Supplements

K-means clustering combined with latent structure analysis identifies 9 depression subtypes with distinct medication patterns in Traditional Chinese Medicine

This study analyzed 4434 prescriptions from 3537 publications and used K-means clustering with latent structure analysis to classify depression patients into 9 subtypes. The most common symptoms were insomnia and depressed mood, and the most frequently used herbs were Radix Bupleuri, Radix Paeoniae Alba, Poria, Rhizoma Chuanxiong, and Radix Curcumae. Medication patterns varied by subtype, with blood-activating herbs used for Clusters 1, 2, and 6; qi-regulating herbs for Clusters 3, 4, 5, 8, and 9; and qi-supplementing herbs for Cluster 7.

Last updated: Jul 7, 2026•0 RCTs•📖 Read as article →

Evidence Score

Evidence Score32/100
Human RCT☆☆☆☆☆
Meta-analysis☆☆☆☆☆
Mechanism★★★★★
Safety★★★★☆
Confidencelow

Study Evidence

Study 1. [Depression Syndrome Typing and Medication Pattern Analysis Through Unsupervised Clustering Combined With Latent Structure Dual Analysis].

observational

Zhu H, Yu C, Li X, Wang R, Chen Y, Wang T, Wu W, Yao L · Sichuan da xue xue bao. Yi xue ban = Journal of Sichuan University. Medical science edition (2025)

Participants: N/A
Duration: N/A
Intervention: K-means clustering combined with latent structure analysis for objective classification of depression subtypes and analysis of TCM medication patterns
Outcome: Identification of 9 depression subtypes, their core TCM formulas, herb properties (4 natures, 5 flavors, meridian tropism), therapeutic efficacy categories, and herb co-occurrence patterns
Effect Size: N/A
Population: Depression patients from clinical literature on TCM formulas retrieved from CNKI, Wanfang Data, VIP Database, Sinomed, Web of Science, and PubMed

Result:

PubMed: 40964125DOI: 10.12182/20250460202

Mechanism Graph

Symptom and medication data were standardized from 4434 prescriptions
↓
K-means clustering combined with latent structure analysis classified patients into 9 subtypes
↓
Core symptoms and formulas were identified for each subtype
↓
Herb properties and co-occurrence patterns were analyzed to reveal medication patterns

Limitations

  • ⚠The study is based on literature review rather than primary clinical data, which may introduce publication bias
  • ⚠The subjective nature of TCM syndrome differentiation and lack of unified classification standards may affect reproducibility

Frequently Asked Questions

What are the 9 depression subtypes identified in this study?▼

The study identified 9 subtypes (Clusters 1-9) using K-means clustering, with Cluster 6 being the largest. Each subtype had distinct core TCM formulas such as Zishui Qinggan Decoction, Danzhi Xiaoyao Powder, and Xiaoyao Powder.

Which herbs were most commonly used for depression in this analysis?▼

The most frequently used herbs were Radix Bupleuri (Chai Hu), Radix Paeoniae Alba (Bai Shao), Poria (Fu Ling), Rhizoma Chuanxiong (Chuan Xiong), and Radix Curcumae (Yu Jin).

How were the depression subtypes differentiated in terms of medication?▼

Clusters 1, 2, and 6 were treated mainly with blood-activating and stasis-dissolving herbs; Clusters 3, 4, 5, 8, and 9 with qi-regulating herbs; and Cluster 7 with qi-supplementing herbs. Meridian tropism also varied, targeting spleen, heart, or liver meridians.

What is the significance of this study for clinical practice?▼

The study provides a machine learning-based objective classification system for depression subtypes, clarifying the rationale for syndrome-based treatment and offering theoretical support for clinicians in selecting appropriate TCM formulas.

Products

Affiliate links coming soon. We only recommend products that match the doses and forms used in the cited research.

References

  1. 1.Zhu H, Yu C, Li X, Wang R, Chen Y, Wang T, Wu W, Yao L. "[Depression Syndrome Typing and Medication Pattern Analysis Through Unsupervised Clustering Combined With Latent Structure Dual Analysis].." Sichuan da xue xue bao. Yi xue ban = Journal of Sichuan University. Medical science edition, 2025. PMID: 40964125 DOI: 10.12182/20250460202
Disclaimer: This content is for educational purposes only and is not medical advice. Evidence scores reflect the quality and quantity of available research, not clinical recommendations. Always consult a healthcare professional before starting any supplement or intervention.