Building the Driver Diagram: A Mixed-Methods Approach to Identify Causes of Unplanned Extubations in a Large Neonatal ICU

Background The key driver diagram (KDD) is an important tool used by improvement teams to guide and frame their work. Methods to build a KDD when little relevant literature or reliable local data exist are poorly described. This article describes the process used in our neonatal ICU (NICU) to build a KDD to decrease unplanned extubations (UE) in chronically ventilated infants. Methods Twenty-seven factors hypothesized to be associated with UE in our NICU were identified. An expert panel of 33 staff members completed three rounds of a modified Delphi process administered through an online interface. After the third round, panel members provided suggestions for interventions to target all factors meeting criteria for consensus. These qualitative data were analyzed by inductive thematic analysis. A follow-up survey to all panel members was used to assess the feasibility of this process for future use. Results After three Delphi rounds, 14 factors met consensus and eight main interventions were identified through thematic analysis. These data were used to build a KDD for testing. All participants who completed the follow-up survey (20/20) stated willingness to participate in this process in the future and 18/20 (90%) stated they would be "more willing" or "much more willing" to support interventions developed using this process. Conclusion A novel mixed-methods approach was used to generate a KDD combining a Delphi process with thematic analysis. This approach provides improvement teams a rigorous and reproducible method to understand local context, generate consensus KDDs, and improve local buy-in for improvement interventions.