Orthopedics · Part Four · Seeing and Ruling Out

27PART IV

Lesson 27 / 44

Why MRI Misleads: What the Scan Records and What the Person Carries

A lumbar MRI is one of the most useful imaging studies in spine care. It also misleads more people than almost any test in medicine, because the picture and the person are not the same thing.

A lumbar MRI misleads when a finding on the image is read as the cause of the pain. The scan is accurate about anatomy and silent about protection. In a randomized trial of 380 primary care patients, rapid MRI and plain radiographs produced nearly identical disability at 12 months, and the MRI group had 10 lumbar operations against 4. The Unified Model of Tone reads the image as a static value sitting inside a population range while the dynamics have drifted.

Early imaging and work disability at one year, workers with sprains

adjusted relative risk 2.03

One patient, ten imaging centers, three weeks

average miss rate 43.6 percent

Early MRI without an indication, added medical cost

12,948 to 13,816 dollars per case

New findings of clinical significance after a first severe episode

2 of 51 people scanned

What a lumbar MRI records

Magnetic resonance imaging maps the water and fat content of tissue. That is why it resolves discs, ligaments, nerve roots and the epidural fat around them better than any other everyday study. It records anatomy at one moment, lying still and unloaded. The way a region behaves under load leaves no signal on the image.

How a finding turns into a diagnosis

A radiology report grades structure in words such as bulge, degeneration and stenosis. Those words then travel: into the referral letter, into what the person believes about their spine, and into the next decision about treatment. The route from an incidental finding to an operation runs through language before it runs through anatomy.

01What a scan answers

A lumbar MRI answers three narrow questions

A lumbar MRI exists to determine whether neural tissue is compressed, whether structural changes explain the symptoms, and whether serious pathology can be excluded. It shows discs, ligaments, nerve roots and the surrounding soft tissue with a clarity no other everyday study offers.

A scan is not a diagnosis. It is one piece of a larger clinical puzzle, and its role is to support reasoning rather than replace it. Understood as an answer to a specific question, imaging becomes a precise instrument. Treated as a verdict, it becomes a source of unnecessary alarm.

The size of the reading problem is set by how often the study finds something. Among 246 adults presenting with acute low back pain or radiculopathy, a herniation was identified in 60 percent at the first examination (Modic 2005). Herniation type, size and behavior over time bore no relationship to outcome.

The question the scan was ordered to answer

A lumbar MRI is read against the question that produced it. A study ordered to exclude infection is read differently from one ordered to localize a compressed nerve root. The narrow question decides which findings carry weight and which are furniture. Reading it well begins with knowing exactly what it was ordered to find.

Two failures follow from skipping that step. Over-interpretation blames a finding that was already there. Under-interpretation writes down a finding that matters as mild. The positive case for ordering a scan, and the criteria that decide when one changes management, sit in What MRI Is Really For.

02Findings

What the research shows

From two randomized trials, three occupational and primary care cohorts, a five-year prospective imaging study, a ten-center reliability study and a national small-area analysis.

Equal outcomes, more operations
Among 380 primary care patients randomized to rapid MRI or plain radiographs, mean 12-month disability was 9.34 against 8.75 on the modified Roland questionnaire. Ten patients in the MRI arm had lumbar spine operations against four in the radiograph arm (Jarvik 2003). The image changed what was done without changing where people ended up.
No difference at one year in 5239 older adults
Among 5239 adults aged 65 and over with a new visit for back pain, 1174 had early radiographs and 349 had early MRI or CT. Neither group differed from matched controls on disability at 12 months, the adjusted difference for radiographs being 0.10 of 24 points (Jarvik 2015). Seeing the anatomy sooner did not change the year.
Scans given without indication cost more
After removing every case in which early MRI might have been indicated, 37 percent of nonspecific low back pain cases and 79.9 percent of radiculopathy cases had still been scanned within 30 days. Those groups had much lower rates of going off disability and medical costs 12,948 to 13,816 dollars higher (Webster 2013). Cost tracked the reading rather than the injury.
Twice the risk of disability at one year
Among 1226 Washington State workers, early imaging carried an adjusted relative risk of 2.03 for work disability benefits at one year in those with sprains, from 1.33 to 3.11 (Graves 2012). For workers with radiculopathy the figure was 1.31 and the interval crossed one.
Ten centers, one patient, poor agreement
One patient scanned at 10 MRI centers within three weeks generated 49 distinct reported findings. None appeared in all 10 reports and 32.7 percent appeared only once, for a Fleiss kappa of 0.20 (Herzog 2017). A report is an interpretation rather than a measurement.
Knowing the result changed nothing
Of 246 people with acute low back pain or radiculopathy, half received their MR results within 48 hours and half stayed blinded to them. Improvement at six weeks was 60 percent unblinded against 67 percent blinded, and general health improved more in the blinded group (Ash 2008). Reassurance is a common reason for scanning, and here it did not arrive.
New pain, old images
Of 200 adults with no history of serious back trouble followed for five years, 51 were scanned after a new severe episode. Forty-three of them, 84 percent, had images unchanged or improved from baseline (Carragee 2006). The finding usually predates the pain it gets blamed for.
Where scans are common, surgery is common
Across 306 hospital referral regions, rates of advanced spinal imaging varied 5.5-fold and accounted for 22 percent of the variability in spine surgery rates (Lurie 2003). The correlation was 0.46, and scans taken in the surgical patients explained little of it.

03Blaming the wrong finding

Over-interpretation blames a change that was present before the pain

Over-interpretation is the first great error, and it is the reason so many people leave an appointment more frightened than the anatomy warrants. Of 200 adults imaged before and after their first severe episode, 84 percent showed no new change at all. Many findings on a lumbar MRI are common in people who have no pain. Disc degeneration, small bulges, mild stenosis and signal changes show up routinely in symptom-free spines, and Findings in People Without Pain carries the age-stratified prevalence table.

The strongest test of that point followed people from before their pain started. Carragee and colleagues imaged 200 adults with no significant history of back trouble, then tracked them for five years. Fifty-one had a new severe episode and a repeat scan within 6 to 12 weeks. Forty-three of the 51, 84 percent, had images that were unchanged from baseline or had improved on it (Carragee 2006).

Two of the 51, both with a primary radicular complaint, had new findings of probable clinical significance. Among those same 51, the most common progressive changes were loss of disc signal and facet arthrosis, each in 10 percent, with increased endplate changes in 4 percent. Those are progressive age changes and not acute events. Finding an abnormality does not prove it is the pain generator.

A degenerated disc on a screen is not automatically the cause of an ache in the back, any more than gray hair is the cause of a headache. Structural change and symptoms are two separate facts, and only careful correlation can link them.

What the report does to the reader

This is where the body can enter a protective, threat-driven state. A person who believes a scan proves their spine is crumbling will guard, brace and avoid movement long after the tissue itself has settled. The words on a report can shape a nervous system as powerfully as the tissue underneath them.

That effect has been measured on the labels alone. A randomized online experiment put six diagnostic labels to 1447 people with and without low back pain, and a content analysis of the free-text answers followed (O'Keeffe 2022). Lumbar sprain, non-specific low back pain and episode of back pain lowered the perceived need for imaging, surgery and second opinions against disc bulge, degeneration and arthritis.

Those three labels also lowered perceived seriousness and raised recovery expectations. In the free-text answers, thoughts of poor prognosis and of surgery clustered on disc bulge, degeneration and arthritis, while good prognosis clustered on the other three.

Adding context to the report changes prescribing. Across 237 lumbar MRI reports documenting uncomplicated degenerative changes, 71 included a statement giving the prevalence of the same findings in people without symptoms. Patients whose reports carried that statement had 0.29 times the odds of receiving a narcotic prescription (McCullough 2012). Rates of injection, surgical consultation and surgery were similar between the groups.

04Calling a finding mild

Under-interpretation lets a finding that matters be written down as mild

The average miss rate across 10 imaging centers reading one spine was 43.6 percent. The opposite mistake is under-interpretation, where a finding that genuinely matters is dismissed or waved away as mild. Meaningful nerve compression can be overlooked, progressive pathology can be missed, and the diagnosis of an important condition can be delayed when subtle abnormalities are ignored.

The pattern a careful clinician watches for is a descriptive label that understates a clinical picture. A report of mild central canal narrowing at two adjacent lumbar levels can sit alongside gradually progressive tingling in both legs, with calf and hamstring cramping that worsens after prolonged sitting. The label describes anatomy at rest. The presentation describes a canal running out of room under load.

Ten reports on one spine

Reports vary far more than the word mild suggests. Herzog and colleagues sent one patient with low back pain and right L5 radicular symptoms to 10 different MRI centers within three weeks. Across the 10 reports, 49 distinct findings were named. Not one of them appeared in all 10 reports, and 32.7 percent appeared only once (Herzog 2017).

Agreement across the reports gave a Fleiss kappa of 0.20. Each examination carried an average of 12.5 interpretive errors, with an average of 10.9 false negatives out of 25 reference findings and 1.6 false positives. Sensitivity averaged 56.4 percent and the miss rate 43.6 percent. Under-interpretation is the more common error in that data, by a wide margin.

The skill is to weigh each finding against the person in front of the clinician, so that a real problem is neither manufactured nor missed. A scan does not diagnose a person. It answers a question. The competence lies in knowing which question was asked, and in refusing to read the image apart from the human being it belongs to.

05The two spaces

Nearly every meaningful lumbar finding is described against two spaces

Most lumbar studies exist to answer one central question: is neural tissue being compressed, and if so, where. The neural structures occupy two major regions, the central canal and the neural foramina, and a finding is reported by which of them it encroaches upon.

The central canal carries the cauda equina and the descending nerve roots. Its compromise can produce bilateral symptoms, weakness, neurogenic claudication and, in severe cases, progressive neurological deficit. The neural foramina are the openings through which the exiting nerve roots leave the spine, and compression there tends to produce radiculopathy, dermatomal pain, sensory loss and weakness.

The roots follow a predictable, almost mathematical map. At L4-L5 the L4 root exits through the foramen while the L5 root descends through the canal, so a foraminal herniation at that level usually affects L4 and a paracentral herniation usually affects L5. Knowing this map is what lets a clinician match a symptom to a level rather than guess.

Herniation on its own is too common to carry the verdict. Among 246 people presenting with acute symptoms, herniation appeared in 57 percent of the low back pain group and 65 percent of the radiculopathy group, a difference that did not reach significance. What did separate the two groups was stenosis and nerve root compression, more frequent with radiculopathy at P below 0.006 (Modic 2005).

The map is what makes a scan worth reading. It is also why a scan read without a symptom pattern says almost nothing: every level offers something to name.

06The path to surgery

Early imaging without an indication moves people toward surgery

The randomized evidence counts the cost of an unnecessary scan in operations. Jarvik and colleagues took 380 primary care patients whose physicians had already ordered lumbar radiographs and randomized them to rapid MRI instead. At 12 months the two arms were level on disability, pain and function. Ten patients in the MRI arm had lumbar spine operations against four in the radiograph arm (Jarvik 2003).

Mean costs ran 2,380 dollars in the MRI arm against 2,059 in the radiograph arm, a difference of 321 dollars whose interval crossed zero. The authors concluded that substituting rapid MRI for radiographs offers little additional benefit and may raise costs through the extra spine operations patients undergo.

The same pattern appears at population scale. Across 306 hospital referral regions in the United States, spine surgery rates varied six-fold and advanced spinal imaging rates varied 5.5-fold. Areas with higher MRI rates had higher spine surgery rates, with a correlation of 0.46, and imaging explained 22 percent of the variability in surgery rates (Lurie 2003). A simulation showed that scans taken in the people who went on to surgery accounted for only a small part of that link.

Removing the indicated scans sharpens the picture. Webster and colleagues drew a nationally representative sample of workers with acute disabling occupational low back pain, then used the medical records to exclude every case in which early MRI might have been warranted. In the 555 that remained, 37 percent of the nonspecific cases and 79.9 percent of the radiculopathy cases had still been scanned within 30 days (Webster 2013).

Those groups showed much lower rates of going off disability and medical costs 12,948 to 13,816 dollars higher. Even among workers with 30 days or less of lost time after the scan, costs ran 7,643 to 8,584 dollars higher.

What the null results say

The cohort data splits, and both halves belong here. Among 1226 Washington State workers followed for a year, 18.6 percent had early imaging. For those with sprains it carried an adjusted relative risk of 2.03 for work disability benefits at one year, from 1.33 to 3.11. For those with radiculopathy the figure was 1.31, with an interval from 0.84 to 2.05 (Graves 2012).

The largest cohort found nothing in either direction. Among 5239 adults aged 65 and over with a new primary care visit for back pain, 1174 had early radiographs and 349 had early MRI or CT, each matched one to one against controls. The radiograph group scored 8.54 at 12 months against 8.74 for controls. The adjusted difference was 0.10, with a 95 percent interval running from 0.71 below zero to 0.50 above (Jarvik 2015).

The MRI and CT group scored 9.81 against 10.50 for controls. The adjusted difference was 0.51, with an interval from 1.62 below zero to 0.60 above. Early imaging did not make the year better in older adults, and it did not make it worse. Read alongside the workers’ compensation data, the harm of an unindicated scan runs through what people do next rather than through the scan itself.

07Image meets examination

A scan earns its place when it is read against the examination

Interpretation in isolation is where both errors are born, so the reading has to be integrated with the clinical picture. A portal-of-entry clinician holds information a scan can never contain: the mechanism, the symptom pattern, the neurological examination, the functional limitations, and the way a complaint behaves through the day.

This practice owns no X-ray or MRI unit. It reasons about whether a scan is warranted, refers out for the study, reads the images personally against the radiology report, and coordinates with radiology and medicine on what follows.

Ordering imaging when it is indicated, interpreting it in three dimensions, and correlating it with what the body actually shows is a competence and a way of thinking. That is the strength of the portal-of-entry model. A chiropractor trained to read a scan does not need to own the machine to add value.

Who a person sees first

The choice of first clinician predicts whether an early scan happens at all. Among 1830 Washington State workers interviewed a median of three weeks after filing a back injury claim, 19.8 percent received an MRI within six weeks. Starting with a surgeon carried a 78 percent greater likelihood of early MRI, an incident rate ratio of 1.78 from 1.08 to 2.92 (Graves 2012b).

Starting with a chiropractor carried an incident rate ratio of 0.53, from 0.42 to 0.66. That study counted who ordered scans rather than who recovered, and the counting is the point. Guideline-concordant restraint is a measurable behavior, and The Clinicians Advantage sets out what the examination contributes that the image cannot.

The magnetic resonance image and the person are reunited into a single understanding, and the essential question never changes: does this anatomy explain this presentation. When the answer is yes, a scan clarifies. When the answer is no, the finding was incidental, and the search continues with a calmer, clearer mind.

08Claims removed here

Two claims from the earlier version were removed

A worked description of one individual came off the page. It walked through a single person with mild stenosis at two levels and progressive leg symptoms. The general clinical pattern replaced it, because these pages teach how the reasoning works rather than how one case went.

The claim that low back pain is one of the most common reasons people seek care came off, because no source cited here measured consultation rates. Conservative First carries the guideline and utilization picture.

What the sources here do measure is what follows once a scan is ordered. Across 306 hospital referral regions, the use of advanced spinal imaging tracks spine surgery rates at a correlation of 0.46 (Lurie 2003).

09The model on imaging

What the Unified Model of Tone claims about a static image

Everything above is established science, including the results that came back null. What follows is this model’s reading, stated as ours rather than drawn from the papers cited.

Reference ranges describe populations, and a person can sit comfortably inside a population range while having drifted far from their own functional baseline. An MRI grade works the same way. Mild, moderate and severe are positions in a distribution built from many spines. None of them reads the spine in front of the clinician.

Our model holds that disturbance announces itself in the dynamics before it appears in the values: in variability, coupling, recovery time, responsiveness, temporal coordination, threshold and transition capacity. Every static number can still sit inside its reference range while that happens. An image records values. It has no access to dynamics.

That is why a scan and a patient disagree so often. The disagreement is a fact about what imaging measures rather than a complaint about imaging.

Compensated distortion looks normal

Subclinical dysfunction is compensated distortion of tone, a distortion the system is still hiding. Our model reads a quiet image in a painful person that way. A region holding a compensation can produce a picture indistinguishable from one that is not, because the compensation lives in how the region behaves rather than in its shape at rest.

The published data fits that reading. Herniation type, size and behavior over time bore no relationship to outcome across 246 people (Modic 2005). Of images taken after a first severe episode, 84 percent were unchanged or improved from baseline (Carragee 2006). The anatomy held still while the experience changed.

The prediction

From that follows a claim the imaging literature does not make. Take people matched on the same MRI grade at the same lumbar level. Our model predicts that four measures recorded together in those people will separate the ones in pain from the ones without it, and will share one underlying factor rather than varying independently, with compensation deciding how far each one moves.

The four are pressure pain threshold at the symptomatic segment, active lumbar range of motion, resting heart rate variability, and time to return to baseline after a standardized load test. Repeated flexion or a timed walk will serve as the load. Our model further predicts the direction of change under an input that restores regulation. People who begin high and people who begin low both move toward the middle, and the spread narrows.

This is a claim about what an image measures rather than a claim about what treatment does. If pressure pain threshold, active lumbar range of motion, resting heart rate variability and time to return to baseline after a load test are shown to move together, the unification claim is confirmed.

10The tone reading

How a misread scan expresses tone

Every topic in this library expresses all of tone. In the reading of a lumbar MRI three aspects carry the signature, because one image lands differently depending on the words wrapped around it and the state it meets.

Prediction

A label sets the expectation. Among 1447 people, the words disc bulge and degeneration raised the perceived need for surgery and lowered recovery expectations.

Time course

The finding usually predates the pain. Of 51 people scanned after a first severe episode, 43 had images unchanged or improved from their own baseline years earlier.

Gain

Threat turns the amplifier up. Adding prevalence data to the report cut the odds of a narcotic prescription to 0.29 across 237 lumbar studies.

The remaining foundations run through the reading as well. Constraint: a person who believes the spine is crumbling guards and braces, and the guarding narrows what the region can do. Input quality: the report is the input, and 10 centers reading one spine produced 49 different findings. Coupling: fear, breathing and paraspinal tone rise together once a finding is read as a verdict. Set point: the resting level of protection decides how much a mild finding costs the person carrying it. Load: a canal graded mild lying still can behave differently under a standing, walking spine. Oscillation: symptoms that change through the day are reporting the regulator rather than the picture. These are readings of one organization rather than separate systems, which is the core claim of the Unified Model of Tone.

11Across the library

How this page relates to the rest of the library

Reading an image against a person is the thread running through the whole of Part Four.

Findings in People Without Pain

The age-stratified prevalence of disc bulge, degeneration and stenosis in people with no symptoms at all.

What MRI Is Really For

The positive case: appropriateness criteria and the situations where a scan changes management.

The Clinicians Advantage

What a history and examination contribute that no image holds, including the challenge-and-return test.

Excluding the Dangerous on Imaging

When imaging is the correct instrument, and how modality sensitivity decides which study answers which question.

Pain Is Not Tissue Damage

Why a protective output can run high while the tissue on the scan is unremarkable.

Conservative First

The guideline pathway that puts the least invasive input capable of carrying the message first.

Low Back Pain

Low back pain read as a regulatory state, with the instruments used to measure it.

12Frequently asked

Questions patients ask about MRI and back pain

Do I need an MRI for back pain?

Usually not at first. Guidelines reserve early imaging for red flags and for persistent radicular pain after a month of conservative management. In a randomized trial of 380 primary care patients, rapid MRI and plain radiographs gave nearly identical disability at 12 months, and 10 people in the MRI arm had lumbar operations against 4. Among 5239 adults over 65, early imaging made no difference to disability at one year. A careful history and examination guide the early decisions better than a scan does.

Why can an MRI be misleading?

Because it records anatomy and says nothing about protection. Disc degeneration, bulges and mild stenosis are common in people with no pain at all, so a finding can be blamed for pain it is not causing. Of 200 adults followed for five years, 51 were scanned after a new severe episode. Forty-three of the 51 had images unchanged or improved from their own baseline, and only two had a new finding of probable clinical significance. The picture and the person stay separate facts until the examination links them.

Should I be worried about degeneration on my scan?

Often not. Disc degeneration and bulges are extremely common and frequently painless, and they accumulate with age in people who feel completely well. What matters is whether the finding fits the symptoms, the examination and the way the complaint behaves through the day. In one study, herniation type, size and change over time bore no relationship to outcome across 246 people with acute low back pain or radiculopathy. The word on the report is a description, not a prognosis.

Does having an early scan make surgery more likely?

The association is consistent across study designs. In the randomized trial of 380 primary care patients, 10 of those given rapid MRI had lumbar spine operations against 4 given radiographs. Across 306 US hospital referral regions, advanced imaging rates varied 5.5-fold and explained 22 percent of the variation in spine surgery rates, at a correlation of 0.46. In workers scanned within 30 days without an indication, medical costs ran 12,948 to 13,816 dollars higher and fewer people came off disability.

Can two radiologists read the same lumbar MRI differently?

Routinely. One patient with low back pain and right L5 radicular symptoms was scanned at 10 different MRI centers within three weeks. The 10 reports named 49 distinct findings, not one of which appeared in all 10, and 32.7 percent appeared only once. Agreement gave a Fleiss kappa of 0.20, average sensitivity was 56.4 percent, and the average miss rate was 43.6 percent. A radiology report is a skilled interpretation rather than a fixed measurement of the spine, which is why it is read against the examination.

How is imaging handled if the practice has no scanner?

The practice owns no X-ray or MRI unit. A scan is ordered only when the history and examination raise a question an image can answer. The referral goes to the appropriate facility, and the films come back to be reviewed alongside the radiology report. The next step is agreed with radiology or with the treating physician. Among 1830 injured workers, those whose first provider was a chiropractor had an incident rate ratio of 0.53 for receiving an early MRI, with an interval from 0.42 to 0.66.

What does the Unified Model of Tone say about MRI?

That an image is a static value inside a population range, while the disturbance sits in the dynamics. Mild and severe are positions in a distribution built from many spines, and a person can sit inside that range while having drifted far from their own functional baseline. From that the model makes a measurable prediction. Among people matched on the same scan grade, pressure pain threshold, active lumbar range of motion, resting heart rate variability and recovery time after a load test will share one underlying factor.

13The sources

References

1
Jarvik JG, Hollingworth W, Martin B, Emerson SS, Gray DT, et al. Rapid magnetic resonance imaging vs radiographs for patients with low back pain: a randomized controlled trial. JAMA. 2003. PMID 12783911
2
Jarvik JG, Gold LS, Comstock BA, Heagerty PJ, Rundell SD, et al. Association of early imaging for back pain with clinical outcomes in older adults. JAMA. 2015. PMID 25781443
3
Webster BS, Bauer AZ, Choi Y, Cifuentes M, Pransky GS. Iatrogenic consequences of early magnetic resonance imaging in acute, work-related, disabling low back pain. Spine (Phila Pa 1976). 2013. PMID 23883826
4
Graves JM, Fulton-Kehoe D, Jarvik JG, Franklin GM. Early imaging for acute low back pain: one-year health and disability outcomes among Washington State workers. Spine (Phila Pa 1976). 2012. PMID 22415000
5
Graves JM, Fulton-Kehoe D, Martin DP, Jarvik JG, Franklin GM. Factors associated with early magnetic resonance imaging utilization for acute occupational low back pain: a population-based study from Washington State workers compensation. Spine (Phila Pa 1976). 2012. PMID 22020590
6
Modic MT, Obuchowski NA, Ross JS, Brant-Zawadzki MN, Grooff PN, et al. Acute low back pain and radiculopathy: MR imaging findings and their prognostic role and effect on outcome. Radiology. 2005. PMID 16244269
7
Ash LM, Modic MT, Obuchowski NA, Ross JS, Brant-Zawadzki MN, et al. Effects of diagnostic information, per se, on patient outcomes in acute radiculopathy and low back pain. AJNR Am J Neuroradiol. 2008. PMID 18467522
8
Carragee E, Alamin T, Cheng I, Franklin T, van den Haak E, et al. Are first-time episodes of serious LBP associated with new MRI findings?. Spine J. 2006. PMID 17088193
9
Herzog R, Elgort DR, Flanders AE, Moley PJ. Variability in diagnostic error rates of 10 MRI centers performing lumbar spine MRI examinations on the same patient within a 3-week period. Spine J. 2017. PMID 27867079
10
McCullough BJ, Johnson GR, Martin BI, Jarvik JG. Lumbar MR imaging and reporting epidemiology: do epidemiologic data in reports affect clinical management?. Radiology. 2012. PMID 22357893
11
Lurie JD, Birkmeyer NJ, Weinstein JN. Rates of advanced spinal imaging and spine surgery. Spine (Phila Pa 1976). 2003. PMID 12642771
12
O'Keeffe M, Michaleff ZA, Harris IA, Buchbinder R, Ferreira GE, et al. Public and patient perceptions of diagnostic labels for non-specific low back pain: a content analysis. Eur Spine J. 2022. PMID 36198841

12 primary sources, each linked to its record. Figures quoted on this page were checked against the published abstract.

Related evidence

← All 44 lessons