Ubel P A, Merz J F, Shea J, Asch D A
Veterans Affairs Medical Center, Philadelphia, PA, USA.
J Investig Med. 1997 Dec;45(9):561-6.
To explore how preliminary trial data affect the general public's stated willingness to enter a randomized clinical trial.
We asked 165 prospective jurors to imagine that their physicians wanted them to enroll in a clinical trial. We then presented them with scenarios portraying preliminary trial results--for example, 9 out of 10 patients get better with drug A and 5 out of 10 get better with drug B--and asked after each scenario, whether they would choose to be part of the trial. We designed the scenarios to test how stated willingness to enter the trial would be influenced by: 1) the difference in effectiveness of the 2 treatments based on the patients enrolled thus far; and 2) by the chance that the difference in effectiveness was random (i.e., the P value). The subjects' willingness to enter the trial at various decision points was analyzed using logistic regression.
Fewer subjects were willing to enter the trial as preliminary data indicated either an increasing difference in the effectiveness of the two treatments or an increasing statistical significance of that difference. For example 75% of subjects were willing to enter the research trial before any preliminary data were presented, but this number fell to 49% when subjects were presented with preliminary data showing that 9 out of 10 patients improved with one treatment and 5 out of 10 with the other. Multivariable logistic regression revealed that higher P values (odds ratio = 4.29; P < 0.001; 95% CI: 2.22-8.28) and smaller differences in effectiveness (odds ratio = 0.02; P < 0.001; 95% CI: 0.00-0.07) implicit in preliminary data presented to subjects made subjects less likely to agree to enter clinical trials. After adjustment for other relevant variables, male gender was associated with increased willingness to enter the trial.
A subjects' willingness to enter the hypothetical trial was influenced by preliminary data. Fewer subjects were willing to enter the trial as the differences in benefit between 2 treatment groups increased. However, the majority of subjects were willing to enter the hypothetical trial even when preliminary evidence strongly favored one treatment over another. Given the importance of informed consent in entering patients in clinical trials, these results should be confirmed in actual trial settings.
探讨初步试验数据如何影响公众表示参与随机临床试验的意愿。
我们让165名潜在陪审员想象他们的医生希望他们参加一项临床试验。然后向他们展示描绘初步试验结果的情景——例如，10名患者中有9名使用药物A后病情好转，10名患者中有5名使用药物B后病情好转——并在每个情景展示后询问他们是否会选择参与试验。我们设计这些情景以测试参与试验的意愿如何受到以下因素的影响：1）根据目前已入组患者情况，两种治疗方法有效性的差异；2）有效性差异是随机的可能性（即P值）。使用逻辑回归分析受试者在不同决策点参与试验的意愿。
随着初步数据显示两种治疗方法有效性差异增大或该差异的统计学显著性增加，愿意参与试验的受试者减少。例如，在未展示任何初步数据之前，75%的受试者愿意参与研究试验，但当向受试者展示初步数据显示10名患者中有9名使用一种治疗方法后病情改善，10名患者中有5名使用另一种治疗方法后病情改善时，这一比例降至49%。多变量逻辑回归显示，向受试者展示的初步数据中较高的P值（比值比 = 4.29；P < 0.001；95%置信区间：2.22 - 8.28）和较小的有效性差异（比值比 = 0.02；P < 0.001；95%置信区间：0.00 - 0.07）使受试者更不太可能同意参与临床试验。在对其他相关变量进行调整后，男性参与试验的意愿增加。
受试者参与假设试验的意愿受到初步数据的影响。随着两个治疗组之间益处差异的增加，愿意参与试验的受试者减少。然而，即使初步证据强烈支持一种治疗方法优于另一种治疗方法，大多数受试者仍愿意参与假设试验。鉴于知情同意在让患者参与临床试验中的重要性，这些结果应在实际试验环境中得到证实。
