An Insider’s Look at What Happened When Chiropractic Was Evaluated Scientifically, and Why No One Remembers (J. Michael Menke, August 27, 2026, Skeptic)
During the 1980s and 1990s chiropractic researchers launched clinical trials, comparative-effectiveness studies, and collaborations with universities and medical centers.
Then, the larger, better-designed studies made the clinical picture go soft. Patients improved—sometimes substantially—but the improvement proved stubbornly hard to pin on manipulation. Natural history, spontaneous recovery, reassurance, exercise, attention, expectancy, and plain contextual healing kept blurring the causal thresholds.2 Manipulation generally performed about as well as any other conservative care including placebo, and when superior, it closely tracked the false positive rate expected under null hypothesis significance testing.3, 4 In fact, to this day you will see chiropractic research conclusions proudly assert that manipulation is no less effective than other treatments for low back pain, including controls—a damnation by faint self-praise.
The problem was not only scientific but psychological, cultural, and institutional, which is why supporters and critics could read the same trials as Rorschach blots and come away convinced of opposite conclusions. When one body of evidence underwrites both conclusions, the fault is rarely bias; more often the evidence was never built to answer the question asked of it.5 The remedy is unglamorous: decide in advance which single measure could settle the question, weigh several rival hypotheses at once,6 and recognize when further information can no longer move the decision.7 In short: name in advance the result that would change your mind. Chiropractic, though, thrived in the ambiguity. When outcomes were positive, manipulation worked; when they were negative, prejudice or incompetence was clearly afoot.
So the center of gravity in chiropractic research quietly retreated from the awkward clinical effectiveness question, towards safer, fundable territory: biomechanics, sensorimotor physiology, adverse-event surveillance, health-services research, and broad biopsychosocial models.
Does spinal manipulation itself produce large, reliable, clinically transformative effects beyond natural history and contextual healing?
I eventually tried to answer that question—at scale, assembling a meta-analysis of spinal manipulation for low back pain: a hundred and twenty randomized trials, some nineteen thousand patients, modeled against the natural history of recovery. The results did not rescue the story I had spent my youth defending. Once natural history and contextual healing were given their due, the slight effect we detected turned out to be natural history carrying manipulation for acute pain and a faint placebo carrying it for chronic pain.8, 9 Years later, a contemporary from those reform days, chiropractor, influencer, and fellow researcher Greg Plaugher, remarked to me that “evidence-based chiropractic” was an old and abandoned paradigm.10