Lavender for Depression
Standardized oral lavender oil (Silexan) shows a small, fairly consistent positive signal for mild-to-moderate depression, but the evidence base is thin and entirely manufacturer-funded, so it is not an established treatment and should not replace antidepressants or psychotherapy.
Why this grade7-layer evidence engine
The grade reflects a genuine but fragile signal. Only one dedicated mild-to-moderate major-depression RCT exists (PMID 38558147, n=498), where 80 mg/day Silexan beat placebo by just 2.17 MADRS points, a statistically significant but modest effect. The rest of the support comes from anxiety trials, including a mixed anxiety-depression RCT (PMID 26718792, n=318) and a meta-analysis of depressive symptoms pooled from anxiety studies (PMID 35262795), where depression was a secondary or post-hoc endpoint rather than the primary target.
Two structural weaknesses cap the evidence at weak. Every trial was funded by the manufacturer (Dr. Willmar Schwabe GmbH) with multiple author conflicts of interest, and much of the depression signal may simply reflect Silexan's established anxiety effect spilling over rather than an independent antidepressant action. There is no Cochrane-grade synthesis and no replication by independent teams.
Regulators and clinics do not endorse it for depression. The FDA lists lavender oil only as a GRAS flavoring agent, and the WHO and NCCIH treat the depression evidence as preliminary; the American Psychiatric Association and ADAA say nothing about lavender, pointing patients instead to medication, psychotherapy, TMS, or ECT. A theoretical interaction with sedatives adds further caution.
Scoring transparency
All scores computed by a 7-layer evidence engine — fully auditable▸View the full decision path (audit trail)
- compute_raw_score — 加權公式: L2×0.30 + L3×0.25 + L5×0.25 + L11×0.10 + L1×0.10 = 0.551
- tier_from_score — 依分數區間映射至 tier letter
- apply_hec_rules — 無高階證據可裁決
- tier_strict_requirement_check — | B→C 因 scope.conflation_risk=true 且 L11 獨評較低 (B7-2 tier cap)
- detect_disputes — 偵測到 0 個 hard + 1 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status