Quality Assurance
Background
Quality assurance (QA)
◦ Refers to a program for the systematic monitoring and evaluation of the various aspects
of vascular testing to ensure that standards of quality are being met
◦ Statistics are used to prove the diagnostic accuracy of vascular ultrasound testing.
Statistics
The science of making effective use of numerical data relating to groups of
individuals or experiments
Deals with collection, analysis, and interpretation of data
Used in planning of the collection of data
Comparing Diagnostic Tests
The “Gold Standard”
◦ Used to compare one form of a newer diagnostic test with another well-established and
reliable test
◦ In vascular imaging, gold standard is typically angiography.
◦ Ideally, vascular ultrasound would match gold standard 100% of the time.
Comparing Diagnostic Tests—(cont.)
True positives
◦ Number of studies performed by ultrasound, which state that disease is present and
agree with the gold standard
True negatives
◦ Number of studies performed by ultrasound, which that disease is NOT present and
agree with the gold standard
Comparing Diagnostic Tests—(cont.)
False positives
◦ Studies that are reported as positive by ultrasound but are found to be negative by the
gold standard
False negatives
◦ Ultrasound states a study is negative; however, gold standard identifies disease
Overall goal is to have high degree of true positives and true negatives, with no or
very few false negatives and false positives.
Sensitivity
Measures the proportion of actual positive studies that are correctly identified
The ability to correctly identify disease when disease is truly present
Calculated by taking the true positives and dividing these by the all-positive results
as determined by gold standard
Specificity
Important to not only detect presence of disease but confirm absence of disease
The ability of a test to identify something as normal
Of the gold standard negatives how many were correctly called negative by
noninvasive studies
Calculated by dividing the number of true negatives by all negative results as
identified by the gold standard
Reliability or Accuracy
The consistency of obtaining similar results under similar circumstances
Accuracy over a period of time
◦ A laboratory is reliable when the results of the tests produced are consistently accurate.
◦ How well does the non-ivasive test both detect and rule out disease
◦ Number of correct non-invasive studies divided by number of all non-invasive studies
Positive Predictive Value
Proportion of patients with positive tests that are correctly identified
Provides the probability that a positive test truly reflects the underlying disease
Calculated as the number of true positives divided by all the positive studies as
identified by ultrasound (true positives and false positives)
Negative Predictive Value
The proportion of negative test results when there is no underlying disease
Probability that a negative test reflects the absence of disease
Calculated by dividing the true negatives by all negative test results (true negatives
and false negatives)
The Chi-Square Test
A statistical test that compares the difference between what you expect and what
you observe
◦ The narrower the difference between what is expected and what is observed, the
greater the accuracy of the test.
True positives, true negatives, false positives, and false negatives can be entered
into a Chi-square to determine accuracy, sensitivity, specificity, positive predictive
value, and negative predictive value.
The Chi-Square Test—(cont.)
Basic design of a Chi-square table comparing observed results (angiography) with expected results
(duplex ultrasound). The cells of the Chi-square are identified, indicating the appropriate placement of
true positive, false positive, true negative, and false negative data.
The Chi-Square Test—(cont.)
To use Chi-square:
◦ Boxes A & B are used to calculate
positive predictive value.
◦ Boxes C & D are used to calculate
negative predictive value.
Figure 34-2. Using the Chi-square table to calculate
positive predictive value (PPV) and negative
predictive value (NPV).
The Chi-Square Test—(cont.)
To use Chi-square:
◦ Boxes A & C are used to calculate
sensitivity.
◦ Boxes B & D are used to calculate
specificity.
Using the Chi-square table to calculate sensitivity
and specificity.
The Chi-Square Test—(cont.)
To use Chi-square:
◦ For accuracy, use boxes A & D
divided by all boxes.
◦ Accuracy is determined both by
ability to find disease and the
absence of disease.
Using the Chi-square table to calculate the overall
accuracy.
The Chi-Square Test—(cont.)
Absolute value for accuracy must be between values of sensitivity and specificity.
◦ Accuracy value also must be between the negative and positive predictive values.
The Chi-Square Test—(cont.)
