Term
| What does sensitivity and specificity measure? |
|
Definition
-sensitivity=(TP/TP+FN)=animals with disease that test positive
-specificity=(TP/TN+FP)=animals with desease that test negative |
|
|
Term
| List reasons for doing lab work: |
|
Definition
1. resolving differential diagnoses
2. screening for disease
3. assessing the severity of disease
4. increasing prognostic accuracy
5. determination of drug toxicities/levels
6. evaluating a response to therapy |
|
|
Term
| Important info regarding interpretation of lab work: |
|
Definition
-RARELY CAN A DIAGNOSTIS OR PROGNOSIS BE MADE ON THE BASIS OF ANY SINGLE TEST RESULT!
-LAB RESULTS MUST BE INTEGRATED WITH HISTORY, PE, RADIOGRAPHIC FINDINGS, ETC. |
|
|
Term
| What knowledge of analytical and statistical principles is critical for answering the question: is the blood work abnormal of is the patient abnormal? |
|
Definition
1. reference intervals
2. sources of analytical variation and error |
|
|
Term
| How are reference intervals constructed? |
|
Definition
-(40-120) animals that reflect patient population used
-generally looking for "healthy" animals, no medications except parasite control
-factors to consider: genetic background/breed, sex and reproductive status, diet and other husbandry conditions, age |
|
|
Term
| For most clinical RI generation, a range of breeds, sexes, ages, and living conditions are utilized to get a "general purpose" interval that can be used for a large potential population of patients. Name some examples of cases that justify the use of different RIs. |
|
Definition
-greyhounds
-warm bloods vs. cold blood horses
-strains of rodents may require their own reference intervals
-age can impact RIs, so if working with neonatal or pediatric paitents, use RIs specifically generated for those populations |
|
|
Term
| Construction of RIs: impact of analytical methodology: |
|
Definition
-RIs are specific to the way the test was performed
-impacted by kind of machine, reagents, etc.
-Use RI generated for your analyzer using the chemicals that you are using (in some cases may be necessary to used published RIs) |
|
|
Term
| True of false: chemistry RIs are less impacted by methodology than hematology RIs. |
|
Definition
| False: in general, hematology RIs are less impacted by methodology than chemistry RIs |
|
|
Term
| How are RIs constructed (numerically)? |
|
Definition
| -mean +/- SD OR central 95% of animals if data are non-parametric |
|
|
Term
| By definition___________% of normal patients will be outside reference intervals. |
|
Definition
|
|
Term
| The statistical probability that a completely normal patient will have normal results on the average serum biochemical profile containing 12 tests is only _____________. |
|
Definition
|
|
Term
| Do reference intervals = "normal values" |
|
Definition
| NO. Normal patients can have values outside of RI, and 5% do by definiction. Sick patients may have values within RI. |
|
|
Term
| List sources of analytical variation and error. |
|
Definition
1. precision and accuracy
2. laboratory errors and artifacts |
|
|
Term
| Why are accuracy and precision required in the laboratory. |
|
Definition
| Accuracy and precision in the laboratory are required to ensure that abnormal values truly reflect the status of the patient and is not the result of poor laboratory technique. |
|
|
Term
| Define precision and accuracy. |
|
Definition
-accuracy=agreement between laboratory results and its "true" value
-precision=agreement between replicate measurements
*good labs perform validation studies and have quality assurance programs to make sure that test results are acceptably accurate and precise |
|
|
Term
|
Definition
-a little is OK and unavoidable
-if you run the same sample through the same analyzer several times, you'll get slightly different numbers, even if the analyzer is acceptably precise and accurate
-variation depends on technique and analyte
-good sample collection/handling minimizes variation
-i.e. PCV can vary 1-3 points, WBC count can vary (1-2 thousand), and platelets vary the most (i.e. when monitoring patients with thrombocytopenia, careful when analyzing changes) |
|
|
Term
| "Inter-assay variation" must be considered when interpreting changes in patient data |
|
Definition
| -b/c precision is rarely if ever perfect,if you run the same sample through an analyzer several times, there will be some variation in the resulting data |
|
|
Term
| List preanalytical sources of error: |
|
Definition
1. inappropriate test choice: not validated for spp or sample type, poor analytical or diagnostic sensitivity or specificity
2. collection of sample: traumatic venipucture, urine collected off a surface, contamination of cytology sample by bacteria
3. sample labeling
4. transport
5. lab labeling errors
6. lab processing errors
7. hemolysis, lipemia, icterus cause interference
8. failure of separation of serum from cells
9. post prandial effects
10. effects of medications; analytical (interferes with machine measurement) and physiologic (changes occur in patient in response to medication) |
|
|
Term
| List analytical sources of error. |
|
Definition
1. technical problems with analyzers
2. quality assurance issues
3. in well run labs with good analyzer maintenance and quality assurance, errors are by people in test choice and sample collection (NOT by analyzers) |
|
|
Term
| Why is bad data worse than no data? |
|
Definition
-leads to inappropriate dx
-pursuit of nonexistent problems
-delayed dx of problem
-DO NOT ACCEPT DATA THAT SEEMS INCONGRUOUS WITH CLINICAL PICTURE WITHOUT DOUBLE CHECKING!! |
|
|
Term
| Remember: normal lab test results do not rule out disease! |
|
Definition
|
|
