Kim Min Jung, Prieto-Merino David, Nicholas Jennifer, Allen Luke, Burton Matthew J, Bastawrous Andrew, Macleod David
Faculty of Epidemiology and Population Health, London School of Hygiene & Tropical Medicine, London, UK.
Universidad de Alcalá, Madrid, Spain.
J Eval Clin Pract. 2025 Aug;31(5):e70197. doi: 10.1111/jep.70197.
Quality improvement (QI) in health service programmes aims to make small, incremental changes to increase reach and efficiency. Simple, low-risk programmatic changes can improve services, particularly when supported by robust evidence. However, in health service contexts, there is tension between the need for swift decision-making and the high research standards for conducting methodologically rigorous trials. Randomized trials are rarely used to evaluate these changes due to high costs and long timelines, especially when the changes are expected to result in marginal improvements. Instead, health service programmes frequently introduce changes informed by anecdotal evidence or less robust evaluation methods such as before-and-after comparisons.
In this paper, we present a narrative review of the concepts underlying Bayesian adaptive trial designs for conducting QI research, highlighting their use in the commercial sector and exploring opportunities for cross-industry learning and future application in healthcare settings.
Relevant studies were selected based on their contextual relevance to the topic, in keeping with the narrative review approach.
Given that programmatic changes typically yield modest improvements, we recommend that adaptive trial designs can strike a balance between obtaining reliable results and avoiding overly large sample sizes. We review how interim analysis and early stopping can be integrated into trials, allowing the level of rigour to be adjusted according to the proramme specifications.
Adaptive trial designs hold significant promise for enhancing the QI efforts. To ensure that adaptive trial designs can be successfully integrated into health service contexts, tradeoffs should be made between methodological rigour and resource constraints.
卫生服务项目中的质量改进（QI）旨在进行微小的渐进式变革，以扩大覆盖范围并提高效率。简单、低风险的项目变革可以改善服务，特别是在有有力证据支持的情况下。然而，在卫生服务环境中，快速决策的需求与进行方法严谨的试验所需的高研究标准之间存在矛盾。由于成本高和时间长，随机试验很少用于评估这些变革，尤其是当预期变革只会带来微小改善时。相反，卫生服务项目经常根据轶事证据或不太可靠的评估方法（如前后比较）来引入变革。
在本文中，我们对用于开展质量改进研究的贝叶斯适应性试验设计的基本概念进行了叙述性综述，强调其在商业领域的应用，并探索跨行业学习的机会以及在医疗环境中的未来应用。
根据与主题的背景相关性选择相关研究，符合叙述性综述方法。
鉴于项目变革通常只会带来适度改善，我们建议适应性试验设计可以在获得可靠结果和避免样本量过大之间取得平衡。我们回顾了如何将中期分析和早期终止纳入试验，使严谨程度能够根据项目规范进行调整。
适应性试验设计在加强质量改进工作方面具有巨大潜力。为确保适应性试验设计能够成功融入卫生服务环境，应在方法严谨性和资源限制之间进行权衡。
