Comparing implementation strategies for optimizing depression care: A randomized control trial.

ImportanceLess than a third of depressed primary care patients experience clinical improvement, in part due to a lack of focus on treatment optimization (e.g., intensification). ObjectiveTo compare the impact of implementation and behavioral science informed system and multi-level strategies on population-wide treatment optimization in integrated/collaborative care model (CoCM) settings. DesignComparative effectiveness randomized controlled trial Setting5 Primary care clinics with a mature integrated/CoCM Participants44 primary care physicians and their patients with elevated depressive symptoms eligible for treatment optimization ExposuresSystem-level strategy (i.e., enhanced usual care [EUC]) focused on staff and behavioral health provider (BHP) activation vs. multi-level strategy (intervention) involving BHP activation, primary care provider (PCP) behavioral support and a patient activation/psychoeducation tool (DepCare) Main outcomes and measuresPatient optimization (e.g., filling a new, intensified/augmented, or previously nonadherent antidepressant and/or completing a new integrated/CoCM visit) during the 4 months following an index visit and PCP optimization (e.g., placing a referral for any integrated/CoCM service and/or initiating, intensifying, switching and/or combining antidepressant medications) at an index visit. We used multilevel logistic regression analysis (level 1 is the patient with an eligible visit, level 2 the PCP) to test our hypotheses. Odds ratios (ORs) and 95% CIs were based on these analyses. ResultsThere were 605 eligible patients with 757 visits in the post-implementation period. The mean age was 48 (SD=17); 486 (80%) were female, 15% Black, 51% Hispanic and 32% Spanish speaking; 41% were on an antidepressant. Patient treatment optimization in the intervention vs. EUC arms was 39.1% vs. 44.9% (OR=0.78; 95% CI 0.50, 1.22, p =0.27). Pre- vs. post-implementation, patient treatment optimization increased from 30.0% to 39.1% (p=0.10) and 30.4% to 44.9% (p=0.001) in the intervention and EUC arms (p=0.22 for differential change). There were similar trends in PCP optimization behaviors. There was low fidelity to the DepCare tool. Conclusions and relevanceOur study demonstrates little added benefit of a multi-level over a system-level strategy as it relates to treatment optimization, with only system-level strategies demonstrating pre-post improvements. Negative unintended impacts of multi-level, particularly clinician targeted, strategies should be explored. Key PointsO_ST_ABSQuestionC_ST_ABSIs a theory-informed system-level strategy better than a multi-level strategy for improving population wide depression treatment optimization in integrated primary care settings? FindingsIn this comparative effectiveness randomized control trial of 2 implementation strategies for improving depression treatment optimization in integrated care settings, a multi-level strategy was no better than a system-level strategy for improving patient and clinician treatment optimization behaviors. Only the system-level strategy exhibited significant pre-post improvement in patient optimization. MeaningThis is the first study to combine implementation and behavioral science to target treatment optimization in integrated care settings. We suggest that multi-level strategies that include clinician behavioral support may not be helpful and even harmful for improving population wide outcomes.