See Article by Ferreira et al In classic cardiovascular phase III trials, individuals satisfying a well-defined set of inclusion and exclusion criteria are randomly assigned to receive one of 2 or more treatments in a nonadaptive parallel-arm design. If the trial is positive, for example in favor of treatment A (a new treatment) over treatment B (standard of care), it is concluded that the new treatment benefits such a patient population. A more concrete hypothetical trial may demonstrate that treatment A significantly reduces the composite end point of death or heart failure–related hospitalizations by 20% relative to treatment B. A subjective interpretation that may follow such trial results is that a patient that meets trial criteria will experience a benefit from the treatment (ie, a reduction in adverse outcomes). This interpretation is flawed. In reality, some individuals benefit, some individuals are harmed, and some individuals experience neither benefit nor harm, but the number of individuals who benefit exceeds the number of those who are harmed, such that the net rate of the end point is ≈20% lower in a group of patients receiving treatment A relative to treatment B (for the purposes of our discussion, we will ignore the confidence interval). Explicitly acknowledging this difference in interpretations has increasingly important implications in the era of precision medicine. Implantable cardioverter defibrillator (ICD) trials can further illustrate these concepts. Consider the SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial),1 which demonstrated that among patients with heart failure with reduced ejection fraction (HFrEF), New York Heart Association class II or III symptoms, and a left ventricular ejection fraction ≤35%, ICD therapy decreased the risk of death by 23%. Patients who meet indications for ICDs based on well-designed phase III trials may experience infection, bleeding, or mechanical complications as a direct consequence of …

中文翻译：

---

拥抱精准医疗的漫漫长路


请参阅 Ferreira 等人的文章 在经典的心血管 III 期试验中，满足一组明确的纳入和排除标准的个体被随机分配接受非适应性平行臂设计中的 2 种或多种治疗中的一种。如果试验是积极的，例如支持治疗 A（一种新治疗）而不是治疗 B（护理标准），则可以得出结论，新治疗有益于此类患者群体。一项更具体的假设试验可能表明，相对于治疗 B，治疗 A 显着降低了死亡或心力衰竭相关住院治疗的复合终点 20%。此类试验结果可能遵循的主观解释是，符合试验标准的患者将体验治疗带来的益处（即不良后果的减少）。这种解释是有缺陷的。现实中，有的人受益，有的人受到伤害，有的人既没有受益也没有受到伤害，但受益的人数超过了受到伤害的人数，这样终点的净利率约为20%接受治疗 A 的患者组相对于治疗 B 组患者的死亡率较低（出于我们讨论的目的，我们将忽略置信区间）。在精准医学时代，明确承认这种解释上的差异具有越来越重要的意义。植入式心律转复除颤器 (ICD) 试验可以进一步说明这些概念。 考虑一下 SCD-HeFT（心力衰竭猝死试验）1，该试验表明，在射血分数降低 (HFrEF) 的心力衰竭患者中，纽约心脏协会 II 级或 III 级症状，且左心室射血分数 ≤ 35 %，ICD 治疗将死亡风险降低了 23%。根据精心设计的 III 期试验，符合 ICD 指征的患者可能会因……的直接后果而经历感染、出血或机械并发症。