Accuracy of the Speed’s and Yergason’s Tests in Detecting

Biceps Pathology and SLAP Lesions: Comparison With

Arthroscopic Findings
Richard Holtby, M.B., B.S., F.R.C.S.C., and Helen Razmjou, B.Sc.(P.T.), M.Sc., Cred. M.D.T.

Purpose: The purpose of this study was to explore and describe reasons for variation in diagnostic
accuracy of clinical tests using Yergason’s and Speed’s tests in predicting biceps tendon pathology
and SLAP lesions. Shoulder arthroscopy was used as the gold standard. Type of Study: Prospective
blinded study of consecutive patients with a wide spectrum of shoulder conditions. Methods: One
hundred fifty-two subjects (65 women and 87 men) with complaints of shoulder pain were examined.
Fifty subjects (16 women and 34 men) ranging in age from 24 to 79 years (mean age, 50 years; SD ⫽
14.36) met the criteria for surgery. The validity of the Yergason’s and Speed’s tests was evaluated
against findings at surgery. Results: The surgical findings related to biceps pathology and SLAP
lesions were as follows: 2 bicipital tendonitis, both associated with significant rotator cuff pathology;
10 biceps partial tears; and 2 complete ruptures. Fifteen patients had SLAP lesion type I, 12 type II,
and 1 type IV. The sensitivity, specificity, and positive and negative predictive values were 43%,
79%, 60%, and 65% for Yergason’s test and 32%, 75%, 50%, and 58% for Speed’s test, respectively.
The likelihood ratios were 1.28 and 0.91 for Speed’s test and 2.05 and 0.72 for Yergason’s test. These
ratios were provided to assist clinicians in calculating the probability of biceps pathology and SLAP
lesions for a single patient with a different history-specific prevalence of having the pathology.
Conclusions: Although Speed’s and Yergason’s tests are moderately specific, they do not generate
a large change in the post-test probability and are unlikely to make a significant change in the pretest
diagnosis. Clinicians should understand that clinical examination tests do not perform consistently
and have variable predictive values in different patient populations and settings. Level of Evidence:
Level I diagnostic study: testing of previously developed criteria in a series of consecutive patients
(using surgery as gold standard). Key Words: Arthroscopy—Validity—Biceps—SLAP—Speed’s—
Yergason’s.

T he accuracy of some of the clinical examination

tests of the shoulder has been questioned lately in
the literature.1-4 New shoulder tests have been de-
scribed and validated in certain patient populations.5-7
However, some readers have been left with the mis-
From The University of Toronto, Orthopaedic and Arthritic
Institute, Sunnybrook & Women’s College Health Sciences Centre,
Toronto, Ontario, Canada.
Supported by research funds of the Orthopaedic and Arthritic
Foundation, Orthopaedic and Arthritic Institute, Toronto, Ontario.
Address correspondence and reprint requests to Richard Holtby,
M.B., B.S., F.R.C.S.C., Orthopaedic and Arthritic Institute, 43
Wellesley Street East, Toronto, Ontario M4Y 1H1, Canada. E-
mail: richard.holtby@sw.ca
© 2004 by the Arthroscopy Association of North America
0749-8063/04/2003-3573$30.00/0
doi:10.1016/j.arthro.2004.01.008
conception that these tests perform consistently and
have fixed characteristics in different patient popula-
tions and settings. In a study8 that examined the gen-
eral practitioners’ ability to estimate disease probabil-
ities, it was found that most practitioners recognized
the correct definitions for sensitivity and positive pre-
dictive values but did not apply them correctly. Im-
portantly, sensitivity and specificity values are rela-
tively independent of prevalence of the disease but
predictive values are significantly affected by the
spectrum of the patients being examined.9 Patients
selected based on extensive exclusion criteria are ex-
pected to have a much higher chance of having the
target disorder. Therefore, predictive values of clinical
tests in these type of studies cannot be used in settings
that have a much lower prevalence or pretest proba-

Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 20, No 3 (March), 2004: pp 231-236 231
232 R. HOLTBY AND H. RAZMJOU
bility of the disease. The more variation in the popu-
lation being studied, the greater the potential to know
how the test will perform in various settings.
The clinical usefulness of a shoulder test is defined
by its measurement properties such as reliability and
validity and also its ability to move the clinician closer
to the correct diagnosis. Traditionally, validity indexes
have been presented as sensitivity and specificity,
which do not describe the ability of a test to revise the
initial probability of disease. The likelihood ratio ap-
proach would help integrate diagnostic information
into individual clinical decision making and obtain a
more accurate estimate of the risk of disease.
For clinicians to confidently rely on clinical tests,
more information on performance measures of these
tests is needed. Information on validity of the tests that
examine the biceps tendon is fairly limited. Two tests
have been described, Yergason’s test10 and Speed’s
test.11 Originally, both tests were noted to be specific
to bicipital tendonitis, and their significance in detect-
ing SLAP lesions was unappreciated. With recent
identification of SLAP lesions, some surgeons have
examined the value of Speed’s test for detecting this
clinical entity.1,12 However, the usefulness of Yerga-
son’s test in detecting SLAP lesions has not been
examined to date. The purpose of this study was to
explore and describe reasons for variation in diagnos-
tic accuracy of clinical tests. Yergason’s and Speed’s
tests were used as examples of clinical tests in pre-
dicting biceps tendon pathology and SLAP lesions.
Shoulder arthroscopy was used as the gold standard.
METHODS
Subjects
This was a prospective study of consecutive patients
referred by family physicians or other orthopaedic
surgeons for assessment of shoulder problems. Pa-
tients who fulfilled the following criteria underwent
surgery: (1) persistent pain and functional disability
for more than 6 months, not responsive to adequate
conservative treatment, with a positive impingement
test confirmed with local anesthesia or clinical or
investigative signs of rotator cuff tears, labrum, or
biceps lesions; (2) symptoms referred to the acromi-
oclavicular joint lasting more than 6 months with
radiographic changes in the joint; or (3) symptomatic
shoulder instabilities. Patients with a history of previ-
ous shoulder surgery or upper extremity fractures
were excluded. All subjects gave written informed
consent before participation. The study received ethics
approval from the Human Ethics Research Board of
the Sunnybrook & Women’s College Health Sciences
Centre, Toronto, Canada.
Yergason’s and Speed’s tests were performed as a
part of a formal examination of the shoulder. Previous
magnetic resonance images (MRI) or radiographs
taken on the day of assessment were examined by the
surgeon after completion of the data collection form.
The results of the clinical tests were recorded sepa-
rately from the patient’s file, and the orthopaedic
surgeon had no access to the original data collection
forms when completing the surgical data collection
forms.
Surgical techniques for visual examination or repair
of the biceps or labral lesions involved the following
steps: With the patient under general anesthesia in the
lateral position, the arm was suspended from a robot
traction device with 12 lb of weight. The arthroscope
was inserted into the glenohumeral joint through the
posterosuperior portal to inspect the articular surfaces,
anterior and inferior labrum, long head of the biceps,
and articular surface of the rotator cuff. The criteria
for true positive findings related to the biceps tendon
included fibrillation, adhesion in the bicipital groove,
subluxation, or dislocation, partial tear, or complete
rupture of the tendon. The criteria for true-positive
findings related to the labrum were detachment of the
biceps anchor from the superior glenoid (type II
SLAP) or a superior labrum bucket handle tear with
extension into the biceps tendon (type IV SLAP).
Superior labral frying with an intact biceps tendon and
bucket handle labral tears with an intact biceps (type I
and III) were not considered positive surgical findings
for the biceps tests. Partial tears of the long head of the
biceps were debrided. No subject of this cohort had a
biceps tenodesis. SLAP lesions were debrided or re-
paired as appropriate. Treatment of associated rotator
cuff pathology involved resecting the subacromial
bursa, anterior acromioplasty, or resection of the lat-
eral end of the clavicle. Rotator cuff repairs were
performed either open or arthroscopically as indi-
cated.
Surgical findings were recorded for true and false
positives and true and false negatives for biceps pa-
thology and labral lesions. Prevalence, sensitivity,
specificity, accuracy, and positive and negative pre-
dictive values were calculated for Speed’s and Yerga-
son’s tests (Tables 1 and 2). Likelihood ratios were
provided to assist clinicians in calculating the post-test
probability of disease for a single patient with a dif-
ferent pretest probability of the disease. Likelihood
 ACCURACY OF SPEED’S AND YERGASON’S TESTS 233

TABLE 1. Speed’s Test: Clinical and Surgical Correlation

Present Surgically Absent Surgically Total

Positive clinically (a ⫽ 7) (b ⫽ 7) (a ⫹ b ⫽ 14)

Negative clinically (c ⫽ 15) (d ⫽ 21) (c ⫹ d ⫽ 36)
Total (a ⫹ c ⫽ 22) (b ⫹ d ⫽ 28) (a ⫹ b ⫹ c ⫹ d ⫽ 50)

True Positive
Positive clinical examination: Pain in the anterior shoulder or glenohumeral joint. The examiner applies a downward pressure to the
upper extremity with 90° of shoulder elevation, full supination, and with the elbow completely extended.
Positive surgical findings: Bicipital tendonitis, subluxed, dislocated, partial tear, or complete rupture of the biceps tendon, or SLAP
lesions type II or IV.
True Negative
Negative clinical examination: No symptoms.
Negative surgical findings: Normal biceps tendon, no SLAP lesion type II or IV.
Prevalence 100% ⫻ (a ⫹ c)/N ⫽ 22/50 ⫽ 44%
Sensitivity 100% ⫻ (a/a ⫹ c) ⫽ 7/22 ⫽ 32%
Specificity 100% ⫻ (d/b ⫹ d) ⫽ 21/28 ⫽ 75%
Accuracy ⫽ a ⫹ d/N ⫽ 28/50 ⫽ 56%
Positive predictive value 100% ⫻ (a/a ⫹ b) ⫽ 7/14 ⫽ 50%
Negative predictive value 100% ⫻ (d/c ⫹ d) ⫽ 21/36 ⫽ 58%
Positive likelihood ratios ⫽ sensitivity/1-specificity ⫽ 1.28
Negative likelihood ratios ⫽ 1-sensitivity/specificity ⫽ 0.91

ratios were interpreted based on the guidelines pro-
posed by Jaeschke et al.13 Therefore, positive likeli-
hood ratios greater than 10 and negative likelihood
ratios less than 0.1 were considered to generate sig-
nificant changes from the pretest to post-test probabil-
ity. Positive likelihood ratios between 5 and 10 and
negative ratios between 0.2 and 0.1 result in a mod-
erate shift in post-test probability. Positive ratios be-
tween 2 and 5 and negative ratios between 0.5 and 0.2
generate small shifts in probability. Ratios that are

TABLE 2. Yergason’s Test: Clinical and Surgical Correlation

Present Surgically Absent Surgically Total

Positive clinically (a ⫽ 9) (b ⫽ 6) (a ⫹ b ⫽ 15)

Negative clinically (c ⫽ 12) (d ⫽ 22) (c ⫹ d ⫽ 34)
Total (a ⫹ c ⫽ 21) (b ⫹ d ⫽ 28) (a ⫹ b ⫹ c ⫹ d ⫽ 49)

True Positive
Positive clinical examination: Pain in the bicipital groove area or glenohumeral joint. The patient’s elbow is flexed to 90° and stabilized
against the thorax with the forearm pronated and the examiner resists supination.
Positive surgical findings: Bicipital tendonitis, subluxed, dislocated, partial tear, or complete rupture of the biceps tendon, or SLAP
lesions type II or IV.
True Negative
Negative clinical examination: No symptoms.
Negative surgical findings: Normal biceps tendon, no SLAP lesion type II or IV.
Prevalence 100% ⫻ (a ⫹ c)/(N) ⫽ 21/49 ⫽ 43%
Sensitivity 100% ⫻ (a/a ⫹ c) ⫽ 9/21 ⫽ 43%
Specificity 100% ⫻ (d/b ⫹ d) ⫽ 22/28 ⫽ 79%
Accuracy ⫽ a ⫹ d/N ⫽ 31/49 ⫽ 63%
Positive predictive value 100% ⫻ (a/a ⫹ b) ⫽ 9/15 ⫽ 60%
Negative predictive value 100% ⫻ (d/c ⫹ d) ⫽ 22/34 ⫽ 65%
Positive likelihood ratios ⫽ sensitivity/1 ⫺ specificity ⫽ 43/21 ⫽ 2.05
Negative likelihood ratios ⫽ 1⫺sensitivity/specificity ⫽ 57/79 ⫽ 0.72
234 R. HOLTBY AND H. RAZMJOU
close to 1 do not result in important changes in post-
test probability.
RESULTS
One hundred fifty-two subjects (65 women and 87
men) were examined. Fifty subjects (16 women and
34 men) ranging in age between 24 and 79 (mean age,
50, SD ⫽ 14.36) met the criteria for surgery. The
spectrum of pathology varied from rotator cuff or
bicipital tendonitis to massive rotator cuff tears, SLAP
lesions type IV, or biceps rupture. The time between
clinical examination and surgery averaged 23 weeks.
Surgical findings related to biceps pathology and
SLAP lesions were 2 bicipital tendonitis associated
with significant rotator cuff pathology, 10 partial tears,
and 2 complete ruptures. Fifteen patients had SLAP
lesion type I, 12 type II, and 1 type IV. The SLAP
lesion type IV was associated with a partial tear of the
biceps. One of the type II SLAP lesions was associ-
ated with a partial tear of the biceps and one with a
complete rupture. The sensitivity and specificity of
Yergason’s test was 43% and 79%, respectively. The
sensitivity and specificity of Speed’s test was 32% and
75%, respectively. The accuracy values were 56% and
63% for Speed’s and Yergason’s tests, respectively.
The positive likelihood ratio for Speed’s test was 1.28,
and negative likelihood ratio was 0.91. The positive
likelihood ratio for Yergason’s test was 2.05 and neg-
ative likelihood ratio was 0.72 (Tables 1 and 2).
DISCUSSION
This study meets the methodologic standards re-
ported by Sackett et al.9 The original sample (152
subjects) and the surgical patients (50 subjects) were
consecutive patients who met the criteria for each
component of the study. Arthroscopy (diagnostic or
therapeutic), which is an accurate way of examining
shoulder pathology, was performed on patients with a
broad spectrum of disease. The orthopaedic surgeon
assessing was blinded to the results of the specific
clinical tests at the time of surgery, and more than half
of patients with a negative clinical test result under-
went the gold standard examination. This indicates
reduced verification bias that is normally associated
with studies that use invasive gold standards such as
surgery and limit the application of this invasive ex-
amination to those who only have the positive test
results.
In the present study, the sensitivity values of
Speed’s and Yergason’s tests were both low, indicat-
ing a high percentage of false-negative results. Spec-
ificity values are much higher, indicating a low num-
ber of false-positive results. Likelihood ratios
obtained for Speed’s test are close to 1, indicating that
both negative and positive clinical tests are unlikely to
assist diagnosis. A positive Yergason’s test generates
a small change in the post-test probability and a neg-
ative test is unlikely to change the pretest diagnosis.
A review of the shoulder diagnostic test literature
shows minimal information on the validity of Yerga-
son’s test as a diagnostic tool for subacromial biceps
impingement syndrome.2 In that study, conventional
radiography and MRI of the shoulder were used as
gold standards. Calis et al.2 reported a poor sensitivity
of 37.0 and moderately high specificity of 86.1% for
Yergason’s test. The same study reported a sensitivity
of 68.5 and specificity of 55.5 for Speed’s test.2 The
criteria for true radiologic findings were increased
signal intensity, irregularity and thinning, and com-
plete disruption of the tendon.
In a different study, conducted by Bennett,1 the
validity of Speed’s test was examined against arthro-
scopy. The criteria for true surgical findings were
macroscopic inflammation or tearing of the biceps
tendon at any level, evidence of a SLAP lesion, or
complete avulsion of the biceps tendon. A sensitivity
of 90% and specificity of 13.8% were calculated for
the presence of bicipital tendonitis, SLAP lesions, and
biceps avulsions. The result of our study on Speed’s
test is not consistent with the previous study that used
arthroscopy as a gold standard.1 It is not clear if the
treating surgeon in that study was blinded to the
results of the clinical examination at the time of sur-
gery. The Speed’s test was evaluated in a tertiary
clinic, and patients had surgery for various pathologic
conditions, which indicates an appropriate spectrum
of patients. The author believed that the reason the
specificity of Speed’s test was not optimal might have
been the intimate contact between biceps tendon and
portions of the subscapularis and supraspinatus ten-
dons or extension of a Bankart labral injury toward the
anterosuperior biceps-labral complex.1 These pathol-
ogies might have increased the number of false-posi-
tive results and decreased the specificity. In another
study that involved a retrospective review of 102
consecutive patients with type II SLAP lesions, Mor-
gan et al.12 examined the correlation between the
Speed’s test and the anatomic location of type II
SLAP lesions. The sensitivity was reported as 29% for
posterior lesions, 100% for anterior, and 78% for
combined lesions. Specificities of 11%, 70%, and 37%
were reported for posterior, anterior, and combined
 ACCURACY OF SPEED’S AND YERGASON’S TESTS
lesions, respectively. Based on these data, the authors
believed that Speed’s test was not clinically useful for
detecting posterior SLAP lesions. In our study, the
SLAP lesions were not categorized based on location
of the lesion, and therefore we cannot comment on
whether the low validity indexes for Speed’s test are
related to the site of the superior labral lesions.
Many factors affect the transferability of clinical
tests from one setting to another.14 These include
consistency in definition of the pathology, application
of the clinical examination tests, and the spectrum of
the disease in the patients being tested. The conflicting
results between studies that examine the same tests are
not surprising, because clinical tests perform differ-
ently in different settings. Clinicians must understand
the implications of this variability in clinical practice.
If a test is performed under comparable conditions, the
sensitivity and specificity are appropriate indicators of
a test’s performance, because they are not signifi-
cantly affected by prevalence of the disease. Con-
versely, the predictive value of clinical tests may
change between primary care settings (with a low
prevalence of the target disorder) and secondary and
tertiary care settings (with higher probabilities of the
target disease).9,15-18
Variability in prevalence not only affects the pre-
dictive values but also the accuracy of the test. There-
fore, clinicians should not assume that a test that was
reported to have high predictive values in a specific
setting will perform exactly the same in their clinic.
To overcome this problem, likelihood ratios can be
used to provide a more accurate estimate of the post-
test probability or predictive values. The post-test
probability is the probability that a patient has a con-
dition of interest given both the initial clinical infor-
mation and the test results.19 Likelihood ratios essen-
tially combine the benefits of sensitivity and
specificity into one index.9 They indicate by how
much a given diagnostic test result will increase or
decrease the pretest probability of the target disor-
der.9,15 Therefore, clinicians are encouraged to calcu-
late the predictive values of a test for a specific patient
by assigning a history-specific prevalence.18,20,21 This
is the chance of having the target disorder based on the
mechanism of injury and information obtained other
than the test of interest.
To examine the effect of disease prevalence on the
value of clinical tests and to assist clinicians in using
validity concepts in their own setting, we have used
Yergason’s test as an example: A 35-year-old right-
hand– dominant man sustains a significant traction
injury to his right arm. Four months after the injury, he
235
TABLE 3. Formulae for Calculation of Predictive Values
Pretest odds ⫽ prevalence/1 ⫺ prevalence
Post-test odds⫹ ⫽ pretest odds ⫻ likelihood ratios⫹
Post-test odds⫺ ⫽ pretest odds ⫻ likelihood ratios⫺
Positive predictive value ⫽ post-test probability⫹ ⫽ post-test
odds⫹/post-test odds⫹ ⫹ 1
Negative predictive value ⫽ 1 ⫺ post-test probability⫺ ⫽ 1 ⫺
(post-test odds⫺/post-test odds⫺ ⫹ 1)
Accuracy ⫽ (prevalence ⫻ specificity) ⫹ {(1 ⫺ prevalence) ⫻
sensitivity}
continues to have symptoms in the right shoulder. The
symptoms are deep pain in the glenohumeral joint,
feeling of instability in mid range of elevation, diffi-
culty lifting away from the body and some discomfort
with full external rotation at 90° of abduction. Based
on the review of the literature and clinical experience,
the information implies a high chance of damage to
the superior labrum and biceps anchor, probably a
SLAP lesion type II or IV. We assigned a prevalence
of 90% to this particular patient.
By using appropriate formulae (Table 3), the pre-
dictive values for this specific patient are calculated as
follows: Prevalence ⫽ 90% ⫽ 0.9, pretest odds ⫽
0.9/1 ⫺ 0.9 ⫽ 9, positive likelihood ratio (Table 2) ⫽
2.05, post-test odds⫹ ⫽ 9 ⫻ 2.05 ⫽ 18.45. Positive
predictive value ⫽ 18.45/19.45 ⫽ 95%.
This means that if this patient’s Yergason’s test
result is positive, he has a 95% chance of having
biceps pathology, most probably a SLAP lesion type
II or IV. Based on the same formula (Table 3), pretest
odds ⫽ 0.9/1 ⫺ 0.9 ⫽ 9, negative likelihood ratio
(Table 2) ⫽ 0.72, and post-test odds⫺ ⫽ 9 ⫻ 0.72 ⫽
6.48. Therefore, the negative predictive value will be
calculated as: 1 ⫺ (6.48/7.48) ⫽ 13%. This means that
if this patient shows a negative Yergason’s test, there
is a 13% chance that he does not have biceps pathol-
ogy. Accuracy of Yergason’s test is calculated as
Accuracy ⫽ 共0.90 ⫻ 0.79兲
⫹ 兵共1 ⫺ 0.90兲 ⫻ 0.43其 ⫽ 75%
which is considered moderately high for patients with
a high history-specific prevalence of biceps or SLAP
lesions type II or IV.
Acknowledgment: The authors thank Paul Stratford,
Professor, School of Rehabilitation Science at McMaster
University, for his insightful comments, and Terry Leeke,
Data Management Consultant, and Terri Myhr, Research
Manager at The Research Facilitation Office of The Centre
for Research in Women’s Health, for their assistance in data
management and statistical analysis. We also wish to ac-
236 R. HOLTBY AND H. RAZMJOU
knowledge the assistance of Gail Gunnis, Elizabeth Ander-
son, and Gena Hayes for data collection and entry.
REFERENCES
1. Bennett WF. Specificity of the Speed’s test: Arthroscopic
technique for evaluating the biceps tendon at the level of the
bicipital groove. Arthroscopy 1998;8:789-796.
2. Calis M, Akgun K, Birtane M, et al. Diagnostic values of
clinical diagnostic tests in subacromial impingement syn-
drome. Ann Rheum Dis 2000;59:44-47.
3. Lyons AR, Tomlinson JE. Clinical diagnosis of tears of the
rotator cuff. J Bone Joint Surg Br 1992;74:414-415.
4. Walch G, Boulahia A, Calderone S, Robinson AH. The “drop-
ping” and “hornblowers” signs in evaluation of rotator-cuff
tears. J Bone Joint Surg Br 1998;80:624-628.
5. Liu SH, Henry MH, Nuccion SL. A prospective evaluation of
a new physical examination in predicting glenoid labral tears.
Am J Sports Med 1996;24:721-725.
6. Mimori K, Muneta T, Nakagawa T, Shinomiya K. A new pain
provocation test for superior labral tears of the shoulder. Am J
Sports Med 1999;27:137-142.
7. O’Brien SJ, Pagnani MJ, Fealy S, et al. The active compres-
sion test: A new and effective test for diagnosing labral tears
and acromioclavicular joint abnormality. Am J Sports Med
1998;26:610-613.
8. Steurer J, Fischer JE, Bachman LM, et al. Communicating
accuracy of tests to general practitioners: A controlled study.
BMJ 2002;324:824-826.
9. Sackett DL, Straus SE, Richardson WS, et al. Evidence-based
medicine: How to practice and teach EBM, Ed 2. New York:
Churchill Livingstone, 2000;67-93.
10. Yergason RM. Supination sign. J Bone Joint Surg 1931;131:
160.
11. Gilecreest EL, Albi P. Unusual lesions of muscles and tendons
of the shoulder girdle and upper arm. Surg Gynecol Obstet
1939;68:903-917.
12. Morgan CD, Burkhart SS, Palmeri M. Type II SLAP lesion:
Three subtypes and their relationships to superior instability
and rotator cuff tears. Arthroscopy 1998;6:553-565.
13. Jaeschke R, Guyatt GH, Sackett DL. Users’ guides to the
medical literature, III: How to use an article about diagnostic
test. B: What are the results and will they help me in caring for
my patients? JAMA 1994;271:703-707.
14. Irwig L, Bossuyt P, Glasziou P, et al. Evidence base of clinical
diagnosis: Designing studies to ensure that estimates of test
accuracy are transferable. BMJ 2002;324:669-671.
15. Einstein AJ, Bodian CA, Gil J. The relationship among per-
formance measures in the selection of diagnostic tests. Arch
Pathol Lab Med 1997;121:110-117.
16. Sackett DL. A primer on the precision and accuracy of the
clinical examination. JAMA 1992;267:2638-2644.
17. Sackett DL, Haynes. Evidence base of clinical diagnosis: The
architecture of diagnostic research. BMJ 2002;324:529-534.
18. Riddle DL, Stratford PW. Interpreting validity indexes for
diagnostic tests: An illustration using the Berg balance test.
Phys Ther 1999;79:939-943.
19. Radack KL, Rouan G, Hedges J. The likelihood ratio: An
improved measure for reporting and evaluating diagnostic test
results. Arch Pathol Lab Med 1986;110:689-693.
20. Stratford PW, Binkley J. A review of the McMurray test:
Definition, interpretation, and clinical usefulness. J Orthop
Sports Phys Ther 1995;22:116-120.
21. Stratford PW. Applying the results from diagnostic accuracy
studies to enhance clinical decision-making. Phys Theory
Pract 2001;17:153-160.