# Non-invasive respiratory support in paediatric critical care: protocol for a cohort study emulating the FIRST-line support for Assistance in Breathing in Children (FIRST-ABC) step-up randomised clinical trial using routinely collected data from the Paediatric Intensive Care Audit Network (PICANet) database

## TL;DR

This study will use real-world data to emulate a clinical trial comparing two respiratory support methods in children, aiming to confirm and expand on previous findings.

## Contribution

The study introduces a new-user design and combines frequentist and Bayesian methods to emulate a trial using routinely collected data.

## Key findings

- The study will benchmark emulated trial results against the original RCT findings.
- Advanced statistical methods will be used to minimize confounding in observational data.
- Results will inform the advantages and drawbacks of frequentist versus Bayesian approaches.

## Abstract

The development of the target trial emulation (TTE) methodology has enhanced the conduct of non-randomised studies. By leveraging readily available routinely collected data, TTEs offer opportunities for complementing randomised controlled trials (RCTs), providing more precise estimates and improving the external validity of RCTs. To explore this potential, we selected a successfully completed RCT as a case study. In the FIRST-line support for Assistance in Breathing in Children (FIRST-ABC) step-up RCT, high flow nasal cannula (HFNC) was found to be non-inferior to continuous positive airway pressure (CPAP) in terms of time to liberation from respiratory support in the paediatric critical care setting. We will emulate the FIRST-ABC step-up trial using routinely collected data from the Paediatric Intensive Care Audit Network (PICANet) database.

This is a protocol for a TTE that will use longitudinally collected data from the PICANet database. The study aims to emulate the FIRST-ABC step-up RCT using an observational study design in a frequentist framework. We will benchmark the results against the published trial. The study will apply a new-user design by selecting children admitted to paediatric intensive care units that started HFNC or non-invasive ventilatory support (as a surrogate for CPAP). The eligibility criteria and selected outcomes will reflect those of FIRST-ABC within the constraints of the available routinely collected data. We will use advanced quantitative doubly robust methods to minimise the impact of confounding by indication and allow for heterogeneity according to child characteristics. The analysis will be repeated using a Bayesian approach for follow-up research.

The research received ethics approval from the London School of Hygiene & Tropical Medicine Research Ethics Committee. This study will expand the findings from the FIRST-ABC step-up RCT, providing additional insight from a large representative sample using real-world data. The frequentist and Bayesian approaches will enable a discussion about the advantages and drawbacks of the two strategies. The results will be disseminated to the research and clinical community and made accessible to the public. In addition, the study results will be used in future research, which aims to supplement RCTs with additional evidence from a TTE.