Comparing multi-level and system-level strategies for the sustainability-implementation phase of collaborative care models: the Transform DepCare randomized clinical trial

Abstract Background Few studies test strategies for the sustainability-implementation phase of evidence-based practices like collaborative care models (CoCMs). Methods We conducted a randomized control hybrid 3 effectiveness-implementation trial of two theory-informed strategies at 5 primary care settings with mature CoCMs (i.e. already had programs in place): (1) system-level strategy: local technical assistance/learning collaborative meetings for Behavioral Health Providers (BHPs); (2) multi-level strategy: the system-level strategy, primary care provider (PCP) training/reminders/audit-feedback and in small subset of patients an activation/psychoeducation tool (DepCare educational material ). We randomized 44 PCPs. Their patients were eligible if they had elevated depressive +/- anxiety symptoms at a primary care (i.e. index) visit (patients could have ≥ 1 eligible visit with a PCP). We examined reach (e.g. completing a new CoCM visit and/or filling a new, intensified/augmented, or previously nonadherent antidepressant), adoption (e.g. CoCM referral by PCP and/or initiating/intensifying/switching/combining antidepressants) and reach - maintenance (≥2 fills and/or CoCM visits). We used multilevel logistic regression analysis (index visits nested within patients nested within PCPs). We thematically analyzed BHP meeting transcripts to elucidate mechanisms (i.e. mediating processes), determinants (hinderances/enablers), preconditions (necessary factors) and moderators at which strategies influenced outcomes. Results There were 605 CoCM eligible patients with 757 index visits post-strategy deployment. The mean age was 47.7 (SD = 17.3); 80% were female, 15% Black, 51% Hispanic; 41% were on an antidepressant. In the multi-level vs. system-level arm, reach was 39.1% vs. 44.9% (OR = 0.78; 95% CI 0.49, 1.22, p = 0.27); reach-maintenance 31.4% vs. 37.3% (OR = 0.77; 95% CI 0.50, 1.19, p = 0.23) and adoption 40.5% vs. 46.1% (OR = 0.74; 95% CI 0.42, 1.31, p = 0.29). Pre- vs. post reach improved in the multi-level (30.0% to 39.1%; (OR = 1.37, 95%CI 0.94, 2.00) and system-level (30.4% to 44.9%; OR = 1.93, 95%CI 1.31, 2.85) arms, with similar trends in adoption and reach-maintenance. Strategy mechanisms included improved BHP-PCP-patient communication and CoCM adaptability (i.e. number of visits, quality improvement); hinderances PCP resistance; moderators resources and PCP/patient motivation, and preconditions depression/anxiety screening quality, all described in the full text along with effectiveness and implementation (i.e. intervention/strategy fidelity) metrics. Conclusions There was no difference between system- and multi- (system plus PCP) level strategies for CoCM sustainability-implementation. Waning PCP involvement, despite clinician-targeted strategies, remains a key barrier to sustainability. Clinicaltrials.gov: NCT05085886. Submitted 9/28/2021