Touch-based therapies in dementia care: a systematic review and narrative synthesis

Touch-based therapies (massage, acupressure, reflexology/shiatsu, and therapeutic/healing touch) are used in dementia care, but effectiveness remains uncertain. The authors evaluated their impact on behavioural and psychological symptoms of dementia (BPSD) and pain, and extracted pragmatic “dose” and delivery parameters to inform a research blueprint. The authors searched major databases (MEDLINE, CINAHL, PsycINFO, Embase, CENTRAL) for studies from January 2005 to February 2023 involving people with any dementia aetiology/severity in community, residential, or inpatient settings. Eligible designs included randomised, quasi-experimental, and pre–post studies with a comparator (usual care, attention/quiet presence, or sham/light-touch). Data were extracted to a prespecified template; study quality was appraised using CASP tools. Owing to substantial clinical and methodological heterogeneity, the authors conducted a structured narrative synthesis as opposed to meta-analysis.

Thirty-three studies met inclusion: 21 massage, 8 acupressure, 3 therapeutic/healing touch, and 2 reflexology/shiatsu. Most were in long-term care or inpatient settings. Interventions typically used brief, repeated sessions (5–20 minutes, several times per week for 2–6 weeks). The most consistent finding was short-term calming, particularly reductions in agitation immediately post-session or over brief treatment courses, with the clearest pattern for massage and acupressure. Effects on broader neuropsychiatric symptoms (e.g., NPI/NPI-NH domains) and pain were mixed. Where monitored, no serious adverse events were reported; minor transient issues (e.g., brief restlessness, skin sensitivity with aromatherapy oils) were infrequent and acceptability generally high. Risk of bias was mixed (≈49% low, 42% moderate, 9% high), and durability beyond 4–8 weeks was rarely assessed.

Current evidence provides preliminary indications that brief, touch‑based therapies may offer short‑term calming effects when used alongside person‑centred care, although certainty remains low and findings should be interpreted cautiously. The authors propose a pragmatic research blueprint that predefines session length, frequency, and course duration; uses attention/sham controls; adopts core outcomes (e.g., Cohen Mansfield Agitation Inventory (CMAI), Neuropsychiatric Inventory / Neuropsychiatric Inventory adapted for Nursing Homes (NPI/NPI-NH); Pain Assessment in Advanced Dementia (PAINAD) where relevant); ensures blinded assessment; and extends follow-up. The authors recommend that future work should prioritise feasibility/pilot studies, followed by adequately powered trials to determine effectiveness, durability, and scalability for practice.