Flashcards in Day 1 Deck (45):
When considering a lab test, what are some things you should take into consideration before you order it?
1. Is the test just information or diagnostic
2. If 2 tests give you the same information, when should one be used before another?
3. What are disadvantages and possibilities of errors or false results
What do unnessisary tests do?
-Hurt patient financially
-Can be inconvenient
-Wasted or misleading information
-Loss of time
What are multichannel chemical analyzers?
Take a large number of tests and have the results back in a short time
-Came back that this wasn't very helpful for diagnosing disease in asymptomatic patients
Tests should be ordered to _______
-Confirm a diagnosis based on the entire clinical picture (history and clinical presentation)
-Or if the result will alter therapy decisions
This is the ability of a test to detect patients with a specific disease
T/F- False negatives are found in sensitive test
What does a test that has 90% sensitivity mean?
That in 10% of patients with the disease, the test will not detect it
This describes how well the test abnormality is restricted to those patients that have the disease in question (that it picks up on the sign due to that specific disease)
Specific tests sometimes show....
Describe false positives...
A specificity of 90% for disease X indicates that 10% of the test results suggestive of disease X aren't due to disease X
This is the incidence of disease (number of people with the disease) in the population being tested
Describe predictive value
The smaller the number of people with a disease, the lower the proportion of people with an abnormal test, and the greater the proportion of false positives
If you wouldn't believe a test, it has a high/low predictive value
How is predictive value applied to lab test?
Predictive value evaluates the reliability of a positive or negative result
What 2 things are used to establish predictive value?
The sensitivity and specificity of the test in question
T/F- To establish predictive value you need a reference method for diagnosis to be available other than the test being evaluated?
This is what the test in question is compared to
-The best test
If it is not possible to detect patients with a certain disease...
Its not possible to provide a truly accurate calculation of sensitivity, specificity, and predictive value
Reliability of laboratory tests is affected by...
Technical performance in the lab
The measure of how closely the lab can approach the same answer when the test is repeatedly performed with the same sample
What factors influence reproducibility
-Experience of technicians
How is variation of the mean expressed?
When the standard deviation is converted to a percentage of the mean value, this is called?
Coefficient of variation
The higher the coefficient of variation, the higher/lower the accuracy, greater/lesser reproducibility?
Lower accuracy, lesser reproducibility
Reproducibility is demonstrated with the coefficient of variation of ___ %?
What does the coefficient of 4% mean?
That 68% of values are within 1 standard deviation (4% above and 4% below the true value)
+/- ___ standard deviations is used to define acceptable limits?
A deviation of ___% is deemed technically acceptable
This is the correct answer, and the result that the assay should produce
How does the lab figure out the accuracy of a test?
Compare it to the assay
What happens in a random inaccuracy?
-Random inaccuracies may not effect all of the specimens and therefore may have not altered the control specimens
What are examples of random inaccuracies?
-Specimen from the wrong patient
-Clerical errors (patients with the same last name)
-Decimal point errors
-Transcription of results into the wrong charts
If unexpected test results are seen, what should you do first?
Repeat the tests
T/F- Normal range is a concept that considers all tests within the normal range to be normal and anything outside the normal range to be abnormal
What are the 2 premises involved in establishing the normal range?
1. Assume that all patients do not demonstrate clinical signs and symptoms that are normal
2. Assume that test results from those people considered normal will have random distribution
Are any factors that would bias a significant group toward either the high or low side included?
What percentage of values fall in +/- 1 standard deviation?
What percentage of values fall within +/- 2 standard deviations?
What percentage of values fall within +/- 3 standard deviations?
If 2 standard deviations are accepted, what percentage of clinically normal people will have values above and below this range?
2.5% above and 2.5% below
What if clinically normal people have undetected disease?
They will influence the normal range so...
-you have to take the top values or the lowest values and run their labs again using the gold standard tests to see if they are in fact normal- if not, they are outta hereee and you made your normal range more accurate
What factors may effect results in non-diseased patients?
-Specimen storage time
The greater number of tests employed, the greater/lesser the chance that at least one will yield a false result?
The farther away the values are from the normal limits, the greater/lesser the likelihood of a true disease