Clinical Trials

Background: With the advancement in biomedicine, many biologically targeted therapies have been developed. These targeted agents, however, may not work for everyone. Biomarker profiles can be used to identify effective targeted therapies. Our goals are to characterize the molecular signature of individual tumors, offer the best-fit targeted therapies to patients in a study, and identify promising agents for future development.

Methods: We propose an outcome-based adaptive randomization trial design for patients with advanced stage non-small cell lung cancer. All patients have baseline biopsy samples taken for biomarker assessment prior to randomization to treatments. The primary endpoint of this study is the disease control rate at 8 weeks after randomization. The Bayesian probit model is used to characterize the disease control rate. Patients are adaptively randomized to one of four treatments with the randomization rate based on the updated disease control rate from the accumulated data in the trial. For each biomarker profile, high-performing treatments have higher randomization rates, and vice versa. An early stopping rule is implemented to suspend low-performing treatments from randomization.

Results: Based on extensive simulation studies, with a total of 200 evaluable patients, our trial has desirable operating characteristics to: (1) identify effective agents with a high probability; (2) suspend ineffective agents; and (3) treat more patients with effective agents that correspond to their biomarker profiles. Our trial design continues to update and refine the estimates as the trial progresses.

Limitations: This biomarker-based trial requires biopsible tumors and a two-week turn around time for biomarker profiling before randomization. Additionally, in order to learn from the interim data and adjust the randomization rate accordingly, the outcome-based adaptive randomization design is applicable only for trials when the endpoint can be assessed in a relative short period of time.

Conclusion: Bayesian adaptive randomization trial design is a smart, novel, and ethical design. In conjunction with an early stopping rule, it can be used to efficiently identify effective agents, eliminate ineffective ones, and match effective treatments with patients' biomarker profiles. The proposed design is suitable for the development of targeted therapies and provides a rational design for personalized medicine. Clinical Trials 2008; 5: 181—193. http://ctj.sagepub.com

Background: The evaluation and validation of surrogate endpoints have been extensively studied in the last decade. Prentice [1] and Freedman, Graubard and Schatzkin [2] laid the foundations for the evaluation of surrogate endpoints in randomized clinical trials. Later, Buyse et al. [5] proposed a meta-analytic methodology, producing different methods for different settings, which was further studied by Alonso and Molenberghs [9], in their unifying approach based on information theory.

Purpose: In this article, we focus our attention on the trial-level surrogacy and propose alternative procedures to evaluate such surrogacy measure, which do not pre-specify the type of association. A promising correction based on cross-validation is investigated. As well as the construction of confidence intervals for this measure.

Methods: In order to avoid making assumption about the type of relationship between the treatment effects and its distribution, a collection of alternative methods, based on regression trees, bagging, random forests, and support vector machines, combined with bootstrap-based confidence interval and, should one wish, in conjunction with a cross-validation based correction, will be proposed and applied. We apply the various strategies to data from three clinical studies: in opthalmology, in advanced colorectal cancer, and in schizophrenia.

Results: The results obtained for the three case studies are compared; they indicate that using random forest or bagging models produces larger estimated values for the surrogacy measure, which are in general stabler and the confidence interval narrower than linear regression and support vector regression. For the advanced colorectal cancer studies, we even found the trial-level surrogacy is considerably different from what has been reported.

Limitations: In general the alternative methods are more computationally demanding, and specially the calculation of the confidence intervals, require more computational time that the delta-method counterpart.

Conclusions: First, more flexible modeling techniques can be used, allowing for other type of association. Second, when no cross-validation-based correction is applied, overly optimistic trial-level surrogacy estimates will be found, thus cross-validation is highly recommendable. Third, the use of the delta method to calculate confidence intervals is not recommendable since it makes assumptions valid only in very large samples. It may also produce range-violating limits. We therefore recommend alternatives: bootstrap methods in general. Also, the information-theoretic approach produces comparable results with the bagging and random forest approaches, when cross-validation correction is applied. It is also important to observe that, even for the case in which the linear model might be a good option too, bagging methods perform well too, and their confidence intervals were more narrow. Clinical Trials 2008; 5: 194—208. http://ctj.sagepub.com

Background: The primary clinical endpoint in many phase II studies in cancer is a time-to-event outcome subject to potential censoring. The decision in favor of abandoning or continuing investigation of the protocol regimen is typically based on the amount of statistical evidence suggesting an improvement compared to a given historical control. The primary statistical endpoint would typically be the median of the time-to-event distribution of the clinical endpoint. For the purpose of sample size or power calculations, software implementing parametric and nonparametric median tests, is freely available. The main assumptions are those of Exponential survival and a Uniform right-censoring mechanism.

Purpose: The performance of the parametric and nonparametric methods is compared to that of using a binomial endpoint based on dichotomizing the event time at a clinically relevant landmark. As sufficient details on the various methods and related designs for phase II clinical design with survival endpoints are provided, this article should also serve as a comparative reference on these three methods for designing phase II studies in cancer with time-to-event endpoints.

Methods: The statistical performance, by virtue of considering the type I and II error rates, of the three methods is compared by carrying out a comprehensive simulation study.

Results: The parametric method may fail to control the type I error rate if the underlying survival distribution is not Exponential, while the nonparametric method may fail to control the type I error rate as the sample sizes for phase II studies are typically small. Both of these methods may be sensitive to the distribution of the censoring variable.

Limitations: The results provided in this article are mostly discussed in the framework of specific examples and by using specific implementations of the tests. As such the results may not be universally generalizable. The recommended method has some drawbacks if patients are censored before the landmark of interest.

Conclusion: A method that should be considered for the purpose of the statistical design and decision rule for phase II studies in cancer is the employment of a binomial endpoint based on dichotomizing the event time at a clinically relevant landmark. Clinical Trials 2008; 5: 209—221. http://ctj.sagepub.com

Background: Missing outcome data from randomized trials lead to greater uncertainty and possible bias in estimating the effect of an experimental treatment. An intention-to-treat analysis should take account of all randomized participants even if they have missing observations.

Purpose: To review and develop imputation methods for missing outcome data in meta-analysis of clinical trials with binary outcomes.

Methods: We review some common strategies, such as simple imputation of positive or negative outcomes, and develop a general approach involving `informative missingness odds ratios' (IMORs). We describe several choices for weighting studies in the meta-analysis, and illustrate methods using a meta-analysis of trials of haloperidol for schizophrenia.

Results: IMORs describe the relationship between the unknown risk among missing participants and the known risk among observed participants. They are allowed to differ between treatment groups and across trials. Application of IMORs and other methods to the haloperidol trials reveals the overall conclusion to be robust to different assumptions about the missing data.

Limitations: The methods are based on summary data from each trial (number of observed positive outcomes, number of observed negative outcomes and number of missing outcomes) for each intervention group. This limits the options for analysis, and greater flexibility would be available with individual participant data.

Conclusions: We propose that available reasons for missingness be used to determine appropriate IMORs. We also recommend a strategy for undertaking sensitivity analyses, in which the IMORs are varied over plausible ranges. Clinical Trials 2008; 5: 225—239. http://ctj.sagepub.com

Objective: Although diabetes is conveniently assessed by self-report, few validation studies have been performed. Therefore, we studied whether self-report of prevalent and incident diabetes in Women's Health Initiative (WHI) participants was concordant with other diagnostic evidence of diabetes.

Study Design and Setting: A total of 161 808 postmenopausal women aged 50—79 were enrolled at 40 clinical centers in the U.S. in 1993—1998 and followed prospectively. At baseline, prevalent medication treated diabetes was defined as a self-report of physician diagnosis and treatment with insulin or oral antidiabetic drugs. During followup, incident treated diabetes was defined as a self-report of a new physician diagnosis of diabetes treated with insulin or oral drugs. Diabetes self-reports were compared with medication inventories and fasting glucose levels at baseline and during follow-up.

Results: At baseline, self-reported treated diabetes was concordant with the medication inventory in 79% of clinical trial, and 77% of observational study participants. Self-reported incident treated diabetes was concordant with the medication inventory in 78% between baseline and Year 1 in the clinical trials, in 62% between Year 1 and Year 3 in the clinical trials, and in 72% between baseline and Year 3 in the observational study. Over similar periods, 99.9% of those who did not report treated diabetes had no oral antidiabetic drugs or insulin in the medication inventory. At baseline, about 3% not reporting diabetes had fasting glucose >126 mg/dl, and 88% of these subjects subsequently reported treated diabetes during 6.9 years of follow-up.

Limitations: Incident self-reported diabetes treated by lifestyle alone was not determined in WHI. Medication inventories may have been incomplete and fasting glucose may have been lowered by treatment; therefore, concordance with self-reported treatment or fasting glucose ≥ 126 may have been underestimated.

Conclusion: In the WHI, self-reported prevalent and incident diabetes was consistent with medication inventories, and a high proportion of those with undiagnosed diabetes subsequently reported diabetes treatment. Self-reports of `treated diabetes' are sufficiently accurate to allow use in epidemiologic studies. Clinical Trials 2008; 5: 240—247. http://ctj.sagepub.com

Background: Clinical trials with time to event outcomes are often designed utilizing the Cox [1] proportional hazard model with a hazard ratio parameter Δ.

Purpose: The purpose of this article is to demonstrate that a Cox proportional hazard model with a hazard ratio parameter is equivalent to a Cox proportional hazard model with a parameter equal to the probability that a patient given one treatment will have an event earlier than if the same patient were given a different treatment. This probability will subsequently be referred to as θ. Clinically interesting differences between the treatment arms are easier for researchers to quantify in terms of θ in situations where they have a difficult time with the hazard ratio, allowing better communication between the statistician and the researcher.

Methods: The problem and its solution are demonstrated mathematically. The utility of the Cox proportional hazard model in terms of θ is illustrated through a Lymphoma clinical trial example.

Results: The Cox proportional hazard model with parameter θ is shown to be equivalent to the Cox proportional hazard model with a hazard ratio parameter Δ. A table of typical hazard ratios Δ is presented with their equivalent θ values. In the appendix the mathematical derivations are developed and an unbiased estimate of θ is provided using Gehan's [2] generalization of the Wilcoxon statistic.

Limitations: The equivalence of the Cox proportional hazard model in terms of the probability θ and the hazard ratio Δ is established only for continuous failure times with a single binary covariate. Conditions under which approximate equivalence holds with multiple covariates are discussed in the Appendix.

Conclusions: The probability θ provides a natural parameterization for the Cox proportional hazard model, affords a tool to conceptualize treatment differences, and provides a method to improve communication between statisticians and researchers. Clinical Trials 2008; 5: 248—252. http://ctj.sagepub.com

Background: To date, no cross-national RCT has addressed the mechanisms underlying the relative success of pharmacological and psychotherapeutic interventions for depression. A multi-site clinical trial that includes psychotherapy as one of the treatments presents numerous challenges related to cross-site consistency and communication.

Purpose: This report describes how those challenges were met in the study ``Depression: The Search for Treatment Relevant Phenotypes'', being carried out at the University of Pittsburgh and the University of Pisa, Italy.

Methods: Implementing the study required the investigators to address methodological and practical challenges related to the different requirements of the two Institutional Review Boards (IRBs), psychotherapy training, independent evaluator training, patient recruitment, development of common tools for data entry, quality control and generation of weekly reports of patient progress as well as establishing a similar clinical and research framework in two countries with substantially different health care systems.

Results: By having bilingual investigators and staff members who spent time at one another's sites, making use of frequent conference-call staff meetings and being flexible within the bounds of the sometimes contradictory requirements of the IRBs, the investigators were able to meet the human subjects protection requirements of both institutions, surmount language barriers to consistent therapist and evaluator training and develop common tools for study management. As a result, recruitment goals were met at both sites and retention rates were high. One instance of inconsistent implementation of the protocol was corrected within the first year.

Limitations: This study was conducted in two Western cultures by researchers with long-standing collaboration. Our findings may not be generalizable to other countries or research settings.

Conclusions: The implementation of a cross-national protocol and the adoption and maintenance of common procedures is possible when investigators are aware of the challenges this may present and are proactive in trying to address them. Clinical Trials 2008; 5: 253—261. http://ctj.sagepub.com

Background: The time between the diagnosis of cancer and a planned definitive surgical procedure offers a strong and direct approach for assessing the impact of interventions (including lifestyle interventions) on the biology of the target tissue and the tumor. Despite the many strengths of presurgical models, there are practical issues and challenges that arise when using this approach.

Purpose:/Methods We recently completed an NIH-funded phase II trial that utilized a presurgical model in testing the comparative effects of flaxseed supplementation and/or dietary fat restriction on the biology and biomarkers associated with prostatic carcinoma. Herein, we report the rationale for our original design, discuss modifications in strategy, and relay experiences in implementing this trial related to the following topics: (1) subject accrual; (2) subject retention; (3) intervention delivery; and (4) retrieval and completion rates regarding the collection of paraffin-embedded and fresh frozen prostate tissue, blood, urine, ejaculate, anthropometric measures and survey data.

Results: This trial achieved its accrual target, i.e., a racially-representative (70% white, 30% minority) sample of 161 participants, low rates of attrition (7%); and collection rates that exceeded 90% for almost all biospecimens and survey data. While the experience gained from pilot studies was invaluable in designing this trial, the complexity introduced by the collection of several biospecimens, inclusion of a team of pathologists (to provide validated readings), and shifts in practice patterns related to prostatectomy, made it necessary to revise our protocol; lessons from our experiences are offered within this article.

Conclusions: While our experience specifically relates to the implementation of a presurgical model-based trial in prostate cancer aimed at testing flaxseed-supplemented and fat-restricted diets, many of the lessons learned have broad application to trials that utilize a presurgical model or dietary modification within various cancer populations. Clinical Trials 2008; 5: 262—272. http://ctj.sagepub.com

Background: The methods and purposes of randomization are often poorly understood by participants in clinical trials. Individual misunderstandings can be compounded in community-based intervention trials, especially in research-naïve communities. Randomizing entire communities to intervention or control status risks creating the perception that control communities are being denied desirable services, ultimately undermining trust in the research process.

Purpose: To develop a randomization scheme for an HIV prevention trial of a community-level intervention that would be credible to the communities involved while maintaining the scientific integrity of the intervention trial at a rural site in Zimbabwe.

Methods: Project staff developed strong partnerships with community stakeholders and embedded randomization into the trial's community preparedness processes. Local idioms were used to explain the concept, purpose, and mechanics of randomization. Actual allocation of communities to intervention or control status took place at a public lottery conducted by local chiefs.

Results: The Project obtained the endorsement of its randomization of eight rural communities by local political stakeholders and community members.

Limitations: This case study may not generalize to other settings.

Conclusions: By developing strong community partnerships, and communicating randomization through local idioms, community based intervention trials conducted in resource-poor environments can successfully mitigate risks inherent in randomizing communities to control status. Clinical Trials 2008; 5: 273—276. http://ctj.sagepub.com

Data Monitoring Committees (DMC) and Institutional Review Boards (IRB) each can play important roles in protecting the rights and interests of research participants and ensuring the integrity of research. While both IRBs and DMCs have unique responsibilities, promoting a robust human research participant protection program requires integration of these efforts. Communication about DMC actions and considerations to IRBs is arguably an important component of integration. We sought to explore whether and how DMCs actions are currently communicated to IRBs as a basis for recommendations to improve such communications. Evidence of communication was sought in files related to research conducted by faculty, staff, and students affiliated with the Johns Hopkins Medical Institutions and the Johns Hopkins Bloomberg School of Public Health. Overall, we found a lack of consistency in the way that DMC actions are communicated to IRBs. While national policy encourages that actions taken by DMCs be communicated to IRBs and there is not a clear consensus about either who should be responsible for disseminating the information to the IRB or what information ought to be considered standard. Such standards promise to enhance the quality of ethics oversight of research. Clinical Trials 2008; 5: 277—282. http://ctj.sagepub.com