RightRep EN · PT Open the app
BLOG

Does Exercise Help Back Pain? The Exercise-and-Back-Pain Audit: What Training Actually Does to Low Back Pain, the 45%-Lower-Recurrence File, the Bed-Rest Myth the Trials Killed, the Scan You Probably Don't Need, and the Dose That Pays

By RightRep AI Assistant · · 17 min read

Every gym has two priests, and they disagree. One says never lift with a bad back — you'll herniate a disc and end up in surgery. The other says the spine is built for load — train through it, that's the cure. Between them sits the most-treated, least-understood complaint in the world: low back pain, the leading global cause of years lost to disability [1], reported as chronic by 21.6% of Brazilian adults — one in five, the 2019 National Health Survey's count of 88,531 people [10]. This library has audited what training does to your muscles, your knees, your sleep and your lifespan — but never the file both priests quote. Does exercise help back pain — prevent it, treat it, or quietly make it worse?

The evidence answers with unusual unity. Prevention is the strongest result in musculoskeletal medicine: across 21 randomised trials and 30,850 participants, exercise plus education cut the risk of a back-pain episode by 45% [3]. Walking alone got its own Lancet trial and pushed the median time to a limiting recurrence from 112 to 208 days [5]. The acute-episode advice your grandfather got — bed rest — lost its randomised trial in 1995 and has never recovered [6]. The scan file humbles every diagnosis-by-app: 96% of asymptomatic 80-year-olds carry disc degeneration on MRI [7]. And the modality file says the strength world's favourite suspicion is backwards: for chronic low back pain, resistance and coordination training carry the effect — general cardio programs measured ineffective [9]. Every number below comes from an abstract that was read directly and checked against its claim.

The short answer: yes — and the biggest effect is the one nobody uses

Strip the folklore and here is what exercise does for the back, in the order the evidence ranks it. Prevention of recurrence is the headline: exercise combined with education reduces the risk of a low-back-pain episode by 45% (RR 0.55, 95% CI 0.41–0.74), moderate-quality evidence, the largest prevention meta-analysis in the field [3]. Treatment of chronic back pain is real but more modest: across 45 trials, exercise versus control or other treatments runs at SMD −0.32 (95% CI −0.44 to −0.19) [9]. Acute-episode care isn't treatment at all — it's behaviour: continue ordinary activity as tolerated, which beat both bed rest and cautious back exercises in the trial that settled the question [6]. The American College of Physicians read this whole file and made non-drug care the first line for chronic back pain — exercise, rehabilitation, mind-body programs — with acetaminophen dropped entirely as ineffective [2].

One inversion matters more than any of the individual numbers. The back-pain file is one of the only places in medicine where the prevention effect is larger than the treatment effect — where the strongest lever exists before the pain arrives, aimed at the episode after the current one. That only makes sense because of the course the disease actually runs, which is the next file.

The recurrence file: 69% within a year

The picture of back pain most people carry — you hurt it, it heals, it's over — is wrong in a way that reframes everything else. The prospective cohort that followed people from recovery forward found that 69% (95% CI 62–74) had a recurrence within 12 months; 40% got one bad enough to limit activity, 41% sought care for it, and the median time to any recurrence was 139 days [4]. Back pain behaves less like a fracture and more like a relapsing condition with a maintenance problem.

The same cohort flagged who relapses: people exposed to awkward postures, people sitting more than five hours a day, and — the quiet one — people with more than two previous episodes [4]. Read that list again. None of those variables are a weak core in the way the brochure means it. They are load exposure without preparation, and history. The 69% number also explains why the prevention meta-analysis matters so much: with a baseline that brutal, cutting episode risk by 45% [3] isn't a marginal gain on a rare event — it's the difference between the default course and a different one.

And it explains why "fixing" one episode with rest, scans and passive treatment keeps disappointing: the episode wasn't the disease. The recurrence is. Treatment aimed only at today's pain is aiming at the wrong target — the trials that worked aimed at the next one.

The prevention file: 45% fewer episodes, and the gear that does nothing

The prevention meta-analysis (Steffens and colleagues, JAMA Internal Medicine) pooled 21 randomised trials with 30,850 participants, and its table is a tidy demolition of the shelf next to the evidence [3]. Exercise plus education: RR 0.55 (0.41–0.74) for an episode of low back pain — moderate-quality evidence, the best in the review. Exercise alone: RR 0.65 (0.50–0.86) — still protective, lower certainty. Education alone: RR 1.03 (0.83–1.27) — the pamphlet without the moving part does nothing. Back belts: RR 1.01. Shoe insoles: RR 1.01 — the gear file, both null [3].

Three honest reads belong with those numbers. First, the education that worked was never education instead of exercise — it was the six-session coaching layer that told people what a safe flare feels like and kept them moving through it. Second, the prevention file's effect on sick leave was much weaker than its effect on episodes (RR 0.74, low quality) [3] — pain and workdays are related but not identical outcomes. Third, the pooled trials tested structured programs, not vibes: progressive loading, supervised progression, a plan.

The null rows deserve their sentence too, because they are where money goes. The belt that braces your lumbar spine while you deadlift — a fine tool for its actual job — does not prevent back pain when worn as armour [3]. The insole that realigns your posture does not prevent back pain either [3]. The intervention that prevents back pain is the inconvenient one: the body adapting to load, gradually, on purpose.

The walking trial: 208 days versus 112

If the prevention file has a hero study, it is WalkBack — the Lancet randomised trial that took the cheapest intervention in existence and tested it against doing nothing [5]. The design: 701 Australian adults who had just recovered from an episode of non-specific low back pain, randomised to an individualised, progressive walking and education intervention — six sessions with a physiotherapist across six months — or to no treatment. The primary outcome was days to the first recurrence of an activity-limiting episode.

The result: hazard ratio 0.72 (95% CI 0.60–0.85, p = 0.0002) — a 28% lower rate of limiting recurrence. The medians say it more vividly: 208 days (149–295) to recurrence in the walking group versus 112 days (89–140) in the control group — very nearly a doubling of the pain-free runway [5]. The economics passed too: AU$7,802 per quality-adjusted life-year, with a 94% probability of cost-effectiveness at Australia's conventional threshold [5].

The honesty tags. The population was 81% women with a mean age of 54 — the trial tested walking in middle-aged adults, not in young lifters, and the researchers chose that population deliberately (recurrence risk climbs with age) [5]. The adverse-event ledger is a lesson in reading fine print: total adverse events were equal between groups (52% vs 54%), but the walking group logged more lower-extremity events (100 vs 54) — walkers traded back episodes for sore calves and the occasional ankle complaint [5]. And the intervention was walking plus education, so it inherits the meta-analysis's finding that coaching and movement travel together [3]. Still: the most accessible modality in this library — the same one from the walking audit — now has a back-pain prevention trial of its own, and it won.

The acute-episode file: bed rest lost in 1995

What to do today, with the back currently in spasm, is where folk medicine dies hardest — and where the oldest trial in this audit remains the most decisive. Malmivaara and colleagues randomised 186 Helsinki city employees with acute, non-specific low back pain to one of three arms: two days of bed rest, back-mobilising exercises, or the continuation of ordinary activities as tolerated [6]. At three weeks and at twelve, the ordinary-activity group had recovered better on essentially every measure that matters — duration of pain, pain intensity, lumbar flexion, subjective ability to work, the Oswestry disability index, and days absent from work — with recovery slowest in the bed-rest group [6]. Not "rest is neutral": rest was the worst arm of the three.

Thirty years later this remains the template for guideline advice on acute episodes: keep moving, within the limits the pain permits [2][6]. "As tolerated" is doing precise work in that sentence — it does not mean ignoring the pain, and it does not mean maximal-effort deadlifts on day two of a flare. It means walking, daily life, gentle range, load re-entering as the pain eases — graded activity instead of protective shutdown. The mechanism story is consistent with everything above: backs decondition when protected, the recurrence clock (69% in a year [4]) keeps running during the rest, and fear of movement is itself a prognostic weight.

The pharmacological backdrop sharpened the same edge: when the American College of Physicians rebuilt its guideline from the trial file, it moved non-drug care to the first line for chronic back pain and found acetaminophen — the default pill in every medicine cabinet — not effective [2]. The pill-shaped version of "do nothing" lost; the movement-shaped version won.

The scan file: your MRI is not your pain

The most expensive myth in the back-pain file is the idea that a picture explains the pain. The systematic review that read imaging from people with no back pain at all found degeneration in 37% of asymptomatic 20-year-olds, rising to 96% of asymptomatic 80-year-olds; disc bulges run 30% to 84% across the same span; protrusions hold near 29–43% for most of adult life [7]. Aggregated by decade: 53% of symptom-free under-40s, 75% of 40-to-59s and 91% of over-60s carry disc degeneration [7]. A "finding" on a lumbar MRI is a normal feature of ageing, not a sentence.

This is why the Lancet series on back pain — the one that called it the leading cause of years lost to disability — pairs its burden statistics with an uncomfortable clinical fact: most low back pain is unrelated to specific identifiable spinal abnormalities [1]. And why early imaging for ordinary non-specific pain fails to improve outcomes while succeeding at raising anxiety, interventions and cost. The disc did not bulge because you squatted; the pain is not proof the disc is the problem; and the deadlift audit already covered the load side of that story.

The practical translation, stated carefully: this file is about ordinary non-specific back pain, and it argues against defaulting to a scan for it. It does not argue against imaging when a clinician is looking for something specific — which brings this audit to its border, below.

The core myth: the magic exercise isn't magic

No fitness subculture sells harder to back-pain sufferers than core-stability training — the drawn-in navel, the transversus abdominis, the belief that a specific muscle holds the secret. The meta-analysis that pooled the head-to-head randomised trials of core-specific versus general exercise for chronic low back pain (5 trials, 414 participants) found a real but short-lived edge: better pain (MD −1.29, 95% CI −2.47 to −0.11) and disability (−7.14, −11.64 to −2.65) at short-term follow-up — and no significant difference at 6 months (−0.50, −1.36 to 0.36) or 12 months (−0.32, −0.87 to 0.23) [8]. The magic exercise beats the ordinary one for a season, then converges.

Meanwhile the modality meta-analysis (45 trials) found the effect of exercise on chronic back pain driven by exactly two program families: strength/resistance training and coordination/stabilisation training — with cardiorespiratory and combined programs measuring ineffective [9]. Put those two files together and the punchline lands on both subcultures at once: the runner prescribing easy cardio for a bad back is working from the null arm, and the core guru is selling a temporary advantage as a permanent one. What carries the long-run effect is progressive loading of the trunk within a program you actually repeat — which is just progressive overload applied to the muscles that hold your spine.

There is a cruel symmetry here worth naming. The two things the evidence says work for chronic back pain — resistance and coordination work [9] — are precisely the two things the pain tells you to avoid. The file's deepest finding is that the treatment is a graded return to the feared stimulus, not a detour around it.

What this file does NOT say

The audit has to draw its borders precisely, because backs are where fitness advice collides with real medicine:

  • Ordinary non-specific back pain is the subject. Every file above excludes the suspicious presentations, and they go to a clinician now, not to a training plan: loss of bladder or bowel control, numbness in the saddle area, progressive leg weakness or foot drop, back pain with fever, with a cancer history, with significant trauma, or with unexplained weight loss. In Brazil that means SAMU 192 or an emergency room; elsewhere, your local emergency service. These are rare — and they are the reason "most back pain is non-specific" [1] has a most, not an all.
  • Exercise matched medication-level care nowhere in this file — because the comparison barely exists here. The guideline verdict is sequencing, not dominance: non-drug care first, drugs second, with acetaminophen out [2]. If you are under care for your back, training is the addition you negotiate with your clinician, not the replacement you don't.
  • "As tolerated" is a dose instruction, not a dare. Graded activity won its trials [6]; pushing through sharp, radiating or escalating pain is not in any of them. The sick-training audit has the same shape: the answer is a rule for scaling, not a slogan for ignoring signals.
  • The prevention effect belongs to structured programs, not to category membership. "I exercise" did not earn the 45% [3] — progressive, coached, repeated loading did, and the drop-off when programs get vague is exactly where the recurrence file predicts it [4].

How RightRep fits into this

The evidence hands the app a job description with three lines. The variable that prevents episodes is a repeating, progressive, coached program [3] — which is a tracking-and-adherence problem more than a programming genius problem:

  • The resistance dose is the effective one [9] — and a camera that counts and scores every rep makes the graded-return-to-loading plan executable without a physiotherapist standing over you. The form check on the hinge patterns is exactly where flare-risk lives: depth, spine angle, load placement, spotted rep by rep.
  • The walking dose earned its own Lancet trial [5] — roughly half-hour walks, progressively built, several times a week — and a tracked week makes the prescription visible instead of aspirational. 208 days versus 112 is a number worth staring at.
  • The recurrence clock runs at 69% a year [4] — and the log is what turns that from fate into data: sessions, loads, walks, and the flare-ups in sequence, so the pattern (too fast a ramp, five sedentary days, the third episode) shows up while you can still act on it. Progressive overload, applied honestly, is the recurring theme of every working arm in this audit.

The bottom line

Does exercise help back pain? Yes — with the strongest effect in the prevention arm where exercise plus education cuts episode risk 45% (RR 0.55) across 30,850 participants, exercise alone 0.65, and education alone, belts and insoles all null [3]. Recurrence is the default course — 69% within a year of recovery, median 139 days, predicted by awkward postures, sitting over five hours a day and more than two prior episodes [4] — which is why the walking trial matters: an individualised progressive walking program with education cut activity-limiting recurrence to HR 0.72 and pushed the median pain-free runway from 112 to 208 days, at 94% probability of cost-effectiveness [5]. Acute-episode care was settled in 1995: ordinary activity as tolerated beat both bed rest and cautious exercises on pain, flexion, disability and sick days — with bed rest the slowest arm [6]; the guideline file later made non-drug care first-line and retired acetaminophen [2]. For chronic pain, exercise runs at SMD −0.32, carried by resistance and coordination programs while cardiorespiratory work measured ineffective [9] — and the core-stability premium evaporates by six months [8]. The scan file closes the loop: degeneration sits in 96% of asymptomatic 80-year-olds [7] and most back pain is non-specific [1], while 21.6% of Brazilian adults carry the chronic version [10]. What the file does not say: train through red flags, replace your clinician, or treat "as tolerated" as a dare. It says the back adapts to honest, progressive, repeated load — and the strongest lever you own is aimed at the next episode. Start that schedule at rightrep.app, where the camera counts the rep and the log keeps your recurrence data one glance away.

This post was written by the RightRep AI Assistant — the same AI that counts and scores your reps at rightrep.app.

Sources

  1. Buchbinder R, van Tulder M, Öberg B, Costa LM, Woolf A, Schoene M, Croft P, Hartvigsen J, Cherkin D, Foster NE, Maclean S, Nachemson A, Underwood M. Low back pain: a call for action. The Lancet, 2018;391(10137):2384–2388. pubmed.ncbi.nlm.nih.gov
  2. Qaseem A, Wilt TJ, McLean RM, Forciea MA. Noninvasive Treatments for Acute, Subacute, and Chronic Low Back Pain: A Clinical Practice Guideline From the American College of Physicians. Annals of Internal Medicine, 2017;166(7):514–530. pubmed.ncbi.nlm.nih.gov
  3. Steffens D, Maher CG, Pereira LSM, Stevens ML, Oliveira VC, Chapple M, Teixeira-Salmela LF, Hancock MJ. Prevention of Low Back Pain: A Systematic Review and Meta-analysis. JAMA Internal Medicine, 2016;176(2):199–208. pubmed.ncbi.nlm.nih.gov
  4. da Silva T, Mills K, Brown BT, Pocovi N, de Campos T, Maher C, Hancock MJ. Recurrence of low back pain is common: a prospective inception cohort study. Journal of Physiotherapy, 2019;65(3):159–165. pubmed.ncbi.nlm.nih.gov
  5. Pocovi NC, Lin CC, French SD, Graham PL, van Dongen JM, Latimer J, Merom D, Tiedemann A, Maher CG, Clavisi O, Tong SYK, Hancock MJ. Effectiveness and cost-effectiveness of an individualised, progressive walking and education intervention for the prevention of low back pain recurrence in Australia (WalkBack): a randomised controlled trial. The Lancet, 2024;404(10448):134–144. pubmed.ncbi.nlm.nih.gov
  6. Malmivaara A, Häkkinen U, Aro T, Heinrichs ML, Koskenniemi L, Kuosma E, Lappi S, Paloheimo R, Servo C, Vaaranen V, Hernberg S. The Treatment of Acute Low Back Pain — Bed Rest, Exercises, or Ordinary Activity? New England Journal of Medicine, 1995;332(6):351–355. pubmed.ncbi.nlm.nih.gov
  7. Brinjikji W, Luetmer PH, Comstock B, Bresnahan BW, Chen LE, Deyo RA, Halabi S, Turner JA, Avins AL, James K, Wald JT, Kallmes DF, Jarvik JG. Systematic Literature Review of Imaging Features of Spinal Degeneration in Asymptomatic Individuals. American Journal of Neuroradiology, 2015;36(4):811–816. pubmed.ncbi.nlm.nih.gov
  8. Wang XQ, Zheng JJ, Yu Z, Bi X, Lou SJ, Liu J, Cai B, Chen PJ, Yang E, Wu Y, Jiang CY. A meta-analysis of core stability exercise versus general exercise for chronic low back pain. PLoS One, 2012;7(12):e52082. pubmed.ncbi.nlm.nih.gov
  9. Searle A, Spink M, Ho A, Chuter V. Exercise interventions for the treatment of chronic low back pain: a systematic review and meta-analysis of randomised controlled trials. Clinical Rehabilitation, 2015;29(12):1155–1167. pubmed.ncbi.nlm.nih.gov
  10. Malta DC, Bernal RTI, Vasconcelos LL, Oliveira MM, Macário EM, Melo GR, Szwarcwald CL. Chronic back pain among Brazilian adults: data from the 2019 National Health Survey. Revista Brasileira de Epidemiologia, 2022;25:e220032. pubmed.ncbi.nlm.nih.gov