Ottawa Ankle and Foot Rules Interpreter
Ottawa Ankle and Foot Rules Interpreter
Apply both Ottawa rules to an acute ankle injury, with the exclusions that define the population they were validated in — and the predictive values worked out from the published 2×2 tables, because near-perfect sensitivity and 25 to 49% specificity is the whole story.
Ottawa ankle and foot rules
Eligibility + zone + tenderness + weight bearingAn adult, alert and with no distracting injury, presenting the same day with malleolar zone pain, bone tenderness at the tip of the lateral malleolus, no midfoot tenderness, able to take four steps
The two rules, and the arithmetic that follows from them
foot series: pain in the midfoot zone AND (navicular or fifth metatarsal base tenderness OR unable to bear weight)
- four steps, limping allowed
- weight bearing is transferring weight twice onto each foot “Both immediately after the injury and for four steps in the emergency department”, and “Limping counts as bearing weight”. A patient who limps four steps has NOT met the criterion; a patient who walked into the department but could not stand at the roadside HAS met it
- the distal 6 cm, and no further
- malleolar tenderness means the posterior edge or tip of the lateral or medial malleolus — “the distal 6 cm of the posterior fibula” and “the distal 6 cm of the posterior tibia”. Ligamentous and anterior joint-line tenderness are not criteria
- the exclusions are part of the rule
- under 18, pregnancy, isolated skin injury, arrival with radiographs, injury over ten days old, re-attendance — and in common use intoxication or altered mental status, diminished sensation and distracting injury. Outside those, the rule has no published performance
- PPV and NPV depend on prevalence, and are computed here
- PPV = sens × prev / (sens × prev + (1 − spec) × (1 − prev)) and NPV = spec × (1 − prev) / (spec × (1 − prev) + (1 − sens) × prev). At the validation cohort’s own numbers the ankle rule’s PPV is 19.6% and its NPV 100%; at the 2022 meta-analysis’s 0.91 and 0.25 the PPV is 11.9% and the NPV 96.2% at 10% prevalence, and 17.6% and 94.0% at 15%
- high sensitivity, low specificity, and what that is for
- a rule of this shape is built to make a negative result informative and accepts that a positive one is not. Four positives in five in the validation cohort had no fracture. The BMJ review concluded that use “should reduce the number of unnecessary radiographs by 30-40%”, which is the claim the design supports
- derivation and validation
- Stiell I et al, “Decision rules for the use of radiography in acute ankle injuries. Refinement and prospective validation”, JAMA 1993;269(9):1127–1132: 1,032 patients in refinement and 453 in prospective validation, in two Canadian emergency departments. The paper itself was not reachable for this page and its figures here are taken from a structured appraisal that reproduces both 2×2 tables cell for cell
Worked example
An adult, alert and with no distracting injury, presenting the same day with malleolar zone pain, bone tenderness at the tip of the lateral malleolus, no midfoot tenderness, able to take four steps
Inside the validated population, so the first rule does not fire
Malleolar zone pain, so the ankle rule applies; no midfoot pain, so the foot rule does not
Malleolar bone tenderness is present, so the ankle rule's second criterion is met and the ankle rule is positive — weight bearing did not need to fail
Now the arithmetic. In Stiell's validation cohort 255 of 453 patients were rule-positive and 50 had a fracture: PPV = 50 / 255 = 19.6% at a fracture prevalence of 50 / 453 = 11.0%. Specificity = 198 / 403 = 49.1%
Take the tenderness away and keep everything else and the answer becomes no criterion met. In that cohort the rule missed none of 50 fractures, so the NPV was 198 / 198 = 100% — but at the 2022 adults-only meta-analysis's sensitivity of 0.91 and specificity of 0.25, per 1,000 patients at 10% prevalence the 2x2 is 91 true positives, 9 false negatives, 225 true negatives and 675 false positives, giving PPV 11.9% and NPV 96.2%
Push the prevalence to 15% with the same sensitivity and specificity and, per 10,000, the 2x2 is 1,365 / 135 / 2,125 / 6,375 — PPV 17.6% and NPV 94.0%. Six missed fractures in every hundred negatives. The rule has not changed; only the population has
Add midfoot pain and navicular tenderness to the original case and both rules are positive
The two published 2×2 tables, and the predictive values computed from them
| Validation cohort, 453 patients | Fracture | No fracture | Derived |
|---|---|---|---|
| Ankle rule positive | 50 | 205 | PPV = 50 / 255 = 19.6% |
| Ankle rule negative | 0 | 198 | NPV = 198 / 198 = 100% |
| Ankle rule, totals | 50 (prevalence 11.0%) | 403 | sensitivity 100% (95% CI 93 to 100%), specificity 198 / 403 = 49.1% (44 to 54%) |
| Foot rule positive | 19 | 90 | PPV = 19 / 109 = 17.4% |
| Foot rule negative | 0 | 344 | NPV = 344 / 344 = 100% |
| Foot rule, totals | 19 (prevalence 4.2%) | 434 | sensitivity 100% (95% CI 83 to 100%), specificity 344 / 434 = 79.3% (75 to 83%) |
Four published performance estimates for one rule, and what each implies at the bedside
| Source and cohort | Sensitivity | Specificity | PPV and NPV at 10% prevalence |
|---|---|---|---|
| Stiell 1993 validation, ankle rule, 453 patients | 100% (93 to 100%) | 49% (44 to 54%) | PPV 17.9%, NPV 100% — computed at 10% for comparability, not the cohort’s own 11.0% |
| BMJ 2003 systematic review, 27 studies, 15,581 patients | “almost 100%”; 99.6% (98.2 to 100.0) where applied within 48 hours | “modest”; “ranged from 10% to 79%” across studies | Pooled negative likelihood ratio 0.08 (0.03 to 0.18) for the ankle and 0.08 (0.03 to 0.20) for the mid-foot; no single pooled sensitivity or specificity is given |
| Beckenkamp et al, as quoted in the 2022 review, ankle and midfoot together | 99.4% (97.9 to 99.8) | 32.3% (28.8 to 42.3) | PPV 14.0%, NPV 99.8% |
| Gomes 2022 meta-analysis, adults and ankle fractures only, 15 studies, 8,560 patients | 0.91 (0.89 to 0.92) | 0.25 (0.24 to 0.26) | PPV 11.9%, NPV 96.2% |
| Am Fam Physician 2003 commentary on the BMJ review | “97.6 percent (96.4 to 98.9)” | “31.5 percent (23.8 to 44.4)” | Neither figure appears in the BMJ paper’s own text. Printed here as a disagreement, not used as a number |
| Seaberg 1998, Ottawa KNEE rule, 750 of 934 patients, 87 fractures | 97% (90 to 99%) | 27% (23 to 30%) | Included because the knee rule is a different rule for a different joint; the Pittsburgh rule on 745 of the same patients gave 99% (94 to 100%) and 60% (56 to 64%) |
Built to be sensitive, and that is why four positives in five find nothing
The Ottawa rules are two structured examinations. An ankle series is indicated when there is pain in the malleolar zone together with bone tenderness at the posterior edge or tip of either malleolus, or an inability to bear weight. A foot series is indicated when there is pain in the midfoot zone together with tenderness at the navicular or the base of the fifth metatarsal, or an inability to bear weight. Weight bearing means four steps, twice onto each foot, both immediately after the injury and again in the department, and limping counts as bearing weight.
The rules were designed to be sensitive at the cost of specificity, and the arithmetic that follows from that design is the most useful thing on this page. In the prospective validation cohort of 453 patients, the ankle rule was positive in 255 and 50 of those had a fracture: a positive predictive value of 19.6% at an 11.0% fracture prevalence, with specificity 49%. Four positives in five had no fracture. The foot rule was more specific at 79% and yet had a lower positive predictive value, 17.4%, because midfoot fracture was four times rarer in the same cohort. A predictive value quoted without its prevalence is not a number, and this is the cleanest demonstration of why.
The sensitivity is where the sources disagree, and the disagreement matters. The 2003 BMJ systematic review of 27 studies and 15,581 patients described “a sensitivity of almost 100%” and reported pooled negative likelihood ratios of 0.08 for both the ankle and the mid-foot. A 2022 meta-analysis restricted to adults and to ankle fractures, 15 studies and 8,560 patients, found a pooled sensitivity of 0.91 (95% CI 0.89 to 0.92) and specificity of 0.25, and its authors note explicitly that this is lower than earlier reviews. At 0.91 and 0.25, the negative predictive value is 96.2% at a 10% fracture prevalence and 94.0% at 15% — around four to six fractures in every hundred patients the rule calls negative. Those are not the same rule in practice as one with 99.6% sensitivity, and a reader is entitled to see both.
Two things this page will not do. It will not say that imaging can be omitted, because the published figures are the performance of a rule in a cohort and the decision about one patient belongs to the clinician in front of them. And it will not apply the rule where the exclusions bite — under 18, intoxicated or with altered sensorium, with diminished sensation from any cause including regional anaesthesia, with other distracting injuries, or presenting more than about ten days later. Each of those exclusions exists because the rule is a physical examination and each of them corrupts the examination. The commonest way to misuse a decision rule is not to misread its criteria but to apply it to someone it was never tested on.
A decision rule is not a diagnosis and it is not an authorisation. It is a structured physical examination with a published sensitivity and specificity in named cohorts, and those two numbers are properties of those cohorts rather than facts about the patient in front of you. Applying a rule outside the population it was validated in — which is what its exclusion list defines — leaves it with no established performance at all. Whether to image is the clinician’s judgement, and this page is one input to it. This page states which published category the entered findings fall in and what the literature reports for that category in the cohorts it was measured in. It renders no clinical or surgical decision: not whether to operate, not which implant, not whether to take or withhold an image, and not whether a limb can take weight. Those are the treating clinician’s, and the classification is one input among many.
Frequently asked questions
What are the Ottawa ankle rules?
Two rules. An ankle radiograph series is indicated if there is pain in the malleolar zone and either bone tenderness at the posterior edge or tip of the lateral or medial malleolus, or an inability to bear weight. A foot series is indicated if there is pain in the midfoot zone and either tenderness at the navicular or the base of the fifth metatarsal, or an inability to bear weight.
What counts as being unable to bear weight?
Being unable to transfer weight twice onto each foot — four steps — both immediately after the injury and again in the department. Limping counts as bearing weight, so a patient who limps four steps has not met the criterion. A patient who walked into the department but could not stand immediately after the injury has met it.
How sensitive and specific are the rules?
In the 453-patient prospective validation cohort, sensitivity was 100% for both rules, with specificity 49% for the ankle rule and 79% for the foot rule. A 2003 systematic review of 27 studies and 15,581 patients reported sensitivity as “almost 100%”, a modest specificity ranging from 10% to 79% across studies, and pooled negative likelihood ratios of 0.08. A 2022 meta-analysis of 15 studies and 8,560 adults found a pooled sensitivity of 0.91 and specificity of 0.25, lower than the earlier reviews.
If the rule is negative, is a fracture excluded?
No. A negative rule means no published criterion was met. In the validation cohort the rule missed none of 50 ankle fractures and none of 19 midfoot fractures, so the negative predictive value in that cohort was 100% — but computed from the 2022 meta-analysis’s sensitivity of 0.91 and specificity of 0.25, the negative predictive value is 96.2% at a 10% fracture prevalence and 94.0% at 15%. Whether to image remains a clinical judgement about the patient in front of you.
Why is the positive predictive value only about 20%?
Because the rules were deliberately built for sensitivity and accept poor specificity. At the validation cohort’s 11.0% fracture prevalence and 49% specificity, 255 of 453 patients were rule-positive and 50 had fractures. The foot rule shows the same point more sharply: it is more specific (79%) and has a lower positive predictive value (17.4%), because midfoot fracture is rarer. The 2003 review’s own claim for the rules is a 30 to 40% reduction in unnecessary radiographs, not that a positive result predicts a fracture.
Do the rules apply to children or to intoxicated patients?
Not as derived. The derivation excluded patients under 18, pregnancy, isolated skin injuries, referrals arriving with radiographs, injuries more than ten days old and re-attendances for the same injury; the exclusion list in common use adds intoxication or altered mental status, diminished sensation from a neurological deficit including regional anaesthesia, and multiple painful or distracting injuries. The paediatric literature is separate and the 2022 meta-analysis quoted here is adults only.
Is there an Ottawa knee rule page?
Not yet. The knee rule is a different rule for a different joint with its own validation, and the figures worth knowing are from Seaberg’s head-to-head comparison in 934 patients: the Ottawa knee rule, applied to 750 of them with 87 fractures, had a sensitivity of 97% (95% CI 90 to 99%) and a specificity of 27% (23 to 30%), while the Pittsburgh rule on 745 patients with 91 fractures had 99% (94 to 100%) and 60% (56 to 64%). The authors concluded the Pittsburgh rule was more specific without a loss of sensitivity.
Related calculators
References
- Stiell I, et al. Decision rules for the use of radiography in acute ankle injuries. Refinement and prospective validation. JAMA. 1993;269(9):1127–1132. Read as the structured appraisal at wikem.org/wiki/EBQ:Ottawa_Ankle_Rule and its companion study page; the paper itself was not reachable. Two prospective Canadian emergency department cohorts: 1,032 patients in the refinement stage and 453 in the validation stage. Ankle rule: an ankle series is needed with pain in the malleolar zone AND bone tenderness at the posterior edge or tip of the lateral malleolus, OR at the posterior edge or tip of the medial malleolus, OR inability to bear weight. Foot rule: “Foot x-ray series needed if” “Pain in the midfoot” AND “Inability to bear weight both immediately and in the ED (4 steps)” OR “Tenderness at the navicular” OR “Tenderness at the base of the 5th metatarsal”. Exclusions as listed: “Age <18 years old”, “Pregnant”, “Isolated injuries of the skin”, “Referred from outside hospital with radiographs”, “Injury occurred >10 days previously”, “Returned for reassessment of the same injury”. Validation 2×2, ankle: 50 true positives, 205 false positives, 0 false negatives, 198 true negatives — sensitivity 100% (95% CI 93% to 100%), specificity 49% (44% to 54%). Validation 2×2, foot: 19, 90, 0, 344 — sensitivity 100% (83% to 100%), specificity 79% (75% to 83%). Every predictive value on this page is computed from those two tables.
- Accuracy of Ottawa ankle rules to exclude fractures of the ankle and mid-foot: systematic review. BMJ. 2003;326(7386):417. Full text read at bmj.com. “32 studies met the inclusion criteria and 27 studies reporting on 15 581 patients were used for meta-analysis.” “The pooled negative likelihood ratios for the ankle and mid-foot were 0.08” — 95% CI 0.03 to 0.18 for the ankle and 0.03 to 0.20 for the mid-foot. “The instrument has a sensitivity of almost 100% and a modest specificity”; specificity across studies “ranged from 10% to 79%”; sensitivity “99.6% (95% confidence interval 98.2% to 100.0%)” where the rule was applied within 48 hours; “its use should reduce the number of unnecessary radiographs by 30-40%”. The paper gives NO single pooled sensitivity or specificity with a confidence interval, which matters because secondary sources quote them as though it did.
- Gomes YE, Chau M, Banwell HA, et al. Diagnostic accuracy of the Ottawa ankle rule to exclude fractures in acute ankle injuries in adults: a systematic review and meta-analysis. BMC Musculoskelet Disord. 2022;23:885. Fifteen studies, 8,560 adult patients, 13 countries. Pooled sensitivity 0.91 (95% CI 0.89 to 0.92) and pooled specificity 0.25 (0.24 to 0.26) — the authors note their sensitivity is lower than in earlier reviews. They conclude the rule “could be used confidently to rule out ankle fractures” but that its lower specificity “increas[es] the likelihood of false-positive outcomes”. The same paper quotes Beckenkamp et al’s combined ankle-and-midfoot figures of 99.4% (97.9% to 99.8%) sensitivity and 32.3% (28.8% to 42.3%) specificity.
- Seaberg DC, Yealy DM, Lukens T, Auble T, Mathias S. Multicenter comparison of two clinical decision rules for the use of radiography in acute, high-risk knee injuries. Ann Emerg Med. 1998;32(1):8–13. Abstract read at em-consulte.com/article/613232. 934 patients enrolled. The Pittsburgh rule applied to 745 patients with 91 fractures (12.2%), missed one, sensitivity 99% (95% CI 94% to 100%) and specificity 60% (56% to 64%). The Ottawa knee rule applied to 750 patients with 87 fractures (11.6%), missed three, sensitivity 97% (90% to 99%) and specificity 27% (23% to 30%). Cited here because a reader looking for the knee rule should be told what the head-to-head found; the knee rule itself is not implemented on this page.
- Ottawa Rules Simplify the Management of Ankle Injuries. Am Fam Physician. 2003;68(4):745–746. A commentary on the BMJ review above, which reports its results as a sensitivity of “97.6 percent (95 percent confidence interval [CI], 96.4 to 98.9)” and a specificity of “31.5 percent (95 percent confidence interval [CI], 23.8 to 44.4)”. Neither figure appears in the BMJ paper’s own text, and the commentary attributes both to a table it does not reproduce. Printed here as a disagreement rather than used as a number.
- Ottawa ankle rules reference guide, casrai.org/guides/ottawa-ankle-rules. The malleolar tenderness landmarks as “the distal 6 cm of the posterior fibula” and “the distal 6 cm of the posterior tibia”; weight bearing as being unable to transfer weight twice onto each foot, four steps, “Both immediately after the injury and for four steps in the emergency department”, with “Limping counts as bearing weight”; and the wider exclusion list in common use — intoxication or altered mental status, diminished sensation from a neurological deficit, and multiple painful or distracting injuries — which the derivation paper’s own list does not name in those words.
Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/