Search

The pharmaceutical industry and its global regulators have routinely used frequentist statistical methods, such as null hypothesis significance testing and pvalues, for evaluation and approval of new treatments. The clinical drug development process, however, with its accumulation of data over time, can be well suited for the use of Bayesian statistical approaches that explicitly incorporate existing data into clinical trial design, analysis and decision-making. Such approaches, if used appropriately, have the potential to substantially reduce the time and cost of bringing innovative medicines to patients, as well as to reduce the exposure of patients in clinical trials to ineffective or unsafe treatment regimens. Nevertheless, despite advances in Bayesian methodology, the availability of the necessary computational power and growing amounts of relevant existing data that could be used, Bayesian methods remain underused in the clinical development and regulatory review of new therapies. Here, we highlight the value of Bayesian methods in drug development, discuss barriers to their application and recommend approaches to address them. Our aim is to engage stakeholders in the process of considering when the use of existing data is appropriate and how Bayesian methods can be implemented more routinely as an effective tool for doing so.

Background: Reasons for treatment discontinuation are important not only to understand the benefit and risk profile of experimental treatments, but also to help choose appropriate strategies to handle intercurrent events in defining estimands. The current case report form (CRF) commonly in use mixes the underlying reasons for treatment discontinuation and who makes the decision for treatment discontinuation, often resulting in an inaccurate collection of reasons for treatment discontinuation. Methods and results: We systematically reviewed and analyzed treatment discontinuation data from nine phase 2 and phase 3 studies for insulin peglispro. A total of 857 participants with treatment discontinuation were included in the analysis. Our review suggested that, due to the vague multiple-choice options for treatment discontinuation present in the CRF, different reasons were sometimes recorded for the same underlying reason for treatment discontinuation. Based on our review and analysis, we suggest an intermediate solution and a more systematic way to improve the current CRF for treatment discontinuations. Conclusion: This research provides insight and directions on how to optimize the CRF for recording treatment discontinuation. Further work needs to be done to build the learning into Clinical Data Interchange Standards Consortium standards. Clinical Trials: Clinicaltrials.gov numbers: NCT01027871 (Phase 2 for type 2 diabetes), NCT01049412 (Phase 2 for type 1 diabetes), NCT01481779 (IMAGINE 1 Study), NCT01435616 (IMAGINE 2 Study), NCT01454284 (IMAGINE 3 Study), NCT01468987 (IMAGINE 4 Study), NCT01582451 (IMAGINE 5 Study), NCT01790438 (IMAGINE 6 Study), NCT01792284 (IMAGINE 7 Study).

An overarching framework is proposed to guide the design of phase 2 studies in central nervous system disorders. Archetypes are considered for scenarios where dose response is highly relevant in clinical practice, as in the symptomatic treatment of acute disorders. Archetypes for scenarios where dose response is less relevant, as in disease modification for neurodegenerative disorders, are beyond the scope of this article. Primary design archetypes are determined by axes of development that are defined by optimism for success (probability of efficacy) and signal detection (magnitude of the anticipated effect size). The fast-to-registration primary archetype uses a dose-response study as the first efficacy, that is, proof of concept (PoC), study and is appropriate when the prospects for signal detection and the optimism for efficacy are higher. These conditions may exist when the anticipated effect size is large and when either testing a drug with a proven mechanism of action or when a favorable biomarker result was obtained in phase 1. The fast-to-PoC primary archetype tests one dose arm to establish PoC before assessing dose response and is appropriate when the optimism for efficacy and the prospects for signal detection are lower. These conditions may exist when testing a drug with a novel mechanism and/or the anticipated effect size is smaller. Secondary archetypes are used to mitigate the trade-offs between the quick-kill fast-to-PoC approach and the quick-win fast-to-registration approach, and are key areas where adaptive designs can be beneficial.

Active comparators are often included in phase 2 studies as a positive control to assess assay sensitivity, but inclusion of a positive control does not necessarily improve decision making. Examples show that positive controls are more useful in assessing assay sensitivity as the probability the test drug is effective decreases and as power for the contrast of the positive control versus placebo increases. These results suggest that a positive control should be powered at a minimum of 807o, and preferably at 90%. Analogously, a positive control can be used in an estimation framework to assess whether the study performed as expected, thereby indicating whether or not the test drug was assessed under the anticipated conditions. In so doing, the sample would need to be sufficiently large to ensure reliable estimation. The key point again is that results of the positive control must be reliable if they are to be useful, and adding a small sample of patients in a positive control arm can do more harm than good. It is also important to recognize that including a positive control only allows assessment, but not improvement, of assay sensitivity. Actual clinical trial data are used to suggest that two smaller two-arm studies of test drug and placebo instead of one larger study that also includes a positive control may improve assay sensitivity with little to no increase in the total sample size.

An increasing need exists to understand the benefits of a test drug compared with standard of care (SoC) earlier in development. Even if a drug is superior to placebo, it may not be worthwhile to continue development unless it has advantages over SoC. However, efficacy and safety comparisons versus SoC in early phase studies can be challenging. Analytic studies were conducted to illustrate that in common scenarios simply randomizing a few patients to SoC will frequently yield misleading results. It may take samples sizes at least fivefold greater to achieve reliable comparisons of a test drug with SoC than when comparing versus placebo. Therefore, it is important that the rate of false positive and false negative results be quantitatively evaluated before determining the sample size and the criteria upon which the test drug and SoC will be compared. Because test drugs often have no benefit, comparing a test drug with SoC may unnecessarily use resources that could be devoted to investigating other drugs. Moreover, it can be difficult to construct valid comparisons of a test drug versus SoC without experience with the test drug regarding appropriate dosing, patient population, and so on. An example with actual clinical trial data is used to illustrate how the trade-off between the need and the difficulties in comparing a test drug with SoC in phase 2 can be mitigated using a literature database of placebo-controlled studies to construct a historical comparison.

An increasing need exists to understand the benefits of a test drug compared with standard of care (SoC) earlier in development. Even if a drug is superior to placebo, it may not be worthwhile to continue development unless it has advantages over SoC. However, efficacy and safety comparisons versus SoC in early phase studies can be challenging. Analytic studies were conducted to illustrate that in common scenarios simply randomizing a few patients to SoC will frequently yield misleading results. It may take samples sizes at least fivefold greater to achieve reliable comparisons of a test drug with SoC than when comparing versus placebo. Therefore, it is important that the rate of false positive and false negative results be quantitatively evaluated before determining the sample size and the criteria upon which the test drug and SoC will be compared. Because test drugs often have no benefit, comparing a test drug with SoC may unnecessarily use resources that could be devoted to investigating other drugs. Moreover, it can be difficult to construct valid comparisons of a test drug versus SoC without experience with the test drug regarding appropriate dosing, patient population, and so on. An example with actual clinical trial data is used to illustrate how the trade-off between the need and the difficulties in comparing a test drug with SoC in phase 2 can be mitigated using a literature database of placebo-controlled studies to construct a historical comparison.

Active comparators are often included in phase 2 studies as a positive control to assess assay sensitivity, but inclusion of a positive control does not necessarily improve decision making. Examples show that positive controls are more useful in assessing assay sensitivity as the probability the test drug is effective decreases and as power for the contrast of the positive control versus placebo increases. These results suggest that a positive control should be powered at a minimum of 807o, and preferably at 90%. Analogously, a positive control can be used in an estimation framework to assess whether the study performed as expected, thereby indicating whether or not the test drug was assessed under the anticipated conditions. In so doing, the sample would need to be sufficiently large to ensure reliable estimation. The key point again is that results of the positive control must be reliable if they are to be useful, and adding a small sample of patients in a positive control arm can do more harm than good. It is also important to recognize that including a positive control only allows assessment, but not improvement, of assay sensitivity. Actual clinical trial data are used to suggest that two smaller two-arm studies of test drug and placebo instead of one larger study that also includes a positive control may improve assay sensitivity with little to no increase in the total sample size.

An overarching framework is proposed to guide the design of phase 2 studies in central nervous system disorders. Archetypes are considered for scenarios where dose response is highly relevant in clinical practice, as in the symptomatic treatment of acute disorders. Archetypes for scenarios where dose response is less relevant, as in disease modification for neurodegenerative disorders, are beyond the scope of this article. Primary design archetypes are determined by axes of development that are defined by optimism for success (probability of efficacy) and signal detection (magnitude of the anticipated effect size). The fast-to-registration primary archetype uses a dose-response study as the first efficacy, that is, proof of concept (PoC), study and is appropriate when the prospects for signal detection and the optimism for efficacy are higher. These conditions may exist when the anticipated effect size is large and when either testing a drug with a proven mechanism of action or when a favorable biomarker result was obtained in phase 1. The fast-to-PoC primary archetype tests one dose arm to establish PoC before assessing dose response and is appropriate when the optimism for efficacy and the prospects for signal detection are lower. These conditions may exist when testing a drug with a novel mechanism and/or the anticipated effect size is smaller. Secondary archetypes are used to mitigate the trade-offs between the quick-kill fast-to-PoC approach and the quick-win fast-to-registration approach, and are key areas where adaptive designs can be beneficial.

The use of the world wide web for clinical trials changes the processes of performing clinical research in several fundamental ways. Greatly improved security, monitoring capability, and accuracy and timeliness of study conduct can be achieved while lowering cost. Data quality is enhanced while co-ordinating centre effort is reduced. The web provides a natural environment for linking the various components of clinical research, leading to new levels of simplicity and efficiency. It also enhances opportunities for recruitment of study investigators and patients. Other information technology tools and databases can be used to assist in this regard as well. Web-based trials change the relationship of the investigator site to the study and the site to the co-ordinating centre. Different roles and responsibilities lead to simplified processes and more and higher quality data. Many standard co-ordinating centre activities, such as randomization, protocol implementation and amending, document tracking, adverse event reporting, site monitoring, report generation and data analysis are all fundamentally changed in a web-based trial. Opportunities are enhanced to identify potential investigators and support their successful study conduct. As the role of investigator sites is changed in web-based research, more primary care medical providers can be attracted to research, providing more typical patients to studies than those sometimes available through more traditional research sites, especially those at academic study sites. Other activities can now be co-ordinated electronically with the advent of the web. The Institutional Review Board (IRB) can use online tools to control investigator participation, resulting in improved study efficiency and patient safety. A web-based research pharmacy provides tremendous efficiencies in managing and distributing study medications. Financial payments to the sites can be performed and recorded electronically, or even administered based on timeliness and quality of the data. Our early experience with web-based trials indicates that there can be tremendous gains in study efficiency and accuracy by restructuring processes, roles and responsibilities through a comprehensive centralized, web-based trial. The future appears bright for web-based clinical trials. Copyright © 2001 John Wiley & Sons, Ltd.

The use of the world wide web for clinical trials changes the processes of performing clinical research in several fundamental ways. Greatly improved security, monitoring capability, and accuracy and timeliness of study conduct can be achieved while lowering cost. Data quality is enhanced while co‐ordinating centre effort is reduced. The web provides a natural environment for linking the various components of clinical research, leading to new levels of simplicity and efficiency. It also enhances opportunities for recruitment of study investigators and patients. Other information technology tools and databases can be used to assist in this regard as well. Web‐based trials change the relationship of the investigator site to the study and the site to the co‐ordinating centre. Different roles and responsibilities lead to simplified processes and more and higher quality data. Many standard co‐ordinating centre activities, such as randomization, protocol implementation and amending, document tracking, adverse event reporting, site monitoring, report generation and data analysis are all fundamentally changed in a web‐based trial. Opportunities are enhanced to identify potential investigators and support their successful study conduct. As the role of investigator sites is changed in web‐based research, more primary care medical providers can be attracted to research, providing more typical patients to studies than those sometimes available through more traditional research sites, especially those at academic study sites. Other activities can now be co‐ordinated electronically with the advent of the web. The Institutional Review Board (IRB) can use online tools to control investigator participation, resulting in improved study efficiency and patient safety. A web‐based research pharmacy provides tremendous efficiencies in managing and distributing study medications. Financial payments to the sites can be performed and recorded electronically, or even administered based on timeliness and quality of the data. Our early experience with web‐based trials indicates that there can be tremendous gains in study efficiency and accuracy by restructuring processes, roles and responsibilities through a comprehensive centralized, web‐based trial. The future appears bright for web‐based clinical trials. Copyright © 2001 John Wiley & Sons, Ltd.

Analyses of dose response studies should separate the question of the existence of a dose response relationship from questions of functional form and finding the optimal dose. A well-chosen contrast among the estimated effects of the studied doses can make a powerful test for detecting the existence of a dose response relationship. A contrast-based test attains its greatest power when the pattern of the coefficients has the same shape as the true dose response relationship. However, it loses power when the contrast shape and the true dose response shape are not similar. Thus, a primary test based on a single contrast is often risky. Two (or more) appropriately chosen contrasts can assure sufficient power to justify the cost of a multiplicity adjustment. An example shows the success of a two-contrast procedure in detecting dose response, which had frustrated several standard procedures. Copyright © 2000 John Wiley & Sons, Ltd.

Analyses of dose response studies should separate the question of the existence of a dose response relationship from questions of functional form and finding the optimal dose. A well‐chosen contrast among the estimated effects of the studied doses can make a powerful test for detecting the existence of a dose response relationship. A contrast‐based test attains its greatest power when the pattern of the coefficients has the same shape as the true dose response relationship. However, it loses power when the contrast shape and the true dose response shape are not similar. Thus, a primary test based on a single contrast is often risky. Two (or more) appropriately chosen contrasts can assure sufficient power to justify the cost of a multiplicity adjustment. An example shows the success of a two‐contrast procedure in detecting dose response, which had frustrated several standard procedures. Copyright © 2000 John Wiley & Sons, Ltd.

Currently, bioavailability is assessed using the ratio of mean pharmacokinetic parameters (e.g., area under the plasma concentration-time profile) of the test formulation to the reference formulation. This estimate may be markedly different from the central tendency of individual ratios even if the parameter is normally distributed. Since an estimate of the location of individual bioavailability is more meaningful than the ratio of mean pharmacokinetic parameter in many applications, there is a desire to study individual ratios. Because the underlying distributions of individual subject ratios tend to be either skewed or heavy-tailed, two robust procedures are proposed to estimate bioavailability. The first approach is based on the generalized maximum likelihood estimator with a t-distribution weight function. It has been shown that thisestimator is more efficient than a number of traditional and nonparametric estimators for a family of distributions with different shape parameters. For the comparison purpose, another approach based on the sample median is also presented. A numerical example is provided to illustrate these procedures.

The pharmacokinetics of the terfenadine active metabolite, metabolite I, was examined in ten healthy elderly adults and ten younger adults after single‐dose oral administration of 120‐mg terfenadine. All subjects successfully completed the study without reporting sedation or other adverse events. Absorption was rapid in both the young and elderly. The mean Cmax was the same for both groups, 501 ng/mL, and occurred at 2.3 hours in the young subjects and 2.5 hours in elderly subjects. However, the apparent clearance was reduced by about 25% in the elderly. After correcting clearance for bodyweight, this difference was not statistically significant.

A two-sample test statistic for detecting shifts in location is developed for a broad range of underlying distributions using adaptive techniques. The test statistic is a linear rank statistics which uses a simple modification of the Wilcoxon test; the scores are Winsorized ranks where the upper and lower Winsorinzing proportions are estimated in the first stage of the adaptive procedure using sample the first stage of the adaptive procedure using sample measures of the distribution's skewness and tailweight. An empirical relationship between the Winsorizing proportions and the sample skewness and tailweight allows for a 'continuous' adaptation of the test statistic to the data. The test has good asymptotic properties, and the small sample results are compared with other populatr parametric, nonparametric, and two-stage tests using Monte Carlo methods. Based on these results, this proposed test procedure is recommended for moderate and larger sample sizes.

The goal of drug development must be to eliminate from further development those compounds that are not worthy of further investment and to eliminate them as soon as possible. Issues in dealing with early clinical trials in man and questions regarding their necessity are presented. The argument is made that early Phase I and Phase II trials, or at least the reasons for running them, must be standardized rigidly across all drugs in a portfolio and that early Phase II trials must be made definitive in the patient population of interest. An example of such a scheme is given. A list of intangible benefits of a standardized process is also offered.

The hemodynamic effects and serum levels of piroximone (MDL 19,205), a new inotropic agent with vasodilating properties, were measured in 10 patients with chronic severe congestive failure during a constant 48-h infusion. The initial five patients (group A) received piroximone at 10 μg/kg/min; however, because a sustained increase in heart rate ≥25 from baseline developed in two patients and an episode of paroxysmal supraventricular tachycardia developed in another, the last 5 patients (group B) received an 8 μg/kg/min infusion. Because the steady-state hemodynamic alterations of group A prior to the onset of tachyarrhythmias were similar to those of group B, these results were combined. A significant increase in cardiac output from 3.65 ± 0.31 (SE) to 5.20 ± 0.49 L/min and decrease in pulmonary capillary wedge pressure (27 ± 2 to 20 ± 2 mm Hg), right atrial pressure (18 ± 2 to 11 ± 2 mm Hg), and systemic vascular resistance (1811 ± 172 to 1293 ± 80 dynes · s · cm−5) occurred (all p < 0.05) without a significant change in mean arterial pressure. The peak plasma piroximone level was lower in the eight patients who did not develop a sustained increase in heart rate ≥25 above baseline (2.1 ± 0.5 μg/ml; range 1.6–2.9 μg/ml) than in the two who did (5.0 and 5.8 μg/ml). The latter two patients had the highest serum creatinine levels in the study population. In conclusion, intravenous (i.v.) piroximone administered as a constant i.v. infusion produced favorable hemodynamic alterations in patients with chronic severe heart failure without untoward effects at an infusion rate of 8 μg/kg/min.

Efficiencies of variety of two-sample tests are examomed for a broad class of distributions. Two new test statistics are introduced. and their potential use as part of an adaptive procedure discussed. Recommendations are made as to the utility of the various test statistics when the underlying distribution is unknown.

Crossover designs are widely accepted for use in clinical studies since each subject serves as his own control. Several incomplete Latin square designs are proposed and investigated for a bioavailability study that compares the individual components of a triple ingredient product with the coadministered components in all possible combinations to detect potential drug-drug interactions. The power of each design and the relative efficiency are evaluated. Finally, recommendations are made for this study based on a number of theoretical and practical criteria.

Previously, the connection between bioequivalence tests and their associated confidence intervals was not well understood. In fact, the α-level two one-sided tests approach was thought to be associated with a 1−2α confidence interval. In this paper, we build up connections which allow us to construct 1−α confidence intervals from α-level tests and vice versa. When applied to the α-level two one-sided tests, the resultant 1−-α confidence intervals are properly contained in Westlakes 1−-α symmetric interval (1976). Our approach is readily generalized to different settings, including the nonparametric setting, the ratio parameter setting, and the repeated confidence interval setting.