Tuina is the manual therapy of Chinese medicine: the practitioner pushes, presses, kneads and moves soft tissue and joints by hand. The World Health Organization’s Benchmarks for training in tuina (2010) take the name from two of its movements, pushing (tui) and grasping (na), and give its older names as an mo, an qiao and qiao mo. The same document dates the word tuina in the literature to the Ming dynasty (1368–1644), records an mo in the Huang Di Nei Jing, and says paediatric tuina became a field of study of its own in the Ming.
The hand techniques
The WHO’s benchmarks for the practice of tuina (2020) set out eight major manipulations, drawing on the Chinese textbook Science of tuina. Each is defined by the part of the hand that carries the force and how it moves.
| Manipulation | What the hand does | Rate or count |
|---|---|---|
| One-finger pushing | The tip or side of the thumb presses while the forearm sways the wrist in and out | 120–160 a minute |
| Rolling | The back of the hand on the little-finger side rolls over the area as the forearm sways | 120–160 a minute |
| Kneading | The thumb, the ball of the thumb or the heel of the palm turns slowly in place | 120–160 a minute |
| Pushing | The thumb, palm or elbow moves slowly in one direction along a channel or a muscle | 5–15 times |
| Rubbing | The palm or the ball of the thumb moves straight back and forth, pivoting at the shoulder | 100–120 a minute |
| Grasping | Thumb and fingers lift the skin and alternately knead and pinch the tissue | 5–15 times |
| Pressing | A finger, palm or elbow presses an acupoint or area with rising force | 5–15 times |
| Palm-twisting | Both palms clasp the area and twist quickly in opposite directions while moving up and down it | 5–15 times |
The benchmarks add that the practitioner may apply a tuina medium, a liniment or ointment, to the skin, and that the selection, force and length of each manipulation are set by the patient’s condition and response. Acupressure covers pressure on acupoints given as a treatment of its own.
In China and in training
The 2010 WHO benchmarks record a tuina department as one of the four clinical departments of medicine under the Tang dynasty (618–907), the first tuina school opening in Shanghai in 1956, and higher education in tuina at universities and colleges of Chinese medicine across China. They say tuina in China is used for soft-tissue injuries and, alongside conventional care, for fractures and dislocations, and that it is used within acupuncture, internal medicine, gynaecology and paediatrics. The full-time programme they describe takes three years and at least 2,060 hours, 500 of them supervised clinical practice.
The WHO’s benchmarks for the training of tuina (2020) updated that with two levels for people without a medical background: a basic level of 1,504 hours and an advanced level of 2,354, each with 400 or 500 hours of supervised practice. Paediatric tuina (64 hours) appears only at the advanced level. People already trained in conventional or traditional medicine take an adapted course.
What reviews and trials found
| Source | What it covered | What it found |
|---|---|---|
| Huang and colleagues, Complementary Therapies in Medicine, 2026 | Chronic non-specific low back pain: 21 trials, 1,612 people, tuina against non-tuina treatments; searched to April 2025 | Less pain (standardised mean difference −1.06, 95% CI −1.37 to −0.76) and better function, both at low certainty; no difference in muscle fitness or serious adverse events |
| Yang and colleagues, Medicine, 2023 | The same condition: 15 trials, 1,390 people | Less pain and better function, no difference in quality of life, all at low certainty; 6 trials reported adverse events, none serious |
| Gross and colleagues, Cochrane, 2024 | Massage of any kind for neck pain, tuina among them: 33 trials, 1,994 people, searched to October 2023 | Little or no difference from placebo in pain at up to 12 weeks (3.4 points better on a 0 to 100 scale, 95% CI 8.2 better to 1.3 worse; low certainty); adverse events too poorly reported to analyse |
| Pach and colleagues, Journal of Alternative and Complementary Medicine, 2018 | Chronic neck pain, Berlin: 92 people, 6 tuina sessions over 3 weeks against a waiting list | Less pain at 4 weeks (−22.8 mm on a 100 mm scale, 95% CI −31.7 to −13.8) and at 12 weeks (−17.9 mm); no serious adverse events |
| Wang and colleagues, Frontiers in Neurology, 2024 | Insomnia: 23 trials, 1,780 people | Lower sleep-quality index scores (−2.34, 95% CI −2.94 to −1.74) than other treatments; 2 trials reported minor adverse events |
| Gao and colleagues, BMC Complementary and Alternative Medicine, 2018 | Acute diarrhoea in children: 26 trials, 2,644 children, paediatric massage against drugs | A higher share judged effective (risk ratio 1.20, 95% CI 1.14 to 1.27) and a shorter time to recovery as the trials defined it, at low quality because of high risk of bias |
| Lu and colleagues, Health and Quality of Life Outcomes, 2021 | Acute diarrhoea in children aged 0 to 6: 86 children, paediatric tuina or sham tuina, both added to usual care | Fewer episodes of diarrhoea on day 3 (risk ratio 0.73, 95% CI 0.59 to 0.91), not significant after adjusting for the children’s characteristics; no difference in days of diarrhoea |
| Ma and colleagues, BMC Complementary Medicine and Therapies, 2025 | Knee osteoarthritis: 140 people, 8 sessions over 3 weeks of tuina or of manual physical therapy | No difference on the WOMAC osteoarthritis index at 4 weeks or 16 weeks; tuina ahead at 8 weeks (−4.33, 95% CI −8.34 to −0.31); no serious adverse events |
The two low back pain reviews searched Chinese and English databases and rated every result at low certainty. A 2023 review of the outcome measures in 17 low back pain trials of tuina, by Zhou and colleagues, judged all 17 of poor methodological quality and found that 41% measured safety and 12% quality of life. The Cochrane neck review rated its placebo comparison low certainty partly because most of those trials gave a single session; Pach’s trial, which it includes, was delivered by a physiotherapist with two years of tuina training and compared tuina with no treatment rather than with a sham.
Cochrane’s 2015 review of massage for low back pain lists tuina among the massage styles it covers, though of its 25 trials it places one in China and one in Hong Kong. Gao’s paediatric review compared massage with drugs, and its main outcome was the share of children each trial judged effective; Lu’s trial is the sham-controlled test of the same question. Low back pain and Neck pain set these against the needling evidence for each condition.
With needling, and against it
Some trials give tuina on top of acupuncture and compare the pair with acupuncture alone, which measures what the tuina adds. In a 2024 network meta-analysis of peripheral facial paralysis, Deng and colleagues pooled 22 trials, 1,814 people, and found acupuncture with tuina ahead of acupuncture alone on a facial grading score and on time to recovery, calling the evidence insufficient. A 2022 review of insomnia by Luo and colleagues pooled 29 trials, 2,688 people, of acupuncture with tuina against drugs or acupuncture alone. Wang’s insomnia review compared tuina, alone or combined, with drugs or acupuncture alone in a subgroup analysis. None of these designs has a sham arm, so neither the combination nor the comparison isolates the effect of either treatment from the time and attention each brings.
Injuries
The WHO practice benchmarks rule tuina out in acute heart or cerebrovascular disease, severe infection, acute abdominal pain, a tendency to bleed and intoxication, over the abdomen or lower back in pregnancy or menstruation, and over infected skin, tumours or injured tendons and ligaments. They name three kinds of incident: injury to the skin (burning, blisters, bleeding), to the tissues (bruising, haematoma) and to the bones (fracture, dislocation), and advise against joint-moving manipulations in elderly people and people with a history of dislocation.
A 2014 review of case reports from 2003 to 2013 by Yin and colleagues, which searched for tuina by name along with massage and chiropractic, found 40 reports of 138 adverse events, 60% of the reports from China. Disc herniation, soft-tissue injury, nerve damage, spinal cord injury, tearing of the vertebral arteries and fractures were the commonest, most followed spinal manipulation with rotational movements, and most patients recovered fully; the authors note that case reports cannot show whether the treatment caused the event.
A 2011 case report by Ying and colleagues describes blood in the urine, attributed to a bruised kidney, after a forceful lumbar pulling manipulation for a herniated disc, which settled in 10 days; their review of earlier reports of tuina and massage names haematoma as a common complication and, among the severe ones, ruptured internal organs. In children, a 2020 review of paediatric massage of every kind by Karkhaneh and colleagues found that half of 60 studies said nothing about adverse events, and linked abdominal massage in premature infants with 17 cases of twisted bowel, 4 of them fatal, without settling whether the massage caused them. Acupuncture for children covers the needling side.
What the reviews could not settle
Almost every comparison above is against no treatment, drugs or another active therapy, so the reviews cannot separate tuina’s effect from the attention and touch that come with it, and the one pooled placebo comparison, the Cochrane neck review's, covers massage of every kind rather than tuina alone. Fewer than half the low back pain trials measured safety, and the serious harms are known from case reports, which give no count of how many people were treated. The reviews do not say which manipulations, how much force or how many sessions matter, and the low back pain reviews leaned on Chinese-language databases; Where the trials come from covers why the country of a trial bears on reading its result. A protocol entry on AcuiQ names its modality where the study recorded one.