Efficacy of doll therapy intervention on behavioral and psychological symptoms of dementia: A randomized single‐blind controlled trial

Background Doll therapy (DT) is a non‐pharmacological intervention recommended for the treatment of behavioural and psychological symptoms of dementia (BPSD), producing benefits on agitation, aggression and wandering. By recognition of the doll as a real baby, challenging behaviours, viewed as requests for care and protection (i.e. attachment behaviours), are replaced with caregiving behaviors. In order to overcome methodological issues of the available studies, we have designed a single‐blind randomised controlled trial on efficacy of DT in managing BPSD and reducing professional caregivers’ perceived stress linked to BPSD (ClinicalTrials.gov, NCT03224143). Method The object presentation procedure (doll or cube, a non‐anthropomorphic object) is structured with the goal of recreating a situation of separation from a known figure and interaction with the environment in order to partially recreate the prototypical phases of the ‘strange situation’. Analysis of covariance (ANCOVA) compared the intervention group and the active control group on the net change (post‐treatment minus pre‐treatment) in NPI‐NH total score and NPI‐NH Distress score. Pre‐treatment NPI‐NH total score and psychotropic medications were used as covariates in these analyses. Result One hundred thirty‐four women living in 26 nursing homes in the Canton Ticino area (Switzerland) were randomized, 129 completed the study. The intervention group (n= 64) and the active control group (n= 65) were comparable: age [(mean (SD)=86.9 (5.9) and mean (SD)=88.4 (5.5), respectively; z=‐1.276, p=0.202)], baseline NPI‐NH total score [(mean (SD)=33.84 (13.28) and mean (SD)= 31.55 (12.1), respectively; z=‐1.05; p=0.292)], baseline NPI‐NH‐Distress total score [(mean (SD)=12.08 (5.4) and mean (SD)=10.6 (5.04), respectively; z=‐1.86, p=0.063), Global Deterioration Scale score [(mean (iqr)=6 (5‐7) and mean (iqr)=6 (5‐7), respectively; z=‐1.146, p=0.252). Based on ANCOVA results, the intervention group reported a significantly higher decrease on NPI‐NH total score [(adjusted mean (SE)=‐ 11,63 (1,0)] with respect to the active control group [adjusted mean (SE)=‐ 4.19 (1.0); F(1)=43.433, p=