Diagnosis:Testing the Test

1
DiagnosisTesting the Test

• Verma Walker
• Kathy Davies

2

Journal of Pediatric Gastroenterology
Nutrition. 35(1)39-43, 2002 Jul.
13 C-urea breath test with infrared
spectroscopy for diagnosing helicobacter
pylori infection in children and adolescents.
BACKGROUND AND OBJECTIVE
Studies support the accuracy of 13C-urea
breath test for diagnosing and confirming
cure of Helicobacter pylori infection in
children. Three methods are used to assess
13CO2 increment in expired air mass
spectrometry, infrared spectroscopy, and
laser-assisted ratio analysis. In this
study, the 13C-urea breath test performed
with infrared spectroscopy in children and
adolescents was evaluated
3
METHODS Seventy-five patients (6 months to 18
years old) were included. The gold standard for
diagnosis was a positive culture or positive
histology and a positive rapid urease test.
Tests were performed with 50 mg of 13C-urea
diluted in 100 mL orange juice in subjects
weighing up to 30 kg, or with 75 mg of 13C-urea
diluted in 200 mL commercial orange juice for
subjects weighing more than 30 kg. Breath
samples were collected just before and at 30
minutes after tracer ingestion. The 13C-urea
breath test was considered positive when delta
over baseline (DOB) was greater than 4.0
RESULTS Tests were positive for H. pylori in 31
of 75 patients. Sensitivity was 96.8,
specificity was 93.2, positive predictive value
was 90.9, negative predictive value was 97.6,
and accuracy was 94.7.
CONCLUSIONS 13C-urea breath test performed with
infrared spectroscopy is a reliable, accurate,
and noninvasive diagnostic tool for detecting H.
pylori infection.
4

• Gold Standard
• Investigation Positive n Negative n
• Histology Positive 28 0
• Negative 3 44
• RUT Positive 30 0
• Negative 1 44
• Culture Positive 22 0
• Negative 9 44
• 13C-UBT Positive 30 3
• Negative 1 41

5
(No Transcript)
6
Sensitivity

• the proportion of truly diseased persons, as
  measured by the gold standard, who are
• identified as diseased by the test under study.
• True Positives/(True Positives False Negatives)
• a/(ac)
• Sensitivity Snout Rules Out

7
Specificity

• The proportion of truly non-diseased persons, as
  measured by the gold standard, who are so
  identified by the diagnostic test under study.
• True Negatives/(False Positive True Negative)
• d/(bd)
• Specificity Spin Rules In

8
(No Transcript)
9
Predictive Values

• In screening and diagnostic tests, the
  probability that a person with a positive test is
  a true positive (i.e., does have the disease), or
  that a person with a negative test truly does not
  have the disease. The predictive value of a
  screening test is determined by the sensitivity
• and specificity of the test, and by the
• prevalence of the condition for which
• the test is used.

10

• Positive Predictive Value
• True Positive/(True Positive False Positive)
• a/(ab)
• Probability that a person with positive test is a
  true positive
• (does have the disease)
• Negative Predictive Value
• True Negative/(True Negative False Negative)
• d/(dc)
• Probability that a person with a negative test
• truly does not have the disease

11
Using Predictive Values

• Keep clinical significance in mind
• Terminal or rare disease
• Impact of false negative on patient outcome
• Benefit of testing to patient
• Population tested is high or low risk?
• Alternative Tests for screening

12
Likelihood Ratios

• The likelihood ratio for a test result compares
  the likelihood of that result in patients with
  disease to the likelihood of that result in
  patients without disease
• Positive LR (a/ac)/(b/bd)
• sensitivity / (1-specificity)
• Negative LR (c/ac)/(d/bd)
• (1-sensitivity) / specificity

13
Impact on Disease Likelihood

• LR gt10 or lt0.1 cause large changes
• in likelihood
• LR 5-10 or 0.1-0.2 cause moderate changes
• LR 2-5 or 0.2-0.5 cause small changes
• LR between lt2 and 0.5 cause
• little or no change

14
Ruling In Out

• Does patient have disease ?
• Higher Positive LR means disease is likely to be
  present if test is positive
• Does patient not have disease?
• Lower Negative LR means that
• disease is not likely present or
• cause of patient current condition

15

• Prevalence
• Proportion of persons with a particular disease
  within a given population at a given time.
  Probability that a person selected at random will
  have disease.
• (ac) / (abcd)
• Pre-test odds
• Odds that a person will have the disease
  calculated before test is complete.
• prevalence / (1-prevalence)
• Post-test odds
• Measures impact of test result on odds of
  disease being present
• pre-test odds LR
• Post-test probability
• Chances of disease after factoring in test
  results
• post-test odds / (post test odds1)

16
Nomogram
17
(No Transcript)
18
Clinical Implications

• One test is not a diagnosis
• Implications of false positive
• Further testing may be needed
• Numbers may be significant but not
• clinically relevant

19
Number Meanings

• 100,000 men studied for coronary artery disease
• Uric Acid Factor in prediction
• Developed CA disease uric acid7.8 mg/L
• Did not develop CA disease uric acid 7.7 mg/L
• P Value 0.05 significant
• Problems?

20
Number Meanings

• Large study found significant difference for
  very small difference in values
• Unlikely that uric acid will be useful as
  clinical predictor
• When test is performed, difference
• is less than any lab error

21
Purposes of Statistics

• Estimate relationships between variables, cause
  effect and differences in magnitude
• Measure the significance of the results do the
  numbers have any clinical meaning?
• Adjust for the impact of confounding
• variables on results

22
Bibliography Center for Evidence Based Medicine.
Ed. Douglas Badenoch, Olive Goddard, Bridget
Burchell, Sept. 2002. NHS Research and
Development. 1 Oct. 2002 lthttp//www.minervation.c
om/cebm/docs/likerats.htmlgt Evidence Based
Medicine Tool Kit. Ed. Jeanette Buckingham, Bruce
Fisher, Duncan Saunders. Nov. 2000. University of
Alberta. 5 Sept. 2002 lthttp//www.med.ualberta.ca/
ebm/ebm.htmgt Kawakami, Elisabete. 13C-Urea
Breath Test with infrared spectroscopoy for
diagnosing Helicobacter pylori infection in
children and adolescents. Journal of Pediatric
Gastroenterology and Nutrition 2002 35(1)
39-43. Riegelman, Richard. Studying a Study and
Testing a Test How to read the Medical
evidence. 4th Edition Lippincott, Williams
Wilkins, 2000 Schwartz, Alan. EBM and Decision
Tools Diagnostic Test Cutoffs
lthttp//araw.mede.uic.edu/cgi-bin/cutoff.cgigt