{
  "abstract": "In recent years, cancer drug randomised controlled trials (RCTs) have been criticised for employing a control arm treatment that is inferior or suboptimal to the existing standard of care.1 2 Using an inferior treatment as the control arm in an RCT undermines both the scientific validity and the ethics of trials and has been recognised as an important bias in cancer clinical trials by the European Society for Medical Oncology-Magnitude of Clinical Benefit Scale3 as well as the Common Sense Oncology initiative.4",
  "authors": [
    {
      "affiliations": [
        "University Hospital Antwerp, Edegem, Flanders, Belgium"
      ],
      "name": "Laure-Anne Teuwen"
    },
    {
      "affiliations": [
        "Queen’s University, Kingston, Canada"
      ],
      "name": "Bishal Gyawali"
    }
  ],
  "full_text": "In recent years, cancer drug randomised controlled trials (RCTs) have been criticised for employing a control arm treatment that is inferior or suboptimal to the existing standard of care.1 2 Using an inferior treatment as the control arm in an RCT undermines both the scientific validity and the ethics of trials and has been recognised as an important bias in cancer clinical trials by the European Society for Medical Oncology-Magnitude of Clinical Benefit Scale3 as well as the Common Sense Oncology initiative.4Scientifically, using an inferior control arm compromises the validity and clinical utility of the results. The question that needs answering is whether the new treatment improves outcomes compared with the existing standard. An RCT that employs an inferior control arm does not answer this question. Ethically, no patients should be harmed as a result of trial participation, but arguably, patients randomised to the control arm of such RCTs are harmed because they would receive better treatments outside of the trial.The declaration of Helsinki, which even these RCTs that employ inferior control treatments nevertheless cite, states that the benefits, risks, burdens and effectiveness of a new intervention must be tested against the best proven intervention(s).5 In practice, however, what constitutes the best proven intervention is at times interpreted rather liberally, with 11% of RCTs published in major oncology journals and 17% of pivotal cancer trials leading to US Food and Drug Administration (FDA) approval employing suboptimal control treatments.1 2One of the most common ‘explanations’ offered to justify the use of inferior control treatments in cancer drug RCTs is that the RCTs were conducted in low- and middle-income countries (LMICs) where the superior treatments are not affordable. This cross-sectional analysis of 98 oncology RCTs recruiting patients in Brazil—a middle-income country—provides some important context to this discussion.6 In this study, the authors assessed the appropriateness of control arms against three benchmarks: treatments offered through Brazil’s public health system (SUS), Brazilian private insurance standards and the National Comprehensive Cancer Network (NCCN) guidelines. Compared with SUS, none of the RCTs had an inferior control arm. However, when evaluated against Brazilian private insurance and NCCN standards, 16.3% and 18.3% of trials, respectively, used inferior control arms. These findings suggest that although international standards were not always met, patients enrolled in these RCTs in Brazil received care that met or exceeded the minimum standard available in their local setting. The authors highlight that while these discrepancies may not compromise patient care locally, they do raise concerns about the broader interpretability of the trial findings. However, several caveats must be considered while interpreting these results.First, this analysis provides only a snapshot in time. A longitudinal analysis could reveal if the use of suboptimal control arms has worsened over time. The study does not distinguish trials intended for drug registration from those designed for other purposes, or superiority trials from non-inferiority. Non-inferiority design RCTs using suboptimal control arms are even more problematic, as they risk ‘proving’ that a new drug is not worse than a treatment that is already below standard. In addition, the study does not evaluate whether the control arms match the standard of care at the time of trial initiation—an important consideration, given that guidelines can change rapidly. When the standard of care changes early on in the trial, RCTs should ideally be protocol amended to update the control arm treatment.7Second, not all control arms classified as ‘inferior’ in this analysis are necessarily so. For instance, we don’t agree with the classification of docetaxel monotherapy in late-line non-small cell lung cancer as inferior control, although the NCCN guidelines recommend sotorasib or ramucirumab plus docetaxel, because both of these agents have extremely weak evidence despite inclusion in the NCCN guidelines. By contrast, use of androgen deprivation therapy (ADT) monotherapy (rather than ADT plus androgen receptor pathway inhibitor (ARPI) or docetaxel) as control treatment in a trial of high-risk hormone-sensitive prostate cancer is clearly inferior although this was considered acceptable by SUS standards.8 These two categories should be clearly distinguished in our discussion of suboptimal comparators in cancer drug RCTs. If health budgets are constrained, resources should be allocated first to drugs like ARPIs or docetaxel, which offer meaningful clinical benefit, rather than to high-cost agents with marginal efficacy such as sotorasib or ramucirumab.Third, it is unacceptable to use inferior control arms in RCTs simply because the trial is being conducted in an LMIC. Although this study does not report what percentage of the NCCN-recommended control and experimental drugs were actually available in Brazil at the time, prior work shows that many LMICs that participated in drug approval trials lacked access to these new drugs several years after FDA approval.9 In that sense, research parasitism looms around the corner: using LMIC populations to conduct trials that would not meet ethical standards in high-income countries (HICs) to secure regulatory approvals in HICs without ensuring access to the tested drugs for the LMIC patients.Fourth, the adequacy of the control arm extends beyond the trial period. For earlier line therapy trials, what treatments the control arm patients received postprogression can meaningfully influence the outcomes. The authors argue that RCTs are important for improving drug access in LMICs, a point we fully agree with. However, when crossover (when indicated) and adequate postprogression therapies are not mandated, access even to patients who participated in the trial is limited. More LMICs should indeed participate in RCTs, but the focus and priorities must shift. Rather than allocating limited resources to facilitate enrolment in trials with suboptimal control arms, efforts should support ‘groundshot’ type trials that are designed to generate context-relevant evidence that can genuinely improve access and outcomes.10Using an inferior control arm undeniably is not in patients’ interests—not current patients who get randomised to an inferior treatment, nor future patients because the trial failed to answer whether the new treatment is better than the current standard. Using patients in LMICs to get away with an unacceptable control arm risks exploitation, misinterpretation of trial results and, most importantly, failure to improve cancer outcomes where they are needed most.",
  "title": "Inferior control arms in oncology trials in LMICs: contextualised or compromised?",
  "uid": "ceb0f188-5f5c-5cc8-a573-e87ee90e429c"
}