Statistical Considerations on Subgroup Analysis in Clinical Trials

Updates on the Global Prevalence and Etiology of Constrictive Pericarditis: A Systematic Review

Constrictive pericarditis is a rare disease with poorly understood epidemiology. A systematic literature search was adopted to assess the region- and period-specific traits of constrictive pericarditis through Pubmed, EMBASE, and Scopus. Case reports and studies including less than 20 patients were excluded. The risk of bias was assessed through the Study Quality Assessment Tools developed by the National Heart Lung Blood Institute by 4 reviewers. Patient demographics, disease etiology, and mortality were the primary assessed outcomes. One hundred thirty studies with 11,325 patients have been included in this systematic review and meta-analysis. The age at diagnosis of constrictive pericarditis has markedly increased after 1990. Patients from Africa and Asia are considerably younger compared with those from Europe and North America. Moreover, there are differences in etiology, as tuberculosis remains the dominant cause of constrictive pericarditis in Africa and Asia but has been surpassed by history of previous chest surgery in North America and Europe. The human immunodeficiency virus affects 29.1% of patients from Africa diagnosed with constrictive pericarditis, a feature that is not observed on any other continent. The early mortality rate after hospitalization has improved. The variances of age at diagnosis and etiology of constrictive pericarditis should be considered by the clinician during the work-up of cardiac and pericardial diseases. An underlying human immunodeficiency virus infection complicates a significant portion of constrictive pericarditis cases in Africa. Early mortality has improved across the world but remains high.

Group-by-Treatment Interaction Effects in Comparative Bioavailability Studies

Comparative bioavailability studies often involve multiple groups of subjects for a variety of reasons, such as clinical capacity limitations. This raises questions about the validity of pooling data from these groups in the statistical analysis and whether a group-by-treatment interaction should be evaluated. We investigated the presence or absence of group-by-treatment interactions through both simulation techniques and a meta-study of well-controlled trials. Our findings reveal that the test falsely detects an interaction when no true group-by-treatment interaction exists. Conversely, when a true group-by-treatment interaction does exist, it often goes undetected. In our meta-study, the detected group-by-treatment interactions were observed at approximately the level of the test and, thus, can be considered false positives. Testing for a group-by-treatment interaction is both misleading and uninformative. It often falsely identifies an interaction when none exists and fails to detect a real one. This occurs because the test is performed between subjects in crossover designs, and studies are powered to compare treatments within subjects. This work demonstrates a lack of utility for including a group-by-treatment interaction in the model when assessing single-site comparative bioavailability studies, and the clinical trial study structure is divided into groups.

Inference on tree-structured subgroups with subgroup size and subgroup effect relationship in clinical trials

When multiple candidate subgroups are considered in clinical trials, we often need to make statistical inference on the subgroups simultaneously. Classical multiple testing procedures might not lead to an interpretable and efficient inference on the subgroups as they often fail to take subgroup size and subgroup effect relationship into account. In this paper, built on the selective traversed accumulation rules (STAR), we propose a data-adaptive and interactive multiple testing procedure for subgroups which can take subgroup size and subgroup effect relationship into account under prespecified tree structure. The proposed method is easy-to-implement and can lead to a more interpretable and efficient inference on prespecified tree-structured subgroups. Possible accommodations to post hoc identified tree-structure subgroups are also discussed in the paper. We demonstrate the merit of our proposed method by re-analyzing the panitumumab trial with the proposed method.

Intravenous tirofiban following successful reperfusion in intracranial large artery atherosclerotic stroke: A secondary analysis of a randomized clinical trial

OBJECTIVE

This study aimed to investigate whether treatment with adjunct intravenous tirofiban is associated with improved outcomes following successful reperfusion in patients with intracranial atherosclerotic stroke.

METHODS

Patients with intracranial large artery atherosclerotic (LAA) stroke and an expanded Treatment in Cerebral Ischemia angiographic score of 2b50 to 3 from the Effect of Intravenous Tirofiban versus Placebo Before Endovascular Thrombectomy on Functional Outcomes in Large Vessel Occlusion Stroke (RESCUE BT) trial were included. The primary outcome was the difference in proportion of independent functional outcome (modified Rankin score of 0-2 at 90 days). Safety outcomes included the rates of symptomatic intracranial hemorrhage (sICH) and 90-day mortality.

RESULTS

Among the 382 patients with intracranial LAA stroke and successful reperfusion, 175 patients (45.8%) were treated with intravenous tirofiban and 207 (54.2%) with placebo. The proportion of patients with independent functional outcome at 90 days was 54.3% (95 out of 175) with tirofiban and 44.0% (91 out of 207) with placebo (adjusted odds ratio [aOR], 1.58; 95% CI, 1.02-2.44; p = 0.04). Intravenous tirofiban was not significantly associated with an increased risk of sICH (12/175 [6.9%] vs. 11/207 [5.3%]; aOR, 1.41; 95% CI, 0.59-3.34; p = 0.44) or 90-day mortality (21/175 [12.0%] vs. 34/207 [16.4%]; aOR, 0.71; 95% CI, 0.38-1.31; p = 0.27).

INTERPRETATION

Among patients with acute intracranial LAA stroke and successful reperfusion following endovascular thrombectomy, adjunct intravenous tirofiban was associated with a higher rate of independent functional outcome, without higher rates of sICH or mortality. Confirmatory randomized trials in these patients are desirable.

Methylprednisolone in acute traumatic spinal cord injury: case-matched outcomes from the NASCIS2 and Sygen historical spinal cord injury studies with contemporary statistical analysis

OBJECTIVE

Methylprednisolone (MP) to treat acute traumatic spinal cord injury (ATSCI) remains controversial since the release of the second National Acute Spinal Cord Injury Study (NASCIS2) in 1990. As two historical studies, NASCIS2 and Sygen in ATSCI, used identical MP dosages, it was possible to construct a new case-level pooled ATSCI data set satisfying contemporary criteria and able to clarify the effect of MP.

METHODS

The new pooled data set was first modernized by excluding patients with injury levels caudal to T10, lower-extremity American Spinal Injury Association (ASIA) motor scores (LEMSs) ≥ 46, Glasgow Coma Scale scores ≤ 11, and age < 15 or > 75 years, and then standardized to the ASIA grading and scoring format. A new updated NASCIS2 data set from this pooled data set contained 31.6% fewer patients than the 1990 NASCIS2 data set.

RESULTS

In the new pooled data set, recovery of LEMSs from baseline to 26 weeks, the primary outcome variable, was separated statistically into five different injury severity cohorts (p < 0.0001). The severity cohorts contained groups with severe floor (62.9%) and ceiling (10.7%) effects, which do not contribute to drug effects. The new NASCIS2 data set duplicated the p value for MP versus placebo in the sub-subgroup analysis of MP initiated ≤ 8 hours (the subgroup) and recovery of motor function on only the right side of the body (a further subgroup within the ≤ 8-hour subgroup), presented as the positive MP effect in the original NASCIS2 reporting. However, current statistical interpretation considers results seen only in post hoc sub-subgroups, without multi-test corrections, to be random effects without clinical significance. The combined case-level pooled data set from the NASCIS2 and Sygen studies increased the MP group from 106 to 431 patients, creating a new MP combined group. This new data set served as a surrogate for a contemporary MP study and found that administration of MP did not enhance ASIA motor score improvement in the lower extremities at 26 weeks. Secondary analysis of descending ASIA motor and sensory cervical neurological levels in cervical ATSCI patients at 26 weeks also found no MP drug effect.

CONCLUSIONS

Analysis of both the new updated NASCIS2 data set and the new case-matched pooled data set from two historical ATSCI studies revealed that administration of MP after spinal cord injury did not demonstrate any enhancement in neurological recovery at 26 weeks. The results of this analysis warrant review by clinical guideline groups.

Bayesian Statistics for Medical Devices: Progress Since 2010

The use of Bayesian statistics to support regulatory evaluation of medical devices began in the late 1990s. We review the literature, focusing on recent developments of Bayesian methods, including hierarchical modeling of studies and subgroups, borrowing strength from prior data, effective sample size, Bayesian adaptive designs, pediatric extrapolation, benefit-risk decision analysis, use of real-world evidence, and diagnostic device evaluation. We illustrate how these developments were utilized in recent medical device evaluations. In Supplementary Material, we provide a list of medical devices for which Bayesian statistics were used to support approval by the US Food and Drug Administration (FDA), including those since 2010, the year the FDA published their guidance on Bayesian statistics for medical devices. We conclude with a discussion of current and future challenges and opportunities for Bayesian statistics, including artificial intelligence/machine learning (AI/ML) Bayesian modeling, uncertainty quantification, Bayesian approaches using propensity scores, and computational challenges for high dimensional data and models.

Effectiveness of adjunct telephone-based postnatal care on maternal and infant illness in the Greater Accra Region, Ghana: a randomized controlled trial

Introduction

Globally, postnatal care (PNC) is fraught with challenges. Despite high PNC coverages in Ghana’s Greater Accra Region (GAR), maternal and newborn health outcomes are of great concern. In 2017, neonatal and post-neonatal mortality rates in GAR were 19 and 13 per 1000 live births respectively despite PNC coverages of 93% for at least one PNC and 87.5% for PNC within 48 hours post-delivery. Telephone follow-up has been used to improve health outcomes in some settings, however, its usefulness in improving maternal and infant health during the postnatal period is not well known in Ghana. We assessed effectiveness of telephone-based PNC on infant and maternal illness in selected hospitals in GAR.

Methods

An open-label, assessor-blinded, parallel-group, two-arm superiority randomized controlled trial with 1:1 allocation ratio was conducted from September 2020 to March 2021. Mother-baby pairs in intervention arm, in addition to usual PNC, received midwife-led telephone counselling within 48 hours post-discharge plus telephone access to midwife during postnatal period. In control arm, only usual PNC was provided. Descriptive and inferential data analyses were conducted to generate frequencies, relative frequencies, risk ratios and 95% confidence intervals. Primary analysis was by intention-to-treat (ITT), complemented by per-protocol (PP) analysis.

Results

Of 608 mother-baby pairs assessed for eligibility, 400 (65.8%) were enrolled. During 3 months follow-up, proportion of infants who fell ill was 62.5% in intervention arm and 77.5% in control arm (p = 0.001). Maternal illness occurred in 27.5% of intervention and 38.5% of control participants (p = 0.02). Risk of infant illness was 20% less in intervention than control arm in both ITT analysis [RR = 0.8 (95%CI = 0.71–0.92] and PP analysis [RR = 0.8 (95%CI = 0.67–0.89)]. Compared to controls, risk of maternal illness in intervention arm was 30% lower in both ITT [RR = 0.7 (95%CI = 0.54–95.00)] and PP analysis [RR = 0.7 (95%CI = 0.51–0.94)].

Conclusion

Telephone-based PNC significantly reduced risk of maternal and infant illness within first 3 months after delivery. This intervention merits consideration as a tool for adoption and scale up to improve infant and maternal health.

Trial registration

This trial was retrospectively registered with the International Standard Randomized Controlled Trial Number (ISRCTN) Registry with number ISRCTN46905855 on 09/04/2021.

Outcomes of Participants With Diabetes in the ISCHEMIA Trials

BACKGROUND

Among patients with diabetes and chronic coronary disease, it is unclear if invasive management improves outcomes when added to medical therapy.

METHODS

The ISCHEMIA (International Study of Comparative Health Effectiveness with Medical and Invasive Approaches) trials (ie, ISCHEMIA and ISCHEMIA-Chronic Kidney Disease) randomized chronic coronary disease patients to an invasive (medical therapy + angiography and revascularization if feasible) or a conservative approach (medical therapy alone with revascularization if medical therapy failed). Cohorts were combined after no trial-specific effects were observed. Diabetes was defined by history, hemoglobin A1c ≥6.5%, or use of glucose-lowering medication. The primary outcome was all-cause death or myocardial infarction (MI). Heterogeneity of effect of invasive management on death or MI was evaluated using a Bayesian approach to protect against random high or low estimates of treatment effect for patients with versus without diabetes and for diabetes subgroups of clinical (female sex and insulin use) and anatomic features (coronary artery disease severity or left ventricular function).

RESULTS

Of 5900 participants with complete baseline data, the median age was 64 years (interquartile range, 57-70), 24% were female, and the median estimated glomerular filtration was 80 mL·min-1·1.73-2 (interquartile range, 64-95). Among the 2553 (43%) of participants with diabetes, the median percent hemoglobin A1c was 7% (interquartile range, 7-8), and 30% were insulin-treated. Participants with diabetes had a 49% increased hazard of death or MI (hazard ratio, 1.49 [95% CI, 1.31-1.70]; P<0.001). At median 3.1-year follow-up the adjusted event-free survival was 0.54 (95% bootstrapped CI, 0.48-0.60) and 0.66 (95% bootstrapped CI, 0.61-0.71) for patients with diabetes versus without diabetes, respectively, with a 12% (95% bootstrapped CI, 4%-20%) absolute decrease in event-free survival among participants with diabetes. Female and male patients with insulin-treated diabetes had an adjusted event-free survival of 0.52 (95% bootstrapped CI, 0.42-0.56) and 0.49 (95% bootstrapped CI, 0.42-0.56), respectively. There was no difference in death or MI between strategies for patients with diabetes versus without diabetes, or for clinical (female sex or insulin use) or anatomic features (coronary artery disease severity or left ventricular function) of patients with diabetes.

CONCLUSIONS

Despite higher risk for death or MI, chronic coronary disease patients with diabetes did not derive incremental benefit from routine invasive management compared with initial medical therapy alone. Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT01471522.

Data-Driven Subgroup Identification in Confirmatory Clinical Trials

Data-driven subgroup analysis plays an important role in clinical trials. This paper focuses on practical considerations in post-hoc subgroup investigations in the context of confirmatory clinical trials. The analysis is aimed at assessing the heterogeneity of treatment effects across the trial population and identifying patient subgroups with enhanced treatment benefit. The subgroups are defined using baseline patient characteristics, including demographic and clinical factors. Much progress has been made in the development of reliable statistical methods for subgroup investigation, including methods based on global models and recursive partitioning. The paper provides a review of principled approaches to data-driven subgroup identification and illustrates subgroup analysis strategies using a family of recursive partitioning methods known as the SIDES (subgroup identification based on differential effect search) methods. These methods are applied to a Phase III trial in patients with metastatic colorectal cancer. The paper discusses key considerations in subgroup exploration, including the role of covariate adjustment, subgroup analysis at early decision points and interpretation of subgroup search results in trials with a positive overall effect.

Effects of Wearable Transcutaneous Electrical Nerve Stimulation on Fibromyalgia: A Randomized Controlled Trial

Purpose

Fibromyalgia is a chronic condition characterized by widespread pain and interference with daily activities. The aim of this study is to assess the benefit of transcutaneous electrical nerve stimulation (TENS) for persons diagnosed with fibromyalgia.

Patients and Methods

Adults meeting diagnostic criteria for fibromyalgia were randomized in a double-blind trial to receive either an active (n=62) or sham (n=57) wearable TENS device for 3-months. Subjects were classified as having lower or higher pain sensitivity by Quantitative Sensory Testing (QST). Patient Global Impression of Change (PGIC, primary outcome) and secondary efficacy measures including Fibromyalgia Impact Questionnaire (FIQR), Brief Pain Inventory (BPI) and painDETECT questionnaire (PDQ) were assessed at baseline, 6-weeks and 3-months. Treatment effects were determined by a mixed model for repeated measures (MMRM) analysis of the intention-to-treat (ITT) population (N=119). A pre-specified subgroup analysis of pain sensitivity was conducted using an interaction term in the model.

Results

No differences were found between active and sham treatment on PGIC scores at 3-months (0.34, 95% CI [−0.37, 1.04], p=0.351) in the ITT population. However, in subjects with higher pain sensitivity (n=60), PGIC was significantly greater for active treatment compared to sham (1.19, 95% CI [0.24, 2.13], p=0.014). FIQR total score (−7.47, 95% CI [−12.46, −2.48], p=0.003), FIQR pain item (−0.62, 95% CI [−1.17, −0.06], p=0.029), BPI Interference (−0.70, 95% CI [−1.30, −0.11], p=0.021) and PDQ (−1.69, 95% CI [−3.20, −0.18], p=0.028) exhibited significant improvements for active treatment compared to sham in the ITT population. Analgesics use was stable and comparable in both groups.

Conclusion

This study demonstrated modest treatment effects of reduced disease impact, pain and functional impairment from wearable TENS in individuals with fibromyalgia. Subjects with higher pain sensitivity exhibited larger treatment effects than those with lower pain sensitivity. Wearable TENS may be a safe treatment option for people with fibromyalgia.

Clinicaltrials.gov Registration

NCT03714425.

Subgroup discovery in non-inferiority trials

Approaches and guidelines for performing subgroup analysis to assess heterogeneity of treatment effect in clinical trials have been the topic of numerous papers in the statistical and clinical literature, but have been discussed predominantly in the context of conventional superiority trials. Concerns about treatment heterogeneity are the same if not greater in non-inferiority (NI) trials, especially since overall similarity between two treatment arms in a successful NI trial could be due to the existence of qualitative interactions that are more likely when comparing two active therapies. Even in unsuccessful NI trials, subgroup analyses can yield important insights about the potential reasons for failure to demonstrate non-inferiority of the experimental therapy. Recent NI trials have performed a priori subgroup analyses using standard statistical tests for interaction, but there is increasing interest in more flexible machine learning approaches for post-hoc subgroup discovery. The performance and practical application of such methods in NI trials have not been systematically explored, however. We considered the Virtual Twin method for the NI setting, an algorithm for subgroup identification that combines random forest with classification and regression trees, and conducted extensive simulation studies to examine its performance under different NI trial conditions and to devise decision rules for selecting the final subgroups. We illustrate the utility of the method with data from a NI trial that was conducted to compare two acupuncture treatments for chronic musculoskeletal pain.

Assessing efficacy in important subgroups in confirmatory trials: An example using Bayesian dynamic borrowing

Assessment of efficacy in important subgroups - such as those defined by sex, age, race and region - in confirmatory trials is typically performed using separate analysis of the specific subgroup. This ignores relevant information from the complementary subgroup. Bayesian dynamic borrowing uses an informative prior based on analysis of the complementary subgroup and a weak prior distribution centred on a mean of zero to construct a robust mixture prior. This combination of priors allows for dynamic borrowing of prior information; the analysis learns how much of the complementary subgroup prior information to borrow based on the consistency between the subgroup of interest and the complementary subgroup. A tipping point analysis can be carried out to identify how much prior weight needs to be placed on the complementary subgroup component of the robust mixture prior to establish efficacy in the subgroup of interest. An attractive feature of the tipping point analysis is that it enables the evidence from the source subgroup, the evidence from the target subgroup, and the combined evidence to be displayed alongside each other. This method is illustrated with an example trial in severe asthma where efficacy in the adolescent subgroup was assessed using a mixture prior combining an informative prior from the adult data in the same trial with a non-informative prior.

The Relative Effects of Prazosin on Individual PTSD Symptoms: Evidence for Pathophysiologically-Related Clustering

Background

The α₁-adrenoreceptor antagonist prazosin has in many but not all studies been found to be effective for PTSD associated nightmares, hyperarousal symptoms, and total symptom severity. The particular efficacy of prazosin for nightmares and hyperarousal symptoms suggests there may be a subset of PTSD symptoms that are more tightly associated with an α₁-adrenoreceptor mediated noradrenergic mechanism, but cross traditional diagnostic symptom clusters. However, the efficacy of prazosin for individual symptoms other than nightmares and sleep disruption has not previously been examined.

Methods

In a post hoc reanalysis of a previously published, randomized controlled trial of twice daily prazosin for PTSD, we examined the relative effect of prazosin on individual items of the CAPS for DSM-IV, and tested whether prazosin responsiveness predicted the partial correlation of the changes in symptom intensity at the level of individual subjects. Results were not adjusted for multiple comparisons.

Results

Prazosin showed the largest effect for distressing dreams, anhedonia, difficulty falling or staying asleep, difficulty concentrating, and hypervigilance. These items were also (a) of higher baseline severity in the underlying population, and (b) more related in how they fluctuated at the level of individual subjects. Covariance analysis did not support a clear cutoff between highly prazosin responsive items and those showing a smaller, not statistically significant response.

Conclusions

In this data set, twice daily prazosin substantially reduced not only nightmares and sleep disruption, but the majority of hyperarousal symptoms, with some evidence of efficacy for avoidance symptoms. The relationship of baseline symptom distribution to which symptoms showed significant response to prazosin reinforces the possibility that differences in a clinical trial's participant populations may significantly influence trial outcome. The pattern of symptom endorsement at the level of individual subjects was consistent with prazosin-responsive items sharing a common pathophysiologic mechanism.

Designing and analyzing clinical trials for personalized medicine via Bayesian models

Patients with different characteristics (e.g., biomarkers, risk factors) may have different responses to the same medicine. Personalized medicine clinical studies that are designed to identify patient subgroup treatment efficacies can benefit patients and save medical resources. However, subgroup treatment effect identification complicates the study design in consideration of desired operating characteristics. We investigate three Bayesian adaptive models for subgroup treatment effect identification: pairwise independent, hierarchical, and cluster hierarchical achieved via Dirichlet Process (DP). The impact of interim analysis and longitudinal data modeling on the personalized medicine study design is also explored. Interim analysis is considered since they can accelerate personalized medicine studies in cases where early stopping rules for success or futility are met. We apply integrated two-component prediction method (ITP) for longitudinal data simulation, and simple linear regression for longitudinal data imputation to optimize the study design. The designs' performance in terms of power for the subgroup treatment effects and overall treatment effect, sample size, and study duration are investigated via simulation. We found the hierarchical model is an optimal approach to identifying subgroup treatment effects, and the cluster hierarchical model is an excellent alternative approach in cases where sufficient information is not available for specifying the priors. The interim analysis introduction to the study design lead to the trade-off between power and expected sample size via the adjustment of the early stopping criteria. The introduction of the longitudinal modeling slightly improves the power. These findings can be applied to future personalized medicine studies with discrete or time-to-event endpoints.

Bayesian Decision-Theoretic Methods for Survival Data using Stochastic Optimization

We introduce a principled method for Bayesian subgroup analysis. The approach is based on casting subgroup analysis as a Bayesian decision problem. The two main innovations are: (1) the explicit consideration of a "subgroup report," comprising multiple subpopulations; and (2) adapting an inhomogeneous Markov chain Monte Carlo simulation scheme to implement stochastic optimization. The latter makes the search for "subgroup reports" practically feasible.

Aerosolized Beractant in neonatal respiratory distress syndrome: A randomized fixed-dose parallel-arm phase II trial

PURPOSE

There is increasing research into novel techniques of administering surfactant to preterm infants (PTIs) with respiratory distress syndrome (RDS) receiving non-invasive respiratory support (NIRS). Although aerosolized surfactant (AS) is promising in PTIs receiving NIRS, the optimal surfactant dose and formulation, drug-device combination and patient profile is not known. The objective of this randomized clinical trial was to investigate the feasibility, safety, efficacy and impact of four dosing schedules of AS using two nebulizers in PTIs with RDS stratified by gestational age (GA).

METHODS

PTIs with RDS receiving pre-defined NIRS for ≤8 h were assigned to 4 A S dosing schedules and 2 nebulizers within three GA strata (I = 24^0/7-28^6/7, II = 29^0/7-32^6/7, III = 33^0/7-36^6/7 weeks). There was no contemporaneous control group; at the recommendation of the Data Monitoring Committee, data was collected retrospectively for control infants.

RESULTS

Of 149 subjects that received AS, the median age at initiation of the 1st dose and duration was 5.5 and 2.4 h respectively. There were 29 infants in stratum I, and 60 each in strata II and III. Of infants <32 weeks GA, 94% received caffeine prior to AS. Fifteen infants (10%) required intubation within 72 h; the rates were not significantly different between GA strata, dosing schedules and nebulizers for infants who received aerosolized surfactant. Compared to retrospective controls, infants who received AS were less likely to need intubation within 72 h in both the intention-to-treat (32% vs. 11%) and the per-protocol (22% vs. 10%) analyses (p < 0.05) with GA stratum specific differences. AS was well tolerated by infants and clinical caregivers. Commonest adverse events included surfactant reflux from nose and mouth (18%), desaturations (11%), and increased secretions (7%).

CONCLUSIONS

We have demonstrated the feasibility, absence of serious adverse events and short-term efficacy of four dosing schedules of AS in the largest Phase II clinical trial of PTIs 24-36 weeks' GA with RDS receiving NIRS (ClinicalTrials.gov NCT02294630). The commonest adverse events noted were surfactant reflux and desaturations; no serious adverse effects were observed. Infants who received AS were less likely to receive intubation within 72 h compared to historical controls. AS is a promising new therapy for PTIs with RDS.

Subgroup-adaptive randomization for subgroup confirmation in clinical trials

A well-known issue when testing for treatment-by-subgroup interaction is its low power, as clinical trials are generally powered for establishing efficacy claims for the overall population, and they are usually not adequately powered for detecting interaction (Alosh, Huque, & Koch [2015] Journal of Biopharmaceutical Statistics, 25, 1161-1178). Hence, it is necessary to develop an adaptive design to improve the efficiency of detecting heterogeneous treatment effects within subgroups. Considering Neyman allocation can maximize the power of usual Z-test (see p. 194 of the book edited by Rosenberger and Lachin), we propose a subgroup-adaptive randomization procedure aiming to achieve Neyman allocation in both predefined subgroups and overall study population in this paper. To verify whether the proposed randomization procedure works as intended, relevant theoretical results are derived and displayed . Numerical studies show that the proposed randomization procedure has obvious advantages in power of tests compared with complete randomization and Pocock and Simon's minimization method.

Flexible Bayesian subgroup analysis in early and confirmatory trials

Subgroup analysis is one of the most important issues in clinical trials. In confirmatory trials, it is critical to investigate consistency of the treatment effect across subgroups, which could potentially result in incorrect scientific conclusion or regulatory decision. There are many challenges and methodological complications of interpreting subgroup results beyond the regulatory setting. For the early phase or proof of concept trials, particularly in basket trials, it is also important to have reliable estimation of subgroup treatment effect in order to guide the next phase go/no-go decision making when large biases can be introduced due to small sample size or random variability. In this paper, we review several recent methods that have been proposed for subgroup analysis in the Bayesian framework to correct for bias. We present simulation results from applying various novel Bayesian hierarchical models for subgroup analysis to a phase II basket trial. For different scenarios considered, we compare the average total sample size, and frequentist-like operating characteristics of power and familywise type I error rate. We compare the precision of the model estimates of the treatment effect by assessing average relative bias and the width of the 95% credible interval for the bias. We also demonstrate flexible Bayesian hierarchical models in a case study of a phase III oncology trial for subgroup treatment effect estimation to help with regulatory decision making. Finally, we conclude our findings in the discussion section and give recommendations on how these methods could be implemented in confirmatory and early phase clinical trials.

Homogeneity assessment for pivotal medical device clinical studies

A question that routinely arises in medical device clinical studies is the homogeneity across demographic subgroups, geographical regions, or investigational sites of the enrolled patients in terms of treatment effects or outcome variables. The main objective of this paper is to discuss statistical concepts and methods for the assessment of such homogeneity and to provide the practitioner a statistical framework and points to consider in conducting homogeneity assessment. Demographic subgroups, geographical regions, and investigational sites are discussed separately as each has its unique issues. Specific considerations are also given to randomized controlled trials, non-randomized comparative studies, and single-arm studies. We point out that judicious use of statistical methods, in conjunction with sound clinical judgment, is essential in handling the issue of homogeneity of treatment effect in medical device clinical studies.

Semi-parametric Bayesian regression for subgroup analysis in clinical trials

Determining whether there are differential treatment effects in subgroups of trial participants remains an important topic in clinical trials as precision medicine becomes ever more relevant. Any assessment of differential treatment effect is predicated on being able to estimate the treatment response accurately while satisfying constraints of balancing the risk of overlooking an important subgroup with the potential to make a decision based on a false discovery. While regression models, such as marginal interaction model, have been widely used to improve accuracy of subgroup parameter estimates by leveraging the relationship between treatment and covariate, there is still a possibility that it can lead to excessively conservative or anti-conservative results. Conceivably, this can be due to the use of the normal distribution as a default prior, which forces outlying subjects to have their means over-shrunk towards the population mean, and the data from such subjects may be excessively influential in estimation of both the overall mean response and the mean response for each subgroup, or a model mis-specification. To address this issue, we investigate the use of nonparametric Bayes, particularly Dirichlet process priors, to create semi-parametric models. These models represent uncertainty in the prior distribution for the overall response while accommodating heterogeneity among individual subgroups. They also account for the effect and variation due to the unaccounted terms. As a result, the models do not force estimates to excessively shrink but still retain the attractiveness of improved precision given by the narrower credible intervals. This is illustrated in extensive simulations investigating bias, mean squared error, coverage probability and credible interval widths. We applied the method on a simulated data based closely on the results of a cystic fibrosis Phase 2 trial.

Sex and Gender Differences Research Design for Basic, Clinical, and Population Studies: Essentials for Investigators

A sex- and gender-informed perspective increases rigor, promotes discovery, and expands the relevance of biomedical research. In the current era of accountability to present data for males and females, thoughtful and deliberate methodology can improve study design and inference in sex and gender differences research. We address issues of motivation, subject selection, sample size, data collection, analysis, and interpretation, considering implications for basic, clinical, and population research. In particular, we focus on methods to test sex/gender differences as effect modification or interaction, and discuss why some inferences from sex-stratified data should be viewed with caution. Without careful methodology, the pursuit of sex difference research, despite a mandate from funding agencies, will result in a literature of contradiction. However, given the historic lack of attention to sex differences, the absence of evidence for sex differences is not necessarily evidence of the absence of sex differences. Thoughtfully conceived and conducted sex and gender differences research is needed to drive scientific and therapeutic discovery for all sexes and genders.

A Bayesian Approach to Intervention-Based Clustering

An important task for intelligent healthcare systems is to predict the effect of a new intervention on individuals. This is especially true for medical treatments. For example, consider patients who do not respond well to a new drug or have adversary reactions. Predicting the likelihood of positive or negative response before trying the drug on the patient can potentially save his or her life. We are therefore interested in identifying distinctive subpopulations that respond differently to a given intervention. For this purpose, we have developed a novel technique, Intervention-based Clustering, based on a Bayesian mixture model. Compared to the baseline techniques, the novelty of our approach lies in its ability to model complex decision boundaries by using soft clustering, thus predicting the effect for individuals more accurately. It can also incorporate prior knowledge, making the method useful even for smaller datasets. We demonstrate how our method works by applying it to both simulated and real data. Results of our evaluation show that our model has strong predictive power and is capable of producing high-quality clusters compared to the baseline methods.

Sources of Safety Data and Statistical Strategies for Design and Analysis: Clinical Trials

BACKGROUND

There has been an increased emphasis on the proactive and comprehensive evaluation of safety endpoints to ensure patient well-being throughout the medical product life cycle. In fact, depending on the severity of the underlying disease, it is important to plan for a comprehensive safety evaluation at the start of any development program. Statisticians should be intimately involved in this process and contribute their expertise to study design, safety data collection, analysis, reporting (including data visualization), and interpretation.

METHODS

In this manuscript, we review the challenges associated with the analysis of safety endpoints and describe the safety data that are available to influence the design and analysis of premarket clinical trials.

RESULTS

We share our recommendations for the statistical and graphical methodologies necessary to appropriately analyze, report, and interpret safety outcomes, and we discuss the advantages and disadvantages of safety data obtained from clinical trials compared to other sources.

CONCLUSIONS

Clinical trials are an important source of safety data that contribute to the totality of safety information available to generate evidence for regulators, sponsors, payers, physicians, and patients. This work is a result of the efforts of the American Statistical Association Biopharmaceutical Section Safety Working Group.

Personalized Medicine Enrichment Design for DHA Supplementation Clinical Trial

Personalized medicine aims to match patient subpopulation to the most beneficial treatment. The purpose of this study is to design a prospective clinical trial in which we hope to achieve the highest level of confirmation in identifying and making treatment recommendations for subgroups, when the risk levels in the control arm can be ordered. This study was motivated by our goal to identify subgroups in a DHA (docosahexaenoic acid) supplementation trial to reduce preterm birth (gestational age<37 weeks) rate. We performed a meta-analysis to obtain informative prior distributions and simulated operating characteristics to ensure that overall Type I error rate was close to 0.05 in designs with three different models: independent, hierarchical, and dynamic linear models. We performed simulations and sensitivity analysis to examine the subgroup power of models and compared results to a chi-square test. We performed simulations under two hypotheses: a large overall treatment effect and a small overall treatment effect. Within each hypothesis, we designed three different subgroup effects scenarios where resulting subgroup rates are linear, flat, or nonlinear. When the resulting subgroup rates are linear or flat, dynamic linear model appeared to be the most powerful method to identify the subgroups with a treatment effect. It also outperformed other methods when resulting subgroup rates are nonlinear and the overall treatment effect is big. When the resulting subgroup rates are nonlinear and the overall treatment effect is small, hierarchical model and chi-square test did better. Compared to independent and hierarchical models, dynamic linear model tends to be relatively robust and powerful when the control arm has ordinal risk subgroups.

Tutorial on statistical considerations on subgroup analysis in confirmatory clinical trials

Clinical trials target patients who are expected to benefit from a new treatment under investigation. However, the magnitude of the treatment benefit, if it exists, often depends on the patient baseline characteristics. It is therefore important to investigate the consistency of the treatment effect across subgroups to ensure a proper interpretation of positive study findings in the overall population. Such assessments can provide guidance on how the treatment should be used. However, great care has to be taken when interpreting consistency results. An observed heterogeneity in treatment effect across subgroups can arise because of chance alone, whereas true heterogeneity may be difficult to detect by standard statistical tests because of their low power. This tutorial considers issues related to subgroup analyses and their impact on the interpretation of findings of completed trials that met their main objectives. In addition, we provide guidance on the design and analysis of clinical trials that account for the expected heterogeneity of treatment effects across subgroups by establishing treatment benefit in a pre-defined targeted subgroup and/or the overall population. Copyright © 2016 John Wiley & Sons, Ltd.

Bayesian analysis of heterogeneous treatment effects for patient-centered outcomes research

Evaluation of heterogeneity of treatment effect (HTE) is an essential aspect of personalized medicine and patient-centered outcomes research. Our goal in this article is to promote the use of Bayesian methods for subgroup analysis and to lower the barriers to their implementation by describing the ways in which the companion software beanz can facilitate these types of analyses. To advance this goal, we describe several key Bayesian models for investigating HTE and outline the ways in which they are well-suited to address many of the commonly cited challenges in the study of HTE. Topics highlighted include shrinkage estimation, model choice, sensitivity analysis, and posterior predictive checking. A case study is presented in which we demonstrate the use of the methods discussed.

Subgroup analyses in confirmatory clinical trials: time to be specific about their purposes

Background

It is well recognized that treatment effects may not be homogeneous across the study population. Subgroup analyses constitute a fundamental step in the assessment of evidence from confirmatory (Phase III) clinical trials, where conclusions for the overall study population might not hold. Subgroup analyses can have different and distinct purposes, requiring specific design and analysis solutions. It is relevant to evaluate methodological developments in subgroup analyses against these purposes to guide health care professionals and regulators as well as to identify gaps in current methodology.

Methods

We defined four purposes for subgroup analyses: (1) Investigate the consistency of treatment effects across subgroups of clinical importance, (2) Explore the treatment effect across different subgroups within an overall non-significant trial, (3) Evaluate safety profiles limited to one or a few subgroup(s), (4) Establish efficacy in the targeted subgroup when included in a confirmatory testing strategy of a single trial. We reviewed the methodology in line with this “purpose-based” framework. The review covered papers published between January 2005 and April 2015 and aimed to classify them in none, one or more of the aforementioned purposes.

Results

In total 1857 potentially eligible papers were identified. Forty-eight papers were selected and 20 additional relevant papers were identified from their references, leading to 68 papers in total. Nineteen were dedicated to purpose 1, 16 to purpose 4, one to purpose 2 and none to purpose 3. Seven papers were dedicated to more than one purpose, the 25 remaining could not be classified unambiguously. Purposes of the methods were often not specifically indicated, methods for subgroup analysis for safety purposes were almost absent and a multitude of diverse methods were developed for purpose (1).

Conclusions

It is important that researchers developing methodology for subgroup analysis explicitly clarify the objectives of their methods in terms that can be understood from a patient’s, health care provider’s and/or regulator’s perspective. A clear operational definition for consistency of treatment effects across subgroups is lacking, but is needed to improve the usability of subgroup analyses in this setting. Finally, methods to particularly explore benefit-risk systematically across subgroups need more research.

Electronic supplementary material

The online version of this article (doi:10.1186/s12874-016-0122-6) contains supplementary material, which is available to authorized users.