Why do people expect antibiotics when they should not?

抗生素耐药性是现代最大的全球健康风险之一。临床初级保健中抗生素处方不当是抗生素耐药性的关键驱动因素之一。过去的研究发现，患者的期望是临床医生决定开抗生素的最有力预测因素之一（Macfarlane, Holmes, Macfarlane & Britten, 1997; Sirota, Round, Samaranayaka & Kostopoulou, 2017）。患者期望和过度处方之间的联系得到了很好的支持，但是这些期望背后的机制仍然不清楚。因此，我提出的研究旨在回答为什么人们在临床上不合适的情况下期望使用抗生素的问题。我将测试一个旨在解释这一现象的心理学理论。具体来说，我将使用基于效用的信号检测理论（Lynn & Barrett, 2014）来衡量人们期望抗生素的最佳时间，并通过操纵模型定义的三个环境参数——收益、基本率、相似性——来改变他们对抗生素期望的最佳标准。例如，该模型预测，当人们在诊断不确定的情况下低估了与抗生素耐药性相关的成本时，对他们来说，期待抗生素是最理想的。让人们意识到这一成本，改变了最佳标准，从而降低了抗生素的预期（Sirota, Thorpe & Juanchich, in prep.）。这种方法已经成功地应用于这种情况（Sirota等人，in prep.），但是缺少定量模型预测的测试。基于这三个参数，我们将根据模型指定的适当的数学函数和高级计算，计算每个可能的标准位置的期望效用。目的是确定哪些参数以及它们在多大程度上影响最佳标准位置，以及何时将不适当的抗生素期望降至最低。利用该模型的框架，我将进一步测试如何以一种可理解和有说服力的方式最好地传达与抗生素过度使用相关的风险（即抗生素耐药性），以减少人们的不适当期望。我将操纵风险表现的不同因素（例如，统计证据与叙事案例，时间和空间接近性，个人与社会影响），同时使用小插图方法，主要目标是找到抗生素耐药性的最佳框架格式，以减少不适当的期望。数据将使用标准的频率分析方法（例如方差分析）、信号检测框架的建模方法和贝叶斯分析来补充分析。拟议的研究对理论和相当大的潜在社会影响具有启示意义。它将提供人们对抗生素期望背后的精确和可测试的机制，这将为我们理解人们对抗生素期望不适合临床的心理原因做出重大贡献。它将提供以证据为基础的战略，制定人们能够接受的关于抗生素耐药性的信息。该提案可以帮助告知国家努力减少对抗生素的期望，从而减少英国的抗生素过度处方，并有助于当前旨在减少抗生素耐药性未来传播的全球努力。

Antibiotic resistance is one of the greatest global health risks of modern times. Clinically inappropriate prescribing of antibiotics in primary care is one of the key drivers of antibiotic resistance. Past research has found that patients' expectations are among the strongest predictors of clinicians' decisions to prescribe antibiotics (Macfarlane, Holmes, Macfarlane & Britten, 1997; Sirota, Round, Samaranayaka & Kostopoulou, 2017). The link between patients' expectations and over-prescribing is well supported, however the mechanisms underlying these expectations still remain unclear. My proposed research, therefore, aims to answer the question of why people expect antibiotics when it is not clinically appropriate. I will test a psychological theory that aims to explain this. Specifically, I will use the utility based signal detection theory (Lynn & Barrett, 2014) to measure when it is optimal for people to expect antibiotics and by manipulating the three environmental parameters defined by the model-payoff, base rate, similarity-change the optimal criterion in terms of their expectations for antibiotics. For instance, the model predicts that when people underestimate the costs linked with the antibiotic resistance in situations of diagnostic uncertainty then it is optimal for them to expect antibiotics. Making people aware of this cost, changes the optimal criterion and hence decreases the expectation of antibiotics (Sirota, Thorpe & Juanchich, in prep.). This approach has been successfully applied to this context (Sirota et al., in prep.), but tests of quantitative model-predictions are missing. Based on the three parameters, we will calculate the expected utility for every possible criterion location, according to the appropriate mathematical functions and advanced computations as specified by the model. The aim is to identify which parameters and to what extent they affect the optimal criterion location, and when inappropriate antibiotic expectations will be minimised. Drawing on the framework of the model, I will further test how the risks linked with antibiotic overuse (i.e., antibiotic resistance) can best be communicated in an understandable and persuasive manner to reduce people's inappropriate expectations. I will manipulate different factors of risk representation (e.g. statistical evidence vs narrative cases, temporal and spatial proximity, personal vs societal impact), while using vignette approaches, with the main goal to find the optimal framing format of antibiotic resistance that will reduce inappropriate expectations. Data will be analysed using standard frequentist methods (e.g. ANOVA), modelling approaches of the signal detection framework and the analyses will be complemented by Bayesian analyses. The proposed research has implications for the theory and considerable potential for societal impact. It will provide exact and testable mechanisms behind people's antibiotic expectations, which will provide substantial contribution to our understanding of the psychological reasons why people expect antibiotics that are not clinically appropriate. It will provide evidence-based strategy to formulate messages about antibiotic resistance to which people will be receptive. The proposal can help inform national efforts to reduce antibiotic expectations and consequently antibiotic over-prescription in the UK and contribute to current global efforts aimed at reducing future spread of antibiotic resistance.