Treatment Effects In Subgroups Research Articles

Clinical effect modifiers of antibiotic treatment in patients with chronic low back pain and Modic changes - secondary analyses of a randomised, placebo-controlled trial (the AIM study)

BackgroundRandomised trials on antibiotic treatment for patients with chronic low back pain and vertebral endplate changes visible on MRI (Modic changes) have shown mixed results. A possible explanation might be a real treatment effect in subgroups of the study populations. The purpose of the present study was to explore potential clinical effect modifiers of 3-months oral amoxicillin treatment in patients with chronic low back pain and type I or II Modic changes at the level of a previous lumbar disc herniation.MethodsWe performed analyses of effect modifiers on data from AIM, a double-blind parallel-group multicentre trial. One hundred eighty patients with chronic low back pain, previous disc herniation, Modic change type I (n = 118) or type II (n = 62) were randomised to 3-months oral treatment with 750 mg amoxicillin (n = 89) or placebo (n = 91) three times daily. The primary outcome was the Roland-Morris Disability Questionnaire (RMDQ) score (possible values 0–24) at 1-year follow-up in the intention-to-treat population. The predefined minimal clinically important between-group mean difference was 4 RMDQ points (not reached in the primary analysis of AIM). Predefined baseline characteristics were analysed as potential effect modifiers, four primary (type I Modic changes, previous disc surgery, positive pain provocation test, high CRP) and five exploratory (disturbed sleep, constant low back pain, short duration of low back pain, younger age, and male) using ANCOVA with interaction terms.ResultsNone of the four primary potential effect modifiers had strong evidence of modifying the treatment effect. In patients younger than 40 years the difference in mean RMDQ score between the treatment groups was − 4.0 (95%CI, − 6.9 to − 1.2), compared to − 0.5 (95%CI, − 2.3 to 1.3) in patients 40 years or older, both in favour of amoxicillin treatment (exploratory analysis).ConclusionsWe did not find evidence for convincing clinical effect modifiers of antibiotic treatment in patients with chronic low back pain and Modic changes. Our results for younger age in these explorative analyses should not affect clinical treatment decisions without confirmation in future studies.Trial registrationClinicalTrials.gov NCT02323412, First registered 23 December 2014.

Bayesian cluster hierarchical model for subgroup borrowing in the design and analysis of basket trials with binary endpoints.

Master protocol designs are often proposed to improve the efficiency of drug development with multiple subgroups. In the basket trial design, different subgroups can have similar biological pathogenesis pathways. Hence, a target therapy can result in similar responses. A good information sharing strategy between different subgroups can potentially improve the efficiency of evaluating treatment efficacy. In traditional hierarchical models, based on the exchangeability assumption, all subgroups are placed into the same sharing pool for cross subgroup information sharing. However, due to the heterogeneity between subgroups, there can be large differences in drug efficacy. Under such cases, strong borrowing across all subgroups is not suitable and no borrowing can be inefficient, because the treatment effect is analyzed in each subgroup separately. We propose a Bayesian cluster hierarchical model (BCHM) to improve the operating characteristics of estimating the treatment effect in multiple subgroups in basket trials. Bayesian nonparametric method is applied to dynamically calculate the number of clusters by conducting a multiple cluster classification based on subgroup outcomes. A hierarchical model is used to compute the posterior probability of the treatment effect, with the borrowing strength determined by the Bayesian nonparametric clustering and the similarities between subgroups. We apply the BCHM to clinical trials with binary endpoints. For treatment effect estimation, the BCHM yields lower mean squared error values, when compared to the independent analyses. In scenarios with a heterogeneous treatment effect, the BCHM provides lower mean squared error values compared to traditional hierarchical models. In addition, we can construct a loss function to optimize the design parameters. BCHM provides a balanced approach and smart borrowing, which yields better results in assessing the treatment effect in different scenarios compared to other conventional methods.

Anticoagulant therapy for septic coagulopathy and disseminated intravascular coagulation: where do KyberSept and SCARLET leave us?

The use of antithrombin and thrombomodulin to restore impaired anticoagulant pathways in septic coagulopathy has been shown to significantly increase the resolution rate of disseminated intravascular coagulation. In KyberSept and SCARLET, two large, international, randomized controlled trials in patients with sepsis, these anticoagulants have not shown significantly reduced mortality. The aim of this assessment was to compare the heterogeneity in responses to treatment in the two trials according to different patient phenotypes. Results of the KyberSept and SCARLET trials reported in original and post‐hoc publications were analyzed and directly compared for treatment effects in various patient subgroups. In both KyberSept and SCARLET, the septic coagulopathy phenotype that benefited most from endogenous anticoagulant supplementation showed markers of excessive activation of coagulation. Interaction between concomitant thromboprophylactic heparin and the endogenous anticoagulants abrogated the efficacy of both antithrombin and thrombomodulin. In both trials, higher disease severity was associated with better treatment outcome. In conclusion, in two landmark studies of endogenous anticoagulants in patients with sepsis, similar findings of beneficial effects in the coagulopathy phenotype and interactions with heparin comedication and disease severity support the potential roles that thrombomodulin and antithrombin might play in treating septic coagulopathy and disseminated intravascular coagulation. Further prospective validation is warranted. Future trial designs to definitively establish the therapeutic relevance of antithrombin and thrombomodulin in septic coagulopathy should focus on involvement of patients characterized by coagulopathy and disease severity as well as interactions between endogenous anticoagulants and exogenous heparin.

1001P - Effect of response to last platinum-based chemotherapy in patients (pts) with platinum-sensitive, recurrent ovarian carcinoma in the phase III study ARIEL3 of rucaparib maintenance treatment

1001P - Effect of response to last platinum-based chemotherapy in patients (pts) with platinum-sensitive, recurrent ovarian carcinoma in the phase III study ARIEL3 of rucaparib maintenance treatment

Subgroup balancing propensity score.

This paper concerns estimation of subgroup treatment effects with observational data. Existing propensity score methods are mostly developed for estimating overall treatment effect. Although the true propensity scores balance covariates in any subpopulations, the estimated propensity scores may result in severe imbalance in subgroup samples. Indeed, subgroup analysis amplifies a bias-variance tradeoff, whereby increasing complexity of the propensity score model may help to achieve covariate balance within subgroups, but it also increases variance. We propose a new method, the subgroup balancing propensity score, to ensure good subgroup balance as well as to control the variance inflation. For each subgroup, the subgroup balancing propensity score chooses to use either the overall sample or the subgroup (sub)sample to estimate the propensity scores for the units within that subgroup, in order to optimize a criterion accounting for a set of covariate-balancing moment conditions for both the overall sample and the subgroup samples. We develop two versions of subgroup balancing propensity score corresponding to matching and weighting, respectively. We devise a stochastic search algorithm to estimate the subgroup balancing propensity score when the number of subgroups is large. We demonstrate through simulations that the subgroup balancing propensity score improves the performance of propensity score methods in estimating subgroup treatment effects. We apply the subgroup balancing propensity score method to the Italy Survey of Household Income and Wealth (SHIW) to estimate the causal effects of having debit card on household consumption for different income groups.

Bayesian inference for a principal stratum estimand to assess the treatment effect in a subgroup characterized by postrandomization event occurrence.

The treatment effect in subgroups of patients is often of interest in randomized controlled clinical trials, as this may provide useful information on how to treat which patients best. When a specific subgroup is characterized by the absence of certain events that happen postrandomization, a naive analysis on the subset of patients without these events may be misleading. The principal stratification framework allows one to define an appropriate causal estimand in such settings. Statistical inference for the principal stratum estimand hinges on scientifically justified assumptions, which can be included with Bayesian methods through prior distributions. Our motivating example is a large randomized placebo-controlled trial of siponimod in patients with secondary progressive multiple sclerosis. The primary objective of this trial was to demonstrate the efficacy of siponimod relative to placebo in delaying disability progression for the whole study population. However, the treatment effect in the subgroup of patients who would not relapse during the trial is relevant from both a scientific and patient perspective. Assessing this subgroup treatment effect is challenging as there is strong evidence that siponimod reduces relapses. We describe in detail the scientific question of interest, the principal stratum estimand, the corresponding analysis method for binary endpoints, and sensitivity analyses. Although our work is motivated by a randomized clinical trial, the approach has broader appeal and could be adapted for observational studies.

Subgroup identification for precision medicine: A comparative review of 13 methods

AbstractNatural heterogeneity in patient populations can make it very hard to develop treatments that benefit all patients. As a result, an important goal of precision medicine is identification of patient subgroups that respond to treatment at a much higher (or lower) rate than the population average. Despite there being many subgroup identification methods, there is no comprehensive comparative study of their statistical properties. We review 13 methods and use real‐world and simulated data to compare the performance of their publicly available software using seven criteria: (a) bias in selection of subgroup variables, (b) probability of false discovery, (c) probability of identifying correct predictive variables, (d) bias in estimates of subgroup treatment effects, (e) expected subgroup size, (f) expected true treatment effect of subgroups, and (g) subgroup stability. The results show that many methods fare poorly on at least one criterion.This article is categorized under: Technologies > Machine Learning Algorithmic Development > Hierarchies and Trees Algorithmic Development > Statistics Application Areas > Health Care

Practical Considerations for Subgroups Quantification, Selection and Adaptive Enrichment in Confirmatory Trials

Subgroup analyses (e.g., baseline information, biomarkers measurements) are commonly encountered and conducted in confirmatory clinical trials to ensure the risk–benefit consistency and appropriate interpretation of the study results. However, there are natural methodological complications that come with multiple analyses which can result in incorrect scientific or regulatory conclusions for subgroup analysis. Typical issues that may arise include, but are not limited to (a) How to make sure subgroup results are reliable and reproducible? (b) How to quantify subgroup reversal effects which may be due to chance finding, or a lack of power in subgroup and/or treatment interaction tests? (c) How to design efficient trials to establish treatment effect in subgroups or/and overall population with proper Type I error control? These are challenges statisticians and clinical trialist regularly face in the drug development process. In this article, we discuss and present different approaches corresponding to these general issues of subgroup analysis, and their impacts and implications on the interpretation of clinical trials as well as the innovations and opportunities in the era of precision medicine. We also conduct simulations to illustrate the operating characteristics of different methodological tools and designs, along with further practical recommendations and guidance from a statistical perspective.

A Bayesian adaptive marker-stratified design for molecularly targeted agents with customized hierarchical modeling.

It is well known that the treatment effect of a molecularly targeted agent (MTA) may vary dramatically, depending on each patient's biomarker profile. Therefore, for a clinical trial evaluating MTA, it is more reasonable to evaluate its treatment effect within different marker subgroups rather than evaluating the average treatment effect for the overall population. The marker-stratified design (MSD) provides a useful tool to evaluate the subgroup treatment effects of MTAs. Under the Bayesian framework, the beta-binomial model is conventionally used under the MSD to estimate the response rate and test the hypothesis. However, this conventional model ignores the fact that the biomarker used in the MSD is, in general, predictive only for the MTA. The response rates for the standard treatment can be approximately consistent across different subgroups stratified by the biomarker. In this paper, we proposed a Bayesian hierarchical model incorporating this biomarker information into consideration. The proposed model uses a hierarchical prior to borrow strength across different subgroups of patients receiving the standard treatment and, therefore, improve the efficiency of the design. Prior informativeness is determined by solving a "customized" equation reflecting the physician's professional opinion. We developed a Bayesian adaptive design based on the proposed hierarchical model to guide the treatment allocation and test the subgroup treatment effect as well as the predictive marker effect. Simulation studies and a real trial application demonstrate that the proposed design yields desirable operating characteristics and outperforms the existing designs.

Ocrelizumab efficacy in subgroups of patients with relapsing multiple sclerosis

ObjectiveThe efficacy and safety of ocrelizumab, versus interferon (IFN) β-1a, for the treatment of relapsing multiple sclerosis (RMS) from the identically designed OPERA I (NCT01247324) and OPERA II (NCT01412333) phase III studies has been reported; here we present subgroup analyses of efficacy endpoints from the pooled OPERA I and OPERA II populations.MethodsPatients with RMS were randomized to either ocrelizumab 600 mg administered by intravenous infusion every 24 weeks or subcutaneous IFN β-1a 44 µg three times per week throughout the 96-week treatment period. Relapse, disability, and MRI outcomes were analyzed for predefined and post hoc subgroups based on demographic and disease characteristics along with prior treatment using appropriate statistical tests to determine the treatment effect in subgroups and treatment-by-subgroup interactions.ResultsThe significant treatment benefit of ocrelizumab, versus IFN β-1a, observed in the overall OPERA I and OPERA II pooled populations was maintained across most subgroup strata for all endpoints, including annualized relapse rate, disability progression, and MRI outputs.ConclusionsThe treatment effect of ocrelizumab versus IFN β-1a, measured by clinical and MRI outcomes, was maintained across most of the subgroups and strata of interest, and the pattern of treatment benefit across all subgroups was consistent with that from the pooled OPERA studies.

Semi-parametric Bayesian regression for subgroup analysis in clinical trials

ABSTRACTDetermining 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.

Prospective multicenter studies in pediatric hydrocephalus.

Prospective multicenter clinical research studies in pediatric hydrocephalus are relatively rare. They cover a broad spectrum of hydrocephalus topics, including management of intraventricular hemorrhage in premature infants, shunt techniques and equipment, shunt outcomes, endoscopic treatment of hydrocephalus, and prevention and treatment of infection. The research methodologies include randomized trials, cohort studies, and registry-based studies. This review describes prospective multicenter studies in pediatric hydrocephalus since 1990. Many studies have included all forms of hydrocephalus and used device or procedure failure as the primary outcome. Although such studies have yielded useful findings, they might miss important treatment effects in specific subgroups. As multicenter study networks grow, larger patient numbers will allow studies with more focused entry criteria based on known and evolving prognostic factors. In addition, increased use of patient-centered outcomes such as neurodevelopmental assessment and quality of life should be measured and emphasized in study results. Well-planned multicenter clinical studies can significantly affect the care of children with hydrocephalus and will continue to have an important role in improving care for these children and their families.

Broadening the Perspective of Cost-Effectiveness Modeling in Chronic Obstructive Pulmonary Disease: A New Patient-Level Simulation Model Suitable to Evaluate Stratified Medicine

ObjectivesTo develop a health economic model that included a great diversity of patient characteristics and outcomes for chronic obstructive pulmonary disease (COPD), which can be used to inform decisions about stratified medicine in COPD. MethodsThe choice of patient characteristics and outcomes to include in the model was based on 3 literature reviews on multidimensional prognostic COPD indices, COPD phenotypes, and treatment effects in subgroups. A conceptual model was constructed including 14 patient characteristics, 7 intermediate outcomes (lung function, physical activity, exercise capacity, symptoms, disease-specific quality of life, exacerbations, and pneumonias), and 3 final outcomes (mortality, quality-adjusted life-years [QALYs], and costs). Regression equations describing the statistical associations between the patient characteristics and intermediate and final outcomes were estimated using the longitudinal data of 5 large COPD trials (19,378 patients). A patient-level simulation model was developed in which individual patients from the baseline population of the 5 trials are sampled and their outcomes over lifetime are predicted based on the regression equations. ResultsThe base-case analysis (single-arm simulation representing treatment with tiotropium) showed that patients had a mean lung function decline of 43 mL/year, 0.62 exacerbations/year, a worsening of their physical activity and quality of life with 1.48 and 1.10 points/year, a life expectancy of 11.2 years, 7.25 QALYs, and total lifetime costs of £24,891. Results for a selection of treatment scenarios and subgroups were shown to demonstrate the potential of the model. ConclusionsWe developed a unique patient-level simulation model that can be used to evaluate COPD treatment options for a variety of subgroups.

Confidence regions for treatment effects in subgroups in biomarker stratified designs

Subgroup analysis has important applications in the analysis of controlled clinical trials. Sometimes the result of the overall group fails to demonstrate that the new treatment is better than the control therapy, but for a subgroup of patients, the treatment benefit may exist; or sometimes, the new treatment is better for the overall group but not for a subgroup. Hence we are interested in constructing a simultaneous confidence interval for the difference of the treatment effects in a subgroup and the overall group. Subgroups are usually formed on the basis of a predictive biomarker such as age, sex, or some genetic marker. While, for example, age can be detected precisely, it is often only possible to detect the biomarker status with a certain probability. Because patients detected with a positive or negative biomarker may not be truly biomarker positive or negative, responses in the subgroups depend on the treatment therapy as well as on the sensitivity and specificity of the assay used in detecting the biomarkers. In this work, we show how (approximate) simultaneous confidence intervals and confidence ellipsoid for the treatment effects in subgroups can be found for biomarker stratified clinical trials using a normal framework with normally distributed or binary data. We show that these intervals maintain the nominal confidence level via simulations.

Estimating heterogeneous treatment effects for latent subgroups in observational studies.

Individuals may vary in their responses to treatment, and identification of subgroups differentially affected by a treatment is an important issue in medical research. The risk of misleading subgroup analyses has become well known, and some exploratory analyses can be helpful in clarifying how covariates potentially interact with the treatment. Motivated by a real data study of pediatric kidney transplant, we consider a semiparametric Bayesian latent model and examine its utility for an exploratory subgroup effect analysis using secondary data. The proposed method is concerned with a clinical setting where the number of subgroups is much smaller than that of potential predictors and subgroups are only latently associated with observed covariates. The semiparametric model is flexible in capturing the latent structure driven by data rather than dictated by parametric modeling assumptions. Since it is difficult to correctly specify the conditional relationship between the response and a large number of confounders in modeling, we use propensity score matching to improve the model robustness by balancing the covariates distribution. Simulation studies show that the proposed analysis can find the latent subgrouping structure and, with propensity score matching adjustment, yield robust estimates even when the outcome model is misspecified. In the real data analysis, the proposed analysis reports significant subgroup effects on steroid avoidance in kidney transplant patients, whereas standard proportional hazards regression analysis does not.

Subgroups from regression trees with adjustment for prognostic effects and postselection inference.

Identification of subgroups with differential treatment effects in randomized trials is attracting much attention. Many methods use regression tree algorithms. This article addresses 2 important questions arising from the subgroups: how to ensure that treatment effects in subgroups are not confounded with effects of prognostic variables and how to determine the statistical significance of treatment effects in the subgroups. We address the first question by selectively including linear prognostic effects in the subgroups in a regression tree model. The second question is more difficult because it falls within the subject of postselection inference. We use a bootstrap technique to calibrate normal-theory t intervals so that their expected coverage probability, averaged over all the subgroups in a fitted model, approximates the desired confidence level. It can also provide simultaneous confidence intervals for all subgroups. The first solution is implemented in the GUIDE algorithm and is applicable to data with missing covariate values, 2 or more treatment arms, and outcomes subject to right censoring. Bootstrap calibration is applicable to any subgroup identification method; it is not restricted to regression tree models. Two real examples are used for illustration: a diabetes trial where the outcomes are completely observed but some covariate values are missing and a breast cancer trial where the outcome is right censored.

Long Term Outcome of Lenalidomide-Dexamethasone (Rd) Vs Melphalan-Lenalidomide-Prednisone (MPR) Vs Cyclophosphamide-Prednisone-Lenalidomide (CPR) As Induction Followed By Lenalidomide-Prednisone (RP) Vs Lenalidomide (R) As Maintenance in a Community-Based Newly Diagnosed Myeloma Population: Updated Analysis of EMN01 Phase III Study

Introduction : Rd and MPR showed to be effective combinations in elderly newly diagnosed multiple myeloma (NDMM) patients (pts). Cyclophosphamide is a less toxic alkylating alternative agent. EMN01 is the first trial to formally compare these three different Lenalidomide-based combinations. Maintenance with Lenalidomide has been recently approved in patients eligible for autologous stem cell transplant (ASCT). Few data are available about the best combination as maintenance in patients not eligible for ASCT.Methods : 662 pts with NDMM were randomized to receive 9 28-day cycles of Rd (lenalidomide 25 mg/day for 21 days; dexamethasone 40 mg on days 1,8,15 and 22 in pts 65-75 years old and 20 mg in those >75 years), MPR (lenalidomide 10 mg/day for 21 days; melphalan orally 0.18 mg/Kg for 4 days in pts 65-75 years old and 0.13 mg/Kg in >75 years pts; prednisone 1.5 mg/Kg for 4 days) or CPR (lenalidomide 25 mg/day for 21 days; cyclophosphamide orally 50 mg/day for 21 days in pts 65-75 years old and 50 mg every other day in >75 years pts; prednisone 25 mg every other day). After induction, pts were randomized to receive maintenance with lenalidomide alone (R; 10 mg/day for 21 days) or with prednisone (RP; R, 10 mg/day for 21 days and P, 25 mg every other day), until disease progression.Results : Pts characteristics were well balanced in all groups; 217 pts in Rd, 217 in MPR and 220 in CPR arms could be evaluated. After a median follow-up of 63.7 months, median PFS was 23.2 months in MPR, 18.9 months in CPR and 18.6 months in Rd (MPR vs CPR p=0.02; MPR vs Rd p=0.08). Median overall survival (OS) was 79.9 months in MPR, 69.4 months in CPR and 68.1 months in Rd (MPR vs CPR p=0.98; MPR vs Rd p=0.64). The most common grade ≥3 adverse event (AEs) was neutropenia: 64% in MPR, 29% in CPR and 25% in Rd pts (p<0.0001). Grade ≥3 non hematologic AEs were similar among arms. At the end of induction, 402 pts were eligible for maintenance, 198 in the RP and 204 in the R groups. PFS from start of maintenance was 22.2 months in the RP group and 17.6 in the R group, with 20% reduced the risk of death/progression for pts receiving RP maintenance (HR 0.81, p=0.07; Figure 1). A subgroup analysis was performed to determine the consistency of RP vs R treatment effect in different subgroups using interaction terms between treatment and cytogenetic abnormalities, ISS, age, sex, induction treatment and response before maintenance (Figure 1). No difference in OS was observed (HR 1.02, p=0.93) but the OS analysis was limited by the low number of events. Median duration of maintenance was 23.0 months in RP pts and 20.5 months in R pts, 14% and 13% of pts discontinued due to AEs, in RP and R groups, respectively.Conclusion : This phase III trial compared 2 different Lenalidomide-containing induction regimens and 2 different Lenalidomide-containing maintenance regimens in an elderly community-based NDMM population. MPR prolonged PFS by approximately 5 months, yet the higher incidence of hematologic toxicity should be carefully considered. The addition of low-dose prednisone to standard lenalidomide maintenance reduced the risk of death/progression by 20%, with a good safety profile. Updated results will be presented at the meeting. [Display omitted] DisclosuresBringhen:Mundipharma: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Janssen: Honoraria; Celgene: Honoraria; Bristol Myers Squibb: Honoraria; Karyipharm: Membership on an entity's Board of Directors or advisory committees. Offidani:celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees; Janssen: Honoraria, Membership on an entity's Board of Directors or advisory committees. Musto:Celgene: Honoraria; Janssen: Honoraria. Gaidano:Gilead: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Roche: Consultancy, Honoraria; AbbVie: Consultancy, Honoraria. De Sabbata:Celgene: Membership on an entity's Board of Directors or advisory committees. Palumbo:Sanofi: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; Amgen: Consultancy, Honoraria, Research Funding; Binding Site: Research Funding; Celgene: Consultancy, Honoraria, Research Funding; Merck: Consultancy, Honoraria, Research Funding; Genmab A/S: Consultancy, Honoraria, Research Funding; Janssen-Cilag: Consultancy, Honoraria, Research Funding; Bristol-Myers Squibb: Consultancy, Honoraria, Research Funding, Speakers Bureau; Takeda: Consultancy, Employment, Equity Ownership, Honoraria, Research Funding. Hájek:Amgen, Takeda, BMS, Celgene, Novartis, Janssen: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Abbvie: Consultancy, Honoraria; Pharma MAR: Consultancy, Honoraria. Boccadoro:Novartis: Honoraria, Research Funding; Bristol-Myers Squibb: Honoraria, Research Funding; Amgen: Honoraria, Research Funding; AbbVie: Honoraria; Mundipharma: Research Funding; Sanofi: Honoraria, Research Funding; Celgene: Honoraria, Research Funding; Janssen: Honoraria, Research Funding.

Bounding, An Accessible Method for Estimating Principal Causal Effects, Examined and Explained

ABSTRACTEstimating treatment effects for subgroups defined by posttreatment behavior (i.e., estimating causal effects in a principal stratification framework) can be technically challenging and heavily reliant on strong assumptions. We investigate an alternative path: using bounds to identify ranges of possible effects that are consistent with the data. This simple approach relies on fewer assumptions and yet can result in policy-relevant findings. As we show, even moderately predictive covariates can be used to substantially tighten bounds in a straightforward manner. Via simulation, we demonstrate which types of covariates are maximally beneficial. We conclude with an analysis of a multisite experimental study of Early College High Schools. When examining the program's impact on students completing the ninth grade “on-track” for college, we find little impact for ECHS students who would otherwise attend a high-quality high school, but substantial effects for those who would not. This suggests a potential benefit in expanding these programs in areas primarily served by lower quality schools.

Understanding Causal Distributional and Subgroup Effects With the Instrumental Propensity Score.

To address issues with measured and unmeasured confounding in observational studies, we developed a unified approach to using an instrumental variable in more flexible ways to evaluate treatment effects. The approach is based on an instrumental propensity score conditional on baseline variables, which can then be incorporated in matching, regression, subclassification, or weighting along with various parametric, semiparametric, or nonparametric methods for the assessment of treatment effects. Therefore, the application of the instrumental propensity score allows different methods for outcome effect evaluations in addition to standard 2-stage least square models while controlling for unmeasured confounders. Several properties of the instrumental propensity score are discussed. The approach is then illustrated using subclassification along with a semiparametric density ratio model and empirical likelihood. This method allows us to evaluate distributional and subgroup treatment effects in addition to the overall average treatment effect. Simulation studies showed that the method works well. We applied our method to a study of the effects of attending a Catholic school versus a public school and found that attending a Catholic school had significant beneficial effects on subsequent wages among a subgroup of subjects.

Hierarchical Bayes approach for subgroup analysis.

In clinical data analysis, both treatment effect estimation and consistency assessment are important for a better understanding of the drug efficacy for the benefit of subjects in individual subgroups. The linear mixed-effects model has been used for subgroup analysis to describe treatment differences among subgroups with great flexibility. The hierarchical Bayes approach has been applied to linear mixed-effects model to derive the posterior distributions of overall and subgroup treatment effects. In this article, we discuss the prior selection for variance components in hierarchical Bayes, estimation and decision making of the overall treatment effect, as well as consistency assessment of the treatment effects across the subgroups based on the posterior predictive p-value. Decision procedures are suggested using either the posterior probability or the Bayes factor. These decision procedures and their properties are illustrated using a simulated example with normally distributed response and repeated measurements.