Hollin Ilene L, Peay Holly L, Bridges John F P
Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, 624 N. Broadway, Rm 689, Baltimore, MD, 21205, USA,
Patient. 2015 Feb;8(1):19-27. doi: 10.1007/s40271-014-0104-x.
Through Patient-Focused Drug Development, the US Food and Drug Administration (FDA) documents the perspective of patients and caregivers and are currently conducting 20 public meetings on a limited number of disease areas. Parent Project Muscular Dystrophy (PPMD), an advocacy organization for Duchenne muscular dystrophy (DMD), has demonstrated a community-engaged program of preference research that would complement the FDA's approach.
Our objective was to compare two stated-preference methods, best-worst scaling (BWS) and conjoint analysis, within a study measuring caregivers' DMD-treatment preferences.
Within one survey, two preference-elicitation methods were applied to 18 potential treatments incorporating six attributes and three levels. For each treatment profile, caregivers identified the best and worst feature and intention to use the treatment. We conducted three analyses to compare the elicitation methods using parameter estimates, conditional attribute importance and policy simulations focused on the 18 treatment profiles. For each, concordance between the results was compared using Spearman's rho.
BWS and conjoint analysis produced similar parameter estimates (p < 0.01); conditional attribute importance (p < 0.01); and policy simulations (p < 0.01). Greatest concordance was observed for the benefit and risk parameters, with differences observed for nausea and knowledge about the drug-where a lack of monotonicity was observed when using conjoint analysis.
The observed concordance between approaches demonstrates the reliability of the stated-preference methods. Given the simplicity of combining BWS and conjoint analysis on single profiles, a combination approach is easily adopted. Minor irregularities for the conjoint-analysis results could not be explained by additional analyses and needs to be the focus of future research.
通过以患者为中心的药物研发，美国食品药品监督管理局（FDA）记录了患者和护理人员的观点，目前正在就有限数量的疾病领域召开20次公开会议。杜氏肌营养不良症（DMD）倡导组织——母项目肌营养不良症（PPMD）已展示了一项社区参与的偏好研究计划，该计划将补充FDA的方法。
我们的目的是在一项测量护理人员对DMD治疗偏好的研究中，比较两种陈述性偏好方法，即最佳-最差尺度法（BWS）和联合分析。
在一项调查中，将两种偏好诱导方法应用于18种潜在治疗方法，这些方法包含六个属性和三个水平。对于每种治疗方案，护理人员确定最佳和最差特征以及使用该治疗方法的意向。我们进行了三项分析，以使用参数估计、条件属性重要性和针对18种治疗方案的政策模拟来比较诱导方法。对于每项分析，使用斯皮尔曼等级相关系数比较结果之间的一致性。
BWS和联合分析产生了相似的参数估计（p < 0.01）；条件属性重要性（p < 0.01）；以及政策模拟（p < 0.01）。在获益和风险参数方面观察到最大的一致性，在恶心和药物知识方面存在差异——使用联合分析时观察到缺乏单调性。
方法之间观察到的一致性证明了陈述性偏好方法的可靠性。鉴于在单个方案上结合BWS和联合分析很简单，很容易采用组合方法。联合分析结果中的微小不规则性无法通过额外分析来解释，需要成为未来研究的重点。
