failure. In 8% of the dogs, severe acute radiation reactions in the final week of treatment resulted in treatment discontinuation. In 7.6% of the dogs, chronic radiation reactions, including bone necrosis and fistula formation, developed. RESULTS: Prognostic factors that independently affected PFS time were histologic type and tumor T stage. Histologic type significantly influenced pattern of failure, but not time to failure, whereas clinical stage significantly influenced time to failure, but not type of failure. CLINICAL IMPLICATIONS: Irradiation was a safe and effective treatment of malignant oral tumors. Because the local efficacy of radiation was influenced only by tumor size, early treatment of oral tumors should improve the prognosis. In dogs without tumor recurrence, systemic metastases, rather than regional metastases, limited long-term survival after radiation therapy.     

The role of eye movements in the missing-letter effect revisited with the rapid serial visual presentation procedure.

When participants read a text while searching for a target letter, they are more likely to miss the target letter embedded in frequent function words than in less frequent content words. This effect is usually observed with a text displayed normally, for which it has been found that frequent function words are fixated for a smaller amount of time than less frequent content words. However, similar pattern of omissions have been observed with a rapid serial visual presentation procedure in which words appear one at a time. These parallel results would demonstrate that fixation duration per se is not the proximal cause of the missing-letter effect only if eye movements are not made during the rapid serial visual presentation procedure. Therefore, the authors performed eye monitoring during the rapid serial visual presentation procedure. Results revealed that, with a rapid serial visual presentation procedure, participants fixated function and content words for almost the entire presentation duration. It is concluded that eye movements are not the proximal cause of the missing-letter effect. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Clinical trial designs for testing of products for dentine hypersensitivity--a review

Dentine hypersensitivity (DH) is a perplexing clinical entity. There is no doubt that patients self-report discomfort arising from various stimuli, but the highly subjective nature of the condition makes it extremely difficult to evaluate DH objectively. This is particularly true when evaluating the efficacy of desensitizing agents in the clinical trial setting. This paper attempts to provide an overview on clinical trial management to evaluate both in-office and over-the-counter (OTC) desensitizing agents. The importance of correct clinical trial design (type, allocation, blinding) with emphasis on patient selection, sample size, statistical power, duration, choice of positive/negative controls, treatment outcomes, and data collection, will be discussed.     

Decision making during a phase III randomized controlled trial

Experiments such as clinical trials should be carried out with specific objectives. For example, in a trial designed to prevent disease, specific considerations should be made concerning the impact of the trial on the health of the target population, including the participants in the trial. These objectives should be assessed continually in light of data accumulating from the trial. Accumulating evidence should be judged in the context of changing circumstances external to the trial, and the trial's design possibly modified. An important type of modification is stopping the trial. This is a sequential decision problem that can be addressed using a Bayesian approach and the methods of dynamic programming. As an example we consider a vaccine trial for the prevention of haemophilus influenzae type b. The objective we consider is minimizing the number of cases of this disease in a Native American population over a specified horizon. We assess the prior probability distribution of vaccine efficacy. We also assess the probability of regulatory approval for widespread use of the vaccine, depending on the data presented to the regulatory officials. In deciding whether to continue the trial we weigh the impact of the possible future results by their (predictive) probabilities. We address the sensitivity of the optimal stopping policy to the prior probability distribution, to the assessed probability of regulatory approval, and to the horizon.     

Data monitoring committees: the moral case for maximum feasible independence

Since data monitoring committees (DMCs) began to be established in the 1970s, they have reviewed the clinical outcome data emerging from ongoing trials. The thesis of this essay is that DMCs most effectively promote the validity and credibility of randomized clinical trials when they perform this function with the maximum feasible independence. The essay argues that four types of individuals and groups should be excluded from the interim monitoring of outcome data: (1) investigators entering patients into a trial; (2) principal investigators or study chairs; (3) representatives of commercial firms connected with a trial; and (4) representatives of regulatory agencies. The role of statistical groups and officials of government funding agencies in the review of outcome data is also discussed.     

Clinical applications of esophageal manometry and pH monitoring

In summary, GERD patients are usually well managed using a careful medical history, endoscopy, and empirical trials of antireflux medications. Extended esophageal pH monitoring is unnecessary in most patients but can be of considerable value in managing patients with typical or atypical symptoms who are refractory to standard therapy for GERD. Furthermore, the test can be useful in documenting abnormal reflux in an individual without esophagitis being evaluated for antireflux surgery. The test is done with compact, portable data loggers, miniature pH electrodes, and computerized data analysis. The pH electrode should be positioned 5 cm above the manometrically defined upper limit of the LES, and patients should undergo the test on an unrestricted diet. In terms of data analysis, the total percentage time of pH < 4 provides as much information as any other scheme of quantifying esophageal acid exposure, but symptom association is essential when evaluating atypical or sporadic symptoms. Enthusiasm for 24-h pH monitoring must, however, be tempered with an analysis of its proven clinical utility in patient management with its utility rightfully compared with that of an empirical trial of anti-reflux therapy. Ambulatory pH monitoring is probably most useful in examining patients without typical reflux symptoms or patients who have either partially or completely failed a trial of anti-reflux therapy. To date, there have not been any prospective, controlled clinical trials evaluating these uses. Suggested clinical indications for ambulatory pH monitoring are listed in Table 5 (53).     

Systematic analysis of the detention phenomenon in mice.

Results of an experiment with 1,033 female CF1 albino mice indicate that detaining Ss for some time in the safe compartment of the apparatus in which they had been conditioned decreased the level of the passive-avoidance CR. The extent of the decrease varied (a) as a direct function of the duration of the detention period, and (b) as an inverse function of the time interval between detention and learning trial and of the strength of the CR. After detention the CR, though weakened, underwent the usual process of incubation. When Ss were exposed to 2 periods of detention, the 2nd period did not produce any additional amnesic effect, although it would have been effective if Ss had been exposed to it alone. (19 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Economic analysis of tirilazad mesylate for aneurysmal subarachnoid hemorrhage. Economic evaluation of a phase III clinical trial in Europe and Australia

This study used data from a multinational phase III randomized, double-blind, vehicle-controlled trial to evaluate the cost-effectiveness of tirilazad mesylate (Freedox) in the treatment of aneurysmal subarachnoid hemorrhage. In men, therapy with 6 mg/kg per day of tirilazad mesylate was associated with significantly increased survival, increased cost of care, and ratios of cost per death averted that compare favorably with the ratios of other life and death interventions. In women, it appeared to have no effects on costs or survival. Further clinical studies may provide additional information about the cost-effectiveness of this intervention.     

Construction of a continuous stopping boundary from an alpha spending function

Lan and DeMets (1983, Biometrika 70, 659-663) proposed a flexible method for monitoring accumulating data that does not require the number and times of analyses to be specified in advance yet maintains an overall Type I error, alpha. Their method amounts to discretizing a preselected continuous boundary by clumping the density of the boundary crossing time at discrete analysis times and calculating the resultant discrete-time boundary values. In this framework, the cumulative distribution function of the continuous-time stopping rule is used as an alpha spending function. A key assumption that underlies this method is that future analysis times are not chosen on the basis of the current value of the statistic. However, clinical trials may be monitored more frequently when they are close to crossing the boundary. In this situation, the corresponding continuous-time boundary should be used. Here we demonstrate how to construct a continuous stopping boundary from an alpha spending function. This capability is useful in the design of clinical trials. We use the Beta-Blocker Heart Attack Trial (BHAT) and AIDS Clinical Trials Group protocol 021 for illustration.