P2.G124: Using a Quality Improvement Initiative to Support a Randomized Control Trial Exploring the Use of Virtual Family-Centered Rounds in a Neonatal Intensive Care Unit

Purpose/Objectives: Virtual family-centered rounds (VFCR) uses telehealth to allow families to participate virtually on rounds. We present a quality improvement (QI) project nested within a cluster RCT of VFCR in a neonatal intensive care unit (NICU). We established four SMART aims with the overall objective to optimize implementation of the trial.

Design/Methods: Our four SMART aims targeted high trial intervention arm subscription to using VFCR (goal 75%), high VFCR intervention protocol adherence (goal 95%), low control arm protocol deviation (goal 3%), and high parent survey response proportions for each survey period (goal 75%). An improvement team was established which comprised research assistants, pediatricians, neonatologists, and neonatology fellows. Additionally, a stakeholder engagement team of NICU parents and healthcare professionals facilitated the trial's design and monitoring. We employed Plan-Do-Study-Act cycles and used run charts to monitor the four outcome measures relevant to each SMART Aim (Figure 1). The creation and implementation of our tests of change were guided by our key driver diagram (Figure 2). Adopted strategies included refinements to the wording of family-facing materials, delegation of roles for launching VFCR encounters, use of weekly gift card incentives to the NICU team for high protocol adherence, calls from NICU providers to intervention arm families who did not receive their virtual invite to VFCR, and optimization of visit launch workflow by leveraging existing EHR software features.

Results: During the 12-month trial period, we met or exceeded 3 out of 4 of our SMART aims. Subscription proportion exceeded our target, achieving 93.5%. Tests of change improved the mean from 80.9% to 95.5%. Mean protocol adherence was 95.8% (Figure 1); tests of change improved the mean from 93.4% to 97.2%. Mean control arm protocol deviation was minimal at 0.6%. Survey response fell below the 75% goal for each survey period despite adjustments in survey delivery and follow-up processes. Survey response at day 0, 30, 60, and 90 days after discharge was 68.8, 59.2, 54.5, and 52.8%, respectively.

Conclusion/Discussion: We successfully applied QI methods to support the implementation of a RCT, achieving significant adherence to trial protocols and survey data collection.