Scientific Evidence on Chinese Medicine: What We Know and What We Don't
The scientific evidence on traditional Chinese medicine is uneven: some techniques, such as acupuncture for certain chronic pain or tai chi for balance, have some support, while for most uses it is limited or insufficient. Studying it is difficult because of the placebo effect, the difficulty of blinding treatments and the variable quality of trials.
Key points
- The evidence varies widely depending on the technique and the health problem: you cannot say that Chinese medicine as a whole works or does not work.
- Acupuncture has some evidence for certain types of chronic pain and for nausea, although its specific effect compared with placebo is modest.
- Trials with sham acupuncture, difficult blinding and small or low-quality studies complicate the interpretation of results.
- The inclusion of traditional medicine in the WHO's ICD-11 is a statistical classification, not an endorsement of efficacy.
- Artemisinin shows that tradition can inspire effective drugs when it is investigated with scientific method.
Does Chinese medicine work? It is a common question, but it is poorly framed. Traditional Chinese medicine (TCM) is not a single intervention, but a set of very different techniques, from acupuncture to herbal formulas or tai chi, applied to very diverse health problems. Scientific research cannot deliver an overall verdict: it has to ask whether a specific technique helps with a specific problem, by how much, and with what risks. On this page we summarize what is known, what is not known and why it is so difficult to study, with the caution that a health topic demands.
How the efficacy of a treatment is evaluated
Evidence-based medicine ranks studies according to their reliability. At the top are systematic reviews of good-quality randomized clinical trials; then come individual trials, observational studies and, finally, case series and expert opinion. Experience accumulated over centuries has historical and cultural value, but it is not enough to demonstrate efficacy: many illnesses improve on their own, and we humans tend to attribute the improvement to the last thing we tried.
When we talk about levels of evidence on this page we use general terms: moderate (several reasonable studies point in the same direction, although uncertainty remains), limited (there are studies, but few, small or with significant flaws) and insufficient (there are no reliable data from which to draw conclusions).
The big challenges: placebo, blinding and sham acupuncture
Studying TCM therapies poses its own difficulties:
- Placebo effect and contextual factors. Prolonged attention, physical contact, ritual and expectations can relieve subjective symptoms such as pain or nausea.
- Blinding. In a drug trial, an identical pill without the active ingredient can be given. With acupuncture it is difficult: the patient feels the needle and the therapist always knows what they are applying.
- Sham acupuncture. Needles that do not penetrate or needling at non-traditional locations are used. But no simulation is completely inert, and in many studies the difference between real and sham acupuncture is small, while both outperform receiving no treatment.
- Quality and publication bias. Many trials are small, short or methodologically weak, and negative results tend to be published less often.
- Individualization. TCM tailors treatment to each person, which makes it harder to standardize the interventions being compared.
Evidence by therapy: summary table
The table offers an approximate and cautious assessment, based on the overall research landscape. It is not a recommendation for use.
| Therapy | Evidence | Comment |
|---|---|---|
| Acupuncture for chronic pain (low back, neck, knee osteoarthritis, headaches) | Moderate | Modest benefit compared with sham; part of the effect seems nonspecific. |
| Acupuncture or acupressure at PC6 for nausea | Moderate | Favorable results in postoperative nausea; small and variable effect. |
| Acupuncture for other indications (fertility, anxiety, addictions...) | Limited or insufficient | Heterogeneous studies with contradictory results. |
| Chinese herbal medicine (traditional formulas) | Limited | Few quality trials; concerns about toxicity, contamination and interactions. |
| Moxibustion | Limited | Scarce data; it has been studied for breech presentation with inconclusive results. |
| Cupping | Limited | Some small studies in pain; high risk of bias. |
| Tuina | Limited | Few rigorous trials; comparable to other forms of massage. |
| Gua sha | Insufficient | Very little quality research. |
| Tai chi | Moderate | May help improve balance and reduce falls in older adults. |
| Qigong | Limited | Possible benefits for well-being and quality of life; small studies. |
| Chinese dietary therapy | Insufficient | Not studied as a system; some advice coincides with a healthy diet. |
| Tongue and pulse diagnosis | Insufficient | Low agreement between practitioners; not validated as a diagnostic method. |
For more details on each technique, see our pages on acupuncture, herbal medicine and qigong and tai chi.
What does the scientific evidence say?
Three references help frame the debate:
WHO, ICD-11, chapter 26 (2019). The World Health Organization added to the International Classification of Diseases a supplementary chapter devoted to traditional medicine diagnoses. Its aim is to allow these diagnoses to be recorded and studied using common criteria. It is worth stressing: classifying is not the same as endorsing the efficacy of any treatment.
NICE guideline NG193 (2021). The UK's National Institute for Health and Care Excellence, in its guideline on chronic primary pain in people aged 16 and over, considers the possibility of offering a single course of acupuncture as one of the options, within a broader treatment plan. It is a recommendation limited to that type of pain and to specific conditions, not a general endorsement.
Artemisinin and the 2015 Nobel Prize. Pharmacologist Tu Youyou received the Nobel Prize in Physiology or Medicine in 2015 for the discovery of artemisinin, an antimalarial obtained from the plant Artemisia annua, whose traditional use inspired her research. It is a valuable example, but it must be understood correctly: success came from isolating the active ingredient and testing it in rigorous trials, not from validating the traditional formula as it was.
What we still don't know
Many questions remain open: what part of acupuncture's effect is specific, whether point location matters, what dose or number of sessions is optimal, or what the long-term safety of many herbs and formulas is. Nor has the existence of Qi (气, qì) or of the meridians as physical structures been demonstrated, although they remain useful as a traditional explanatory framework, as described in what is Chinese medicine.
Safety and precautions
A treatment with little evidence may be reasonable if it is safe and inexpensive, but it becomes problematic if it delays a diagnosis or replaces an effective treatment. Never stop a prescribed medical treatment, tell your doctor about any complementary therapy, and be especially careful with medicinal herbs because of their potential toxicity and interactions. You will find practical advice on our safety page.
Frequently asked questions
Has it been scientifically proven that acupuncture works?
It depends on the problem. For some chronic pain and for nausea there is some evidence of benefit, and the NICE NG193 guideline considers acupuncture an option for chronic primary pain. However, the difference compared with sham acupuncture is usually small, which suggests that part of the effect is due to nonspecific factors. For many other indications the evidence is limited or insufficient.
Does the WHO recognize Chinese medicine?
In 2019 the WHO included a traditional medicine chapter, chapter 26, in the International Classification of Diseases ICD-11. It allows traditional diagnoses to be recorded for statistical and research purposes. That inclusion does not mean that the WHO endorses the efficacy of Chinese medicine treatments, which must be assessed with clinical studies for each specific use.
What is sham acupuncture?
It is a control procedure used in clinical trials. It may consist of retractable needles that do not penetrate the skin or of shallow needling at locations that are not traditional points. Its aim is to separate the specific effect of acupuncture from the effect of the ritual, the attention and the expectations. Because no simulation is perfectly inert, interpreting these trials is not straightforward.
If Chinese medicine works for something, why doesn't everyone use it?
Because for most of its applications there are not enough good-quality studies, and when there are, the benefits are usually modest and comparable to other options. In addition, herbs can carry risks of toxicity and interactions. Some techniques are incorporated as a complement in certain settings, while others still lack support. The key is to evaluate each technique and each use separately.
Health notice. This content is educational and describes concepts of Traditional Chinese Medicine together with the available scientific evidence. It does not replace diagnosis or treatment by a healthcare professional. If symptoms are persistent or severe, consult your doctor. Safety guide.
