Target trial emulation in cystic fibrosis: optimising methods for evaluating antibiotic treatment using registry data

Research question This project address two key research questions: what are the optimal methodological practices for target trial emulation using UK Cystic Fibrosis (CF) Registry data, and what are the long-term impacts of antibiotics in CF? Background CF is an inherited disease affecting over 11,000 people in the UK.

People with CF require multiple long-term treatments, and a key research priority in CF is to investigate the long-term effects of treatments. Such questions are unlikely to be addressed in randomised controlled trials (RCTs) due to feasibility and cost. Real-world data provide an alternative resource for studying treatment effects, but come with challenges.

Target trial emulation (TTE) is a framework for studying treatment effects using real-world data, which applies the study design principles of RCTs to help avoid common biases.

TTE involves first describing the RCT we would like to conduct – the “target trial” – and then specifying how to emulate that trial using real-world data. TTE is increasingly used in many disease areas, but it s use in CF is so far limited.

Aims and objectives To establish best practices for application of TTE methodology using UK CF Registry data to estimate effects of CF treatments. To apply the methods to emulate three existing RCTs that have been conducted assessing the effects of antibiotics.

To extend Aim 2 to questions that have not yet been studied in RCTs, including to estimate longer-term effects of antibiotics and the effects in CFTR modulator users. Methods The UK CF Registry contains longitudinal data on over 99% of the UK CF population.

We will determine optimal methods for using different components of the data in TTE, including data on prescription dates and longitudinal data on clinical variables collected at annual review visits.

We will conduct emulations of three published RCTs, which will allow us to compare the different methodological approaches, and help us to establish the best ways of using the registry to define the variables needed for TTE and to illustrate its applicability in CF. We will then apply the methods to address previously unanswered questions about the impacts of antibiotics in CF.

Statistical analysis approaches will include propensity-score based methods, g-formula and doubly-robust methods.

Timelines for delivery Key outputs during this 16-month project will include practical guidance for the application of TTE using UK CF Registry data, and new evidence on the impacts of long-term use of antibiotics within CF.

Impact and dissemination A primary impact will be enabling well-designed, robust studies of treatment effectiveness using TTE and UK CF Registry data.

We will disseminate our findings to the CF research and clinical communities through an international collaboration focussed on the use of TTE in CF, a study website, conferences, and publications. Our findings on the impacts of long-term use of antibiotics will be relevant to the wider CF community.

We will disseminate to this community via CF Trust channels using material developed based on focus groups. Longer-term impacts will include encouraging use of TTE within patient registries in other disease areas.