The implementation of evidence-based practices such as kangaroo care (KC), a practice provided in clinical neonatal care, is subject to a complex and dynamic interplay of factors. Using traditional methods to capture and assess this complexity and underlying mechanisms related to implementation, has often proven difficult. This study applied system dynamics and network diffusion modeling to map implementation determinants of KC implementation and to examine underlying implementation mechanisms through which these factors influence the average duration of KC within the complex adaptive implementation system of neonatal intensive care units. A causal loop diagram (CLD) was developed through a systems mapping exercise with 32 multiprofessional KC experts and validated through interviews with participants with lived KC experience. The CLD was subsequently converted into a network graph to derive a network diffusion model. The model estimated the influence of cascades initiated by selected variables (‘seeds’) on the outcome variable of Average hours of KC provided. The CLD comprised 73 variables and 155 connections across three interconnected domains: parental engagement, workforce capacity and service delivery context. In the network diffusion model, Nurses per Shift and Availability of KC Education & Counseling initiated the strongest influential cascades on average time of KC provided. This study demonstrates the utility of integrating system dynamics and network science to investigate complex mechanisms underlying implementation processes. Further applications are needed to validate and refine these methods and to strengthen their contribution to research surrounding implementation mechanisms.
Nyantakyi et al. (Fri,) studied this question.