Designing and Implementing a Run-In Phase to Ensure Site Performance for a Multicenter Neonatal Ventilator Trial

Multicenter clinical trials evaluating performance-dependent interventions are particularly vulnerable to variability in technical proficiency and protocol adherence across participating sites. Despite the potential impact of this variability on trial validity, there is little published guidance describing how to prepare and qualify sites before subject randomization. We describe the rationale, design, implementation, and lessons learned from a structured Run-In Phase developed for the multicenter Diaphragmatic Initiated Ventilatory Assist (DIVA) trial and to provide practical recommendations for future performance-based clinical trials.

The DIVA trial incorporated a prospective site-level Run-In Phase before randomization to evaluate each site's ability to consistently deliver non-invasive neurally adjusted ventilatory assist (NIV-NAVA). The Run-In Phase included formation of an expert technical committee, standardized educational materials, boot camp and on-site multidisciplinary training, continuous technical support, and objective assessment of technical competency using predefined performance metrics. Sites were required to demonstrate successful management of a minimum number of patients and adherence to protocol-defined standards before progressing to the randomized phase. Perspectives from principal investigators, site investigators, research coordinators, and the data coordinating center were incorporated to evaluate the operational impact of the Run-In Phase.

A structured Run-In Phase provides a systematic framework for assessing and supporting intervention fidelity, reducing performance-related variability, and preparing sites for multicenter trials involving technically complex, operator-dependent interventions. Although resource intensive, a structured Run-In Phase provides a reproducible framework for preparing sites to deliver technically complex, performance-dependent interventions and for identifying and addressing implementation challenges before initiation of the randomized phase. We recommend broader adoption and standardized reporting of Run-In Phases in multicenter performance-based clinical trials.