Feasibility Study: A Mathematical Language for Complex Healthcare Interventions

随机临床试验是评估医疗保健领域新疗法或竞争性疗法的黄金标准。众所周知，它们用于比较新的药物治疗，其中最常见的试验类型是将新药与标准药物进行比较。然而，随机试验越来越广泛地应用于其他领域，如精神病学、外科和社会护理。在这些领域，要评估的保健干预措施往往更为复杂，由不同组成部分的不同组合组成。例如，精神病学干预的有效性可能取决于书面材料呈现的形式，治疗过程的时间长度和治疗师。计划这样的试验可能比简单地随机分配治疗对象要复杂得多，特别是在随机化有一些限制的情况下，例如，一个中心的所有患者可能必须看同一个治疗师，但可以给不同形式的书面材料。因此，必须仔细选择计划此类试验和分析从中收集的数据的方法，以确保发现和获得所有所需的信息而不带偏见。这种复杂的干预可以用实验统计设计的数学语言进行分类和描述。这被广泛用于工程、实验室科学和其他领域的实验，但很少用于临床试验。这个项目的目的是将不同类型的具有复杂干预的临床试验映射到这个数学框架上。这将使我们能够确定需要哪些方法来设计和分析特定类型的试验。然后，这些方法将被翻译成临床试验人员能够理解的语言，以便它们可以在未来的试验中使用。预计这项工作将导致我们在某些类型的临床试验的实验设计方法上发现一些差距。因此，这将用于指导未来实验设计方法的研究，以及在临床试验中的应用。从长远来看，这项工作将使医疗保健研究人员能够加快对新干预措施的测试，并获得更有力的证据，证明它们将在一系列现实生活场景中发挥作用。

Randomised clinical trials are the gold standard in assessing new, or competing, treatments in healthcare. It is well known that they are used in comparing new pharmaceutical treatments, where the most common type of trial compares a new drug with a standard. However, randomised trials are increasingly widely used in other areas such as psychiatry, surgery and social care. In these areas the healthcare interventions to be assessed are often more complex, being made up of different combinations of components. For example, the effectiveness of a psychiatric intervention might depend on the form in which written material is presented, the length of time of the course of treatment and the therapist. Planning such trials can be more complex than simply randomly allocating subjects to treatments, especially where there are some restrictions in the randomisation, e.g. all patients at one centre might have to see the same therapist, but can be given written material in different forms. Therefore methods for planning such trials and for analysing the data collected from them have to be chosen carefully in order to make sure all the required information is discovered and obtained without bias.Such complex interventions can be classified and described using the mathematical language of the statistical design of experiments. This is widely used in experiments in engineering, laboratory science and elsewhere, but rarely in clinical trials. The aim of this project is to map the different types of clinical trials with complex interventions onto this mathematical framework. This will allow us to identify which methods are needed to design and analyse particular types of trial. These methods will then be translated back into the language understood by clinical triallists, so that they can be adapted to use in future trials.It is expected that this work will lead to us identifying some gaps in the methodology of the design of experiment for some types of clinical trial. This will therefore be used to guide future research in the methodology of the design of experiments, as well as in the application to clinical trials. In the long term, this work will allow healthcare researchers to speed up the testing of new interventions and obtain more robust evidence that they will work in a range of real-life scenarios.