An expert synthesis of the clinical examination of the hand and wrist — what each named test actually measures, how it was described, how accurate it is, and where the evidence contradicts the textbook — for FESSH candidates and consultant peers.
Most hand and wrist diagnoses are made in the clinic room, and the tests that make them carry names rather than numbers. The numbers exist. Systematic reviews and diagnostic-accuracy studies now give sensitivities, specificities and likelihood ratios for the carpal tunnel manoeuvres, the scaphoid shift, the grind test and the collateral-ligament stress tests, and reliability studies show which signs two examiners can agree on123. The pattern that emerges is consistent: no single test is adequate, combinations perform better than their parts, and an eponymous test performed as its author described it behaves differently from the version that has drifted into the textbooks45. This article sets the tests against that evidence, in the order of a structured examination.

§ 01 · THE GENERAL SCREEN — LOOK, FEEL, MOVE

Dominance and the ten per cent rule

Grip strength is measured with a Jamar dynamometer in the standardised seated position — shoulder adducted, elbow flexed to 90°, forearm neutral; the mean of three trials gives the highest test–retest reliability, and the elbow position matters — grip is significantly stronger at 90° of flexion than in full extension67. Age- and sex-stratified norms exist for 20 to 94 years, with peak grip between 25 and 39 and pinch stable to 59 before declining8. The "ten per cent rule" — that the dominant hand is a tenth stronger — holds for right-handers (12.7 per cent) and not at all for left-handers, whose hands are equivalent, and in a third of left-handers the non-dominant grip is the stronger910.

MRC grading

The Medical Research Council scale remains the language of motor assessment; in radial palsy its inter-rater agreement is substantial (weighted κ 0.77–0.78, intra-rater 0.82–0.86) and its correlation with dynamometry 0.7811. Its weakness is the breadth of grade 4, which most modified scales subdivide. of sensation (S0–S4) uses the same letters for a different scale.

Range of motion, total active motion and the Kapandji score

Total active motion (TAM) — the sum of active flexion at the metacarpophalangeal, proximal and distal interphalangeal joints minus the sum of the extension deficits — was recommended for wider adoption by the IFSSH tendon committee in 1983, alongside the Buck-Gramcko system, and Strickland's percentage of interphalangeal motion is the other convention1213. Thumb opposition is scored on Kapandji's ten-stage scale: 1 for the lateral side of the middle phalanx of the index, 3 for the index tip, 6 for the little-finger tip, 8 for its proximal interphalangeal crease and 10 for the distal palmar crease — valid only if the early stages are achievable, because a thumb that crawls across the palm is not opposing14. The score has a strong ceiling effect; the pronation angle plateaus around stage 615.

Rotation and the cascade

Rotational malalignment after a metacarpal or phalangeal fracture is judged with the fingers flexed: in a prospective series of 98 metacarpal fractures a quarter had less than 10° of rotation and five had more, and assessment must include an end-on view of the fingernail because metacarpophalangeal motion is often restricted after the fracture16. The comparison with the other hand is less reliable than it feels: normal little-finger overlap in flexion averaged 25 per cent of the adjacent nail, ranged from 0 to 71 per cent, and differed between a person's own hands by 16 per cent on average17.
Examination notes. Jamar, elbow 90°, mean of three; ten per cent rule right-handers only; MRC κ ≈ 0.78; TAM = sum of active flexion − sum of extension deficits; Kapandji 10 = distal palmar crease, ceiling at stage 6; rotation judged in flexion with an end-on nail view, contralateral overlap varies by 16 per cent.

Lumps, nodules and the table-top

A dorsal wrist ganglion arises from the scapholunate joint; the "occult" ganglion presents as dorsal wrist pain with isolated tenderness over the scapholunate junction and no palpable lump, and ganglionic degeneration of the dorsal scapholunate ligament was found in eight of ten such patients at operation18. A progresses to a cord in about half of patients over eight years and regresses in 12 per cent; onset before 50 predicts progression, and the standard surgical criteria remain a metacarpophalangeal contracture of more than 30° or any proximal interphalangeal contracture19. The — the patient's inability to lay the hand flat — is the classical threshold for considering surgery20.

The infected finger: Kanavel's signs

Kanavel's cardinal signs of flexor sheath infection — tenderness along the sheath, a flexed posture, pain on passive extension and fusiform swelling — date from the early twentieth century and are still the basis of the diagnosis21. Tested against other finger infections in a retrospective series, the four were 91–97 per cent sensitive but only 51–69 per cent specific, and the independent predictors of pyogenic tenosynovitis were sheath tenderness, pain on passive extension and symptoms of under five days (area under the curve 0.91)22. Prognosis follows Pang's classification: no amputations without subcutaneous purulence or ischaemia, 8 per cent with purulence, 59 per cent with ischaemia23. Tenderness over the A1 pulley, a refinement of "tenderness along the sheath", had the highest odds ratio and specificity of any sign in one series24.

§ 02 · NERVE EXAMINATION

The median nerve at the wrist

The three classical provocative tests for are less accurate than their reputation. MacDermid's systematic review of 60 studies gave weighted sensitivities and specificities of 68 and 73 per cent for the , 50 and 77 per cent for the and 64 and 83 per cent for carpal compression, with and or abductor pollicis brevis weakness specific but insensitive1. A 2023 meta-analysis pooled lower figures still — Phalen 0.57 and 0.67, Tinel 0.45 and 0.78 — with single-study sensitivities ranging from 0.12 to 0.97, and concluded that no single manoeuvre is adequate4. Durkan's , direct pressure over the carpal tunnel, outperformed Tinel and Phalen in its original series25, but an independent evaluation found its specificity only 18 per cent26. The — both hands held overhead until symptoms appear — had 75.5 per cent sensitivity and 98.5 per cent specificity in its description27, and the highest pooled diagnostic odds ratio of any manoeuvre (128.6, against 15.8 for Durkan, 7.2 for Phalen and 5.3 for Tinel), on an evidence base of fewer than 900 wrists28.
The examination performs better as a set. The of Graham weights six history and examination items into a probability that correlated 0.71 with expert judgement29, and when that pre-test probability is 0.80 or more — as it was in 73 per cent of referrals — electrodiagnosis changes it by only 0.02 to 0.0630. Inter- and intra-observer reliability is satisfactory for static and moving two-point discrimination, Tinel, Phalen, vibration and manual testing of abductor pollicis brevis, but not for monofilament testing (intra-observer reliability was satisfactory for all seven tests), and hand surgeons and therapists are more reliable than occupational-health examiners2.

The scratch collapse test

Cheng and Mackinnon described the in 2008: the patient resists bilateral shoulder external rotation with the elbows flexed, the skin over the suspected compression site is lightly scratched, and momentary loss of resistance on that side is positive — 64 per cent sensitive for carpal tunnel and 69 per cent for cubital tunnel syndrome in the inventors' hands31. Independent evaluation found a sensitivity of 31 per cent (against 67 per cent for Phalen, 77 per cent for Durkan and 88 per cent for a lax CTS-6) and a specificity of 61 per cent26; a meta-analysis of three level-II studies pooled a sensitivity of 0.32, a specificity of 0.62 and likelihood ratios of 0.75 and 1.03 — a test that does not change the probability of disease32. It is a confirmatory sign at best. Current narrative reviews still rank it above Tinel and elbow flexion for cubital tunnel syndrome33, but that ranking traces back to the inventors' 2008 series, and the independent figures above give no reason to rely on it.

Sensibility: threshold versus innervation density

Under experimental compression of the , threshold tests — and vibration — change first, whereas innervation-density tests (static and moving two-point discrimination) stay normal until nearly all sensory conduction has ceased34. That is why monofilaments are the sensitive test in compression and two-point discrimination the sensitive test after nerve repair. The threshold chosen decides the monofilament's performance: with the conventional 2.83 filament the test was 98 per cent sensitive and 17 per cent specific, and with the 3.22 filament 80 per cent sensitive and 93 per cent specific for moderate-to-severe disease35; in a predominantly male veteran population a conservative protocol was 98 per cent sensitive and 15 per cent specific and a liberal one 13 and 88 per cent, so neither was useful36; across sixteen studies the 3.22 filament in any radial digit was the most sensitive test (0.49–0.96), while thenar atrophy (0.96–1.00), two-point discrimination (0.81–0.98) and grip and pinch strength were the specific ones37. Dellon introduced moving two-point discrimination in 1978 to test the quickly adapting fibre-receptor system, static (Weber) two-point discrimination testing the slowly adapting one38; normal static values in adults are about 3 mm at the fingertips and 5–7 mm in the mid-palm39; in carpal tunnel syndrome the middle finger is the most symptomatic and the worst-discriminating digit (mean static value 6 mm), and only its two-point discrimination correlates, weakly, with electrodiagnosis40. Strauch's Ten Test asks the patient to rate moving light touch as a ratio out of ten against the normal side and is reliable between examiners41.
Examination notes. Pooled Phalen 57–68 / 67–73 per cent, Tinel 45–50 / 77–78 per cent, Durkan 64 / 83 per cent (independent specificity as low as 18 per cent); hand elevation 75 / 98 per cent, highest diagnostic odds ratio; CTS-6 ≥ 0.80 → electrodiagnosis adds little; scratch collapse sensitivity 31 per cent independently and 32 per cent pooled — confirmatory only; monofilaments change first in compression, two-point discrimination last; 3.22 filament for specificity.

The ulnar nerve at the elbow

Novak's provocative series set the benchmarks for : Tinel at the elbow 0.70, elbow flexion at 30 seconds 0.32, pressure over the nerve 0.55 and combined pressure with flexion 0.91, each rarely positive in controls42. The as Buehler and Thayer described it is full elbow flexion with full wrist extension for three minutes, positive in all 13 of their patients43; shorter versions perform worse (25 per cent at five seconds, 36 per cent at ten), and a positive 45-second test did not correlate with the extraneural pressure measured at operation444546. Adding shoulder internal rotation — the arm abducted 90°, the shoulder in maximal internal rotation, the elbow flexed — raised sensitivity to 80 per cent at ten seconds and 87 per cent at five, with specificities of 98–100 per cent4544.

Ulnar motor signs

(flexor pollicis longus substituting for a weak adductor pollicis during key pinch), Jeanne's sign (hyperextension of the thumb metacarpophalangeal joint in the same manoeuvre), the (actively irreducible abduction of the little finger from unopposed extensor digiti minimi with a weak third palmar interosseous), the crossed-finger test, Egawa's sign and the Duchenne claw are the motor vocabulary of ulnar palsy; compared with motor conduction velocity in 26 patients, fasciculations and the finger-flexion sign were the most frequent, and each sign except finger flexion separated slower from faster conduction4748. Wartenberg's sign is also seen during regeneration, and a "pseudo-Wartenberg" occurs without neuropathy4950. The Bouvier manoeuvre — passive block of metacarpophalangeal hyperextension restoring interphalangeal extension — guides the choice of procedure for the claw hand49.

The anterior interosseous nerve

Kiloh and Nevin described isolated anterior interosseous neuritis in 195251; the OK sign fails because flexor pollicis longus and the index profundus cannot flex, but the palsy is frequently incomplete — isolated flexor pollicis longus loss in half of one series52. The "hand of benediction" is not the posture of high median palsy: the index metacarpophalangeal joint still flexes through the interossei and the middle finger flexes through the dual innervation of its profundus and the quadriga53.

The Hoffmann-Tinel sign in regeneration

Hoffmann and Tinel described the tingling sign independently in 19155455; it signals regenerating axons and, advancing distally, the regeneration front, but not the quality of eventual recovery — a misconception as old as the sign56. After open carpal tunnel release the Phalen test remits earlier than the Tinel sign57.

§ 03 · TENDON EXAMINATION

Superficialis, profundus and the little finger

Flexor digitorum superficialis is tested by holding the other fingers in extension and asking for proximal interphalangeal flexion; profundus by blocking the middle phalanx and asking for distal flexion. The little-finger superficialis is functionally deficient in a third of normal hands on the standard test and 16 per cent when the ring finger is allowed to flex with it58 — the tendon is present in every cadaver hand, but its function is independent in only 58 per cent, shared with the ring finger in 21 per cent and absent in 21 per cent, and asymmetric between hands in a quarter of people59. A "positive" superficialis test in the little finger is therefore normal in one person in three, and an isolated profundus injury in such a finger risks loss of all interphalangeal flexion.

The central slip: Elson's test

Closed rupture of the is easily missed until the appears. Elson's test flexes the proximal interphalangeal joint to 90° over the edge of a table and asks the patient to extend it against resistance: absence of extension force at the proximal joint together with rigid extension of the distal joint is an immediate sign of complete rupture, whereas Boyes' test (inability to flex the distal joint with the proximal joint held extended) becomes positive only late60. In a cadaver model Boyes' test was unreliable for both the pre-boutonnière and the passively correctable boutonnière, and was the only manoeuvre that detected the divided slip in both; resisted extension should be tested with the joint flexed to weaken the 61. Immediately after transection the extra resistance to distal-joint flexion is a matter of a newton or so, probably imperceptible clinically, so delayed re-testing is advised to increase the test's sensitivity, especially when pain limits the first examination62. Extensor continuity is also read through tenodesis: wrist flexion should extend the metacarpophalangeal and interphalangeal joints passively if the tendon is intact.

Intrinsic, extrinsic and retinacular tightness

Bunnell's compares passive proximal interphalangeal flexion with the metacarpophalangeal joint extended and flexed; tight intrinsics allow less flexion when the metacarpophalangeal joint is extended, tight extrinsic extensors the reverse63. The mechanics are measurable: moving the metacarpophalangeal joint from 60° of flexion to neutral raises the equilibrium angle of the proximal interphalangeal joint by more than 20° and its stiffness by more than half, an effect that is probably amplified when the intrinsics are tight64. The oblique retinacular ligament test — greater distal interphalangeal flexion with the proximal joint flexed than extended — rests on Landsmeer's ligament, which is variably present and attenuated in some dissections and omnipresent in others; its check-rein role underlies the swan-neck and mallet reconstructions even where its dynamic function is disputed6566.

Quadriga

Verdan's syndrome describes the loss of flexion in the adjacent fingers when one profundus tendon is shortened or tethered, because the profundus tendons share a muscle belly and cross-connections67; in cadavers 10 mm of shortening of the little-finger profundus left tip-to-palm deficits of 10–14 mm in the ring and middle fingers68, and splinting any single finger in extension reduces the total active motion of all the others, worst for the ring finger69.

Sagittal band, trigger finger and mallet

Closed injury is graded by Rayan and Murray into type I without extensor instability, type II with subluxation and type III with dislocation over the metacarpal head; the little finger mostly injures its radial band, in about half of cases with an abduction deformity, and splinting within three weeks gives satisfactory results70. are graded on Quinnell's 1980 scale (from uneven movement, through actively and then passively correctable triggering, to a fixed flexion deformity)71, and the examiner feels for the nodule at the A1 pulley while the finger moves. A mallet finger is an extensor lag at the distal joint; Wehbé and Schneider's review of mallet fractures showed remodelling and near-normal painless motion in all but one of 21 fractures, splinted or operated, with surgery offering no advantage — irrespective of fragment size or subluxation72. Rheumatoid finger deformities are examined and staged by Nalebuff's classifications of the and boutonnière deformities, which group the fingers by the flexion available in each joint position, by passive correctability and by the radiographic state of the joint7374.
Examination notes. Little-finger FDS test positive in a third of normal hands; Elson (PIP 90° over the table, resisted extension, rigid DIP) beats Boyes and may need repeating after a few days; Bunnell: less PIP flexion with the MCP extended = tight intrinsics; quadriga = 10 mm shortening costs 10–14 mm of neighbouring flexion; sagittal band types I–III; mallet fractures remodel.

§ 04 · JOINT STABILITY

The thumb ulnar collateral ligament

Heyman's cadaver and clinical study explains why the thumb metacarpophalangeal joint is stressed in both extension and 30° of flexion: dividing the proper collateral ligament increases valgus laxity mainly in flexion, and laxity in extension rises only when the accessory ligament and volar plate are also divided; clinically, more than 35° of valgus instability in extension consistently indicated a tear of both parts, and a was present in 15 of 17 such thumbs75. The working thresholds — more than 30°, or more than 15° more than the other thumb, for a complete rupture — come from Heyman's review76, but the contralateral comparison is unreliable: in 200 normal thumbs 34 per cent differed by 10° or more between sides in extension and 12 per cent by 15° or more, so the absence of a firm end point is the better criterion77. Ultrasound detects the Stener lesion with pooled sensitivity and specificity of 95 and 94 per cent (MRI 93 and 98 per cent) and is the first-line imaging test78.

Finger collateral ligaments and the cam effect

At the proximal interphalangeal joint the lateral collateral ligament is the primary restraint, its palmar fibres tight in extension and first to fail; lateral stress of more than 20° means complete disruption, whereas less than 20° is split evenly between partial and complete tears held in place by the surrounding soft tissue79. The volar plate ruptures distally under rapid loading and attenuates proximally under slow loading, the basis of hyperextension testing80. The metacarpophalangeal collaterals are tested in flexion, where the cam of the metacarpal head tightens them.

The thumb carpometacarpal joint

The grind test — axial compression with rotation of the thumb metacarpal — is specific but insensitive for basal joint arthritis: 42–53 per cent sensitivity and 80–93 per cent specificity with only moderate inter-rater agreement (κ 0.48) in one study81, and 0.44 and 0.92 with poor agreement (κ 0.31) in Gelberman's prospective evaluation of 129 patients, in which the thumb metacarpal adduction and extension provocative tests were both 0.94 sensitive and 0.93–0.95 specific with excellent reliability82. A negative grind test does not exclude radiographic arthritis.

The scaphoid and the scapholunate ligament

Watson's presses the examiner's thumb on the palmar tubercle of the scaphoid while the wrist moves from ulnar to radial deviation; a painful dorsal subluxation of the scaphoid over the dorsal rim of the radius, often with a clunk on release, is positive83. Against arthroscopy in 447 patients the test was 0.50 sensitive overall, 0.61 sensitive and 0.62 specific when a scapholunate injury was suspected, and 0.66 sensitive for Geissler grade 3 and 4 lesions84. Its limits are well documented: a painless positive shift in 32 per cent of uninjured wrists, correlated with generalised ligamentous laxity and not with carpal malalignment85; positive on clinical examination in 36 per cent of normal subjects, in whom fluoroscopy showed the confounder to be displacement at the capitolunate joint — a "midcarpal shift" — rather than true scaphoid subluxation86; and, against arthroscopy, a positive likelihood ratio of 2.9 and a negative of 0.28, the most useful of seven provocative wrist tests but still only mildly so3. Pain during the shift, rather than the subluxation itself, is the finding that correlates with scaphoid displacement from the lunate on stress radiographs87, and a dorsal wrist ganglion is frequently associated with a positive test, which became negative in 17 of 19 wrists after excision of the ganglion88.
The fracture signs are sensitive and non-specific. In 215 consecutive patients, anatomical , scaphoid tubercle tenderness and pain on longitudinal thumb compression were each 100 per cent sensitive, with specificities of 9, 30 and 48 per cent; combined within 24 hours of injury they kept full sensitivity and raised specificity to 74 per cent89. Mallee's meta-analysis of 25 tests found snuffbox tenderness the most sensitive (0.87–1.00) with a specificity anywhere from 0.03 to 0.98, and combining tests raised the post-test probability while keeping sensitivity90; when the initial radiographs are normal, about 9 per cent of suspected fractures are occult and absent snuffbox tenderness lowers the odds substantially (negative likelihood ratio 0.2) but does not exclude the fracture91.

The lunotriquetral joint

Reagan's stabilises the lunate and displaces the triquetrum dorsally and palmarly; pain, crepitus or laxity suggests a lunotriquetral sprain, which typically follows hyperextension with twisting and presents with ulnar-sided pain, weakness and a click on lateral movement92. Kleinman's shear test pushes the dorsally against a stabilised lunate93; against arthroscopy the ballottement test was 64 per cent sensitive and 44 per cent specific94.

The ulnar side: fovea, DRUJ and the ECU

The — the examiner's thumb pressed into the soft spot between the ulnar styloid and flexor carpi ulnaris, reproducing the patient's pain — was 95.2 per cent sensitive and 86.5 per cent specific for foveal disruption of the radioulnar ligaments or ulnotriquetral ligament injury in Tay and Berger's series of 272 arthroscopies; the two are separated clinically by the stability of the distal radioulnar joint95. Later series are less flattering: of 73 patients with a positive fovea sign, only 21 had a foveal tear at arthroscopy, so the sign is not specific to foveal disruption96. Nakamura's ulnocarpal stress test (axial load in ulnar deviation with forearm rotation) flags ulnar-sided pathology of any kind — abutment, triangular fibrocartilage or lunotriquetral tears, arthritis — rather than a specific lesion97, and Lester's press test — the patient pushes up from a chair — loads the same structures98. Of the three tests for distal radioulnar instability, only the ballottement test reached statistical accuracy after complete release of the triangular ligament in cadavers; the piano-key and ulnocarpal stress tests did not99. Ruland's — resisted radial abduction of the thumb with the forearm supinated, which co-contracts extensor carpi ulnaris — separates extra-articular ECU tendinopathy from intra-articular ulnar pathology and, in the original algorithm, reduced the need for MRI and diagnostic arthroscopy100; against sonographic ECU abnormality its sensitivity was 74 per cent and specificity 86 per cent101.

Midcarpal instability and the pisotriquetral joint

Lichtman's ulnar midcarpal instability presents as a painful clunk reproduced by ulnar deviation, axial compression and pronation with normal radiographs, cinefluoroscopy showing sudden dissociation of the two carpal rows102. Pisotriquetral arthritis is common in the elderly and under-diagnosed: microscopic changes were present in all ten wrists sectioned from ten elderly donors (severe in five), and gross changes in nine of the other ten, while radiographs were inaccurate103; it can be unmasked after carpal tunnel release, and a local anaesthetic and corticosteroid injection doubles as a diagnostic test104.

De Quervain: Finkelstein, Eichhoff and the WHAT test

The test most clinicians perform as "Finkelstein's" is Eichhoff's — thumb clenched in the fist, wrist ulnarly deviated — and the error has produced false positives for three decades5. Finkelstein's own test grasps the patient's thumb and deviates the wrist ulnarly; in asymptomatic wrists it produced significantly fewer false positives and less discomfort than Eichhoff's manoeuvre105. The — wrist hyperflexion with resisted thumb abduction — was 0.99 sensitive but only 0.29 specific against ultrasound, and its authors also use it to detect dynamic instability of the first-compartment tendons after release106.
Examination notes. Thumb UCL: stress in extension and 30° flexion, > 35° in extension = both parts torn, Stener 87 per cent; contralateral comparison unreliable — use the end point; ultrasound first. PIP > 20° = complete LCL tear. Grind 42–53 / 80–93 per cent; adduction and extension tests 0.94 / 0.93–0.95. Watson shift painless positive in a third of normals, LR+ 2.9; scaphoid signs 100 per cent sensitive alone, 74 per cent specific combined. Fovea sign 95 / 87 per cent in the original; DRUJ ballottement the only accurate test; ECU synergy 74 / 86 per cent. Finkelstein grasps the thumb; Eichhoff clenches it.

§ 05 · VASCULAR EXAMINATION

The was described in 1929 and was re-examined when radial artery harvest for coronary bypass became routine107. Its accuracy against Doppler is modest and cut-off dependent: at the conventional six seconds 54.5 per cent sensitive and 91.7 per cent specific, at five seconds 75.8 and 81.7 per cent, at three seconds 100 per cent sensitive and 27 per cent specific, with no cut-off satisfactory108; against ultrasound with plethysmography, 73 per cent sensitive and 97 per cent specific, so that a negative test predicts safe radial harvest109. The modified test times the return of maximal palmar blush after release of the ulnar artery with the radial occluded — nine seconds or less in 94 per cent of 1,010 patients — and the Barbeau plethysmographic classification excludes only 1.5 per cent110. In six patients whose test suggested a non-patent radial artery, fluorescein angiography showed the radial artery perfusing the thumb and thenar eminence and, with the ulnar artery occluded, the entire hand111, and in a series of 237 radial forearm flaps the surgeons who still performed the test reported that its result never changed the reconstructive plan, while the three busiest surgeons had abandoned it without an ischaemic complication112. Capillary refill is slower than the two-second rule implies: median 0.7–1.0 second in the young, up to 1.8 seconds in elderly men, with a two-second limit falsely positive in 29 per cent of the elderly, and cooling prolongs it113.
Examination notes. Allen: six-second cut-off 55 / 92 per cent against Doppler, 73 / 97 per cent against duplex; modified test ≤ 9 seconds in 94 per cent; false positives by angiography; a negative test clears the radial artery, a positive one calls for duplex. Capillary refill upper limit 2.9 seconds in adult women and 4.5 seconds in the elderly.

§ 06 · PUTTING IT TOGETHER

The examination is a set of imperfect tests whose value lies in combination and in the pre-test probability that the history sets. Where the pre-test probability is high, as it is for most carpal tunnel referrals, confirmatory investigation adds little30; where it is intermediate, the tests with the strongest likelihood or diagnostic odds ratios — hand elevation and Durkan at the wrist, combined pressure and flexion or shoulder internal rotation at the elbow, the adduction and extension tests at the thumb base, snuffbox plus tubercle plus compression for the scaphoid — move it furthest28428289. Patient-reported measures give the examination its outcome context: the 's symptom and function scales are reproducible and responsive and correlate only weakly with two-point discrimination and monofilaments114, and the QuickDASH retains the properties of the full DASH in eleven items115. Documentation should record the test as performed — which manoeuvre, how long, against what — because the same eponym covers different tests, and the numbers above belong to the version the author described.