Traditional Use vs Scientific Evidence

By Ella Leung · Written October 2026

Historical records and clinical research are often used to support the same sentence, but they answer different questions. A long record of traditional use tells you what people in a particular time and place believed and practised; by itself, it does not show that a practice produces a specific health effect. Clinical research is built to test whether a treatment causes a measured change — but it can answer only the narrow question it was designed to ask. This page explains the difference, compares common types of evidence, and gives you a checklist for reading claims that blend them.

Contents

1. What historical records can answer
2. What historical records cannot answer
3. What clinical research can answer
4. What clinical research cannot answer
5. Comparing evidence types
6. Reading a case: "used for centuries"
7. Where "traditional use" appears in regulation
8. A checklist for reading a claim
9. Sources and limits

1. What historical records can answer

When someone says a practice has been "used for centuries," they are pointing at a historical record. In the case of Chinese medicine — 中医, zhōngyī, usually rendered in English as "Chinese medicine" or "TCM" — that record includes texts, commentaries, case notes, and living traditions of practice. Practices within this tradition have evolved over a long period and are still studied today [1].

Historical records can answer questions such as:

• What did practitioners in a given period write about a practice?
• How was a term such as 本草 (běncǎo, materia medica) used, and how did that change over time?
• Who used which practice, where, and in what setting?
• How did ideas travel between regions and across centuries?

These are questions about people, documents, and culture. A text such as the 黄帝内经 (Huángdì Nèijīng, Yellow Emperor's Inner Classic) is a historical source; studying its different layers and versions can tell us how ideas were organised and transmitted.

2. What historical records cannot answer

A written record of use is not the same as proof of effect. This distinction is where most confusion begins.

In traditional usage, texts described a practice through ideas such as balance, qi, and correspondences between the body and its environment. Reading those texts tells us how the tradition explained and categorised things. It does not, however, tell us whether the practice caused the outcome attributed to it. Modern research has not established those traditional concepts as measurable physiological mechanisms.

A record of use does not tell you:

• whether the practice actually caused the results credited to it;
• whether people recovered for unrelated reasons;
• whether the practice sometimes caused harm;
• whether successful cases were written down while failures were not.

A long record of traditional use is good evidence of use and of belief. On its own, it is not evidence that the practice works [2].

3. What clinical research can answer

Clinical research asks a narrower, more mechanical question: does this treatment change this measured outcome, compared with something else? The strongest single design for that question is the randomised controlled trial (RCT).

In an RCT, participants are assigned to groups by chance — typically to the treatment under study and to a control (a placebo, a sham, no treatment, or standard care). Outcomes are then measured in both groups and compared. Randomisation is used to balance known and unknown differences between the groups, so that a difference in outcome is more likely to reflect the treatment itself [4].

A single trial can still mislead. Researchers therefore examine all eligible studies in a systematic review. When the studies are sufficiently comparable, they may also pool numerical results in a meta-analysis. Cochrane describes systematic reviews as a reliable way to answer a focused question because they bring relevant studies together and assess them systematically [3].

A concrete example: an individual-patient-data meta-analysis of acupuncture for several chronic pain conditions used data from 39 randomised trials and 20,827 patients [5]. Across the comparisons, acupuncture showed a larger difference against non-acupuncture controls than against sham procedures. The non-acupuncture controls varied across trials and were not necessarily "doing nothing"; sham procedures may also have effects of their own. The study supports asking two separate questions: whether acupuncture outperformed the care used in a given control group, and which parts of the acupuncture encounter might account for the difference.

4. What clinical research cannot answer

Clinical research is powerful, but it is bounded. Common limits include:

• It answers the question it was built to answer. A trial of one preparation in one group does not automatically generalise to other preparations or groups.
• Blinding is difficult for hands-on or talking treatments. It is hard to hide from a patient that needles were inserted. Some trials find real and sham acupuncture perform similarly, raising the question of how much of an effect is specific to the technique [1].
• "Statistically significant" is not the same as "large" or "important." The size of an effect, and how precise the estimate is, matter more than the mere existence of a difference.
• An unanswered question is not a refuted one. A lack of good studies means the question is open, not that the practice has been shown not to work.
• Whole systems are hard to test. WHO notes that non-medicinal therapies such as acupuncture vary widely in practice and pose challenges for randomised trials; it calls for methods that can provide credible evidence for specific conditions [2].

5. Comparing evidence types

The table below is the key material for this page. It maps each common type of evidence to what it can and cannot show; it is a reading aid, not a ranking.

6. Reading a case: "used for centuries"

Consider a generic wellness sentence: "X has been used in Chinese medicine for over a thousand years to support digestion." It reads as one claim, but it is really two, plus a gap.

• A historical claim — that X has been used for a long time. This could be accurate and can be checked against historical sources.
• A health claim — the phrase "to support digestion" is vague. What outcome is meant? How would we measure it? No comparison and no measurement are offered.
• The gap — the sentence lets "traditional use" do the work of evidence. It borrows the authority of age without providing the kind of evidence that age cannot supply.

The practical lesson: a page can be honest about history and still tell you nothing proven about effect. Separate the historical part of a claim from the health part, and ask what evidence supports the second.

7. Where "traditional use" appears in regulation

Regulators sometimes treat "traditional use" as a distinct category. In the European Union, certain herbal medicinal products can go through a "traditional use registration" route. Applicants can rely largely on published safety and use information rather than new clinical trials, but they must show sufficient safety data, plausible efficacy, and at least 30 years of use, including 15 years within the EU. This route applies to products intended for use without a medical practitioner's supervision and not administered by injection; a full quality dossier is still required. The simplified route was introduced in 2004 by Directive 2004/24/EC [6].

This official category shows why "traditional use" is not a synonym for proven clinical benefit. It is a regulatory route with specific safety, quality and use requirements. The separate "well-established use" route requires scientific literature establishing recognised efficacy and acceptable safety after at least ten years of medicinal use within the EU [6].

8. A checklist for reading a claim

• What is actually claimed — a historical fact, an association, or a cause?
• Is a source given, and what type is it — a text, a case report, a trial, a review?
• Is there a comparison, and what is it compared with?
• What outcome was measured, and how large was the difference?
• Who funded the study, and who benefits if you accept the claim?
• Does the page say "traditional use" when it means "proven"?

9. Sources and limits

This page explains how to tell historical records and clinical research apart, so you can read a claim more carefully. It does not diagnose any condition, does not recommend or reject any specific treatment, and does not judge whether a particular herb or practice works for you. That judgment depends on your own circumstances and belongs with a qualified health professional. Where this site describes traditional usage, it describes what a tradition says — not a confirmed modern mechanism or a promise of a health outcome.

Sources

1. National Center for Complementary and Integrative Health, Traditional Chinese Medicine: What You Need To Know.
2. World Health Organization, Traditional medicine: questions and answers.
3. Cochrane, What are systematic reviews?.
4. Oxford Centre for Evidence-Based Medicine, The Oxford Levels of Evidence 2.
5. Vickers AJ and colleagues, “Acupuncture for chronic pain: update of an individual patient data meta-analysis”, Journal of Pain 2018;19(5):455–474.
6. European Medicines Agency, Herbal medicinal products, on traditional-use registration and the well-established-use route.

Related reading

• Chinese Medicine Culture: A Beginner's Guide
• Traditional Chinese Medicine and Classical Chinese Medicine

Continue exploring