Reappraising the Efficacy and Acceptability of Multicomponent Interventions for Caregiver Depression in Dementia: The Utility of Network Meta-Analysis

Abstract

Abstract Background Previous meta-analyses have demonstrated the efficacy of interventions for caregiver depression in dementia. However, they generally lumped the interventions together without accounting for the multicomponent nature of most interventions. It is unknown which combination of components contributed to the efficacy and should be implemented in clinical practice. Objectives To reappraise the interventions for caregiver depression from a recent systematic review, and evaluate whether a network meta-analytic approach offers additional insight into the efficacy and acceptability of multicomponent interventions. Research Design and Methods This study built upon a recent systematic review which identified 33 RCTs on caregiver depression from multiple databases (Medline, EMBASE, CENTRAL, PsycINFO, PSYNDEX). The 33 RCTs had different combinations of components serving at least one of the three key functions: addressing the care needs of persons with dementia (“CN”), addressing caregiving competency (“CC”), and providing emotional support for loss and grief (“ES”). A Bayesian network meta-analysis was conducted to compare the efficacy and all-cause attrition across different combinations of interventions. Results CN–CC combination was significantly more efficacious than treatment-as-usual (standardized mean difference in depression scores, SMD = −0.25, 95% credible interval [CrI] = −0.41 to −0.08) and marginally better than CN (SMD = −0.43, 95% CrI = −0.85 to −0.00). While none of the interventions clearly had both high efficacy and acceptability, three combinations of interventions (CC, CN–CC, and CN–CC–ES) had a reasonable balance between efficacy and acceptability. Discussion and Implications Although interventions for caregiver depression in dementia are effective in general, the different components of interventions may not share the same efficacy and acceptability. In implementing interventions, policymakers may consider addressing CC first, introducing CN in a graded manner, and providing ES only when indicated. Future studies may also consider using network meta-analysis to gain additional insights on how to implement multicomponent interventions in geriatric care.