Corrections and updates | Rule It In, Rule It Out

Corrections

The first edition has not been released. Corrections to the free test library and red-flag pages on this site are listed below.

16 October 2026: wording

The note on the 15 October audit below now says how the checks were done. The red-flag pages now say they are not a complete list and should be used with local protocols; each test page now says it is drafted with AI assistance; and a terms of use page was added. No figures changed.

15 October 2026: full-site audit

Every page on the site was checked against the published studies on the internet, in checks run separately from the drafting (AI-assisted, with the results reviewed by the author). All 133 test citations pointed to the right study and every DOI and PubMed link worked; one set of page numbers was corrected (Coris 2003, listed below). Changes made as a result:

Thomas test: the previous figures (sensitivity 75%, specificity 43%) came from a review that attributed them to the Thomas test; the original study (Narvani 2003) reported them for a different manoeuvre and found the Thomas test neither sensitive nor specific. The card now uses the only published accuracy figures we could find for the test itself (Vigotsky 2016, hip flexor tightness: sensitivity 32%, specificity 57%), with a caution.
Single-leg decline squat: the full paper (Mendonça 2016) was obtained: the target is patellar tendon abnormality on ultrasound; sensitivity 62.5%, specificity 85%, LR− 0.4 added; the test is done on a 30° decline board to 30° of knee flexion.
Piano key test: figures checked against the full thesis (Tuhi 2023): LR+ 1.88 as published (continuity-corrected); the how-to now loads the ulna about 4 cm above the distal radio-ulnar joint, as in the study.
Tinel’s sign at the wrist: likelihood ratios now as published by the review (LR+ 2.10, LR− 0.67; previously our calculated 2.05 and 0.71).
Hornblower’s sign: the how-to previously described the drop sign; it now describes Hornblower’s sign as in Walch 1998. The source did not report an LR−, so none is shown.
Arm squeeze test: full paper obtained (Gumina 2013); LR+ 10.6 as published.
Hook test and biceps crease interval: now labelled for complete distal biceps tendon rupture, which is what the study measured; hook test sensitivity 71.4%, specificity 95.5%.
Spurling’s test: LR− now as published by the review (0.28; previously our calculated 0.51), with a note on the difference.
Sacroiliac tests (Laslett 2005): LR− corrected for thigh thrust (0.18), compression (0.46) and sacral thrust (0.50); Gaenslen’s now shows the published right-side figures (sensitivity 53%, specificity 71%) instead of an average of the two sides.
Web-space tenderness (Morton’s neuroma): the note now gives the counts behind the 100% specificity (one foot without a neuroma) and its confidence interval.
Notes updated where a full paper has since been obtained (O’Brien’s, Clarke’s, TFCC grind, grip strength difference) and cautions added to the Thompson, Matles and palpable Achilles gap tests (single study at high risk of bias).
Strongest tests page: tests whose figures were checked against the abstract only are now marked as such in the rankings.
Other: the home page no longer says every test has all three figures; the lumbar red-flag page now links to the thoracic page for the bedside signs of vertebral fracture; this corrections page is now in the site map and linked from the author, book and contact pages; Coris 2003 page numbers corrected.

13 October 2026 (second update): figures re-checked against the published studies

Every test figure in the library was re-checked against its published study wherever the study could be accessed; cards whose figures could not yet be checked against the full paper say so. Changes:

Source notes: each test card now says whether its figures were checked against the full paper or against the published abstract only, which likelihood ratios we calculated ourselves, and which figures have not yet been checked against the full paper.
Ulnar fovea sign: likelihood ratios now as published (LR+ 1.73, LR− 0.22; previously our calculated 1.71 and 0.23).
Elbow extension test: LR− shown as published in the full paper (0.06, adults and children combined).
Unable to hop and dorsiflexion–external rotation tests: specificity and LR+ recalculated from the published sensitivity and LR− (single-leg hop: specificity 30%, LR+ 1.27; dorsiflexion–external rotation: LR+ 1.92).
Windlass test: the figures apply to the standing version; the how-to now says so.
Cautions added: Hoover test (small unblinded study) and TFCC grind test (only 7 people without the injury); notes on study populations added for several hip, shoulder and lumbar tests.
Red-flag pages: the knee dislocation row (which also covers tibial plateau fracture) now includes compartment syndrome; the hand compartment syndrome and bite-wound rows now cite peer-reviewed sources.
Layout: on phones, comparison tables now show one card per test.

13 October 2026: test library and red-flag pages

A re-check of the book and the library against the cited studies led to these changes:

Empty can, full can, drop arm, external rotation lag, hornblower's and belly-press tests: now labelled as tests for rotator cuff tear. The study behind their figures (Jain 2017) compared people with any rotator cuff tear, partial or full-thickness and in any tendon, with people without a tear; they were previously labelled for single tendons (supraspinatus, infraspinatus or subscapularis). The comparison page for supraspinatus tear now points to the rotator cuff tear page.
Matles test: LR+ now shown as published by the review (6.29; previously our calculated 5.87), with the probability ladder updated.
Passive lumbar extension test: the caution now uses a cited study (Esmailiejah 2018: sensitivity 79% in 52 surgical patients, instability judged at operation), replacing an uncited figure of about 43%. The reference standard (flexion–extension radiographs) is now stated.
Patellar-pubic percussion test: target changed from "occult hip fracture" to "suspected hip fracture after trauma", which is what the studies tested.
Dorsiflexion–external rotation test: "Kleiger's" removed from the test's title, as the study used the dorsiflexion–external rotation stress test; Kleiger's test is kept as an alternative name.
Small figure corrections: pelvic compression LR+ 13 (was 12.94, recalculated from the study's counts), full can LR+ 3.75 (was 3.7, as published).
Verdicts: ankle clonus and inverted supinator sign changed from "best used to rule in" to "treat with caution", in line with their existing cautions; the Thompson test moved out of the strongest-tests rankings into the list of tests to read with caution.
Cautions added or strengthened where a figure rests on a single small study or a design that flatters accuracy: Thompson, Matles and palpable gap (Achilles rupture), Royal London Hospital test, grip strength, slump test, cervical distraction (a 2026 validation found LR+ 1.43), single-leg stance and resisted external derotation.
Biceps crease interval: a positive is now defined as more than 6 cm only; the ratio criterion gave the same result in every patient.
Red-flag pages: rows without a stated action now say how urgently to act; rows added for Achilles tendon rupture and distal biceps rupture; the septic arthritis rows on the shoulder, elbow, hip and knee pages note that fever is often absent.
Source details completed or corrected on several pages (authors, page numbers, links, reference standards), cautions reworded on the arm squeeze and TFCC compression tests, and notes added saying whether each likelihood ratio is as published or calculated.

Feedback

We're happy to add to or improve the special tests library and the book if you have any feedback. Suggestions of newer studies are welcome at info@trokos.com.au. Changes are posted here and noted in the Rule It In newsletter.

Page created 4 October 2026. Updated 16 October 2026.