Boada José N, Boada Carlos, García-Sáiz Mar, García Marcelino, Fernández Eduardo, Gómez Eugenio
Clinical Pharmacology Unit, Pharmacovigilance Centre, University Hospital of Canary Islands, University of La Laguna, La Laguna, Santa Cruz de Tenerife, Spain.
PLoS One. 2008;3(10):e3580. doi: 10.1371/journal.pone.0003580. Epub 2008 Oct 31.
Although several mathematical models have been proposed to assess the risk:benefit of drugs in one measure, their use in practice has been rather limited. Our objective was to design a simple, easily applicable model. In this respect, measuring the proportion of patients who respond favorably to treatment without being affected by adverse drug reactions (ADR) could be a suitable endpoint. However, remarkably few published clinical trials report the data required to calculate this proportion. As an approach to the problem, we calculated the expected proportion of this type of patients.
METHODOLOGY/PRINCIPAL FINDINGS: Theoretically, responders without ADR may be obtained by multiplying the total number of responders by the total number of subjects that did not suffer ADR, and dividing the product by the total number of subjects studied. When two drugs are studied, the same calculation may be repeated for the second drug. Then, by constructing a 2 x 2 table with the expected frequencies of responders with and without ADR, and non-responders with and without ADR, the odds ratio and relative risk with their confidence intervals may be easily calculated and graphically represented on a logarithmic scale. Such measures represent "net efficacy adjusted for risk" (NEAR). We assayed the model with results extracted from several published clinical trials or meta-analyses. On comparing our results with those originally reported by the authors, marked differences were found in some cases, with ADR arising as a relevant factor to balance the clinical benefit obtained. The particular features of the adverse reaction that must be weighed against benefit is discussed in the paper.
NEAR representing overall risk-benefit may contribute to improving knowledge of drug clinical usefulness. As most published clinical trials tend to overestimate benefits and underestimate toxicity, our measure represents an effort to change this trend.
尽管已经提出了几种数学模型来综合评估药物的风险效益，但它们在实际应用中相当有限。我们的目标是设计一个简单、易于应用的模型。在这方面，测量对治疗有良好反应且未受药物不良反应（ADR）影响的患者比例可能是一个合适的终点。然而，极少有已发表的临床试验报告计算该比例所需的数据。作为解决该问题的一种方法，我们计算了这类患者的预期比例。
方法/主要发现：理论上，无ADR的反应者可通过将反应者总数乘以未发生ADR的受试者总数，再将乘积除以所研究的受试者总数得到。研究两种药物时，可对第二种药物重复相同计算。然后，通过构建一个2×2表格，列出有和无ADR的反应者以及有和无ADR的无反应者的预期频数，可轻松计算比值比和相对危险度及其置信区间，并以对数尺度进行图形表示。这些指标代表“经风险调整的净疗效”（NEAR）。我们用从一些已发表的临床试验或荟萃分析中提取的结果对该模型进行了检验。将我们的结果与作者最初报告的结果进行比较时，在某些情况下发现了显著差异，ADR成为平衡所获得临床益处的一个相关因素。本文讨论了必须与益处相权衡的不良反应的具体特征。
代表总体风险效益的NEAR可能有助于增进对药物临床实用性的认识。由于大多数已发表的临床试验往往高估益处而低估毒性，我们的指标旨在改变这一趋势。
