Does Chiropractic Work? What the Research Shows

By Benjamin McCay, DC, CCSP

Published September 3, 2026

The title of this article asks a common question, but scientifically, it is not a very useful one. Chiropractic is a health care profession, not a single treatment, and asking whether it “works” does not tell us what result we are talking about. Are we asking whether treatment reduces pain, improves movement, decreases limitations in daily activities, or helps someone return to work, exercise, or normal life?

It would be just as imprecise to ask whether physical therapy or medicine “works” without identifying the condition being treated, the treatment being used, and the result being measured. A better question is:

For which conditions, in which patients, and for which outcomes does research support treatments commonly provided by chiropractors?

For several common muscle and joint problems, especially low back pain, neck pain, and certain headaches that come from the neck, research supports spinal manipulation and chiropractic care as reasonable conservative treatment options. Before looking at that research, however, it is important to understand an unusual problem researchers face when studying spinal manipulation: it cannot realistically be tested in the same way as a medication.

What Are Researchers Actually Studying?

Much of what is commonly described as chiropractic research actually studies spinal manipulative therapy (SMT). When spinal manipulation is performed by a chiropractor, it may also be called chiropractic manipulative therapy (CMT). This distinction matters because chiropractic is a profession, while spinal manipulation is one type of treatment commonly provided by chiropractors.

Spinal manipulation is also not one standardized procedure. More than 100 chiropractic technique systems have been described, ranging from traditional hands-on manipulation to instrument-assisted, drop table, flexion-distraction, and upper cervical techniques.[1] These approaches can differ in patient positioning, speed, force, contact points, equipment, and the parts of the spine being treated.

This variation matters when interpreting research. Two studies may both say they evaluated “spinal manipulation” while using very different techniques, numbers of visits, or treatment approaches. Some studies try to isolate one specific manipulation, while others evaluate chiropractic care more closely as it is actually delivered in practice. These studies answer different questions.

Spinal Manipulation Cannot Be Studied Like a Pill

The randomized, double-blind, placebo-controlled trial is often considered one of the strongest ways to study a medication. Researchers can make two pills that look identical. One contains the medication and the other contains an inactive substance. Neither the patient nor the clinician giving the medication needs to know which pill was given.

Spinal manipulation is fundamentally different. The clinician obviously knows whether an adjustment was performed, and reliably blinding the patient is also extremely difficult because the patient directly feels what is happening. An adjustment involves physical contact, positioning, movement of a joint, a quick mechanical thrust, and often an audible or palpable joint cavitation, commonly described as a popping or cracking sound.

Researchers have tried to solve this problem by creating placebo, or “sham” (simulated), adjustments. But a good placebo needs to do two things at the same time: it has to feel enough like the real treatment that the patient cannot tell the difference, while also producing no treatment effect of its own. With a hands-on treatment, those goals can conflict.

If the sham treatment is realistic enough to resemble an actual adjustment, it may involve touching the patient, positioning the body, moving joints, applying pressure, stretching tissues, or otherwise stimulating the muscles, joints, and nervous system. Those things may themselves affect pain or movement. If the sham is made inactive enough to avoid those effects, it becomes less like a real adjustment and may be easier for the patient to recognize as a placebo.

A systematic review examining placebo procedures used in studies of low back and pelvic manipulation found significant problems with existing sham treatments. Among 25 placebo-controlled trials, all had a high or unclear risk of bias. Few adequately tested whether patients remained blinded, and the researchers found substantial potential for patients to recognize which group they were in or for the sham treatment itself to produce physical effects.[2]

Researchers have even experimented with placing patients under short duration general anesthesia before either performing or withholding spinal manipulation so the patients truly could not know whether an adjustment had occurred.[3] Although that solves the awareness problem, it creates another major problem: anesthesia changes the conditions under which treatment is normally performed and introduces additional variables into the study.

This illustrates the basic difficulty of trying to study spinal manipulation in a controlled environment. Under normal clinical conditions, it is not realistically possible to create the equivalent of a double-blind medication trial in which neither the doctor nor the patient knows whether an actual manipulation occurred and the placebo both feels identical and produces no physical effect.

That does not mean research on spinal manipulation is useless. It means we need to understand what a particular study can and cannot tell us. Comparing spinal manipulation with exercise, medication, physical therapy, usual medical care, or no treatment may sometimes answer more practical questions than trying to create a perfect placebo adjustment.

What Does “Low Certainty Evidence” Mean?

This research problem is important because systematic reviews of spinal manipulation often describe the available evidence as low certainty or very low certainty. That phrase can easily sound as though researchers found little or no benefit, but that is not what it means.

A certainty rating describes how confident researchers are about the actual amount of benefit produced by a treatment. Evidence can be rated lower because studies involve different types of patients, different treatment techniques, different numbers or frequencies of visits, relatively small groups of patients, inconsistent results, or study designs that may make the results less reliable.

Manipulation research has the additional problems described above: the clinician cannot be blinded, reliably blinding the patient is difficult, and creating a believable sham treatment that has no physical effect is extremely challenging.

A study can therefore show that patients improved while the evidence is still classified as low certainty. Researchers are not necessarily saying that treatment failed. They are saying they are less certain about exactly how much benefit the treatment provides on average.

Low Back Pain

Low back pain is one of the most extensively studied conditions treated with spinal manipulation. In 2026, the Cochrane Collaboration published a major updated review that included 76 studies involving 11,866 patients with chronic low back pain. Compared with sham treatment, patients receiving spinal manipulation had better average pain and functional outcomes. Compared with receiving no treatment, the improvements were larger. When spinal manipulation was compared with other conservative treatments, pain outcomes were generally similar, with some improvement in function favoring manipulation.[4]

The authors still rated much of the evidence as low or very low certainty. The research limitations discussed above help explain why. The studies involved different types of patients, different manipulation techniques, different numbers and frequencies of treatments, and different study designs. The sham-controlled studies also faced the basic problem of trying to create a believable placebo adjustment that had no physical effect of its own.

After reviewing decades of research, the Cochrane authors made an especially important observation: continuing to conduct randomized trials in essentially the same manner is unlikely to substantially increase confidence in the evidence.[4] There may simply be a limit to how precisely a hands-on treatment can be isolated using the traditional placebo-controlled research model.

A different type of study can ask a much more practical question: what happens when chiropractic care is actually added to the care patients are already receiving?

A large study published in JAMA Network Open followed 750 active duty U.S. military personnel with low back pain. Patients received either usual medical care alone or usual medical care plus chiropractic care. Usual medical care could include medication, self care recommendations, physical therapy, or referral to a pain clinic. Chiropractic care commonly included spinal manipulation along with other treatment and exercise recommendations. After six weeks, patients receiving chiropractic care in addition to usual medical care reported significantly greater improvement in both pain and limitations from their low back condition. They were also more satisfied with their care and more likely to report overall improvement.[5]

Rather than asking whether a fake adjustment could perfectly imitate a real one, this study asked a question that is directly relevant to patients: does adding chiropractic care to ordinary medical care improve patient outcomes? In this large trial, it did.

Neck Pain

Neck pain also has good clinical research comparing treatments that patients might actually receive. A study published in Annals of Internal Medicine followed 272 adults with recent neck pain who received spinal manipulation, medication, or a home exercise program with advice. Patients receiving spinal manipulation reported better pain outcomes than those receiving medication at several follow-up periods. Home exercise also performed well, and results between spinal manipulation and home exercise were generally similar.[6]

This does not mean that one of those treatments failed. It suggests that more than one conservative approach can help patients with neck pain. Some patients may respond particularly well to hands-on treatment, others may do well with exercise, and many ultimately benefit from combining improved movement with progressive exercise and rehabilitation. The appropriate choice depends on the individual patient and how that patient responds.

Cervicogenic Headache

The type of headache matters. Migraine, tension-type headache, cervicogenic headache, and headaches caused by another medical condition are different problems, and chiropractic treatment should not be presented as appropriate for every headache simply because the patient also has neck pain.

A cervicogenic headache is a headache that comes from structures in the neck. It is often associated with neck pain, restricted neck movement, or reproduction of the headache during an examination of the neck.

One of the stronger clinical trials in this area included 200 patients with cervicogenic headache and compared manipulative therapy, a specific exercise program, combined manipulation and exercise, and a control group. At 12 months, both manipulative therapy and exercise had significantly reduced headache frequency and intensity, and those improvements were maintained. Combining the treatments was not significantly better than either treatment alone, although somewhat more patients obtained relief with the combined approach.[7]

Again, this illustrates an important principle in muscle and joint care: several treatments may be reasonable for the same condition. Manipulation does not need to outperform exercise for manipulation to be useful, just as exercise does not become ineffective because another treatment produces similar results.

Different Treatment Types Can Be Effective

For most muscle and joint conditions, there is no single conservative treatment that is best for every patient. A person with low back or neck pain might improve with spinal manipulation, exercise, physical therapy, massage, medication, education, or some combination of these approaches.

The appropriate choice may depend on the patient's symptoms, examination findings, severity of pain, ability to move comfortably, ability to exercise, previous response to treatment, medical history, and personal preferences. This is also why chiropractic care and physical therapy should not necessarily be viewed as competing treatments.

A patient with substantial pain, muscle tightness, and restricted movement may initially benefit from treatment aimed at decreasing pain and improving comfortable movement. As movement becomes easier, exercise and rehabilitation can take a progressively larger role in restoring strength, endurance, coordination, and normal function. Another patient may be ready for active rehabilitation immediately. Treatment should fit the patient rather than forcing every patient into the same plan.

Research Studies Groups. Doctors Treat Individuals.

Research is an important part of evidence-based health care because it helps determine which treatments are reasonable to consider and which may have little benefit or unacceptable risk. But research necessarily reports what happened on average among groups of patients, while doctors treat one individual patient at a time.

A treatment that produces a modest average improvement may include some patients who improve substantially, some who improve only slightly, and others who do not respond. Likewise, favorable research does not justify continuing treatment indefinitely when a particular patient is not getting better.

This is where evidence-based care becomes practical. Research helps determine whether a treatment is reasonable to try. Once treatment has begun, the individual patient's measurable response becomes the most important evidence for deciding whether that treatment is reasonable to continue. That response may include improvement, maintenance of meaningful gains, or prevention of a recurring decline in symptoms or function.

The Patient's Response Should Guide Treatment

In clinical practice, treatment should produce some indication that the patient's condition is improving. Depending on the problem being treated, that may mean:

• less pain
• better movement
• fewer or less intense headaches
• improved sleep
• greater tolerance for sitting, standing, or walking
• improved ability to exercise
• easier work and daily activities
• less reliance on hands-on treatment

Not every patient improves at the same rate, and different conditions have different expected recovery times. Treatment should be continued only when there is evidence that it is providing a meaningful benefit. Most often, that means improving pain, movement, function, or the ability to perform normal activities. As recovery progresses, treatment frequency can usually be reduced while normal activity, exercise, rehabilitation, and self-management take on a larger role.

In some cases, reducing or withdrawing care may reveal that treatment was also helping maintain function or prevent a recurring decline. If symptoms or functional limitations consistently worsen when care is withdrawn and improve again when treatment resumes, additional or intermittent care may still provide a meaningful benefit.

At our clinic, successfully performing an adjustment is not itself considered a successful outcome. What matters is how the patient responds. If the patient is improving or demonstrating another meaningful benefit from care, the treatment plan can be progressed accordingly. If the patient is not responding as expected, the diagnosis and treatment plan should be reconsidered rather than simply repeating the same treatment indefinitely. Depending on the situation, that may mean changing the treatment approach, obtaining additional testing or imaging, adding physical therapy, seeking medical treatment, or referring the patient to an appropriate specialist.

That is evidence-based care in practical terms. Research helps determine whether a treatment is reasonable to try, the doctor's examination and judgment help determine how it should be applied, and the patient's goals and preferences also matter. Once treatment begins, however, the patient's measurable response becomes the most important practical evidence for deciding what happens next.

Where Chiropractic Fits in Conservative Care

Spinal manipulation is recognized as a conservative treatment option in major health care guidelines. The American College of Physicians includes spinal manipulation among treatment options that do not involve medication for acute, subacute, and chronic low back pain.[8] The World Health Organization also includes spinal manipulation among treatments that may be offered as part of care for adults with chronic low back pain.[9]

Neither organization suggests that spinal manipulation is the only appropriate treatment, nor should it be. Exercise, physical therapy, massage, medication, education, and other treatments may all have a role depending on the patient's condition. Some patients may also need imaging, injections, surgical consultation, or another type of medical care.

Chiropractic care is best viewed as one part of conservative care for back, neck, and other muscle and joint problems. For an appropriately selected patient, the research supports a reasonable trial of treatment. After that, the question is no longer whether “chiropractic” in general is effective, but whether this particular patient is benefiting from care.

Benefit most commonly means continued improvement in pain, movement, or function. In some patients, however, treatment may also help maintain important gains or prevent a recurring decline. If care is reduced or withdrawn and the patient consistently worsens, then improves again when treatment resumes, that response may support additional or intermittent treatment. If treatment produces no meaningful improvement and does not help maintain function or prevent deterioration, the treatment plan should change.