DIAGNOSIS WORKSHEET
Citation:
The Diagnostic Accuracy of Kernig's Sign, Brudzinski's Sign, and Nuchal
Rigidity in Adults with Suspected Meningitis
Karen E. Thomas, Rodrigo Hasbun, James Jekel, and Vincent J. Quagliarello
Clinical Infectious Diseases. 2002:35 (1 July) pp 46-52.
Are the results of this diagnostic study valid?
Yes – CSF analysis was performed on all
patients involved in the study.
Was there an independent, blind
comparison with a reference (“gold”)
standard of diagnosis?
Was the diagnostic test evaluated in an
appropriate spectrum of patients (like
those in whom it would be used in
practice)?
Was the reference standard applied
regardless of the diagnostic test result?
Yes – Diagnostic tests were evaluated
only in patients with clinical suspicion of
meningitis based on their presenting
complaints.
Yes – All patients received lumbar
puncture regardless of the results of
diagnostic testing.
Was the test (or cluster of tests) validated No – The diagnostic tests were not
in a second, independent group of
performed on a second, independent
patients?
group of patients.
Are the valid results of this diagnostic study important?
CALCULATIONS: Kernig’s Sign
Target disorder
(Meningitis)
Present
Absent
Diagnostic
test result
(Kernig’s
Sign)
Totals
Positive
3
a
8
b
11
a+b
Negative
63
c
163
d
226
c+d
66
a+c
171
b+d
237
a+b+c+d
Totals
Sensitivity = a/(a+c) = 3/66 = 4.55%
Specificity = d/(b+d) = 163/171 = 95.3%
Likelihood ratio for a positive test result = LR+ = sens/(1-spec) = 4.55%/4.68% = 0.97
Likelihood ratio for a negative test result = LR - = (1-sens)/spec = 95.5%/95.3% = 1.0
Positive Predictive Value = a/(a+b) = 3/11 = 27.2%
Negative Predictive Value = d/(c+d) = 163/226 = 72.1%
Pre-test probability (prevalence) = (a+c)/(a+b+c+d) = 66/237 = 27.8%
Pre-test odds = prevalence/(1-prevalence) = 27.8%/72.2% = 0.385
Post-test odds = pre-test odds  LR = 0.385 x 0.97 = 0.373
Post-test probability = post-test odds/(post-test odds +1) = 0.373/1.373 = 27.2%
CALCULATIONS: Brudzinski’s Sign
Target disorder
(Meningitis)
Present
Absent
Diagnostic
test result
(Brudzinski’s
Sign)
Totals
Positive
3
a
8
b
11
a+b
Negative
63
c
162
d
225
c+d
66
a+c
170
b+d
236
a+b+c+d
Totals
Sensitivity = a/(a+c) = 3/66 = 4.55%
Specificity = d/(b+d) = 162/170 = 95.3%
Likelihood ratio for a positive test result = LR+ = sens/(1-spec) = 4.55%/4.68% = 0.97
Likelihood ratio for a negative test result = LR - = (1-sens)/spec = 95.5%/95.3% = 1.0
Positive Predictive Value = a/(a+b) = 3/11 = 27.2%
Negative Predictive Value = d/(c+d) = 162/225 = 72%
Pre-test probability (prevalence) = (a+c)/(a+b+c+d) = 66/236 = 28%
Pre-test odds = prevalence/(1-prevalence) = 28%/72% = 0.389
Post-test odds = pre-test odds  LR = 0.389 x 0.97 = 0.377
Post-test probability = post-test odds/(post-test odds +1) = 0.377/1.377 = 27.4%
CALCULATIONS: Nuchal Rigidity
Target disorder
(Meningitis)
Present
Absent
Diagnostic
test result
(Nuchal
Rigidity)
Totals
Positive
24
a
69
b
93
a+b
Negative
56
c
148
d
204
c+d
80
a+c
217
b+d
297
a+b+c+d
Totals
Sensitivity = a/(a+c) = 24/80 = 30%
Specificity = d/(b+d) = 148/217 = 68.2%
Likelihood ratio for a positive test result = LR+ = sens/(1-spec) = 30%/31.8% = 0.94
Likelihood ratio for a negative test result = LR - = (1-sens)/spec = 70%/68.2% = 1.03
Positive Predictive Value = a/(a+b) = 24/93 = 25.8%
Negative Predictive Value = d/(c+d) = 148/204 = 72.5%
Pre-test probability (prevalence) = (a+c)/(a+b+c+d) = 80/297 = 26.9%
Pre-test odds = prevalence/(1-prevalence) = 26.9%/73.1% = 0.368
Post-test odds = pre-test odds  LR = 0.368 x 0.94 = 0.346
Post-test probability = post-test odds/(post-test odds +1) = 0.346/1.346 = 25.7%
Can you apply this valid, important evidence about a diagnostic test in caring for
your patient?
Is the diagnostic test available, affordable,
- These diagnostic tests are easily
accurate, and precise in your setting?
available and affordable. Based on
the results of this study and the above
calculations, it appears that the tests
are neither precise nor accurate.
Can you generate a clinically sensible
- The study patients were patients
estimate of your patient’s pre-test probability presenting to a busy urban emergency
(from personal experience, prevalence
department in New Haven, CT similar
statistics, practice databases, or primary
to our own.
studies)?
- It is unlikely that the disease
possibilities or probabilities have
 Are the study patients similar to your
changed since the evidence was
own?
 Is it unlikely that the disease possibilities gathered.
- Therefore we should be able to
or probabilities have changed since the
reasonably generate a clinically
evidence was gathered?
sensible estimate of our patient’s pretest probability.
Will the resulting post-test probabilities
- The post-test probabilities for all
affect your management and help your
three diagnostic tests are nearly
patient?
equivalent to their respective pre-test
probabilities.
 Could it move you across a test- Therefore, the results of these three
treatment threshold?
diagnostic tests are unlikely to change
 Would your patient be a willing partner
our management of patients with
in carrying it out?
suspected meningitis
Would the consequences of the test help
- It appears from the above data that
your patient?
the results of these diagnostic tests
will be of little help to our patients.
Additional notes:
Each individual physician performed his or her own version of Brudsinski’s and Kernig’s
Signs. A standardized method was not taught to participating physicians. This was done
intentionally to coincide with practices in a typical emergency department. It is unclear
whether a more rigid standard for these tests would have yielded different results.