Effects of cognitive stimulation therapy on patients with dementia: An umbrella review of systematic reviews and meta-analyses

For nursing students

Study summary

Dementia causes progressive cognitive decline that makes everyday activities harder to manage, and cognitive stimulation therapy (CST) is one of the most widely used non-drug interventions offered to people living with the condition. This 2023 paper, published in Experimental Gerontology by Cao and colleagues, is an umbrella review — a review of reviews. Instead of analyzing individual trials, the authors pulled together existing systematic reviews and meta-analyses on CST to see what the overall body of evidence actually shows, and how trustworthy that evidence is.

The team searched eight databases, including Cochrane Library, PubMed, CINAHL, Web of Science, and several Chinese-language databases (CNKI, Wanfang Data, CBMdisc, VIP), from each database's inception through December 31, 2022. They registered their protocol in advance with PROSPERO, the international registry for systematic review protocols, which adds transparency to the process. To judge how well-conducted each included review was, they used AMSTAR 2, a standardized 16-item checklist for appraising systematic reviews. Reviews that scored 9 or more points (moderate quality or better) were analyzed further using GRADE, a widely used framework for rating the certainty of evidence behind specific outcomes.

Fourteen systematic reviews and meta-analyses met the criteria for inclusion. Most were judged to be of moderate methodological quality on AMSTAR 2 — meaning the field has real reviews to draw on, but not many top-tier ones. The authors mapped out how CST was actually delivered across these studies: who provided it, how often sessions happened, how long programs ran, and where they took place (for example, care homes versus community settings). They also examined evidence across eight separate health outcomes: cognition, depression, behavioral and psychological symptoms, quality of life, activities of daily living (ADL), language and communication, anxiety, and memory.

The clearest and most consistent finding concerned cognition. Eleven of the included reviews, with overall confidence ratings ranging from low to high, consistently reported that CST significantly improves cognitive function in people with dementia, and some of this supporting evidence was rated high quality. This makes cognition the outcome with the strongest and most reproducible evidence base for CST.

The picture was much less clear for other outcomes. Effects on depression, behavioral symptoms, quality of life, and activities of daily living were inconsistent across the reviews, and the evidence behind them was rated only low to moderate quality — meaning we can't yet say with confidence whether CST reliably improves these areas. Language and communication, anxiety, and memory were each examined by only a handful of studies, so there simply isn't enough research volume yet to draw firm conclusions about CST's effect on these outcomes.

The authors concluded that CST is supported as an effective intervention for improving cognitive function specifically, and that multi-component CST programs (those combining several types of activities) tend to outperform single-component programs, provided they are delivered on a regular, ongoing basis. They also called for future systematic reviews in this area to be designed and reported using rigorous quality frameworks like AMSTAR 2, since inconsistent reporting quality was itself a limitation of the current evidence base.

For nursing students, this study is a useful example of evidence hierarchy in action: an umbrella review sits above individual trials and single systematic reviews, letting you see where the evidence is genuinely strong (cognition) versus where it is still emerging or mixed (mood, behavior, function, communication). It's a reminder that a popular, widely recommended therapy can have solid evidence for one outcome while still needing more research for others — and that reading the quality rating behind a claim matters just as much as the claim itself.

For educators

Teaching and appraisal notes

This umbrella review by Cao et al. (2023), published in Experimental Gerontology, synthesizes evidence from systematic reviews and meta-analyses on cognitive stimulation therapy (CST) for people with dementia. Umbrella reviews sit at the top of the evidence hierarchy for intervention questions with a mature review literature, and this study is a good teaching case for how methodological rigor is assessed at that level, using AMSTAR 2 to appraise included reviews and GRADE to rate certainty of evidence for specific outcomes once a quality threshold (AMSTAR 2 score of 9+) was met.

Methodologically, the search strategy spanned eight databases — Cochrane Library, PubMed, CINAHL, Web of Science, and four Chinese-language databases (CNKI, Wanfang Data, CBMdisc, VIP) — searched from inception to December 31, 2022, with a PROSPERO-registered protocol (CRD42022364259). This breadth is a strength worth highlighting to students: including Chinese-language databases alongside the standard English-language sources reduces language bias, a common blind spot in reviews conducted solely in Western databases.

Fourteen systematic reviews/meta-analyses met inclusion criteria; most were rated moderate quality on AMSTAR 2, which is instructive in itself — it signals that even a mature-seeming intervention literature can lack a critical mass of high-quality reviews. The authors mapped intervention characteristics (content, provider, frequency, duration, setting) across the 14 reviews and evaluated evidence across eight outcome domains: cognition, depression, behavioral/psychological symptoms, quality of life, activities of daily living, language/communication, anxiety, and memory.

The central finding for classroom discussion is the split between strong and weak outcome evidence. Eleven reviews, with overall confidence ratings spanning low to high (including some high-quality evidence), consistently supported CST's benefit for cognition specifically. In contrast, evidence for depression, behavioral symptoms, quality of life, and ADL was inconsistent and only low-to-moderate quality, while language/communication, anxiety, and memory were addressed by too few studies to support any conclusion. This is a strong opportunity to discuss GRADE's distinction between the direction of an effect and the certainty behind it — students should be pushed to articulate why an intervention can be "supported" for one outcome and "unclear" for another within the very same evidence base.

The authors' practical conclusion — that multi-component CST delivered regularly outperforms single-component, irregular delivery — is a useful bridge to clinical application discussions: what does "multi-component" and "regular delivery" look like operationally in a long-term care or memory clinic setting, and what are the staffing and training implications of delivering CST with fidelity?

For faculty use, this paper works well in courses on evidence appraisal, gerontological nursing, and non-pharmacological dementia care. Suggested classroom exercises: (1) have students construct a GRADE evidence table for the eight outcome domains from the findings described; (2) have students critique why the authors limited GRADE analysis to only reviews scoring 9+ on AMSTAR 2, and what that threshold decision does to the review's inclusiveness versus rigor trade-off; (3) discuss the implications of 'inconsistent' behavioral-symptom evidence for a nurse deciding whether to recommend CST as an add-on for a resident's behavioral care plan versus purely for cognitive maintenance goals.