Understanding adaptations in the Veteran Health Administration’s Transitions Nurse Program: refining methodology and pragmatic implications for scale-up

Abstract

Background: When complex health services interventions are implemented in real-world settings, adaptations are inevitable. Adaptations are changes made to an intervention, implementation strategy, or context prior to, during, and after implementation to improve uptake and fit. There is a growing interest in systematically documenting and understanding adaptations including what is changed, why, when, by whom, and with what impact. The rural Transitions Nurse Program (TNP) is a program in the Veterans Health Administration (VHA), designed to safely transition a rural veteran from a tertiary hospital back home. TNP has been implemented in multiple cohorts across 11 sites nationwide over 4 years. In this paper, we describe adaptations in five TNP sites from the first cohort of sites and implications for the scale-up of TNP and discuss lessons learned for the systematic documentation and analysis of adaptations. Methods: We used the Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) expanded version of the original Stirman framework to guide the rapid qualitative matrix analysis of adaptations. Adaptations were documented using multiple approaches: real-time database, semi-structured midpoint and exit interviews with implementors, and member checking with the implementation team. Interviews were recorded and transcribed. To combine multiple sources of adaptations, we used key domains from our framework and organized adaptations by time when the adaptation occurred (pre-, early, mid-, late implementation; sustainment) and categorized them as proactive or reactive. Results: Forty-one unique adaptations were reported during the study period. The most common type of adaptation was changes in target populations (patient enrollment criteria) followed by personnel changes (staff turnover). Most adaptations occurred during the mid-implementation time period and varied in number and type of adaptation. The reasons for this are discussed, and suggestions for future adaptation protocols are included. Conclusions: This study demonstrates the feasibility of systematically documenting adaptations using multiple methods across time points. Implementors were able to track adaptations in real time across the course of an intervention, which provided timely and actionable feedback to the implementation team overseeing the national roll-out of the program. Longitudinal semi-structured interviews can complement the real-time database and elicit reflective adaptations.