Sports · Part Two · Assessment and Movement
Lesson 12 / 64
Upper-Extremity Screening
A clean upper-extremity screen reads the shoulder, elbow, and wrist as one kinetic chain, then asks how well the nervous system controls it.
Upper-extremity screening is the sequenced examination of the neck, shoulder, elbow, wrist and hand that separates a structural lesion from a control problem in an overhead athlete. No single shoulder test is accurate enough to name a lesion, and the review that pooled them says so. Clusters do better, and the side-to-side total rotational arc predicts injury where internal rotation loss alone does not. The Unified Model of Tone reads the screen as one regulatory state read across the whole limb.
Shoulder joints
Five articulations screened
Neer test
72 percent sensitive
Apprehension test
96 percent specific
Total rotation deficit
2.6 times elbow risk
Upper-extremity screening.
A sequenced examination of the cervical spine, shoulder, elbow, wrist and hand. Neurovascular status and the neck are cleared first, then inspection, active and passive range of motion, manual muscle testing, and targeted orthopedic tests last.
The test cluster.
Most single shoulder tests are strong on one axis and weak on the other, so a lone result rarely settles anything. Counting how many of several tests fire turns a set of weak signals into one usable probability.
01What the accuracy studies show
The Numbers Behind the Upper-Extremity Screen
Eight findings that set the limits of what a shoulder, elbow or cervical test can establish.
02Clearing the neck first
The Upper-Extremity Screen Clears the Neck and the Nerves Before It Tests the Shoulder
The examiner clears neurovascular status and the cervical spine first, since C5 and C6 refer pain straight into the shoulder. Upper-extremity screening works because it follows that plan and never breaks it. Then come visual inspection, active range of motion, passive range of motion, manual muscle testing, and only then the targeted orthopedic tests.
The shoulder alone is five articulations. Three are true joints: the glenohumeral, the acromioclavicular and the sternoclavicular. Two are functional: the scapulothoracic and the suprahumeral. An examination that reads the glenohumeral joint and calls that the shoulder has covered one articulation of five.
Order matters because the upper extremity is deep, muscular, and hard to palpate, and pain patterns lie. A cuff problem radiates into the deltoid. An AC joint problem mimics instability. The cervical spine masquerades as both.
A disciplined sequence turns an inexact region into a readable one, and it follows how the nervous system governs the limb: proximal stability before distal mobility, scapula before hand.
What clearing the neck actually buys
The neck-first rule has a measured payoff. In 82 patients with suspected cervical radiculopathy or carpal tunnel syndrome, examined against electrophysiologic testing, 13 clinical items reached a likelihood ratio above 2 or below 0.50 Wainner 2003. A cluster of four of them produced a positive likelihood ratio of 30.3, well above any single item. The authors record that the confidence intervals around every estimate in that study were wide.
Upper limb tension test A was the most useful single item for ruling cervical radiculopathy out. Double Crush and the Arm handles that test, the double crush mechanism behind it, and the carpal tunnel and thoracic outlet syndromes it helps sort.
Inspection carries its own neurology. Wasting of both supraspinatus and infraspinatus points to suprascapular nerve entrapment at the suprascapular notch. Isolated infraspinatus wasting points further along the nerve, to the spinoglenoid notch. Atrophy names a nerve before any provocative test is performed.
03Impingement test accuracy
No Single Impingement Test at the Shoulder Is Accurate Enough to Stand Alone
The impingement tests every clinician learns are weaker than their reputation. Pooled across the shoulder examination literature, the Neer test reached 72 percent sensitivity and 60 percent specificity, and the Hawkins-Kennedy test 79 and 59 Hegedus 2012. The painful arc runs the other way, at 53 percent sensitivity and 76 percent specificity.
The earlier pass at the same literature put Neer at 79 and 53, and the Speed test for superior labral pathology at 32 and 61 Hegedus 2008.
A test on those numbers shifts a probability rather than establishing a diagnosis, and a careful clinician reasons from it that way. Sensitivity governs what a negative result is worth, so a sign that misses more than a quarter of true cases clears nobody on its own. Specificity governs what a positive result is worth, and a sign that flags four intact shoulders in ten has not named a lesion.
What a cluster buys that a single test cannot
Counting positives changes the arithmetic. Eight tests were run against arthroscopic findings in a series graded by severity of rotator cuff disease Park 2005. The best combination for any degree of impingement was the Hawkins-Kennedy sign, the painful arc and the infraspinatus muscle test, at a 95 percent post-test probability. For full-thickness tears the best trio was the painful arc, the drop-arm sign and the infraspinatus muscle test, at 91 percent.
A prospective series of 55 patients with surgical diagnosis as the reference standard found the same shape Michener 2009. The painful arc carried a positive likelihood ratio of 2.25, with a 95 percent confidence interval of 1.33 to 3.81. The empty-can test reached 3.90, interval 1.5 to 10.12, and external rotation resistance 4.39, interval 1.74 to 11.07. Three positives out of five tests was the cut point that discriminated best, with an area under the curve of 0.79.
How much the cluster gains is contested. The pooled review that produced the single-test figures above concludes that combinations provide better accuracy, but only marginally Hegedus 2012. Either way the clinician runs several tests, because one is not enough.
Reliability follows the same order. Agreement between examiners was moderate to substantial for the painful arc, the empty-can and external rotation resistance, at kappa 0.45 to 0.67. It was fair for Neer and Hawkins-Kennedy, at 0.39 to 0.40. The tests carrying the most information are the ones two examiners agree on.
04Instability and the labrum
The Instability Tests Rule a Shoulder In and the Labral Tests Barely Move the Estimate
Anterior instability is the one shoulder diagnosis a positive test comes close to settling. Across 363 shoulders examined and then taken to arthroscopy, 46 with traumatic anterior instability, apprehension reached 72 percent sensitivity and 96 percent specificity, likelihood ratio 20.2 Farber 2006. Relocation reached 81 and 92, ratio 10.4.
The criterion decides the number. When pain rather than apprehension was accepted as a positive result, sensitivity, specificity and likelihood ratio all fell. The anterior drawer, scored positive only when it reproduced the athlete's instability symptoms, reached 53 and 85, ratio 3.6. The authors conclude that all three tests are specific but not sensitive.
The rest of the instability examination quantifies rather than diagnoses. The load and shift grades translation of the humeral head on the glenoid, and the sulcus sign grades inferior laxity by the gap below the acromion. Laxity alone is an ordinary finding in an overhead athlete, so symptoms decide whether it counts.
The labral tests do less than any of that. In the pooled analysis, the best sensitivity for a superior labral tear belonged to the relocation test at 52 percent Hegedus 2012. The best specificity belonged to Yergason's test at 95 percent. The best positive likelihood ratio in the entire group was 2.81, from the compression-rotation test.
A ratio of 2.81 moves a pretest probability of 30 percent to roughly 55 percent. That is a nudge, and O'Brien's active compression and Speed's test are read the same way. Each test is a hypothesis the examiner is trying to confirm or kill.
The acromioclavicular joint, where one test is not enough either
The acromioclavicular joint obeys the same rule. Of 1,037 patients with shoulder pain, 38 completed a protocol of 20 clinical tests, imaging and a diagnostic injection, and 28 had confirmed acromioclavicular pain Walton 2004. Tenderness over the joint was the most sensitive finding at 96 percent and the Paxinos test reached 79 percent, both with low specificity.
Pairing a sensitive test with a specific one closed that gap. A positive Paxinos test together with a positive bone scan gave post-test odds of 55 to 1 and a 99 percent post-test probability. Both negative gave a likelihood ratio of 0.03 to 1, which rules the joint out. One positive and one negative left an indeterminate 0.4 to 1.
That indeterminate cell is the third answer the pairing can give. Earlier pooling found that no test for impingement or acromioclavicular pathology reached significant diagnostic accuracy alone Hegedus 2008. Two tests that disagree have told the examiner the answer is not here yet.
05The total rotational arc
The Total Rotational Arc Predicts Injury Where Internal Rotation Loss Alone Does Not
The single most important number in an overhead athlete's shoulder is the total rotational arc, not any one motion alone. External rotation plus internal rotation, measured supine with the arm abducted 90 degrees and the scapula stabilized at the coracoid, gives one figure per side. The comparison that carries information is the difference between the two sides.
The throwing arm trades range. It gains external rotation and gives back internal rotation, an adaptive shift built on humeral retroversion and a tightening posterior capsule. Wilk's total motion concept treats external rotation plus internal rotation as a single figure. The version in wide circulation adds that the sum should be roughly one hundred eighty degrees and symmetric side to side. The prospective data support the symmetry half of that and not the fixed target.
Over three competitive seasons, 170 pitcher-seasons of preseason range of motion were measured across 122 professional pitchers by one examiner with a bubble goniometer, reproducibility 0.81 Wilk 2011. Pitchers with glenohumeral internal rotation deficit, defined as a loss of 20 degrees or more against the other shoulder, were nearly twice as likely to be injured. That association did not reach statistical significance, at P equals .17. Pitchers whose total rotational motion deficit exceeded 5 degrees carried the higher injury rate.
What the deficit does to the joint
The deficit crowds the posterior capsule, drives the labrum and cuff into the arch, and sets up the cascade toward labral tearing and elbow overload. The Shoulder examines rotator cuff and labral epidemiology in overhead athletes, GIRD as a diagnosis, and scapular dyskinesis.
Screening the total arc and the side-to-side difference is how a trainer catches the throwing shoulder going wrong while the athlete still feels fine. The arc difference reached significance in both prospective cohorts of professional pitchers. The internal rotation deficit reached it in neither.
06The medial elbow
The Elbow Screen Reads a Load the Shoulder Delivered
The thrower's elbow fails at the medial side, where the ulnar collateral ligament absorbs valgus load during the late cocking and acceleration phases. The screen probes that ligament with valgus stress at 30 degrees of flexion and with the moving valgus stress test, watching for medial gapping and reproduction of the athlete's pain.
The moving valgus stress test has the best published accuracy in the region and the smallest sample behind it. Across 21 patients taken to surgery for medial elbow pain, the test was 100 percent sensitive at 17 of 17 and 75 percent specific at 3 of 4 O'Driscoll 2005. Pain appeared across a shear range of 120 to 70 degrees of flexion and peaked at 90 degrees. Four control patients cannot fix a specificity estimate, so the sensitivity figure carries the weight.
Valgus extension overload, where the posteromedial olecranon impinges, and little league elbow round out the medial picture. Medial epicondyle tenderness with pain on resisted wrist flexion and forearm pronation points to flexor-pronator tendinopathy. The Elbow carries ulnar collateral ligament injury rates and the valgus load that produces them, and The Wrist, Hand and Fingers takes the distal segment.
The ulnar nerve is cleared every time
The ulnar nerve travels with this territory and gets cleared at every screen. Tested in 32 subjects with electrodiagnostically proven cubital tunnel syndrome against 33 controls, Tinel's sign reached 70 percent sensitivity Novak 1994. The elbow flexion test held for 30 seconds reached 32 percent, and pressure over the nerve reached 55 percent. Combining pressure with elbow flexion reached 91 percent.
Three tests are weak alone and one combination is strong, the same arithmetic the shoulder produced, reached through a different nerve. Controls were rarely positive on any of the four, which is what makes the 91 percent combination usable.
Because the medial elbow and the throwing shoulder share one chain, a shoulder that has lost rotational arc loads the elbow harder, which is why the upper-extremity screen never examines the elbow in isolation. That link has been measured rather than inferred from anatomy.
07Findings that travel
A Finding in One Segment Forecasts Failure in Another
A measurement taken at the shoulder predicted injury at the elbow across 505 examinations of 296 professional pitchers Wilk 2014. Over eight consecutive years, 38 of those players sustained 49 elbow injuries and 8 surgeries, costing 2,551 days. Neither internal rotation deficit nor external rotation insufficiency correlated with any of it.
What did correlate sat one joint above the injury. A total rotation deficit greater than 5 degrees in the throwing shoulder carried 2.6 times the risk of elbow injury. A shoulder flexion deficit of 5 degrees or more carried 2.8 times. The measurement was taken at the shoulder and the failure happened at the elbow.
The pattern continues below the arm. In 347 professional pitchers screened during the last two weeks of spring training, lumbopelvic control was quantified as peak anterior-posterior deviation of the pelvis during a single-leg raise Chaudhari 2014. Pitchers in the worst third for that measure were 3.0 times more likely to miss 30 days or more than pitchers in the best third.
Among the men who missed any time at all, the worst third missed a mean of 98.6 days against 43.8 days in the best third. Nothing in that test touched an arm. The Kinetic Chain handles energy transfer and proximal-to-distal sequencing. The screen adds the consequence: a pelvis measured in March forecasts time lost months later.
The control read the pathology tests miss
The best upper-extremity screen reads the quality of motor control feeding the limb, which no pathology test measures. Range of motion, muscle balance and joint position sense are the raw inputs the nervous system uses to coordinate a throw, a serve or a press. A scapula that fires late or a cuff that switches on a fraction slow leaves the glenohumeral joint to find stability from passive tissue, and passive tissue is what tears. The screen exposes that control gap before the box score does.
Joint position sense error in degrees, and the receptors that generate it, are measured on Proprioception and Joint Position Sense. Cortical Drive and Force carries cortical drive to the limb. The screening question is narrower and answerable in one visit. Does this athlete organize the chain the same way on both sides, and the same way they did in March?
08What we corrected
Five Claims Removed From This Page
This page previously treated the sum of external and internal rotation as a rule, at roughly one hundred eighty degrees. The prospective work is built on side-to-side comparison rather than on a fixed target, and the deficit that carried risk was greater than 5 degrees against the athlete's own other shoulder Wilk 2011. The number that predicts is the difference.
Second, the page listed internal rotation loss as the threshold that flags risk. That association did not reach significance across 170 pitcher-seasons, and it was not correlated with elbow injury across 505 examinations Wilk 2014. Third, the page put a thrower's external rotation beyond one hundred seventy degrees. No source could be traced for that figure and it is gone.
Fourth, the page called apprehension followed by relocation the most sensitive pairing for anterior microinstability. Both tests are specific and not sensitive, at 96 and 92 percent specificity against 72 and 81 percent sensitivity Farber 2006. Fifth, the painful arc was given endpoints of sixty and one hundred thirty degrees, which no source verified.
A pull-quote attributed to Dr. Jason Dulberg had no source in anything he said or wrote, so it is gone. Every figure above is tied to the study that produced it, and every reading of those figures is named as the model's.
09The model's claim
One Regulatory State, Read Across the Whole Limb
Pooled test accuracy, the cluster studies, the two pitching cohorts and the provocative testing at the elbow are established science, cited above. The reading laid over them belongs to the Unified Model of Tone and is stated here as ours.
The model expects remote effects that a purely local view finds puzzling: a plantar fascia that influences headaches, a big toe injury that alters contralateral shoulder gait, cranial work that reaches the sacrum. Applied to the upper limb, that predicts what a region-by-region screen is not built to look for. A finding at one segment should track a finding at another, on the opposite side or higher in the axis, more often than independence allows.
Two published results already have that shape. A total rotation deficit at the shoulder predicted elbow injury at 2.6 times the risk while the internal rotation measure predicted nothing Wilk 2014. Lumbopelvic control predicted days missed across 347 professional pitchers Chaudhari 2014. Neither study set out to test a unification claim, and each is a single instance of a pattern the model expects to be general.
The screen that would settle it
A preseason upper-extremity screen recorded on both arms and the axis, rather than on the complaining side, would settle the question in one squad across one season. It records four numbers. The side-to-side difference in total rotational arc in degrees. Shoulder joint position sense error in degrees on the non-throwing arm. Lumbopelvic anterior-posterior deviation in millimeters during a single-leg raise. The count of the five subacromial tests that fire on the asymptomatic shoulder.
This is a claim about how the upper limb is organized rather than a claim about what treatment does. If total rotational arc difference, non-throwing shoulder joint position sense error, lumbopelvic anterior-posterior deviation and asymptomatic-side cluster count are shown to move together within the same athletes across a season, the unification claim is confirmed.
10The tone reading
The Upper Limb as One Organized Chain
Three signatures of tone show up in measurements an upper-extremity screen already records.
Coupling
A total rotation deficit greater than 5 degrees carried 2.6 times the elbow injury risk. The reading crossed a joint the examiner never stressed.
Constraint
A total arc that collapses below the other side leaves the throwing shoulder fewer ways to organize a pitch, and the elbow absorbs the difference.
Input quality
The apprehension test reached 96 percent specificity only when the examiner scored apprehension rather than pain as the positive result. The question decides the reading.
The other foundations of tone surface at different points in the same examination. Gain shows in a cuff that still contracts at full force while the scapula that should have preceded it fires late. Set-point is the arc measured on that athlete in March, which beats any population target including the one hundred eighty degree rule. Prediction is the feedforward model that stiffens the shoulder before ball release, and it fails before any tissue does. Load is what the arm absorbs when the arc collapses, and forecasting that absorption is the point of the screen. Time course is why a deficit measured at spring training predicted days missed across a whole season rather than in the week that followed. Oscillation is the rhythm underneath a throw, where the useful reading is the structure of the motion rather than its peak. Tone held inside its range keeps the arm able to adapt to the next pitch, and an arm that has drifted outside that range is the one that tears. The foundations named here are defined in the Unified Model of Tone.
11Where this sits
How This Page Relates to the Rest of the Library
Seven neighboring pages carry the parts of this argument the screen only touches.
The same accuracy arithmetic below the waist, where the Ottawa rules do what no shoulder test can.
Where the posterior capsule tightening behind an arc deficit gets its own examination.
What the valgus load actually does to a ligament the moving valgus stress test can only provoke.
Scaphoid, TFCC and finger injuries, which a limb screen can clear but not characterize.
Why one nerve compressed twice explains a limb that tests abnormal in two places at once.
The sequencing that explains how a pelvis measured in March turns into weeks lost later in the season.
What a program does once the screen has named the failing link, and the criteria for progressing it.
12Questions athletes ask
Questions Athletes Ask
What does an upper-extremity screening exam check in an athlete?
It maps the cervical spine, shoulder, elbow, wrist and hand for laxity, impingement and the motion deficits that precede injury. The order runs proximal to distal and neural to structural, so referred pain is excluded before a cuff test is scored. The shoulder alone is five articulations, and the sequence exists to stop an examiner reading one joint and calling it the region. A neurological layer runs through all of it, because muscle wasting points to a specific nerve and a slow scapula points to a control problem.
How accurate are the shoulder special tests on their own?
They are weaker than their reputation. Pooled across the literature the Neer test reaches 72 percent sensitivity and 60 percent specificity, the Hawkins-Kennedy test 79 and 59, and the painful arc 53 and 76. For superior labral tears the best positive likelihood ratio in the whole group is 2.81, which moves a pretest probability of 30 percent to roughly 55 percent. A test on those numbers shifts a probability and settles nothing. The review that pooled them recommends no single test as a diagnosis.
What happens during an overhead athlete's shoulder screen, and is it safe?
It follows a fixed order. Neurovascular status and the cervical spine come first, then inspection, active and passive range of motion, and manual muscle testing. Targeted tests come last: Neer, Hawkins-Kennedy, the empty-can, apprehension with relocation, the compression-rotation test and the total arc measurement. It is non-invasive. Provocative tests are graded and stopped at the first reproduction of symptoms. Beyond ruling out pathology the screen reads motor control, scapular timing and the durability of the chain under load. That control read is what a pathology test never reaches.
What is GIRD, and does losing internal rotation really predict injury?
GIRD is glenohumeral internal rotation deficit, a loss of 20 degrees or more in the throwing shoulder against the other side. The prospective evidence is thinner than the term suggests. Across 170 pitcher-seasons, pitchers with GIRD were nearly twice as likely to be injured, and the association did not reach significance at P equals .17. Across 505 examinations it was not correlated with elbow injury at all. The total rotational arc difference is the measure that carried risk in both cohorts.
Why does the screen clear the cervical spine and nerves first?
Because C5 and C6 refer pain straight into the shoulder, and a cuff that fires slow is often a signaling problem rather than a tissue problem. Clearing the neck separates a true glenohumeral lesion from referred pain. It also has measured value: in 82 patients examined against electrophysiologic testing, a four-item cervical cluster produced a positive likelihood ratio of 30.3 where no single item came close. Wasting of supraspinatus and infraspinatus localizes a nerve at the same time, so the neurology comes first.
Can a shoulder or trunk measurement predict an elbow injury?
The shoulder has, and the trunk predicts time lost. In 505 examinations of 296 professional pitchers, a total rotation deficit greater than 5 degrees in the throwing shoulder carried 2.6 times the risk of elbow injury. A shoulder flexion deficit of 5 degrees or more carried 2.8 times. In 347 pitchers, the worst third for lumbopelvic control were 3.0 times more likely to miss 30 days or more. Neither measurement was taken at the joint that eventually failed, which is why a single-joint screen misses them.
What does the Unified Model of Tone add to an upper-extremity screen?
It predicts where to look. The model holds that a stress at one corner of a structure is answered by the whole structure. A finding at one segment should therefore track a finding on the opposite side or higher in the axis more often than chance allows. That turns a region-by-region examination into a bilateral one, and it makes the asymptomatic arm worth measuring. It also names a specific study worth running: total arc difference, non-throwing shoulder position sense, lumbopelvic deviation and asymptomatic-side cluster count, recorded together across a season.
13The sources
References
12 primary sources, each linked to its record. Figures quoted on this page were checked against the published abstract.
Related evidence