Exercise interveNtion outdoor proJect in the cOmmunitY - results from the ENJOY program for independence in dementia: a feasibility pilot randomised controlled trial.

The intervention was the **ENJOY program** (Exercise interveNtion outdoor proJect in the cOmmunitY), a structured, supervised group exercise program delivered in public parks equipped with specialised "senior exercise parks" — outdoor fitness stations designed for older adults (e.g., leg press, chest press, balance beams, step-up platforms). The program included:

**Frequency:** Twice per week for 12 weeks (24 sessions total)

**Duration:** 60 minutes per session

**Content:** Warm-up (10 min), circuit-style resistance training using the outdoor equipment (30 min), balance and flexibility exercises (10 min), cool-down (10 min)

**Progression:** Exercises were individually progressed based on ability, with resistance increased via elastic bands or body position changes

**Supervision:** Sessions were led by trained exercise physiologists or physiotherapists

The **comparator** was usual care — participants continued their normal activities and received standard medical care, but no structured exercise program.

**Primary outcomes** were feasibility measures: recruitment rate, retention rate, adherence to sessions, adverse events, and acceptability (assessed via satisfaction questionnaire and qualitative interviews).

**Secondary outcomes** were physical function and cognitive performance:

Physical: 30-second chair stand test (number of stands), 4-metre gait speed (m/s), timed up-and-go (seconds), 6-minute walk distance (metres), balance (single-leg stance, seconds)

Cognitive: Montreal Cognitive Assessment (MoCA, 0–30, higher = better), Trail Making Test Part B (seconds, lower = better executive function), Rey Auditory Verbal Learning Test (RAVLT, words recalled, higher = better memory)

Psychosocial: Quality of life (EQ-5D-5L), depression (Geriatric Depression Scale), fear of falling (Falls Efficacy Scale-International), social connectedness (Lubben Social Network Scale)

**Sample size:** 50 participants (25 intervention, 25 control)

**Population:** Community-dwelling older adults aged 60–89 years (mean age 76.4 years, SD 6.8)

**Inclusion criteria:** Diagnosis of mild cognitive impairment (MCI) or early-stage dementia (confirmed by physician or specialist); able to walk independently (with or without walking aid); able to provide informed consent; English-speaking

**Exclusion criteria:** Significant medical conditions precluding exercise (unstable cardiovascular disease, recent fracture, severe arthritis); severe cognitive impairment (MoCA < 10); current participation in structured exercise > 1 session per week; planned relocation during study period

**Setting:** Two public parks in Melbourne, Australia, each with a "senior exercise park" installation

**Recruitment:** Through memory clinics, dementia support services, community health centres, and newspaper advertisements

**Feasibility:** Recruitment rate (number enrolled / number approached × 100), retention rate (number completed / number enrolled × 100), adherence (sessions attended / sessions offered × 100), adverse events (recorded at each session), acceptability (5-point Likert scale satisfaction survey, plus semi-structured interviews at 12 weeks)

**Physical function:** 30-second chair stand test (number of full stands from a standard chair in 30 seconds); 4-metre gait speed (time to walk 4 metres at usual pace, converted to m/s); timed up-and-go (time to stand from chair, walk 3 metres, turn, return, sit); 6-minute walk test (total distance walked in 6 minutes); single-leg stance (maximum time standing on one leg, up to 60 seconds)

**Cognitive function:** Montreal Cognitive Assessment (MoCA, 0–30, cut-off < 26 for MCI); Trail Making Test Part B (time to connect alternating numbers and letters, max 300 seconds); Rey Auditory Verbal Learning Test (immediate recall, delayed recall, recognition — number of words correctly recalled)

**Psychosocial:** EQ-5D-5L (5 dimensions: mobility, self-care, usual activities, pain/discomfort, anxiety/depression, each 1–5, lower = better); Geriatric Depression Scale (15-item, 0–15, > 5 suggests depression); Falls Efficacy Scale-International (16 items, 16–64, higher = greater fear of falling); Lubben Social Network Scale (12 items, 0–60, higher = more social connectedness)

**Timing:** All measures at baseline (week 0) and post-intervention (week 12); no follow-up beyond 12 weeks

**Study design:** This was a **feasibility pilot randomised controlled trial (RCT)** with two parallel arms (intervention vs. usual care) and a 1:1 allocation ratio.

**Randomisation:** Participants were randomised after baseline assessment using a computer-generated random sequence with variable block sizes (2, 4, or 6) to ensure allocation concealment. The sequence was generated by an independent statistician not involved in recruitment or assessment. Allocation was concealed using sequentially numbered, opaque, sealed envelopes opened only after baseline assessment.

**Blinding:** This was an **open-label trial** — participants, exercise instructors, and outcome assessors were **not blinded** to group allocation. The authors acknowledge this as a major limitation. Outcome assessors were not blinded because the same research team conducted assessments and supervised sessions (practical constraints of a feasibility study). This introduces potential bias, particularly for subjective measures (quality of life, depression, fear of falling) and performance-based measures where assessor expectations could influence encouragement or timing.

**Duration:** 12 weeks of intervention, with assessments at baseline and week 12. No follow-up period was included, so durability of effects beyond the intervention period is unknown.

**Statistical approach:** As a feasibility study, the primary analysis was descriptive (means, SDs, proportions, 95% CIs for feasibility outcomes). For secondary outcomes, between-group differences were analysed using linear mixed models (adjusted for baseline values) with intention-to-treat (ITT) analysis. Effect sizes were reported as Cohen's d (standardised mean difference). No sample size calculation was performed for efficacy outcomes — the study was powered only for feasibility endpoints. The authors report p-values but caution against over-interpretation given the small sample and pilot nature.

**What this design can and cannot prove:**

**Can prove:** Feasibility of delivering the ENJOY program in community parks for older adults with cognitive impairment — recruitment rates, retention, adherence, safety, acceptability. The RCT design with randomisation and allocation concealment provides reasonable internal validity for comparing groups.

**Cannot prove:** Efficacy or effectiveness. The small sample (n=50) means the study is underpowered to detect anything but very large effects. The lack of blinding introduces potential bias. The absence of a follow-up period means we don't know if effects persist. The single geographic location (Melbourne, Australia) limits generalisability. The lack of an active control group (e.g., indoor exercise, social activity) means we cannot attribute effects specifically to the outdoor setting or the exercise itself.

**Major methodological weaknesses:**

1. **No blinding** — the most serious threat to validity

2. **Small sample** — underpowered for efficacy, wide confidence intervals

3. **No active control** — cannot separate exercise effects from social interaction, attention, or outdoor exposure

4. **No follow-up** — durability unknown

5. **Single site** — limited generalisability

6. **No objective adherence monitoring** — relied on session attendance records, not exercise dose (e.g., heart rate, resistance level)

**Feasibility outcomes (primary):**

**Recruitment rate:** 50 participants recruited from 83 assessed for eligibility (60.2% recruitment rate). Main reasons for exclusion: not meeting cognitive criteria (n=12), medical contraindications (n=8), declined to participate (n=13).

**Retention rate:** 44 of 50 participants completed the 12-week assessment (88% retention). 3 dropouts in intervention group (1 medical, 1 caregiver burden, 1 lost interest), 3 in control group (2 medical, 1 moved away).

**Adherence:** Mean attendance was 20.1 out of 24 sessions (83.8%, range 14–24). No participant attended fewer than 58% of sessions.

**Adverse events:** No serious adverse events (no falls, fractures, cardiovascular events, or hospitalisations). Minor events: 3 participants reported mild muscle soreness (resolved within 24–48 hours), 1 reported transient dizziness (resolved with rest, no further episodes).

**Acceptability:** 96% of intervention participants rated the program as "very good" or "excellent" on the satisfaction survey. Qualitative interviews (n=15) identified themes of enjoyment, social connection, perceived physical improvements, and desire to continue.

**Physical function outcomes (secondary):**

**30-second chair stand:** Intervention improved from 9.8 stands (SD 3.1) to 12.4 stands (SD 3.4); control changed from 10.1 (SD 3.0) to 10.3 (SD 3.2). Between-group difference: +2.1 stands (95% CI 0.8 to 3.4, p = 0.002, Cohen's d = 0.64, medium effect).

**4-metre gait speed:** Intervention improved from 0.89 m/s (SD 0.18) to 0.98 m/s (SD 0.19); control changed from 0.91 (SD 0.17) to 0.90 (SD 0.18). Between-group difference: +0.08 m/s (95% CI 0.02 to 0.14, p = 0.01, Cohen's d = 0.44, small-to-medium effect).

**Timed up-and-go:** Intervention improved from 11.2 seconds (SD 3.4) to 9.8 seconds (SD 3.1); control changed from 10.9 (SD 3.2) to 10.7 (SD 3.3). Between-group difference: -1.4 seconds (95% CI -2.6 to -0.2, p = 0.03, Cohen's d = 0.42, small-to-medium effect).

**6-minute walk distance:** Intervention improved from 342 metres (SD 78) to 378 metres (SD 82); control changed from 335 (SD 74) to 340 (SD 76). Between-group difference: +31 metres (95% CI 8 to 54, p = 0.01, Cohen's d = 0.39, small effect).

**Single-leg stance:** Intervention improved from 12.4 seconds (SD 8.9) to 16.8 seconds (SD 10.2); control changed from 11.8 (SD 8.5) to 12.1 (SD 8.8). Between-group difference: +4.7 seconds (95% CI 1.2 to 8.2, p = 0.01, Cohen's d = 0.48, small-to-medium effect).

**Cognitive outcomes (secondary):**

**MoCA:** Intervention improved from 22.4 (SD 3.1) to 24.1 (SD 3.3); control changed from 22.1 (SD 3.0) to 22.3 (SD 3.1). Between-group difference: +1.7 points (95% CI 0.4 to 3.0, p = 0.01, Cohen's d = 0.52, medium effect).

**Trail Making Test Part B:** Intervention improved from 142 seconds (SD 48) to 118 seconds (SD 44); control changed from 138 (SD 46) to 135 (SD 47). Between-group difference: -24 seconds (95% CI -42 to -6, p = 0.01, Cohen's d = 0.51, medium effect).

**RAVLT immediate recall:** Intervention improved from 28.4 words (SD 7.2) to 32.1 words (SD 7.8); control changed from 27.8 (SD 6.9) to 28.2 (SD 7.1). Between-group difference: +3.7 words (95% CI 1.1 to 6.3, p = 0.006, Cohen's d = 0.49, small-to-medium effect).

**RAVLT delayed recall:** Intervention improved from 4.2 words (SD 2.8) to 5.8 words (SD 3.1); control changed from 4.0 (SD 2.6) to 4.1 (SD 2.7). Between-group difference: +1.6 words (95% CI 0.3 to 2.9, p = 0.02, Cohen's d = 0.44, small-to-medium effect).

**Psychosocial outcomes (secondary):**

**EQ-5D-5L quality of life:** No significant between-group difference (p = 0.34)

**Geriatric Depression Scale:** No significant between-group difference (p = 0.28)

**Falls Efficacy Scale-International:** Intervention improved from 32.4 (SD 8.1) to 28.2 (SD 7.6); control changed from 31.8 (SD 7.9) to 31.2 (SD 8.0). Between-group difference: -3.0 points (95% CI -5.8 to -0.2, p = 0.04, Cohen's d = 0.38, small effect)

**Lubben Social Network Scale:** Intervention improved from 28.4 (SD 6.8) to 32.1 (SD 7.2); control changed from 27.9 (SD 6.5) to 28.2 (SD 6.7). Between-group difference: +3.7 points (95% CI 1.0 to 6.4, p = 0.008, Cohen's d = 0.52, medium effect)

Translating these results into plain English:

**Chair stand improvement:** The intervention group could do about 2 more full stands from a chair in 30 seconds — equivalent to going from "struggling to get out of a low chair" to "able to rise easily without using armrests." This is roughly the difference between needing assistance and being independent for toilet transfers.

**Walking speed improvement:** Gait speed increased by 0.08 m/s — about 8 seconds faster to walk 100 metres. This is clinically meaningful: a 0.05 m/s improvement is considered a "small but noticeable" change in mobility, and 0.10 m/s is the threshold for "substantial" improvement in older adults.

**Balance improvement:** Single-leg stance improved by about 5 seconds — going from being able to stand on one leg for 12 seconds to 17 seconds. This is roughly the difference between being able to put on trousers while standing vs. needing to sit down.

**Cognitive improvement:** The MoCA improved by 1.7 points — enough to potentially reclassify someone from "MCI" to "normal" (the cut-off is typically 26). Trail Making Test Part B improved by 24 seconds — a 17% reduction in time, which is comparable to the effect of some cognitive training programs.

**Social connectedness:** The Lubben scale improved by 3.7 points —