Low back pain: what actually helps, what doesn’t, and when to worry

First: the symptoms that mean stop reading and get help

Almost all back pain is not dangerous. A small number of presentations are, and they are worth knowing cold, because the treatment for them is not physical therapy — it's an emergency room.

• Numbness in the area that would touch a saddle — groin, genitals, buttocks, inner thighs — or new loss of sensation during sex, or new erectile dysfunction alongside back pain.
• New trouble with your bladder — can't start, can't empty, leaking, or you can't feel that you're full or that urine is passing.
• New loss of bowel control, or no feeling when you wipe.
• New weakness or numbness in both legs.
• New or worsening weakness in either leg — a foot that drags or catches, a leg that gives way — especially if it is getting worse over hours or days. This one counts even on one side only.
• Fever with back pain, or back pain after a recent spinal injection, procedure, or infection elsewhere.
• Sudden tearing back or abdominal pain with faintness, especially over 65 with a smoking or vascular history — call 911.

The first four are the warning signs of cauda equina syndrome, where nerves at the base of the spinal cord are being compressed. Time matters — delayed treatment can mean permanent loss of bladder, bowel, and sexual function. Do not wait for an appointment with me or anyone else.

These aren't emergencies, but they need a medical workup rather than an exercise program:

• A fall, crash, or other significant injury — or even a minor one if you're over 65, have osteoporosis, or have taken corticosteroids long-term. A spinal compression fracture can happen from something you wouldn't think to mention.
• Any personal history of cancer with new back pain. Of all the warning signs studied, this is the one that most changes the odds.
• Unexplained weight loss alongside the pain.
• Pain that doesn't ease when you lie down, or that reliably wakes you at night.

Here's the honest part most articles skip: these warning signs are better at ruling danger in than ruling it out. Studies comparing them against MRI find they catch fewer true cases than you'd hope. That's precisely why I'm giving you the list instead of just promising to screen for it. A clinician's screen is a net with holes. You are the one who's with your body all day.

What back pain usually is

Roughly 90 to 95% of low back pain is classified as "non-specific," which is medicine's way of saying no single structure can be confidently blamed. That sounds unsatisfying. It is actually good news: it means the pain is usually not a sign that something is broken or wearing out.

Why your scan probably isn't the answer

A large review pooled imaging of people with no back pain at all. Among symptom-free 20-year-olds, 37% already had disc degeneration and 30% had a disc bulge. By age 80 those numbers reach 96% and 84%. Findings like these are closer to grey hair than to injury — near-universal, age-related, and often unrelated to whether anything hurts.

I want to be fair rather than glib, because the "scans are meaningless" line has become its own kind of overstatement: the same research group found degenerative changes are more common in people with back pain than without. The accurate version is that these findings are common, expected with age, and a poor explanation for any one person's pain — not that they never matter.

Which is why major guidelines, including the American College of Physicians and the Choosing Wisely campaign, recommend against imaging ordinary back pain in the first six weeks absent warning signs. Cohort studies consistently find people who get early scans end up with more surgery, more opioids, and no better outcomes. Worth flagging honestly: those are observational studies, not randomized trials, so "associated with" is the right phrase, not "causes." But the pattern is consistent enough that the guidelines moved.

What actually helps

The largest review of exercise for chronic low back pain pooled 249 trials and over 24,000 people. Exercise beat no treatment, with pain roughly 15 points better on a 100-point scale, and a smaller effect on function. That is a real, moderate-certainty benefit — and it is not a miracle. I'd rather you hear the actual number than a promise.

A common claim is that no type of exercise beats any other. That's close to true but oversimplified: a network analysis did find small advantages for some approaches. The differences were modest and the certainty low, so the practical conclusion stands — the best exercise is the one you'll actually keep doing — but it's not because everything is identical.

The most striking recent result came from the RESTORE trial, published in The Lancet in 2023: nearly 500 people with disabling chronic back pain, treated with an approach combining movement retraining with addressing the fear and beliefs around the pain. It substantially outperformed usual care, and the benefits held at one year. Adding movement-sensor biofeedback added nothing. The caveats are real: it couldn't be blinded, it was one country, and the therapists received specific training. But it is the strongest signal in this field in years, and it points at something PT has undersold: how you understand your back changes how it behaves.

What doesn't help as much as people think

• Bed rest. Staying active is modestly better. Note the word modestly — the evidence says rest is not better, not that it wrecks you.
• Opioids for chronic back pain. A 12-month randomized trial in chronic back and arthritis pain found them no better than non-opioid medication for function, with worse side effects. If you're currently prescribed them, that's a conversation with your prescriber, not something to change after reading a blog post.
• Acetaminophen for a fresh episode. A trial of 1,652 people found median recovery of 17 days on it and 16 days on placebo.
• Ultrasound, TENS, and back braces for chronic pain — no good evidence of benefit. Heat does have modest support early on.
• Spinal fusion for ordinary, non-specific back pain. The UK's NICE guidance says not to offer it outside a trial. This is not a statement about surgery for cauda equina, fracture, tumor, infection, instability, or stubborn nerve compression — where it can be exactly the right call.

One more, because it's my own profession's sacred cow: the idea that back pain comes from a specific weak deep abdominal muscle, and that retraining it is the fix. Core exercise works — about as well as other exercise. What the trials don't support is the mechanism story. You were probably not injured by a late-firing transversus abdominis. Keep doing the exercise; drop the fragile-spine narrative that came with it.

How long this actually takes

You'll read that 90% of back pain resolves in six weeks. That's a misreading of the research. What the pooled data actually shows: marked improvement over about six weeks, then a plateau — with low-to-moderate pain still present for many people at twelve months. And roughly a third to just under half have another episode within a year.

I tell patients this deliberately, because the alternative is worse. If you're promised a cure in six weeks and you're still sore at four months, you conclude something is seriously wrong with you, or that treatment failed. Neither is usually true. Recurrence is normal, it is not evidence of damage, and it does not mean you're back at the start.

Hurt is not the same as harm

This applies after the warning signs at the top of this page have been ruled out. With that boundary in place: pain is not a damage meter. It's a protective signal, and its volume is turned up or down by sleep, stress, mood, fear, and what you believe the pain means. That is not the same as saying pain is imaginary — your pain is real and it hurts. It means the volume knob has more inputs than tissue damage alone.

Practically: fear of movement, expecting the worst, and low confidence in recovery are among the better predictors of who stays in pain. Addressing them isn't a consolation prize when there's "nothing to find" — it's treating one of the actual drivers. Explaining pain works best alongside exercise, by the way; as a lecture on its own, the evidence is thin.

Can this be treated over video?

Mostly, yes — and I'll give you the honest state of the evidence rather than a sales pitch. Reviews comparing telerehabilitation to in-person care for back pain find comparable results for pain and disability. But that comparison rests on a small number of modest trials, and the authors themselves say it isn't enough to declare telehealth equivalent. Most studies compare it to usual care, not to excellent in-person physical therapy. This is a place where practice is ahead of the data.

What makes it work in practice is that the treatments with the best evidence for back pain — graded exercise, coaching, education, load management, consistency — are things that transfer over video well. What doesn’t transfer is a physical neurological exam — testing reflexes and strength by touch. So the rule here is simple and I'd rather state it than bury it: if anything on the red-flag list appears, you go to an emergency department, not to your next appointment with me. That escalation path is part of the care plan, not an afterthought.

The short version

Know the symptoms above that mean emergency, and the ones that mean get checked first — and act on them without hesitating. Past that, most back pain is not dangerous, your scan is probably describing normal aging, and the things that help are unglamorous: keep moving, load the tissue progressively, sleep, and stop treating your spine as fragile. The effects are moderate, not magical — but they beat the alternatives, and they don't carry the harms.

And if you've been told your back is crumbling, degenerating, or out of place — by anyone — it's worth a second opinion from someone willing to show you what the evidence actually says.