背景:质量调整生命年(QALYs)是在卫生技术评估和卫生服务研究中广泛使用的关键结果指标。QALY结合了生活的数量和质量,生活质量的计算依赖于不同健康状况的价值。这样的健康状态值捕获了人群的偏好,并且通常是通过数值启发方法建立的。证据表明,这些价值得分受到使用方法的影响,并且个人报告在激发自己的偏好时存在认知困难。来自其他领域的证据进一步表明,个人可能更喜欢使用不同的启发技术,并且这种偏好可能会受到他们的计算能力的影响。在这项研究中,我们探讨了使用MACBETH(通过基于分类的评估技术测量吸引力)非数字偏好激发方法进行健康状况评估。
方法:开发了一种基于MACBETH(仅需要定性判断)的偏好启发新协议,并在网络调查格式中进行了测试。葡萄牙普通人群(n=243)的样本,使用MACBETH协议和时间权衡(TTO)协议的变体,评估了25个EQ-5D-3L健康状态,为了比较和理解受访者对不同方案的偏好和不一致评价的差异。受访者回答了简短的算术测试,和收集的基本社会经济信息。
结果:结果表明,从MACBETH和TTO变体得出的平均值具有很强的相关性;然而,几个健康状态的值存在实质性差异。在两种方法的受访者回答中发现了大量和相似数量的逻辑不一致。根据测试,具有较高计算能力的参与者更喜欢用MACBETH表达价值判断,而水平较低的参与者对这两种方法大多漠不关心。对于具有较高计算能力的个体,观察到MACBETH和TTO变体评估之间的相关性更高。
结论:结果表明,使用非数值偏好启发方法值得研究。当不需要探索和使用明确考虑不确定性时,算术测试更适合于偏好启发。需要进一步的行为研究,以充分了解在不同的环境中使用这些方法的潜力(例如,在不同的评估环境中以及在面对面和非面对面的环境中),以及探讨识字对评估和受访者偏好的影响。
BACKGROUND: Quality Adjusted Life Years (QALYs) are a key outcome measure widely used within health technology assessment and health service research studies. QALYs combine quantity and quality of life, with quality of life calculations relying on the value of distinct health states. Such health states\' values capture the preferences of a population and have been typically built through numerical elicitation methods. Evidence points to these value scores being influenced by methods in use and individuals reporting cognitive difficulties in eliciting their preferences. Evidence from other areas has further suggested that individuals may prefer using distinct elicitation techniques and that this preference can be influenced by their numeracy. In this study we explore the use of the
MACBETH (Measuring Attractiveness by a Categorical Based Evaluation Technique) non-numerical preference elicitation approach for health states\' evaluation.
METHODS: A new protocol for preference elicitation based on
MACBETH (only requiring qualitative judgments) was developed and tested within a web survey format. A sample of the Portuguese general population (n=243) valued 25 EQ-5D-3L health states with the
MACBETH protocol and with a variant of the time trade-off (TTO) protocol, for comparison purposes and for understanding respondents\' preference for distinct protocols and differences in inconsistent evaluations. Respondents answered to a short numeracy test, and basic socio-economic information collected.
RESULTS: Results show that the mean values derived from
MACBETH and the TTO variant are strongly correlated; however, there are substantial differences for several health states\' values. Large and similar numbers of logical inconsistencies were found in respondents\' answers with both methods. Participants with higher levels of numeracy according to the test preferred expressing value judgments with MACBETH, while participants with lower levels were mostly indifferent to both methods. Higher correlations between
MACBETH and TTO variant evaluations were observed for individuals with higher numeracy.
CONCLUSIONS: Results suggest that it is worth researching the use of non-numerical preference elicitation methods. Numeracy tests more appropriate for preference elicitation when no explicit considerations of uncertainty are made need to be explored and used. Further behavioural research is needed to fully understand the potential for using these methods in distinct settings (e.g. in different evaluation contexts and in face-to-face and non-face-to-face environments), as well as to explore the effect of literacy on assessments and on respondents\' preferences.