"Wrong, but Useful": Negotiating Uncertainty in Infectious Disease Modelling
"Wrong, but Useful": Negotiating Uncertainty in Infectious Disease Modelling
复制标题
DOI：
10.1371/journal.pone.0076277
复制
发表时间：
2013-10-16
期刊：
影响因子：
3.7
通讯作者：
Latham, Sophia M.
中科院分区：
文献类型：
--
作者：
Christley, Robert M.;Mort, Maggie;Latham, Sophia M.
传染病动力学模型要为传染病控制政策提供信息，模型必须能够预测；然而，众所周知，这样的预测永远不会是完美的。然而，共识是，尽管模型是不确定的，但有些模型可能会为有效的行动提供信息。这假定可以确定模型的质量，以便充分评估模型的不确定性，并决定是否应该“使用”模型，或者以何种方式或在什么条件下“使用”模型。我们利用一系列数据检查了建模中的不确定性：对科学家、决策者和顾问的采访，以及对政策文件、科学出版物和对主要牲畜流行病的重大调查报告的分析。我们表明，传染病模型中的不确定性话语是多层次的，灵活的，偶然的，嵌入在上下文中，并在谈判模型可信度中起着关键作用。我们认为，模型的可用性和稳定性是在网络和围绕它的话语中发生的协商的结果。这种协商采用了一系列话语手段，使传染病建模中的不确定性具有可塑性，可以适应“解释灵活性”。模型在面对不确定性时的效用是这种灵活性的功能，它允许的协商，以及模型输出在决策过程中被框架和解释的上下文。我们认为，与其主要基于对质量的信念，模型的有用性和权威性有时可能主要基于其在其所处的广泛社会和政治环境中的功能地位。
For infectious disease dynamical models to inform policy for containment of infectious diseases the models must be able to predict; however, it is well recognised that such prediction will never be perfect. Nevertheless, the consensus is that although models are uncertain, some may yet inform effective action. This assumes that the quality of a model can be ascertained in order to evaluate sufficiently model uncertainties, and to decide whether or not, or in what ways or under what conditions, the model should be 'used'. We examined uncertainty in modelling, utilising a range of data: interviews with scientists, policy-makers and advisors, and analysis of policy documents, scientific publications and reports of major inquiries into key livestock epidemics. We show that the discourse of uncertainty in infectious disease models is multi-layered, flexible, contingent, embedded in context and plays a critical role in negotiating model credibility. We argue that usability and stability of a model is an outcome of the negotiation that occurs within the networks and discourses surrounding it. This negotiation employs a range of discursive devices that renders uncertainty in infectious disease modelling a plastic quality that is amenable to 'interpretive flexibility'. The utility of models in the face of uncertainty is a function of this flexibility, the negotiation this allows, and the contexts in which model outputs are framed and interpreted in the decision making process. We contend that rather than being based predominantly on beliefs about quality, the usefulness and authority of a model may at times be primarily based on its functional status within the broad social and political environment in which it acts.