Strategies for Improving Memory: A Randomized Trial of Memory Groups for Older People (HOA & aMCI)

Paper Overview

• Journal: Journal of Alzheimer’s Disease, Vol. 49, 2016, pp. 31–43

• DOI: 10.3233/JAD-150378

• Study type: Randomized, controlled, cross-over clinical trial

• Focus: Effectiveness of a 6-week group-based memory-strategy intervention (LaTCH Memory Group) for

  • Healthy Older Adults (HOA)

  • Older adults with amnestic Mild Cognitive Impairment (aMCI)

• Sample: $n<em>{HOA}=113$, $n</em>{aMCI}=106$

• Primary outcomes: Knowledge & use of memory strategies, self-reported memory ability, objective memory tests, wellbeing

Background & Rationale

• Population ageing ➔ governments promote cognitive-health maintenance to reduce dementia burden

• aMCI characteristics

  • Objective memory loss ($\ge 1.5\,\text{SD}$ below age norms)

  • Rapid progression risk to Alzheimer’s disease $\approx 5\text{–}15\%\,/\text{year}$ vs $1\text{–}2\%$ in HOA

  • Prevalence in 70–90 y: $\le 21\%$

• Existing approaches

  • Computerised brain-training ⇢ mixed transfer to daily life

  • Mnemonic/compensatory strategies show promise; need real-world generalisation

• Behaviour-change theory

  • Knowledge ➔ Skill acquisition ➔ Self-efficacy ➔ Real-world implementation

  • Group settings foster peer modelling & motivation

Participants & Eligibility

• General inclusion

  • Age >60, community-dwelling, English proficiency

• HOA specific

  • No clinical memory concerns; all 4 episodic-memory screens >-1.5\,\text{SD}

• aMCI specific (Petersen/Winblad criteria)

  • Subjective or informant concern

  • $\le -1.5\,\text{SD}$ on ≥1 delayed-recall screen

  • CDR <1; ADL independence; MMSE $\ge 24$; no dementia

• Exclusions: Major co-morbidities, untreated sensory loss

• Recruitment flow

  • Screened $n=378$ ➔ excluded $n=138$ ➔ declined/unavailable $n=21$ ➔ enrolled $n=219$

  • Retained at 6-mo follow-up: $80\%$ overall; attrition HOA $11\%$ vs aMCI $31\%$

Study Design

• Two-ar block randomisation within diagnosis

  • Early-Intervention group: training immediately

  • Late-Intervention (wait-list control)

• Timeline

  • Pre-test ➔ 6-week course ➔ 2-week wait ➔ Early Post-test

  • Cross-over (late group receives course) ➔ 2-week wait ➔ Late Post-test

  • 6-month Follow-up for all

Intervention: LaTCH Memory Group

• Format: Six × 2-h weekly sessions; mixed HOA + aMCI; up to 12 participants; manualised

• Delivery team: Neuropsychologist / Occupational therapist + co-facilitator

• Core components

  • Education on brain health, ageing, lifestyle factors

  • Multi-strategy training:
    • Internal mnemonics – semantic association, imagery, retrieval practice, implementation intentions
    • External aids – diaries, smartphones, pill boxes, timers
    • Task organisation & goal-planning

  • Coping & self-efficacy modules (stress, confidence, peer support)

  • Weekly home assignments & written hand-outs

Outcome Measures

1. Strategy Knowledge

   • Strategy Repertoire Test (8 everyday scenarios; scores 0–2/response; $\alpha=0.78$)

2. Strategy Use

   • MMQ-Strategy (19 items; factors: Internal vs External strategies; $\alpha=0.83$)

3. Memory Ability

   • MMQ-Ability (20 items; $\alpha=0.93$)

   • CAMPROMPT (6 prospective-memory tasks)

   • CVLT-II Long-Delay Recall (episodic learning)

4. Wellbeing

   • MMQ-Contentment (18 items; $\alpha=0.95$)

Statistics Overview

• Intention-to-Treat; Missing-at-Random ➔ Expectation-Maximisation imputation

• Early-Post: ANCOVA (covariate = pre-test) within each diagnosis

• Late-Post & 6-mo: ANOVA on gain scores (pre-test baseline)

• Effect size reported: $\eta^{2}$ (small≈.01, medium≈.06, large≈.14)

• Power calculation: $n=45$/arm for medium effect, $\beta=0.20$

Key Results (Early-Post Intervention Effects)

HOA (large/moderate effects)

• Strategy Knowledge $\eta^{2}=0.20$

• Internal Strategy Use $\eta^{2}=0.18$; External $0.07$

• Self-rated Memory Ability $0.06$

• Wellbeing $0.11$

• Prospective Memory $0.02$ (small)

• No change on CVLT-II

aMCI (smaller effects)

• Strategy Knowledge $0.06$

• Internal Strategy Use $0.08$; External ≈0

• Wellbeing $0.05$

• No early gains on Memory Ability, CAMPROMPT, or CVLT-II

Long-Term Gains (Late-Post & 6-Month)

• Strategy Knowledge sustained only in HOA (gain ≈ $\Delta=7.35$ points)

• Strategy Use (internal & external) sustained in both groups, but HOA > aMCI

• Self-rated Memory Ability: short-term gain in HOA, none sustained

• CAMPROMPT: gains $\approx+5$ pts in both groups sustained at 6 mo

• CVLT-II: small sustained gain in HOA; transient only in aMCI

• Wellbeing: sustained to 6 mo in HOA; transient in aMCI

Interpretation & Significance

• Knowledge ➔ Behaviour: Intervention successfully moved participants along the behaviour-change continuum, especially HOA

• Real-world translation: Prospective-memory test (allows strategies) showed durable improvement; traditional list-learning less sensitive

• aMCI challenges

  • Semantic-knowledge degradation may limit retention of strategy repertoire

  • Suggest need for booster or extended sessions, or adjunct goal-oriented individual therapy

• Group dynamics fostered peer modelling, normalisation of memory lapses, and enhanced self-confidence

Limitations

• Wait-list (not active) control ➔ Cannot isolate effects of therapeutic contact or group support

• Possible test-retest/practice on neuropsych measures despite alternate forms

• Higher attrition in aMCI (31 %) may bias long-term estimates

• Six-week dose may be insufficient for consolidated gains in aMCI

Practical & Ethical Implications

• Low-cost, easily disseminated program suitable for community or outpatient settings

• Aligns with public-health goals of prolonging independence and delaying institutional care

• Encourages ethical responsibility to provide early, empowering interventions rather than passive monitoring of decline

Suggested Future Directions

• Incorporate $\ge$ 1 booster session or spaced retrieval format

• Blend group program with personalised, goal-driven training (e.g., GREAT model)

• Compare against active control (e.g., social or health-education group) to disentangle non-specific effects

• Longer follow-up (12–24 mo) & functional outcomes (IADL scales, caregiver burden)

Quick-Reference Numerical Highlights

• Risk of AD conversion in aMCI: $5\text{–}15\%\,/\text{yr}$ vs $1\text{–}2\%$ HOA

• Effect sizes (Strategy Knowledge): HOA $\eta^{2}=0.20$; aMCI $0.06$

• Prospective-memory gain at 6 mo: HOA $\approx+5.05$ pts; aMCI $\approx+4.09$ pts

• Attrition by 6 mo: HOA $11\%$; aMCI $31\%$

• Sample power target: $n=45$/arm (achieved $\ge52$)

Mnemonic Examples Shared in Sessions

• Name retrieval: visualize a facial feature and create a semantic link ("Bill ➔ dollar bill on his hat")

• Goal planning / implementation intention: "When I finish lunch, I will call my granddaughter."

• External aid pairing: set smartphone alarm labelled "Pick up grandchildren" synchronized with GPS location of supermarket

Connections to Prior Literature

• ACTIVE trial: showed small but persistent memory-test gains ($d=0.23$) over 5–10 y; current study parallels benefits but extends to aMCI

• Semantic-encoding training (Kirchhoff et al.) and associative-pair mnemonics (Hampstead et al.) support internal strategy modules

• Self-efficacy frameworks (Bandura; West et al.) underpin contentment findings

Core Takeaways for Exam Revision

• Group-based multi-strategy memory training yields sizeable benefits in strategy knowledge/use and wellbeing, especially in HOA.

• aMCI gains are modest and less durable; booster or tailored components likely needed.

• Prospective memory is an ecologically valid indicator of functional transfer.

• Behaviour-change theory (knowledge ➔ use ➔ performance) is a central conceptual scaffold for interpreting intervention effects.