(PS3-B33) Examining Part C Providers’ Self-reported Modifications to a Social Communication Intervention During a Hybrid Type I Effectiveness-implementation Trial

Evidence-based social communication interventions are still not consistently available to young autistic children in community settings (Hume et al., 2021). To reduce this science-to-service gap, we need to “translate” interventions to community systems while maintaining effectiveness. Providers often modify EBPs when implementing them in community practice, yet these modifications are rarely well documented (Wiltsey-Stirman et al., 2019). The objective of this study was to document the nature of Early Intervention (EI) providers’ self-reported modifications to Caregiver-Implemented Reciprocal Imitation Teaching (CI-RIT).

Within a hybrid type 1 effectiveness-implementation trial (“RISE”), EI providers in the Washington, Michigan, Illinois, and Massachusetts Part C systems were trained in CI-RIT, an evidence-based intervention for social communication, and provided RIT as part of their regular practice. Providers completed a survey (informed by the FRAME and MADI adaptation frameworks) on the “nature” of the modifications they made to CI-RIT for a specific family after three months of use (n=39). Providers reported on modifications to CI-RIT made across their EI caseloads 12-months post-training (n = 19). Providers rated the extent to which they modified various aspects of CI-RIT (e.g., shortened/lengthened, added/removed content or coaching, changed the order of sessions/coaching) on a Likert scale from 1 (“never”) to 5 (“always”).

Providers reported modifying CI-RIT to various extents (range=1-5, Table 1). When reporting on caseload compared to family-specific modifications, providers reported making more modifications to RIT, M=2.89 vs. M=2.08, t(56)=-2.53, p=.01, and more often adjusting the order of RIT lessons, t(56)=-2.67, p=.02.

On family-specific surveys, providers reported more often modifying how they coached caregivers in CI-RIT than the lesson content, t(38)=-3.41, p< .001. E.g., they more often changed the order of the coaching activities than the order of the RIT lessons themselves, t(38)=-4.38, p< .001. Providers were more likely to modify CI-RIT by increasing than by decreasing aspects; they more often added than removed/skipped RIT content, t(38)=3.14, p< .01, and coaching strategies, t(38)=3.25, p< .01. Providers reported more often increasing the number of RIT sessions beyond 8 than decreasing to below 7, t(38)=3.10, p< .01. However, there was no difference to the frequency with which providers reported they lengthened vs. shortened the duration of CI-RIT sessions, t(37)=-1.50, p=.14.

Providers seemed to modify coaching style more than RIT content and to increase time teaching CI-RIT rather than decreasing it, which differs from prior work on RIT implementation (Ibañez et al., 2021). Providers reported more modifications to CI-RIT across their caseload relative to modifications reported for a specific family, which highlights the importance of collecting both kinds of modifications data. Future directions include using mixed methods and the MADI framework to examine self- vs. researcher-rated modifications, reasons why providers make modifications, and whether modifications relate to proximal implementation outcomes (CI-RIT feasibility, acceptability, and social validity).