Psychiatry · Medical Literature
Dose-response efficacy of horticultural therapy for geriatric depression: a systematic review and meta-analysis of randomized controlled trials.
Abstract
Horticultural therapy (HT) is a promising non-pharmacological intervention for geriatric depression. However, the optimal "dose" required for maximum clinical efficacy remains poorly defined. To evaluate the efficacy of HT in alleviating geriatric depression and to quantify the dose-response relationship between intervention intensity and clinical outcomes. A systematic search was conducted across eight databases (including PubMed, Embase, and Cochrane Library, etc) for randomized controlled trials (RCTs) published up to 24 February 2026. Standardized mean differences (SMD) were pooled using a random-effects model. The dose-response relationship was modeled using restricted cubic splines (RCS). A total of eight RCTs involving 585 participants were initially identified. Systematic sensitivity analysis identified one study as a significant statistical outlier. After its exclusion, the final meta-analysis included seven studies with 435 participants. HT was associated with a significant reduction in geriatric depression scores (SMD = -0.52; 95% CI: -0.86 to -0.18; p = 0.0078). Heterogeneity remained moderate (I2 = 65.6%). The RCS model revealed an exploratory, hypothesis-generating non-linear dose-response relationship (p = 0.076), suggesting that the antidepressant benefit might peak at a cumulative dose of approximately 700-800 min. However, this finding should be interpreted with caution due to the limited number of studies. Subgroup analysis indicated that the GDS-30 scale showed a marginally significant trend toward higher sensitivity to treatment effects than the GDS-15 (p = 0.055). No significant publication bias was detected via Egger's test (p = 0.936) and Begg's test (p = 0.652). HT is an effective intervention for geriatric depression. A cumulative dose of 700-800 min represents a potential exploratory therapeutic window for maximizing efficacy while minimizing potential fatigue in frail older adults. PROSPERO, CRD420261323722.
Clinical relevance
In older adults (aged ≥60 years) with depressive symptoms, does horticultural therapy (HT) significantly reduce depression severity compared to routine care or no intervention, and is there a quantifiable dose-response relationship between intervention intensity and clinical outcome? __SNAPSHOT__ {"studyDesign":"Systematic review and meta-analysis of randomized controlled trials (RCTs), with restricted cubic spline (RCS) dose-response modeling; PROSPERO pre-registered (CRD420261323722); conducted in accordance with PRISMA 2020","setting":"International; studies sourced from eight electronic databases (PubMed, Embase, Web of Science, Cochrane Library, PsycINFO, CNKI, WanFang, VIP); no language restrictions applied; institutional affiliations include Shandong Provincial Hospital Affiliated to Shandong First Medical University, China","sampleSize":"Initial pool: 8 RCTs, n=585; after sensitivity analysis excluding one statistical outlier: 7 RCTs, n=435","population":"Older adults aged ≥60 years with depressive symptoms, including both community-dwelling individuals and those in residential care facilities; consistent with prior literature indicating mild-to-moderate severity in this population; patients with major depressive disorder were not specifically described as included or excluded in the available abstract","intervention":"Horticultural therapy (HT), defined broadly to include active gardening activities and passive sensory stimulation with plants, guided by trained professionals; intervention intensity (dose) quantified as a composite of session duration, frequency, and total program length for RCS modeling","comparator":"Routine care, no intervention, or waiting-list control","primaryOutcome":"Severity of geriatric depression, measured by validated scales (e.g., Geriatric Depression Scale GDS-15/GDS-30); pooled as standardized mean difference (SMD)","followUp":"Not reported in the available abstract; full program durations across constituent RCTs not reported in accessible text","fundingSource":"Not reported in the available abstract or retrieved article metadata"}
Key findings
In the primary analysis of 7 RCTs (n=435) after exclusion of a pre-specified statistical outlier, horticultural therapy was associated with a statistically significant reduction in geriatric depression scores (SMD = −0.52; 95% CI: −0.86 to −0.18; p = 0.0078). This represents a moderate effect size by conventional Cohen's d benchmarks. Heterogeneity was moderate (I² = 65.6%), indicating meaningful between-study variability that limits interpretive precision. The number needed to treat (NNT) cannot be calculated from the SMD alone without reference to an absolute event rate or a defined responder threshold, neither of which was available in the accessible abstract. A dose-response analysis using restricted cubic spline (RCS) modeling was performed to characterize the relationship between intervention intensity and depression reduction; however, the abstract was truncated before reporting the specific RCS findings (e.g., inflection points, optimal dose thresholds, or nonlinearity test results), and these values are therefore not reported here. No data on adverse events associated with HT were reported in the available text. The initial pool of 8 RCTs (n=585) included one study identified as a significant statistical outlier by sensitivity analysis; its exclusion altered both the point estimate and confidence interval, underscoring the fragility of the pooled result to individual study influence.
Practice implications
This meta-analysis provides statistically significant evidence that horticultural therapy reduces depressive symptom scores in older adults compared to routine care or no treatment, with a pooled SMD of −0.52. This effect size falls in the moderate range, though its translation to clinically meaningful improvement (e.g., achieving remission, reducing caregiver burden, or improving functional status) cannot be confirmed from the available abstract, as no absolute risk data, responder rates, or clinical threshold crossings were reported. The study is the first meta-analysis to apply RCS dose-response modeling to HT in geriatric depression, representing a methodological advance; however, the specific dose-response curve parameters and optimal 'dose prescription' were not recoverable from the truncated abstract. Findings are most applicable to older adults with mild-to-moderate depressive symptoms in community or residential care settings; applicability to major depressive disorder, cognitively impaired populations, or Western healthcare settings is uncertain given the predominance of Asian studies in this literature. The moderate heterogeneity (I² = 65.6%) and the sensitivity of results to a single outlier study suggest that confidence in the point estimate should be tempered. This summary is based only on the published abstract and retrieved open-access text; the complete results section—including full RCS outputs, subgroup analyses, risk-of-bias assessments, and publication bias testing—was not accessible for review, which is an important limitation. __LIMITATIONS__ Very small evidence base: only 7 RCTs with 435 participants included in the primary analysis, providing limited statistical power and precision,Moderate heterogeneity (I² = 65.6%) that was not fully explained in the accessible abstract; sources of between-study variance (e.g., HT modality, setting, severity, control condition) remain incompletely characterized,Result fragility: exclusion of a single outlier study was necessary to achieve statistical significance, raising concern about robustness,Dose-response modeling (RCS) requires aggregated study-level dose data rather than individual patient data, introducing ecological bias and limiting the precision of any optimal dose estimate,Outcome measurement heterogeneity: multiple depression scales (GDS-15, GDS-30, and potentially others) were pooled via SMD, which may obscure clinically important scale-specific differences,Likely geographic and cultural homogeneity of constituent trials (predominantly East Asian settings), limiting generalizability to Western or ethnically diverse populations,Absence of long-term follow-up data; durability of any antidepressant effect of HT after program completion is unknown,Blinding of participants to HT is inherently impossible (performance bias); blinding of outcome assessors was not assessable from the available abstract,Funding source not reported; potential conflicts of interest cannot be assessed,Abstract truncation prevented full appraisal of RCS results, subgroup analyses, GRADE certainty ratings, and publication bias assessment __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ This pre-registered meta-analysis of 7 RCTs (n=435) found that horticultural therapy was associated with a moderate reduction in geriatric depression scores (SMD = −0.52; 95% CI: −0.86 to −0.18; p = 0.0078), but the result is sensitive to outlier exclusion and carries moderate heterogeneity, limiting confidence. HT may be a reasonable adjunctive, non-pharmacological option for older adults with mild-to-moderate depressive symptoms, particularly in settings where engagement is feasible, but it should not replace established evidence-based treatments pending larger, more rigorous trials. Clinicians wishing to incorporate HT should note that the optimal dose parameters could not be confirmed from available data and that evidence is primarily derived from Asian care settings.