Using real world data and rapid learning to drive improvements in lung cancer survival

Background Radiotherapy plays a major role in curative-intent cancer treatment, but rarely considers the effect of comorbidities on patient outcome. Radiotherapy treatments evolve quickly, typically without formal evaluation of clinical impact. Conventional clinical trials are challenging for rapid and ongoing technical changes, and often under-represent patient groups (e.g. those with co-morbidities). Rapid-learning proposes using real-world data to evaluate such changes in practice and optimise them to the health status of local patient populations in successive learning cycles. This unmet need was highlighted in NHS England s 2019 pledge to develop a Learning Healthcare System for Radiotherapy Services . Recent data suggests limiting heart dose during lung cancer radiotherapy will improve survival. Aims 1. Use rapid-learning to introduce a heart dose-limit for lung cancer radiotherapy and optimise the limit to give the best clinical outcome. 2. Determine the potential and challenges of using rapid-learning across the NHS for the evidence-based introduction and optimisation of changes in radiotherapy treatments. Methods Work-package 1 will determine the ethical, legal and regulatory acceptability of using rapid-learning to change radiotherapy practice. Conceptual analysis will be combined with stakeholder engagement, using citizens juries to reach a consensus position. Work-package 2 will implement the statistical approaches needed to assess clinical impact using real-world radiotherapy data with a rapid-learning methodology. The heart dose-limit data we collect will be used to develop an individualised cardiac dose-limit risk model. Work-package 3 will demonstrate clinical effectiveness of rapid-learning by prospectively optimising heart dose-limits for lung cancer radiotherapy. At least two learning cycles will use real-world data to evaluate the effect of a new dose-limit on patient survival and treatment-related toxicity, and then successively refine the dose-limit. Work-package 4 will generate health-economic evidence of the effect of introducing a heart dose-limit, and evaluate the impact different rapid-learning resource uses will have on this. A patient-level decision-analytic model-based cost-effectiveness analysis will be undertaken using patient reported quality-of-life data from the clinical demonstration. Work-package 5 will investigate if rapid-learning is translatable. Case studies of historic changes in radiotherapy practice will be developed and the statistical approaches adapted for them. Robustness-to-bias analysis will indicate the susceptibility of evidence from different rapid-learning designs to bias in each case. Work-package 6 will investigate the feasibility and barriers to widespread implementation of rapid-learning across NHS radiotherapy departments. Evidence will be synthesised from structured interviews with healthcare professionals, professional societies, regulatory bodies and opinion-formers. Work-package 7 will achieve a consensus position on the role of rapid-learning in the NHS. It will aim to build awareness of rapid-learning in relevant communities and develop pathways to expand the use of the concept beyond the programme. Impact We expect our programme to improve the short-term survival of lung cancer patients (by 15-20%) and establish rapid-learning as a new NHS evaluation framework for technical changes in radiotherapy. We will deliver guidelines and practical tools (software, manuals) to guide the future use of both heart dose-limits and rapid-learning in radiotherapy. An integrated programme of engagement and communication will aim to maximise community involvement and uptake.