The PACBACK trial: can early self-management prevent chronic low back pain?

Most low back pain research asks a narrower question than it sounds like: given that someone already has chronic pain, what helps? Our broader guide to low back pain covers that ground, built on guidelines from NICE and the American College of Physicians. PACBACK asks a different and harder question: once someone has new back pain and looks likely, based on risk factors, to develop a chronic problem, can anything change that trajectory before it happens? Large trials designed specifically to answer that are rare, which is why PACBACK drew attention even though its own effect sizes were modest.

PACBACK (Prevention of Acute to Chronic Back pain) is a 2 by 2 factorial randomized clinical trial run out of research clinics at the University of Minnesota and the University of Pittsburgh, enrolling from November 2018 through May 2023 with follow-up concluding in June 2024.^[3] It randomized 1,000 adults (mean age 47, 58% female) with acute or subacute low back pain who scored at moderate to high risk of developing chronic, disabling pain on the STarT Back screening tool.^[1] Participants were assigned to one of four groups for up to 8 weeks: spinal manipulation therapy alone (201 people), clinician-supported biopsychosocial self-management alone (305 people), the two combined (193 people), or guideline-based medical care (301 people). Physical therapists and chiropractors delivered the manipulation and self-management arms. Ninety-three percent of participants completed the trial.^[1]

The first published paper, in JAMA, tracked low back disability (the Roland-Morris Disability Questionnaire) and pain intensity averaged over a year of follow-up.^[1] Disability differed significantly across the four groups (mean scores: self-management 4.7, manipulation 5.5, combined 4.8, medical care 5.9), while pain intensity did not differ significantly between groups at all.^[1] Compared with medical care, self-management alone reduced disability by a mean of 1.2 points (95% CI, -1.9 to -0.5) and the combined arm by 1.1 points (95% CI, -1.9 to -0.3); manipulation alone showed no significant difference from medical care (mean difference -0.4, 95% CI -1.2 to 0.4).^[1]

That split matters. A treatment can genuinely improve how much back pain limits someone’s life without measurably lowering how much pain they report feeling, and that is close to what this trial found.

The second paper, in JAMA Internal Medicine, reported on the trial’s original prevention question more directly: a low back pain impact score built by the NIH Task Force on Chronic Low Back Pain, combining pain, function, and interference with daily life on an 8 (best) to 50 (worst) scale, averaged over 10 to 12 months.^[2] Group differences were again small but statistically significant: self-management reduced the impact score by 1.7 points versus medical care (95% CI, -2.7 to -0.6), and the combined arm by 1.3 points (95% CI, -2.5 to 0); manipulation alone was not significantly different from medical care (mean difference -0.3, 95% CI -1.5 to 1.0).^[2] On a responder analysis, 64% of the self-management group reached a 50% or greater reduction in impact, versus 55% on medical care, and 12 percentage points fewer participants in the self-management group went on to report chronic pain that frequently interfered with their regular activities.^[2]

A secondary, exploratory mediation analysis in the same paper found that changes in psychosocial factors (confidence managing symptoms, fear of movement, and similar measures) at 6 months statistically explained about 76% of self-management’s effect at one year.^[2]That is consistent with the trial’s underlying theory: that addressing the psychological and behavioral side of pain, not just the tissue, is what changes the trajectory. It is a secondary, hypothesis-generating finding within a single trial, not independent confirmation of the mechanism.

Across both papers, spinal manipulation delivered on its own did not significantly outperform usual medical care on any primary outcome, and adding it to supported self-management did not improve on self-management alone.^[1][2] That is a real, specific finding from a large trial, worth taking seriously. It is not the same as evidence that manipulation is harmful or that it provides no value at all: participants who received it reported better secondary outcomes, including lower health care utilization, less medication use, and greater satisfaction, than the medical-care group.^[1]

The trial’s own authors describe these differences as small.^[2] A 1.7-point difference on a 42-point impact scale, or 1.2 points on a disability questionnaire, is not a dramatic individual-level change, and the trial considered a 30% within-patient reduction the threshold for a clinically important difference on its impact measure.^[2] The reason the finding still counts as meaningful is the pattern across measures: the group-level averages, the higher responder rate (64% versus 55%), and the lower rate of frequently interfering chronic pain all point the same direction, which is different from a single marginal p-value on one outcome.^[2] It is a real signal worth taking seriously, and it is also genuinely modest, not a treatment that reliably prevents chronic back pain in any given individual.

Everyone in PACBACK had acute or subacute low back pain (recent onset, not long-standing) and screened at moderate to high risk of chronicity on a validated tool before enrollment.^[1] That is a specific, at-risk population, not everyone with a sore back. If your pain has already been present for months, this trial was not designed to tell you what helps; the treatments in our general guide to chronic low back pain remain the more relevant starting point. Participants were also recruited into dedicated research clinics at two academic medical centers, treated by physical therapists and chiropractors trained to a study protocol, which may not fully generalize to a typical primary-care visit.

PACBACK does not overturn existing guidance so much as add a prevention-focused data point underneath it. Guidelines from NICE and the American College of Physicians already favor staying active, non-drug treatment, and patient-involved approaches over passive care for people who already have low back pain.^[2] PACBACK suggests that delivering a structured, coached version of that same active, self-management-oriented approach early, to people flagged as being at risk, may modestly reduce how often acute pain turns into a chronic, high-impact problem. It is a complementary finding, not a competing one.

Nothing in this trial changes the standard advice on when back pain needs prompt medical assessment rather than self-management.

PACBACK is a genuinely rare kind of trial: large, randomized, and built specifically to test prevention rather than treatment of chronic low back pain. It found that clinician-supported self-management, delivered early to people at elevated risk, modestly reduces disability and the odds of pain becoming a chronic, high-impact problem, without significantly changing pain intensity itself. It also found that spinal manipulation alone, despite being widely used for this purpose, did not outperform usual medical care on the outcomes this trial measured. Both findings come from a single, well-conducted trial that deserves independent replication before being treated as settled. See how we grade evidence strength across topics in our editorial and evidence standards.