Does manual therapy or spinal manipulation help back pain?

Short answer

It helps, but only a little. It cut pain by 0.7 points out of 10 more than fake manipulation, and showed almost no difference from other non-surgical treatments such as exercise.

Does manual therapy or spinal manipulation help back pain?
'Put the misaligned bones back and the pain goes away' — is that right?
Many people with back pain turn to manual therapy or spinal manipulation. We checked how much pushing and twisting the spine by hand reduces pain, using studies that compared it with fake treatment and other treatments.
Compared with fake manipulation?
The pain difference was 0.7 points out of 10
A 2026 analysis pooled 76 trials (11,866 people) on back pain lasting over 3 months. In 16 trials (1,570 people) comparing it with fake manipulation that only mimicked the moves, pain fell 0.7 points more out of 10 after a month.
Compared with doing nothing?
Pain fell 1.4 points more
In 4 trials (325 people) comparing it with no treatment, pain fell 1.4 points more out of 10 after a month, and daily movement improved a lot. The gap was twice as large as against fake manipulation.
Compared with other treatments like exercise?
A 0.5-point difference, almost none
In 31 trials (4,109 people) comparing it with other non-surgical treatments such as exercise, the pain difference was 0.5 points out of 10. A 2019 analysis of 47 trials with 9,211 people also found a difference of only 0.3 points versus recommended treatments.
What about back pain that just started?
Pain fell about 1 point over 6 weeks
In 15 trials (1,699 people) on back pain less than 6 weeks old, spinal manipulation reduced pain by about 1 point out of 10. It helped, but not by much.
Any side effects?
No serious ones, but minor discomfort was common
The trials reported no serious side effects linked to manipulation. But brief increases in pain, muscle stiffness and headaches reached 50-67% in large case series of people who had it.
How does it compare with other treatments overall?
Few treatments beat fake ones at all
A 2025 analysis of 301 trials against fake treatments found that for back pain lasting over 3 months, only 5 treatments, including exercise and spinal manipulation, beat the fake. All 5 had small effects.
Bottom line: it works, but only a little
It does not 'realign bones and make pain vanish'
Spinal manipulation cut pain 0.7 points out of 10 more than fake manipulation, and about 1 point for pain that had just started. It showed almost no difference from other non-surgical treatments like exercise. It is best seen as one of several options with similar effects.