Clinical library · Clinical reasoning
Red Flags
Screening for serious pathology is not optional. But the specific red flags taught in every physiotherapy programme are mostly expert opinion, guidelines disagree with each other almost completely, and only a handful have usable diagnostic accuracy.
In one line. Features from the history or examination that raise the probability of serious underlying pathology — fracture, malignancy, infection, cauda equina syndrome, vascular events — and should prompt further investigation or referral rather than continued conservative treatment.
This page does not tell you to stop screening. It tells you what the individual items on the checklist are actually worth, because using a list of 46 items as though each carried equal weight is how both missed diagnoses and unnecessary imaging happen.
Read this correctly
Nothing on this page reduces the obligation to screen, and nothing here should be used to justify not referring a patient you are worried about. The finding is that the evidence behind individual checklist items is weak — which argues for clinical judgement and for knowing which few items carry real weight, not for abandoning the process. Where you are uncertain, refer.
What the evidence shows
| Question | Finding | Source and quality |
|---|---|---|
| Do guidelines agree on which red flags to use? | No. 46 discrete red flags across 16 guidelines from 15 countries | Evidence for the accuracy of recommended red flags was lacking; eight guidelines based their choice on consensus or previous guidelines and five gave no reference at all [2] |
| Agreement in neck pain guidelines | Median Fleiss' kappa of 0 across 29 guidelines and 114 red flags | Red flags were mainly supported by expert opinion. The review states specific recommendations cannot be provided, except the Canadian C-Spine rule for post-traumatic fracture [3] |
| Which flags are informative for vertebral fracture? | Trauma (+LR 1.93 to 12.85); older age over 70 (+LR 11.19, 95% CI 5.33 to 23.51); corticosteroid use | 14 studies. Meta-analysis was not possible due to heterogeneity; results come from single studies, and risk of bias was low in only one study for reference standard and flow and timing [1] |
| How common is serious pathology? | 2.5% to 5.1% (prospective) and 0.7% to 7.4% (retrospective) in the emergency department | 22 studies, 41,320 patients. Higher than reported prevalence in primary care. Vertebral fracture 0.0–7.2%, spinal cancer 0.0–2.1%, infection 0.0–1.9% [4] |
| Which flags raised likelihood in the emergency setting? | Suspicion or history of cancer (spinal cancer); intravenous drug use, indwelling vascular catheter, other infection site (epidural abscess) | Diagnostic accuracy for most red flags was reported by only a single study; further prospective validation is needed [4] |
| Is there an agreed framework? | An international framework for red flags for potential serious spinal pathologies has been published | Cited here for the framework itself; this page does not reproduce figures from it that it has not verified [5] |
Where it misleads
1. A long checklist implies a precision that does not exist
Across 16 low back pain guidelines, 46 discrete red flags were endorsed, with no consensus on which to use and evidence of accuracy rarely provided. Eight guidelines derived their list from consensus or from previous guidelines; five provided no supporting reference at all. [2] In neck pain the situation is starker: median agreement between 29 guidelines was a Fleiss' kappa of zero. [3] Two guidelines picked at random agree about as well as chance.
2. Most individual flags have never been tested
The Cochrane review could not perform a meta-analysis because of data heterogeneity, and reports that only a small number of the red flags investigated may be informative, from single studies. [1] The emergency department review makes the same point: diagnostic accuracy for most red flags was reported by only a single study. [4] "Taught everywhere" and "validated" are not the same status.
3. Low prevalence makes even a good flag mostly false-positive
Serious pathology requiring immediate or urgent treatment ran at 2.5% to 5.1% in prospective emergency department studies — and the authors note this is higher than the prevalence reported in primary care. [4] At primary care prevalence, a flag with a positive likelihood ratio of 2 leaves the post-test probability still very low. This is the arithmetic behind over-imaging.
4. The few flags that do work are worth knowing individually
For vertebral fracture, trauma, older age above 70 and corticosteroid use carried informative positive likelihood ratios — older age over 70 at 11.19 (95% CI 5.33 to 23.51). [1] For spinal cancer, suspicion or history of cancer; for epidural abscess, intravenous drug use, an indwelling vascular catheter or another infection site. [4] These are a different order of evidence from the rest of the list, and a combination of them matters more than any single item.
5. Imaging is not the only response to a flag
Where a flag is present, the appropriate response may be a further question, an examination finding, a blood test, or a conversation with the referring doctor. Treating "flag present" as "image now" is what drives the over-imaging problem — see imaging in low back pain.
What the evidence supports — and what it does not
Supported
- Screening for serious pathology as part of assessment. Prevalence is low but not zero. [4]
- Trauma, age over 70 and corticosteroid use as informative for vertebral fracture. [1]
- History of cancer, intravenous drug use, indwelling catheter and other infection site as raising likelihood. [4]
- The Canadian C-Spine rule for post-traumatic cervical fracture. [3]
- Using an explicit framework rather than an ad hoc list. [5]
Not supported
- Treating all listed red flags as equivalent. [1][2]
- Any particular published list. No consensus between guidelines. [2][3]
- Most individual flags in neck pain. Mainly expert opinion, agreement kappa 0. [3]
- Imaging on the basis of a single weak flag. [1][4]
- Assuming primary care prevalence matches emergency department figures. [4]
How certain is this?
Evidence grade: Low, and that is the point.
The grade describes the diagnostic accuracy of individual red flags, not the wisdom of screening. The Cochrane review is methodologically sound and its finding is largely negative: meta-analysis was impossible, risk of bias was low in only one study for two of the four domains, and only a small number of flags appeared informative. [1]
The two guideline reviews are consistent with each other and with the Cochrane finding. [2][3] Their conclusion — that red flags are mainly supported by expert opinion and that specific recommendations cannot be provided — is a strong claim from a good method.
The prevalence figures come from 22 studies and 41,320 patients, which is a substantial base, though the ranges are wide and the setting is emergency rather than outpatient physiotherapy. [4]
What would change the grade: prospective diagnostic accuracy studies of defined combinations of flags in primary care and direct-access physiotherapy populations, which is where most physiotherapists actually screen.
Common questions
Should I stop using red flags?
No. The evidence undermines the checklist's individual items, not the obligation to screen. Serious pathology requiring urgent treatment occurred in 2.5% to 5.1% of emergency department low back pain presentations, [4] and missing it matters enormously. What the evidence argues for is weighting: know the few flags with real likelihood ratios, treat the rest as prompts for thought rather than triggers for action, and refer when uncertain.
Which red flags actually carry weight?
For vertebral fracture: trauma (+LR 1.93 to 12.85), age over 70 (+LR 11.19), and corticosteroid use. [1] For spinal cancer: suspicion or history of cancer. For epidural abscess: intravenous drug use, indwelling vascular catheter, another infection site. [4] For post-traumatic cervical fracture, the neck pain review recommends the Canadian C-Spine rule specifically. [3]
Why do the guidelines I have read disagree?
Because most red flags are expert opinion rather than evidence. A review of 16 low back pain guidelines found 46 discrete red flags with no consensus, evidence of accuracy rarely provided, and five guidelines giving no supporting reference at all. [2] In neck pain, agreement across 29 guidelines had a median Fleiss' kappa of zero. [3]
If a flag is present, should I order imaging?
Not automatically, and physiotherapists in most settings are not ordering it anyway. The response to a flag is proportionate escalation: further questioning, examination, or referral to the doctor who can investigate. Given low prevalence, a single weak flag moves the probability very little, which is precisely how over-imaging happens. [4] See imaging in low back pain.
Does this apply to neck pain too?
It applies more strongly. The neck pain guideline review found 114 red flags across 12 guidelines that provided any, spanning fracture, cancer, infection, myelopathy, artery dissection and more, with median agreement of zero and mainly expert-opinion support. Its only specific recommendation was the Canadian C-Spine rule. [3] See neck pain and cervical radiculopathy.
References
- Han CS, Hancock MJ, Downie A, et al. Red flags to screen for vertebral fracture in people presenting with low back pain. Cochrane Database of Systematic Reviews. 2023 Aug 24;8(8):CD014461. doi:10.1002/14651858.CD014461.pub2 PMID 37615643 Cochrane diagnostic test accuracy review
- Verhagen AP, Downie A, Popal N, et al. Red flags presented in current low back pain guidelines: a review. European Spine Journal. 2016 Sep;25(9):2788–802. doi:10.1007/s00586-016-4684-0 PMID 27376890 Review of clinical practice guidelines
- Feller D, Chiarotto A, Koes B, et al. Red flags for potential serious pathologies in people with neck pain: a systematic review of clinical practice guidelines. Archives of Physiotherapy. 2024 Jan-Dec;14:105–115. doi:10.33393/aop.2024.3245 PMID 39639931 Systematic review of clinical practice guidelines
- Galliker G, Scherer DE, Trippolini MA, et al. Low Back Pain in the Emergency Department: Prevalence of Serious Spinal Pathologies and Diagnostic Accuracy of Red Flags. The American Journal of Medicine. 2020 Jan;133(1):60–72.e14. doi:10.1016/j.amjmed.2019.06.005 PMID 31278933 Systematic review
- Finucane LM, Downie A, Mercer C, et al. International Framework for Red Flags for Potential Serious Spinal Pathologies. Journal of Orthopaedic & Sports Physical Therapy. 2020 Jul;50(7):350–372. doi:10.2519/jospt.2020.9971 PMID 32438853 International framework
About this resource
Using this in clinic
Every figure here is traceable to its source.
This page separates the obligation to screen, which is not in question, from the diagnostic accuracy of individual checklist items, which mostly is. Where a value could not be verified against the paper it came from, it is not on this page, and the omission is stated rather than filled with a number from a secondary source.