Abstract:BACKGROUND:The use of suboptimal controls in randomized trials of new cancer drugs can produce potentially unreliable clinical efficacy results over the current standard of care and expose patients to substandard therapy. We aim to investigate the proportion of randomized trials of investigational cancer drugs that used a suboptimal control arm and the number of trial participants at risk of exposure to suboptimal treatments in China. The association between the use of a suboptimal control and concluding statistical significance on the primary endpoint was also examined. METHODS AND FINDINGS:This observational study included randomized controlled trials (RCTs) of cancer drugs that were authorized by specific Chinese institutional review boards between 2016 and 2021, supporting investigational new drug applications of these drugs in China. The proportion of trials that used a suboptimal control arm and the total number of trial participants at risk of exposure to suboptimal treatments were calculated. In a randomized trial for a specific condition, a comparator was deemed suboptimal if it was not recommended by clinical guidelines published in priori or if there existed a regimen with a higher level of recommendation for the indication. The final sample included 453 Phase II/III and Phase III randomized oncology trials. Overall, 60 trials (13.2%) adopted a suboptimal control arm. Among them, 58.3% (35/60) used comparators that were not recommended by a prior guideline for the indication. The cumulative number of trial participants at risk of exposure to suboptimal treatments totaled 18,610 by the end of 2021, contributing 15.1% to the total number of enrollees of all sampled RCTs in this study. After adjusting for the year of ethical approval, region of participant recruitment, line of therapy, and cancer site, second-line therapies (adjusted odds ratio [aOR] = 2.7, 95%CI [1.2, 5.9]), adjuvant therapies (aOR = 8.9, 95% CI [3.4, 23.1]), maintenance therapies (aOR = 5.2, 95% CI [1.6, 17.0]), and trials recruiting participants in China only (aOR = 4.1, 95% CI [2.1, 8.0]) were more likely to adopt a suboptimal control. For the 105 trials with publicly available results, no statistically significant difference was observed between the use of a suboptimal control and concluding positive on the primary endpoint (100.0% [12/12] versus 83.9% [78/93], p = 0.208). The main limitation of this study is its reliance on clinical guidelines that could vary across cancer types and time in assessing the quality of the control groups. CONCLUSIONS:In this study, over one-eighth of randomized trials of cancer drugs registered to apply for regulatory approval in China used a suboptimal comparator. Our results highlight the necessity to refine the design of randomized trials to generate optimal clinical evidence for new cancer therapies.

Intention-to-treat Analysis Controls the Bias when Suffer from Noncompliance and Missing Outcomes in Randomized Controlled Trial Study

Understanding the Intention-to-treat Principle in Randomized Controlled Trials

Intent-to-treat Analysis of Cluster Randomized Trials when Clusters Report Unidentifiable Outcome Proportions.

Using modified intention-to-treat as a principal stratum estimator for failure to initiate treatment

INTENTION TO TREAT ANALYSIS IN CLINICAL RESEARCH: BASIC CONCEPTS FOR CLINICIANS

Exploring Statistical Methods in Clinical Trials When Some of the Patients are Treated against the Randomization

Response to Comment on "controlling for Ascertainment Bias May Introduce Control Group Bias in Prospective Nonrandomized Trials".

Series 2-19-2009 Multiple Imputation Methods for Treatment Noncompliance and Nonresponse in Randomized Clinical Trials

Causal survival analysis: A guide to estimating intention-to-treat and per-protocol effects from randomized clinical trials with non-adherence

Estimating treatment effect in randomized trial after control to treatment crossover using external controls

Use of Suboptimal Control Arms in Randomized Clinical Trials of Investigational Cancer Drugs in China, 2016-2021: an Observational Study

Generalizing the intention-to-treat effect of an active control against placebo from historical placebo-controlled trials to an active-controlled trial: A case study of the efficacy of daily oral TDF/FTC in the HPTN 084 study

Supplementary Table S3 from Proportional Hazards Violations in Phase III Cancer Clinical Trials: A Potential Source of Trial Misinterpretation

Assessing efficacy in non‐inferiority trials with non‐adherence to interventions: Are intention‐to‐treat and per‐protocol analyses fit for purpose?

Placebo Adherence as a Negative Control Exposure

Conditional and Unconditional Treatment Effects in Randomized Clinical Trials: Estimands, Estimation, and Interpretation

Applying the estimand and target trial frameworks to external control analyses using observational data: a case study in the solid tumor setting

Analysis approaches to address treatment nonadherence in pragmatic trials with point-treatment settings: a simulation study

Bounds On Causal Effects In Three-Arm Trials With Non-Compliance

A novel statistical method for comparing effectiveness of two treatments-simulated randomized controlled trials

ASSESSING THE HETEROGENEITY OF TREATMENT EFFECTS BY IDENTIFYING THE TREATMENT BENEFIT RATE AND TREATMENT HARM RATE