12 Design and Analysis of Clinical Trials

Ling Chen and Kathryn Trinkaus

QUESTIONS

Each of the numbered items below is followed by lettered answers. Select the ONE lettered answer that is BEST in each case unless instructed otherwise.

Question 12.1 Which of the following is TRUE concerning a clinical trial?

A. The sample size of a clinical trial is fixed and determined before the trial begins.

B. Clinical trials are prospectively planned experiments involving only animals.

C. Using tumor registry data to compare the survival rates of prostate cancer patients treated with surgery to those of patients who received radiotherapy is an example of clinical trial.

D. Clinical trials test a clearly stated hypothesis using a predefined analysis plan.

Question 12.2 Phase I trials of cytotoxic agents are designed to determine a dose that is appropriate for use in phase II trials. Which of the following statements is TRUE about phase I trials?

A. Phase I trials involve the first use of a drug, device, or procedure in humans in a new disease setting or a new combination of therapies.

B. Patients for phase I trials are usually not chosen on the basis of their likelihood of having a favorable response to the experimental treatment.

C. Phase I trials do not require statistical justification because the sample sizes are usually small.

D. A and B.

Question 12.3 Which of the following regarding the conventional 3 + 3 phase I designs is TRUE?

A. Participants are often those with end-stage disease who have many treatment options.

B. Such studies are usually performed by starting with a high dose.

C. Cohorts of three to six patients are treated at each dose level, and the initial dose is chosen to cause little or no toxicity.

D. The optimal dose recommended for phase I is chosen using a dose-toxicity curve generated from a statistical model fitted to the full data.

Question 12.4 Traditional phase I trials are designed to determine a dose that is appropriate for use in phase II trials. Which of the following statement on traditional phase I trials is TRUE?

A. Many patients may receive subtherapeutic doses and it does not allow for dose escalation in the individual patient.

B. All phase I trials are completed in a short time period.

C. Dose escalation in the individual patient is allowed.

D. Phase I trial provides substantial information regarding cumulative toxicity from the study drug

Question 12.5 A drug, device, or other treatment may be considered for phase II testing when:

A. The maximum tolerable dose is known but there is genuine doubt about efficacy and patients with the condition of interest and a likely favorable outcome can be recruited as study participants.

B. A standard treatment exists for comparison.

C. Accrual is rapid enough to allow recruitment of 20 to 30 patients.

D. An appropriate control group can be recruited at the same time as the experimental group.

Question 12.6 which of the following statements regarding phase II trials is CORRECT?

A. Phase II trials are traditionally performed in patients with a wide variety of tumor types.

B. Two-stage designs allow for early elimination of ineffective drugs.

C. Patient eligibility should not be restricted by biomarker testing.

D. Phase II trials cannot help develop predictive biomarker assays for the study drug.

Question 12.7 Phase II trials of a single agent are designed to determine whether the experimental drug has any antitumor activity against a given type of tumor and the antitumor effect is often measured as an objective response based on tumor shrinkage. However, such a trial is NOT appropriate for the question of patient welfare because:

A. The welfare of patients with cancer is often measured by survival or symptom control, but there is no necessary relationship between such beneficial end points and tumor response.

B. Survival is inherently a comparative end point that should be assessed in a prognostically comparable set of patients.

C. Comparing the survival times of responders and nonresponders is not a valid means of showing a treatment is beneficial to patients because responders may have more favorable prognostic factors than nonresponders.

D. All of the above.

Question 12.8 Patients in a phase II trial are often recruited in a two-stage manner that allows early stopping if the null hypothesis is unlikely to be rejected, thus saving patients from unnecessary exposure to inactive agents. Which of the following is TRUE of a two-stage design?

A. The optimal two-stage design minimizes the average sample size and thus optimizes the protection of patients given an inactive agent.

B. Simon two-stage designs are often applied in phase II trials on molecularly targeted drugs and therapeutic cancer vaccines.

C. If both stages are completed, the minimax design usually enrolls more patients compared with the optimal design.

D. Simon two-stage designs are often used in phase II trials comparing treatment groups to an internal control group to draw conclusions about survival.

Question 12.9 Unlike cytotoxic agents that work by killing tumor cells, cytostatic agents work via biologic effects that modify the environment of tumor growth; that is, inhibition of a molecular target. Which of the following is TRUE regarding early-phase trials of cytostatic agents such as molecularly targeted drugs and cancer therapeutic vaccines?

A. Phase I trials on cytostatic agents are often designed to identify maximum tolerated dose.

B. Heavily treated patients with end-stage disease greatly benefit from a cancer vaccine.

C. The response rate based on tumor shrinkage is the best primary end point measuring treatment efficacy in phase II trials on cytostatic agents.

D. Because of the nature of low toxicity profiles in cytostatic agents, factorial designs and randomized studies have appealing features and are used to evaluate multiple drugs in a single trial.

Question 12.10 Which of the following best describes Simon phase 2.5 design on molecularly targeted drugs and cancer therapeutic vaccines?

A. The design compares progression-free survival times of the same patient from the current trial with his/her previous trial.

B. The design first treats all eligible patients with two to four courses of the experimental drugs. Patients are then evaluated, and the experimental drug is either continued or discontinued, depending on their response to the initial treatment.

C. The design is similar to a phase III trial except that it allows a relatively large type I (false positive) error and aims to identify a relatively large difference, thus requiring fewer patients.

D. The design will compare time to progression of patients in the study with a prospectively identified and prognostically comparable historical control.

Question 12.11 Which of the following is TRUE of phase III trials?

A. The primary end points should be demonstrably related to patient benefit such as tumor shrinkage.

B. Eligibility criteria should be limited to a group of patients with certain conditions.

C. Conditions of treatment should be suited to community patient care.

D. A quality-of-life measure should be included to measure benefit to patients.

Question 12.12 Which of the following is TRUE of randomized treatment assignment?

A. Trial participants are a representative sample of patients with the disease of interest.

B. Randomization may fail to balance unknown prognostic factors in small studies.

C. Randomization is unnecessary if controls can be matched closely enough to trial participants.

D. Randomization is effective only if patients and treating physicians are unaware of treatment assignment for the duration of the study.

Question 12.13 Randomization is generally unnecessary in which of the following scenarios?

A. Study of patients with a disease that will uniformly and rapidly progress to an end stage.

B. Phase III trial

C. Screening study on multiple cytostatic agents, such as molecularly targeted drugs or cancer therapeutic vaccines

D. Phase II trial on cytotoxic agents using time to progression as the primary efficacy end point

Question 12.14 Which of the following is TRUE of power and hypothesis testing in clinical trials?

A. Statistical power mainly depends on the anticipated size of treatment effect. Trials designed to detect a large effect will demand small sample sizes.

B. Statistical power is also heavily influenced by the prespecified type I error (probability of erroneously claiming a “positive” result), and a two-sided type I error of 0.05 has been widely accepted, especially in phase III trials.

C. Confidence intervals do not provide information about the size and direction of treatment effect, and it is less informative than significance testing.

D. Because data from small randomized trials can be pooled for meta-analysis, the calculation of sample size for a randomized clinical trial is no longer critical.

Question 12.15 Which of the following is TRUE for therapeutic equivalence (or noninferiority) trials?

A. Therapeutic equivalence trials are designed to test similarity, as measured by a clinical end point, between the experimental drugs and the standard treatment.

B. Because it is impossible to prove equivalence in true sense, in practice therapeutic equivalence will be accepted if the difference in efficacy is larger than or smaller than a prespecified amount δ.

C. The traditional superiority trials can be used to prove equivalence

D. A trial designed to demonstrate therapeutic equivalence requires fewer patients because it usually allows a relatively large type I error.

Question 12.16 Therapeutic equivalence trials often express results as an estimate of difference between two treatments with a 95% confidence interval. Which of the following statements about the 95% confidence interval is CORRECT?

A. The confidence interval is a range of values consistent with the hypothesized difference, given the precision with which the trial can measure that difference.

B. The trial estimates that in 5 of 100 cases the true difference will fall outside the 95% confidence interval.

C. It is unlikely that the true difference is larger than the lower limit or smaller than the upper limit of the confidence interval.

D. If the confidence interval includes 0, then the trial has shown that the two treatments do not differ.

Question 12.17 Which one of the following describing the pros and cons of Bayesian methods in planning clinical trials is TRUE?

A. In the traditional statistical (frequentist) methods, the treatment effect is regarded as a fixed but unknown parameter, and the likelihood of the parameter is described using a probability derived from observed frequencies in a defined distribution. In contrast, a Bayesian statistical method assumes that the treatment effect itself is a random quantity drawn from a prior distribution.

B. Bayesian methods are more widely applied in planning phase III trials than phase I or phase II trials.

C. Bayes’ theorem can be considered as a substitute for randomization.

D. Bayesian clinical trials require fewer patients than frequentist trials.

Question 12.18 The intention-to-treat analysis is considered the primary analysis of data from a clinical trial. Which of the following describes an intention-to-treat analysis?

A. An ineligible patient is not included in an intention-to-treat analysis because the study treatment is not appropriate for that patient.

B. All patients who give consent and are randomized are included in the intention-to-treat analysis in the study arm to which they were randomized.

C. Data from patients who are ineligible or who die or withdraw from the study before completion should be analyzed separately from data from eligible patients who complete the trial.

D. Lack of compliance can distort the results from patients who fail to complete the trial, so these data should be dropped from an intention-to-treat analysis.

Question 12.19 An interim analysis is any assessment of data performed during patient enrollment or follow-up period of a trial. Which of the following is TRUE for planning clinical trials with interim analyses?

A. Multiple looks at outcome data over the course of a trial will alter type I and II errors. Unless they are included in the design, multiple looks will cause the operating characteristics of the trial to deviate from the planned values.

B. Interim outcome information is generally available to participating physicians and study leaders.

C. The method of stochastic curtailment is designed to calculate conditional power (the projected probability of rejecting the null hypothesis at the end of study), conditional on the accumulated information, and to stop the trial if the conditional power is small (e.g., 0.2). In such a “futility” analysis, the number of interim analyses does not need to be fixed in advance.

D. A data-safety committee is usually put in charge of the interpretation of interim analyses. The committee also includes the study leaders to notify the patients and medical community in a timely manner.

Question 12.20 One of the most important features that make survival data different from other kinds of data is the presence of censoring. That is, the actual survival times for some patients cannot be observed during the study. Which of the following is TRUE for the analysis of survival data?

A. The usual statistical methods are appropriate for analyzing survival data.

B. Two methods are most frequently used for the estimation of survival curves. The life table method requires relatively large sample size, and the Kaplan–Meier product limit method is appropriate for any sample size.

C. Traditional summary statistics, such as means and standard deviations, can provide a sufficient summary for survival data without censoring.

D. The statistical power of clinical trials using survival as a primary end point will be mainly determined by the total number of patients rather than the actual number of events including death or tumor progression.

Question 12.21 An important assumption in the analysis of survival data is noninformative censoring. That is, censoring is not related to the outcome of interest. For a clinical trial with cancer-specific survival as the primary end point, which of the following patients is more likely to suffer from informative censoring?

A. Alive at the time of study closeout

B. Withdrew from the trial because of disease progression

C. Died of accident

D. Moved out of the region

Question 12.22 Which of the following is characteristic of phase 0 trials?

A. Phase 0 trials do not require an assay for measuring the pharmacodynamic endpoint.

B. Patients receive multiple doses of the experimental drug.

C. Phase 0 trials identify whether the experimental drug has inhibitory effect against its specified molecular target.

D. Drug doses used in these trials are expected to cause significant toxicity.

Question 12.23 Which of the following is TRUE regarding meta-analysis?

A. Meta-analysis combines the results of several studies that address a set of related research hypothesis.

B. Meta-analysis on therapeutic studies usually includes both randomized and nonrandomized trials.

C. Meta-analysis based on summarized p-values is more precise than meta-analysis based on individual patient data.

D. Meta-analysis is an alternative to properly designed and sized randomized clinical trials.

Stay updated, free articles. Join our Telegram channel

Join