Improved precision in the analysis of randomized trials with survival outcomes, without assuming proportional hazards

Covariate adjustment in randomized controlled trials: General concepts and practical considerations

There has been a growing interest in covariate adjustment in the analysis of randomized controlled trials in past years. For instance, the US Food and Drug Administration recently issued guidance that emphasizes the importance of distinguishing between conditional and marginal treatment effects. Although these effects may sometimes coincide in the context of linear models, this is not typically the case in other settings, and this distinction is often overlooked in clinical trial practice. Considering these developments, this article provides a review of when and how to use covariate adjustment to enhance precision in randomized controlled trials. We describe the differences between conditional and marginal estimands and stress the necessity of aligning statistical analysis methods with the chosen estimand. In addition, we highlight the potential misalignment of commonly used methods in estimating marginal treatment effects. We hereby advocate for the use of the standardization approach, as it can improve efficiency by leveraging the information contained in baseline covariates while remaining robust to model misspecification. Finally, we present practical considerations that have arisen in our respective consultations to further clarify the advantages and limitations of covariate adjustment.

More efficient and inclusive time-to-event trials with covariate adjustment: a simulation study

Adjustment for prognostic covariates increases the statistical power of randomized trials. The factors influencing the increase of power are well-known for trials with continuous outcomes. Here, we study which factors influence power and sample size requirements in time-to-event trials. We consider both parametric simulations and simulations derived from the Cancer Genome Atlas (TCGA) cohort of hepatocellular carcinoma (HCC) patients to assess how sample size requirements are reduced with covariate adjustment. Simulations demonstrate that the benefit of covariate adjustment increases with the prognostic performance of the adjustment covariate (C-index) and with the cumulative incidence of the event in the trial. For a covariate that has an intermediate prognostic performance (C-index=0.65), the reduction of sample size varies from 3.1% when cumulative incidence is of 10% to 29.1% when the cumulative incidence is of 90%. Broadening eligibility criteria usually reduces statistical power while our simulations show that it can be maintained with adequate covariate adjustment. In a simulation of adjuvant trials in HCC, we find that the number of patients screened for eligibility can be divided by 2.4 when broadening eligibility criteria. Last, we find that the Cox-Snell [Formula: see text] is a conservative estimation of the reduction in sample size requirements provided by covariate adjustment. Overall, more systematic adjustment for prognostic covariates leads to more efficient and inclusive clinical trials especially when cumulative incidence is large as in metastatic and advanced cancers. Code and results are available at https://github.com/owkin/CovadjustSim .

Estimating the Efficiency Gain of Covariate-Adjusted Analyses in Future Clinical Trials Using External Data

We present a framework for using existing external data to identify and estimate the relative efficiency of a covariate-adjusted estimator compared to an unadjusted estimator in a future randomized trial. Under conditions, these relative efficiencies approximate the ratio of sample sizes needed to achieve a desired power. We develop semiparametrically efficient estimators of the relative efficiencies for several treatment effect estimands of interest with either fully or partially observed outcomes, allowing for the application of flexible statistical learning tools to estimate the nuisance functions. We propose an analytic Wald-type confidence interval and a double bootstrap scheme for statistical inference. We demonstrate the performance of the proposed methods through simulation studies and apply these methods to estimate the efficiency gain of covariate adjustment in Covid-19 therapeutic trials.

Transfusing Convalescent Plasma as Post-Exposure Prophylaxis Against Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Infection: A Double-Blinded, Phase 2 Randomized, Controlled Trial

BACKGROUND

The efficacy of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) convalescent plasma (CCP) for preventing infection in exposed, uninfected individuals is unknown. CCP might prevent infection when administered before symptoms or laboratory evidence of infection.

METHODS

This double-blinded, phase 2 randomized, controlled trial (RCT) compared the efficacy and safety of prophylactic high titer (≥1:320 by Euroimmun ELISA) CCP with standard plasma. Asymptomatic participants aged ≥18 years with close contact exposure to a person with confirmed coronavirus disease 2019 (COVID-19) in the previous 120 hours and negative SARS-CoV-2 test within 24 hours before transfusion were eligible. The primary outcome was new SARS-CoV-2 infection.

RESULTS

In total, 180 participants were enrolled; 87 were assigned to CCP and 93 to control plasma, and 170 transfused at 19 sites across the United States from June 2020 to March 2021. Two were excluded for screening SARS-CoV-2 reverse transcription polymerase chain reaction (RT-PCR) positivity. Of the remaining 168 participants, 12/81 (14.8%) CCP and 13/87 (14.9%) control recipients developed SARS-CoV-2 infection; 6 (7.4%) CCP and 7 (8%) control recipients developed COVID-19 (infection with symptoms). There were no COVID-19-related hospitalizations in CCP and 2 in control recipients. Efficacy by restricted mean infection free time (RMIFT) by 28 days for all SARS-CoV-2 infections (25.3 vs 25.2 days; P = .49) and COVID-19 (26.3 vs 25.9 days; P = .35) was similar for both groups.

CONCLUSIONS

Administration of high-titer CCP as post-exposure prophylaxis, although appearing safe, did not prevent SARS-CoV-2 infection.

CLINICAL TRIALS REGISTRATION

NCT04323800.

Doubly-robust estimator of the difference in restricted mean times lost with competing risks data

In the context of competing risks data, the subdistribution hazard ratio has limited clinical interpretability to measure treatment effects. An alternative is the difference in restricted mean times lost (RMTL), which gives the mean time lost to a specific cause of failure between treatment groups. In non-randomized studies, the average causal effect is conventionally used for decision-making about treatment and public health policies. We show how the difference in RMTL can be estimated by contrasting the integrated cumulative incidence functions from a Fine-Gray model. We also show how the difference in RMTL can be estimated by using inverse probability of treatment weighting and contrasts between weighted non-parametric estimators of the area below the cumulative incidence. We use pseudo-observation approaches to estimate both component models and we integrate them into a doubly-robust estimator. We demonstrate that this estimator is consistent when either component is correctly specified. We conduct simulation studies to assess its finite-sample performance and demonstrate its inherited consistency property from its component models. We also examine the performance of this estimator under varying degrees of covariate overlap and under a model misspecification of nonlinearity. We apply the proposed method to assess biomarker-treatment interaction in subpopulations of the POPLAR and OAK randomized controlled trials of second-line therapy for advanced non-small-cell lung cancer.

Early Outpatient Treatment for Covid-19 with Convalescent Plasma

BACKGROUND

Polyclonal convalescent plasma may be obtained from donors who have recovered from coronavirus disease 2019 (Covid-19). The efficacy of this plasma in preventing serious complications in outpatients with recent-onset Covid-19 is uncertain.

METHODS

In this multicenter, double-blind, randomized, controlled trial, we evaluated the efficacy and safety of Covid-19 convalescent plasma, as compared with control plasma, in symptomatic adults (≥18 years of age) who had tested positive for severe acute respiratory syndrome coronavirus 2, regardless of their risk factors for disease progression or vaccination status. Participants were enrolled within 8 days after symptom onset and received a transfusion within 1 day after randomization. The primary outcome was Covid-19-related hospitalization within 28 days after transfusion.

RESULTS

Participants were enrolled from June 3, 2020, through October 1, 2021. A total of 1225 participants underwent randomization, and 1181 received a transfusion. In the prespecified modified intention-to-treat analysis that included only participants who received a transfusion, the primary outcome occurred in 17 of 592 participants (2.9%) who received convalescent plasma and 37 of 589 participants (6.3%) who received control plasma (absolute risk reduction, 3.4 percentage points; 95% confidence interval, 1.0 to 5.8; P = 0.005), which corresponded to a relative risk reduction of 54%. Evidence of efficacy in vaccinated participants cannot be inferred from these data because 53 of the 54 participants with Covid-19 who were hospitalized were unvaccinated and 1 participant was partially vaccinated. A total of 16 grade 3 or 4 adverse events (7 in the convalescent-plasma group and 9 in the control-plasma group) occurred in participants who were not hospitalized.

CONCLUSIONS

In participants with Covid-19, most of whom were unvaccinated, the administration of convalescent plasma within 9 days after the onset of symptoms reduced the risk of disease progression leading to hospitalization. (Funded by the Department of Defense and others; CSSC-004 ClinicalTrials.gov number, NCT04373460.).

Using Mobile Integrated Health and telehealth to support transitions of care among patients with heart failure (MIGHTy-Heart): protocol for a pragmatic randomised controlled trial

INTRODUCTION

Nearly one-quarter of patients discharged from the hospital with heart failure (HF) are readmitted within 30 days, placing a significant burden on patients, families and health systems. The objective of the 'Using Mobile Integrated Health and Telehealth to support transitions of care among patients with Heart failure' (MIGHTy-Heart) study is to compare the effectiveness of two postdischarge interventions on healthcare utilisation, patient-reported outcomes and healthcare quality among patients with HF.

METHODS AND ANALYSIS

The MIGHTy-Heart study is a pragmatic comparative effectiveness trial comparing two interventions demonstrated to improve the hospital to home transition for patients with HF: mobile integrated health (MIH) and transitions of care coordinators (TOCC). The MIH intervention bundles home visits from a community paramedic (CP) with telehealth video visits by emergency medicine physicians to support the management of acute symptoms and postdischarge care coordination. The TOCC intervention consists of follow-up phone calls from a registered nurse within 48-72 hours of discharge to assess a patient's clinical status, identify unmet clinical and social needs and reinforce patient education (eg, medication adherence and lifestyle changes). MIGHTy-Heart is enrolling and randomising (1:1) 2100 patients with HF who are discharged to home following a hospitalisation in two New York City (NY, USA) academic health systems. The coprimary study outcomes are all-cause 30-day hospital readmissions and quality of life measured with the Kansas City Cardiomyopathy Questionnaire 30 days after hospital discharge. The secondary endpoints are days at home, preventable emergency department visits, unplanned hospital admissions and patient-reported symptoms. Data sources for the study outcomes include patient surveys, electronic health records and claims submitted to Medicare and Medicaid.

ETHICS AND DISSEMINATION

All participants provide written or verbal informed consent prior to randomisation in English, Spanish, French, Mandarin or Russian. Study findings are being disseminated to scientific audiences through peer-reviewed publications and presentations at national and international conferences. This study has been approved by: Biomedical Research Alliance of New York (BRANY #20-08-329-380), Weill Cornell Medicine Institutional Review Board (20-08022605) and Mt. Sinai Institutional Review Board (20-01901).

TRIAL REGISTRATION NUMBER

Clinicaltrials.gov, NCT04662541.

Randomized Controlled Trial of Early Outpatient COVID-19 Treatment with High-Titer Convalescent Plasma

BACKGROUND

The efficacy of polyclonal high titer convalescent plasma to prevent serious complications of COVID-19 in outpatients with recent onset of illness is uncertain.

METHODS

This multicenter, double-blind randomized controlled trial compared the efficacy and safety of SARS-CoV-2 high titer convalescent plasma to placebo control plasma in symptomatic adults ≥18 years positive for SARS-CoV-2 regardless of risk factors for disease progression or vaccine status. Participants with symptom onset within 8 days were enrolled, then transfused within the subsequent day. The measured primary outcome was COVID-19-related hospitalization within 28 days of plasma transfusion. The enrollment period was June 3, 2020 to October 1, 2021.

RESULTS

A total of 1225 participants were randomized and 1181 transfused. In the pre-specified modified intention-to-treat analysis that excluded those not transfused, the primary endpoint occurred in 37 of 589 (6.3%) who received placebo control plasma and in 17 of 592 (2.9%) participants who received convalescent plasma (relative risk, 0.46; one-sided 95% upper bound confidence interval 0.733; P=0.004) corresponding to a 54% risk reduction. Examination with a model adjusting for covariates related to the outcome did not change the conclusions.

CONCLUSION

Early administration of high titer SARS-CoV-2 convalescent plasma reduced outpatient hospitalizations by more than 50%. High titer convalescent plasma is an effective early outpatient COVID-19 treatment with the advantages of low cost, wide availability, and rapid resilience to variant emergence from viral genetic drift in the face of a changing pandemic.

Trial Registration

ClinicalTrials.gov number, NCT04373460.

Randomized controlled trial transfusing convalescent plasma as post-exposure prophylaxis against SARS-CoV-2 infection

BACKGROUND

The efficacy of SARS-CoV-2 convalescent plasma (CCP) for preventing infection in exposed, uninfected individuals is unknown. We hypothesized that CCP might prevent infection when administered before symptoms or laboratory evidence of infection.

METHODS

This double-blinded, phase 2 randomized, controlled trial (RCT) compared the efficacy and safety of prophylactic high titer (≥1:320) CCP with standard plasma. Asymptomatic participants aged ≥18 years with close contact exposure to a person with confirmed COVID-19 in the previous 120 hours and negative SARS-CoV-2 test within 24 hours before transfusion were eligible. The primary outcome was development of SARS-CoV-2 infection.

RESULTS

180 participants were enrolled; 87 were assigned to CCP and 93 to control plasma, and 170 transfused at 19 sites across the United States from June 2020 to March 2021. Two were excluded for SARS-CoV-2 RT-PCR positivity at screening. Of the remaining 168 participants, 12/81 (14.8%) CCP and 13/87 (14.9%) control recipients developed SARS-CoV-2 infection; 6 (7.4%) CCP and 7 (8%) control recipients developed COVID-19 (infection with symptoms). There were no COVID-19-related hospitalizations in CCP and 2 in control recipients. There were 28 adverse events in CCP and 58 in control recipients. Efficacy by restricted mean infection free time (RMIFT) by 28 days for all SARS-CoV-2 infections (25.3 vs. 25.2 days; p=0.49) and COVID-19 (26.3 vs. 25.9 days; p=0.35) were similar for both groups.

CONCLUSION

In this trial, which enrolled persons with recent exposure to a person with confirmed COVID-19, high titer CCP as post-exposure prophylaxis appeared safe, but did not prevent SARS-CoV-2 infection.

TRIAL REGISTRATION

Clinicaltrial.gov number NCT04323800 .

G-computation and doubly robust standardisation for continuous-time data: A comparison with inverse probability weighting

In time-to-event settings, g-computation and doubly robust estimators are based on discrete-time data. However, many biological processes are evolving continuously over time. In this paper, we extend the g-computation and the doubly robust standardisation procedures to a continuous-time context. We compare their performance to the well-known inverse-probability-weighting estimator for the estimation of the hazard ratio and restricted mean survival times difference, using a simulation study. Under a correct model specification, all methods are unbiased, but g-computation and the doubly robust standardisation are more efficient than inverse-probability-weighting. We also analyse two real-world datasets to illustrate the practical implementation of these approaches. We have updated the R package RISCA to facilitate the use of these methods and their dissemination.

Improving precision and power in randomized trials for COVID-19 treatments using covariate adjustment, for binary, ordinal, and time-to-event outcomes

Time is of the essence in evaluating potential drugs and biologics for the treatment and prevention of COVID-19. There are currently 876 randomized clinical trials (phase 2 and 3) of treatments for COVID-19 registered on clinicaltrials.gov. Covariate adjustment is a statistical analysis method with potential to improve precision and reduce the required sample size for a substantial number of these trials. Though covariate adjustment is recommended by the U.S. Food and Drug Administration and the European Medicines Agency, it is underutilized, especially for the types of outcomes (binary, ordinal, and time-to-event) that are common in COVID-19 trials. To demonstrate the potential value added by covariate adjustment in this context, we simulated two-arm, randomized trials comparing a hypothetical COVID-19 treatment versus standard of care, where the primary outcome is binary, ordinal, or time-to-event. Our simulated distributions are derived from two sources: longitudinal data on over 500 patients hospitalized at Weill Cornell Medicine New York Presbyterian Hospital and a Centers for Disease Control and Prevention preliminary description of 2449 cases. In simulated trials with sample sizes ranging from 100 to 1000 participants, we found substantial precision gains from using covariate adjustment-equivalent to 4-18% reductions in the required sample size to achieve a desired power. This was the case for a variety of estimands (targets of inference). From these simulations, we conclude that covariate adjustment is a low-risk, high-reward approach to streamlining COVID-19 treatment trials. We provide an R package and practical recommendations for implementation.

Doubly robust nonparametric instrumental variable estimators for survival outcomes

Instrumental variable (IV) methods allow us the opportunity to address unmeasured confounding in causal inference. However, most IV methods are only applicable to discrete or continuous outcomes with very few IV methods for censored survival outcomes. In this article, we propose nonparametric estimators for the local average treatment effect on survival probabilities under both covariate-dependent and outcome-dependent censoring. We provide an efficient influence function-based estimator and a simple estimation procedure when the IV is either binary or continuous. The proposed estimators possess double-robustness properties and can easily incorporate nonparametric estimation using machine learning tools. In simulation studies, we demonstrate the flexibility and double robustness of our proposed estimators under various plausible scenarios. We apply our method to the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial for estimating the causal effect of screening on survival probabilities and investigate the causal contrasts between the two interventions under different censoring assumptions.

Heterogeneity in design and analysis of ICU delirium randomized trials: a systematic review

BACKGROUND

There is a growing number of randomized controlled trials (RCTs) evaluating interventions to prevent or treat delirium in the intensive care unit (ICU). Efforts to improve the conduct of delirium RCTs are underway, but none address issues related to statistical analysis. The purpose of this review is to evaluate heterogeneity in the design and analysis of delirium outcomes and advance methodological recommendations for delirium RCTs in the ICU.

METHODS

Relevant databases, including PubMed and Embase, were searched with no restrictions on language or publication date; the search was conducted on July 8, 2019. RCTs conducted on adult ICU patients with delirium as the primary outcome were included where trial results were available. Data on frequency and duration of delirium assessments, delirium outcome definitions, and statistical methods were independently extracted in duplicate. The review was registered with PROSPERO (CRD42020141204).

RESULTS

Among 65 eligible RCTs, 44 (68%) targeted the prevention of delirium. The duration of follow-up varied, with 31 (48%) RCTs having ≤7 days of follow-up, and only 24 (37%) conducting delirium assessments after ICU discharge. The incidence of delirium was the most common outcome (50 RCTs, 77%) for which 8 unique statistical methods were applied. The most common method, applied to 51 of 56 (91%) delirium incidence outcomes, was the two-sample test comparing the proportion of patients who ever experienced delirium. In the presence of censoring of patients at ICU discharge or death, this test may be misleading. The impact of censoring was also not considered in most analyses of the duration of delirium, as evaluated in 24 RCTs, with 21 (88%) delirium duration outcomes analyzed using a non-parametric test or two-sample t test. Composite outcomes (e.g., rank-based delirium- and coma-free days), used in 11 (17%) RCTs, seldom explicitly defined how ICU discharge, and death were incorporated into the definition and were analyzed using non-parametric tests (11 of 13 (85%) composite outcomes).

CONCLUSIONS

To improve delirium RCTs, outcomes should be explicitly defined. To account for censoring due to ICU discharge or death, survival analysis methods should be considered for delirium incidence and duration outcomes; non-parametric tests are recommended for rank-based delirium composite outcomes.

TRIAL REGISTRATION

PROSPERO CRD42020141204 . Registration date: 7/3/2019.

Population-level changes in outcomes and Medicare cost following the introduction of new cancer therapies

OBJECTIVE

To examine the population-level impacts of the introduction of novel cancer therapies with high cost in the United States, using immunotherapies in advanced nonsmall cell lung cancer (NSCLC) as an example.

DATA SOURCES

Surveillance, Epidemiology, and End Results data in 2012-2015 linked to Medicare fee-for-service claims until 2016.

STUDY DESIGN

We examined population-level trends in treatment patterns, survival, and Medicare spending in patients diagnosed with advanced NSCLC, the leading cause of cancer death in the United States, between 2012 and 2015. We estimated the percentage of patients who received any antineoplastic therapy within two years of diagnosis, including novel immunotherapies. We compared the trends in overall survival and mean two-year Medicare spending per each patient before and after the introduction of immunotherapies in 2015.

DATA COLLECTION/EXTRACTION METHODS

Not Applicable.

PRINCIPAL FINDINGS

The percentage of patients treated with any antineoplastic therapy remained the same at 46.7% in 2012 and 2015, whereas the use of immunotherapies increased from 0% to 15.2%. The two-year survival rate and median survival increased by 3.3 percentage points (95% CI: 2.0, 4.5) and 0.4 months (CI: 0.0, 0.9), respectively, during the same period. The mean two-year total Medicare spending and outpatient spending per patient increased by $5735 (CI: 3479, 8040) and $7661 (CI: 5902, 9311), respectively, which were largely attributable to the increases in immunotherapy spending by $5806 (CI: 5165, 6459).

CONCLUSIONS

The introduction of lung cancer immunotherapies was accompanied by improvements in survival and increases in spending between 2012 and 2015 in the Medicare population. As novel immunotherapies and other target therapies continue to change the clinical management of various cancers, further efforts are needed to ensure their effective and efficient use, and to understand their population-level impacts in the United States.

Improving Precision and Power in Randomized Trials for COVID-19 Treatments Using Covariate Adjustment, for Binary, Ordinal, and Time-to-Event Outcomes

Time is of the essence in evaluating potential drugs and biologics for the treatment and prevention of COVID-19. There are currently over 400 clinical trials (phase 2 and 3) of treatments for COVID-19 registered on clinicaltrials.gov. Covariate adjustment is a statistical analysis method with potential to improve precision and reduce the required sample size for a substantial number of these trials. Though covariate adjustment is recommended by the U.S. Food and Drug Administration and the European Medicines Agency, it is underutilized, especially for the types of outcomes (binary, ordinal and time-to-event) that are common in COVID-19 trials. To demonstrate the potential value added by covariate adjustment in this context, we simulated two-arm, randomized trials comparing a hypothetical COVID-19 treatment versus standard of care, where the primary outcome is binary, ordinal, or time-to-event. Our simulated distributions are derived from two sources: longitudinal data on over 500 patients hospitalized at Weill Cornell Medicine New York Presbyterian Hospital, and a Centers for Disease Control and Prevention (CDC) preliminary description of 2449 cases. We found substantial precision gains from using covariate adjustment--equivalent to 9-21% reductions in the required sample size to achieve a desired power--for a variety of estimands (targets of inference) when the trial sample size was at least 200. We provide an R package and practical recommendations for implementing covariate adjustment. The estimators that we consider are robust to model misspecification.

Statistical inference for data-adaptive doubly robust estimators with survival outcomes

The consistency of doubly robust estimators relies on the consistent estimation of at least one of two nuisance regression parameters. In moderate-to-large dimensions, the use of flexible data-adaptive regression estimators may aid in achieving this consistency. However, n^1/2 -consistency of doubly robust estimators is not guaranteed if one of the nuisance estimators is inconsistent. In this paper, we present a doubly robust estimator for survival analysis with the novel property that it converges to a Gaussian variable at an n^1/2 -rate for a large class of data-adaptive estimators of the nuisance parameters, under the only assumption that at least one of them is consistently estimated at an n^1/4 -rate. This result is achieved through the adaptation of recent ideas in semiparametric inference, which amount to (i) Gaussianizing (ie, making asymptotically linear) a drift term that arises in the asymptotic analysis of the doubly robust estimator and (ii) using cross-fitting to avoid entropy conditions on the nuisance estimators. We present the formula of the asymptotic variance of the estimator, which allows for the computation of doubly robust confidence intervals and p values. We illustrate the finite-sample properties of the estimator in simulation studies and demonstrate its use in a phase III clinical trial for estimating the effect of a novel therapy for the treatment of human epidermal growth factor receptor 2 (HER2)-positive breast cancer.

Targeted learning ensembles for optimal individualized treatment rules with time-to-event outcomes

We consider estimation of an optimal individualized treatment rule when a high-dimensional vector of baseline variables is available. Our optimality criterion is with respect to delaying the expected time to occurrence of an event of interest. We use semiparametric efficiency theory to construct estimators with properties such as double robustness. We propose two estimators of the optimal rule, which arise from considering two loss functions aimed at directly estimating the conditional treatment effect and recasting the problem in terms of weighted classification using the 0-1 loss function. Our estimated rules are ensembles that minimize the crossvalidated risk of a linear combination in a user-supplied library of candidate estimators. We prove oracle inequalities bounding the finite-sample excess risk of the estimator. The bounds depend on the excess risk of the oracle selector and a doubly robust term related to estimation of the nuisance parameters. We discuss the convergence rates of our estimator to the oracle selector, and illustrate our methods by analysis of a phase III randomized study testing the efficacy of a new therapy for the treatment of breast cancer.