Bala Malgorzata M, Agarwal Arnav, Klatt Kevin C, Vernooij Robin W M, Alonso-Coello Pablo, Steen Jeremy P, Guyatt Gordon H, Duque Tiffany, Johnston Bradley C
Chair of Epidemiology and Preventive Medicine, Department of Hygiene and Dietetics, Jagiellonian University Medical College, Krakow, Poland.
Department of Medicine, McMaster University, Hamilton, Ontario, Canada.
BMJ Nutr Prev Health. 2024 Aug 29;7(2):e000834. doi: 10.1136/bmjnph-2023-000834. eCollection 2024.
This article continues from a prior commentary on evaluating the risk of bias in randomised controlled trials addressing nutritional interventions. Having provided a synopsis of the risk of bias issues, we now address how to understand trial results, including the interpretation of best estimates of effect and the corresponding precision (eg, 95% CIs), as well as the applicability of the evidence to patients based on their unique circumstances (eg, patients' values and preferences when trading off potential desirable and undesirable health outcomes and indicators (eg, cholesterol), and the potential burden and cost of an intervention). Authors can express the estimates of effect for health outcomes and indicators in relative terms (relative risks, relative risk reductions, OR or HRs)-measures that are generally consistent across populations-and absolute terms (risk differences)-measures that are more intuitive to clinicians and patients. CIs, the range in which the true effect plausibly lies, capture the precision of estimates. To apply results to patients, clinicians should consider the extent to which the study participants were similar to their patients, the extent to which the interventions evaluated in the study are applicable to their patients and if all patient-important outcomes of potential benefit and harm were reported. Subsequently, clinicians should consider the values and preferences of their patients with respect to the balance of the benefits, harms and burdens (and possibly the costs) when making decisions about dietary interventions.
本文延续了之前一篇关于评估营养干预随机对照试验偏倚风险的评论。在概述了偏倚风险问题之后，我们现在探讨如何理解试验结果，包括对效应最佳估计值及其相应精度（如95%置信区间）的解读，以及根据患者的独特情况（如患者在权衡潜在的有益和有害健康结果及指标（如胆固醇）时的价值观和偏好，以及干预措施的潜在负担和成本）将证据应用于患者。作者可以用相对指标（相对风险、相对风险降低率、比值比或风险比）来表达健康结果和指标的效应估计值，这些指标在不同人群中通常是一致的；也可以用绝对指标（风险差值）来表达，这对临床医生和患者来说更直观。置信区间是真实效应可能所在的范围，它体现了估计值的精度。为了将结果应用于患者，临床医生应考虑研究参与者与他们的患者的相似程度、研究中评估的干预措施对其患者的适用程度，以及是否报告了所有对患者重要的潜在有益和有害结果。随后，临床医生在就饮食干预做出决策时，应考虑患者在权衡益处、危害和负担（可能还有成本）方面的价值观和偏好。
