Using Consensus Group Methods to Formulate Core Components of the Post‐Incarceration Engagement Program

Abstract

Research ObjectiveFor successful implementation and spread of an intervention, identifying its core components is essential for fidelity, adaptation, replication, and evaluation. However, an intervention’s core components are often gradually refined, as the intervention’s outcomes are repeatedly examined alongside actually implemented components. When empirical evidence is limited, as is often the case with early implementation of innovative interventions, consensus group methods can be used to gather and synthesize expert opinions to determine and refine intervention core components. We used aspects of the Delphi method, hybrid Delphi method, and nominal group technique to determine an intervention’s initial core components.Study DesignThis approach was devised in the context of the VA’s Post‐Incarceration Engagement (PIE) intervention, which uses peer support to provide better linkage to health care and other needed services to veterans recently released from incarceration. Utilizing an iterative approach, a four‐person expert panel of PIE developers/implementers and a two‐person moderator team followed these steps: Panel members individually suggested PIE intervention core components using a worksheet. Moderators reviewed the suggested components, removing duplicates and consolidated them into a preliminary list of core components categorized by theme. Moderators facilitated a panel meeting to discuss and clarify overlapping/distinct suggestions. Moderators drafted an updated list of core components and their definitions, based on the facilitated panel meeting. Panel members individually suggested revisions to the list, using an updated structured worksheet. Moderators reviewed the suggested revisions, incorporating ones that were consistent across the panel and organizing suggestions into the list. To resolve inconsistent suggestions, moderators may facilitate another panel meeting for consensus reaching and then return to Step 4. If not, the latest list is considered to reflect the initial core components of the innovation. Population StudiedThe PIE intervention was conducted by two peer support specialists over a two‐year period in one state. The consensus process included an expert panel of four study team members and two moderators.Principal FindingsThe PIE intervention has three core aims: improve linkage, social support, and skill building for veterans returning to community after incarceration. The consensus effort identified 20 core components at preimplementation, implementation, and sustainment stages. The identified components were sorted into three domains: (i) onboarding of peers; (ii) peers’ veteran‐facing work; and (iii) ongoing peer supervision, coordination, and networking. Iterative review of components determined a final set of 21 core components, refined definitions of roles and tasks, and specified preconditions of implementation.ConclusionsThe iterative hybrid Delphi approach allowed the research team to determine the initial core components of a complex intervention. As scale‐up continues, analyzing modifications made to the intervention during its spread to other states will be essential to determining the stability of core components.Implications for Policy or PracticeWhen implementing an intervention, specifying the core components is critical to (i) assess the success of the implementation and (ii) prioritize components when resources are limited. The iterative hybrid Delphi approach offers steps toward determining core components of interventions that have yet to be tested.Primary Funding SourceDepartment of Veterans Affairs.