Yu Peng, Yang Liu, Zhongxian Guo, Yuhan Zhang, Liyan Sha, Xiaorun Wang, Yang He
8 min
As the global population ages, dementia-related behavioral and psychological symptoms (BPSD)—including agitation, apathy, irritability, and wandering—place a significant burden on both nursing home residents and care systems. This systematic review and meta-analysis synthesized data from 10 studies to evaluate the efficacy of doll therapy (DT) as a nonpharmacological intervention for older adults in institutional settings. By providing dolls as a therapeutic tool, the intervention aims to tap into innate attachment needs, potentially soothing residents and providing a safe, relational focus to alleviate distress.
The researchers conducted a rigorous meta-analysis of randomized controlled trials and quasi-experimental studies, evaluating 573 participants. They analyzed three primary outcome domains: behavior (agitation, apathy, irritability, and wandering), psychology (pleasure, anxiety, and depression), and cognitive function. The study also performed subgroup analyses to determine whether intervention characteristics—such as the type of doll used (empathy vs. lifelike) and the mode of training (caregiver coordination vs. health education)—influenced the therapeutic outcomes.
The meta-analysis found that doll therapy is an effective, nonpharmacological approach for managing behavioral and psychological disturbances. Residents receiving the intervention showed a significant reduction in agitation and other challenging behaviors. Furthermore, the therapy improved positive psychological states while decreasing symptoms of anxiety and depression. A notable finding from the subgroup analysis is that using empathy dolls, combined with active coordination and guidance from caregivers, yielded the most significant improvements. Conversely, the analysis found no significant evidence that doll therapy improves cognitive function, suggesting that the intervention acts primarily on the emotional and behavioral level rather than on core cognitive processes.
These findings support the use of doll therapy as a cost-efficient, sustainable, and humane intervention for institutionalized older adults with dementia. By focusing on attachment-based interaction, caregivers can address the emotional and behavioral root causes of resident distress without relying solely on psychotropic medications. This study provides actionable guidance for nursing homes on how to optimize these programs for better clinical outcomes.
Objective: To explore the effectiveness of doll therapy (DT) on behavior, psychology and cognition among older nursing home residents with dementia. Methods: A systematic review and meta-analysis was conducted. Subgroup analyses were performed to determine whether the intervention characteristics influenced effect sizes. Results: Ten studies met the inclusion criteria and were selected for qualitative and quantitative synthesis. The overall methodological quality was relatively high. DT significantly improved all behaviors [SMD=-0.42, P=0.01], including agitation [SMD=-0.94, P<0.001], apathy, irritability and wandering, and psychological states (i.e., pleasure, anxiety and depression). However, there was no significant difference in the improvement of cognition. Subgroup analyses revealed that the DT process employing empathy dolls and coordinating with caregivers was more beneficial for improving all behaviors (P=0.01; P=0.02). Conclusion: DT significantly reduced behavioral and psychological disturbances among older nursing home residents with dementia. Specifically, administering empathy dolls and coordinating with caregivers may be the most appropriate and effective option.
Alex: How did they turn such varied interactions into evidence they could compare?
Sam: They searched research databases and screened for randomized trials or other controlled intervention studies. The included studies covered 573 older adults, mostly with moderate to severe dementia. Participants needed sufficient movement, sensory function, and communication to interact with a doll.
Alex: That eligibility requirement already limits the reach. The results cannot automatically represent residents who cannot engage with the intervention.
Sam: And the studies did not all deliver the same program. Sessions and intervention periods varied substantially. Controls included standard care, no intervention, and alternatives intended to provide a comparison experience.
Alex: What did one of those alternatives look like? That seems important for separating a doll’s relational role from sensory stimulation.
Sam: In one Swiss study, residents received caregiver-guided sessions with an empathy doll. These dolls are designed to recreate sensations of holding a child and encourage caring responses. The comparison used a soft foam rubber cube covered in colored, velvety fabric.
Alex: That comparison is more informative than simply adding a doll to usual care. But across the review, the different controls mean the pooled result represents several comparisons.
Sam: It does. The authors pooled outcomes when their measurements were sufficiently comparable. Where they could not combine results, they described the individual studies. That distinction matters because not every positive outcome has a pooled effect estimate behind it.
Alex: Let’s take the strongest numerical result first. How much did behavior improve compared with the control conditions?
Sam: Overall behavioral disturbance fell by about four-tenths of a standard deviation relative to controls. That is a standardized mean difference, expressing effects across different measurement scales. The estimate favored doll therapy, and its confidence interval excluded no difference.
Alex: That gives us a common scale, but not a direct count of difficult episodes avoided. Was any particular symptom more clearly affected?
Sam: Agitation showed a larger estimated benefit: about nine-tenths of a standard deviation relative to controls. That estimate came from three studies. It supports a benefit for agitation, but does not tell us how many episodes a particular resident would avoid.
Alex: And what about residents who withdraw, wander, or seem persistently anxious? Were those findings equally well combined?
Sam: No. Individual studies reported improvements in apathy, meaning reduced motivation or engagement, as well as irritability and wandering. Psychological findings included greater pleasure and reduced anxiety or depression. Those results were described separately rather than combined into one pooled psychological effect.
Alex: So the review’s broad positive conclusion rests on different kinds of synthesis. Cognition sounds like the important contrast.
Sam: The main pooled analysis did not find a statistically significant cognitive benefit. That does not establish that doll therapy has no cognitive effect. It means this evidence did not clearly demonstrate one, and the authors call for further investigation.
Alex: The review also compares delivery approaches. How much confidence should we put in its preference for particular dolls and caregiver involvement?
Sam: Subgroup analyses favored empathy dolls over the other main category, lifelike dolls, for overall behavior. They also favored sessions coordinated with caregivers over health-education-based delivery. But these are comparisons across studies, not a direct randomized test of each program ingredient.
Alex: So a promising design clue, rather than proof of the best recipe. What else keeps the conclusions from being firmer?
Sam: The authors rated the outcome evidence moderate or low. Blinding participants and intervention staff was uncommon, although assessor blinding was more frequent. About four in five participants were women, which also limits confidence that responses would be similar across sexes.
Alex: My bigger concern is the caregiver. If someone guides, reassures, and validates a resident throughout a session, that interaction could itself be calming.
Sam: The authors explicitly identify that problem. They could not clearly separate the effects of professional interaction from the effects of the doll. Some studies also relied on observational measures, and there was no standardized way to assess the attachment relationship itself.
Alex: There is an ethical question here too. A potentially helpful intervention still needs to respect an adult’s preferences.
Sam: The review emphasizes autonomy, personal rights, and initial reactions to the doll. Caregivers should recognize preferences and help residents choose, rather than treat participation as automatic. Some participants refused, responded negatively or neutrally, or did not want to continue.
Alex: Given those qualifications, who should read the full review, and where should they begin?
Sam: Researchers developing dementia-care interventions should start with the outcome analyses, separating pooled findings from descriptive reports. Then read the intervention-characteristics discussion and the limitations, especially the caregiver-interaction problem. Care teams considering implementation should also examine the discussion of autonomy and person-centered delivery.
Alex: And for listeners outside dementia care, what is the line to carry away?
Sam: This review supports doll therapy more as a way to ease distress than to improve cognition. The doll may help, but the relationship around it remains part of the unanswered question.
Alex: Thanks for listening.