Statistical analysis plan for the Identification and Treatment of Hypoxemic Respiratory Failure (HRF) and ARDS with Protection, Paralysis, and Proning: a type-1 hybrid stepped-wedge cluster randomized effectiveness-implementation study

Background: The Identification and Treatment of Hypoxemic Respiratory Failure (HRF) and ARDS with Protection, Paralysis, and Proning (TheraPPP) study is a type-1 hybrid stepped-wedge cluster randomized effectiveness-implementation study involving 17 adult Intensive Care Units (ICUs). This study will evaluate the effectiveness and implementation of an evidence-based, stakeholder-informed, multidisciplinary care pathway called Venting Wisely that standardizes the diagnosis and delivery of life-saving therapies for critically ill patients with Hypoxemic Respiratory Failure (HRF) and acute respiratory distress syndrome (ARDS). Objective: To describe a pre-specified statistical analysis plan (SAP) for the TheraPPP study prior to completion of recruitment, electronic data retrieval, and before any analysis has been conducted. Methods and analysis: The Statistical Analysis Plan (SAP) was designed by the principal investigators and senior biostatistician and reviewed in detail by the Venting Wisely Scientific Steering Group before being approved. This statistical analysis plan is reported in accordance with Guidelines for the Content of Statistical Analysis Plans in Clinical Trials. A study specific CONSORT diagram and baseline characteristics table were developed. We estimate a total of 18816 mechanically ventilated patients will be included in this study with 11424 patients pre-implementation and 7392 patients post implementation. Given that ARDS patients are an important subgroup within this study, we estimate that this will generate a sample size of 2688 sustained ARDS patients within our TheraPPP study cohort. The primary clinical outcome is 28-day ventilator free days (VFDs). For the primary analysis, we will compare the mean 28-day VFDs pre-implementation and post-implementation using a mixed effects linear regression model to account for clustering of patients within site. Secondary clinical outcomes will be similarly compared pre-implementation and post-implementation using mixed effects linear or logistic regression models, as appropriate. All models will be adjusted for age, sex, severity of illness (sequential organ failure assessment score on admission) and severity of hypoxemia on admission based on PF ratio, as well as type and size of ICU. Pre-specified subgroups will include patient sex, age, HRF, ARDS, Covid-19 and cardiac surgical status, body mass index (BMI), height, illness acuity, and ICU volume. Ethics and Trial Registration: The study has received ethics approval from the University of Calgary (20-0646) and the University of Alberta (pro00112232). The trial was registered with ClinicalTrials.gov (NCT04744298) prior to the enrollment of any patients on Feb 8, 2021.