Reference interval Flashcards
Define biased results
results are on one side & not near expected result
List reasons why you would consider changing to a new method
- improve accuracy & precision
- reduce reagent/labour cost
- new instrument/equip. eg. bc slow
- new kit
- To measure a new analyte
Types of error
- Constant systematic error: change is constant
- Proportional systematic error: changes due to PERCENTAGE of analyte (conc)
- Random error: no direction
State 5 causes of systematic error.
- Wrong caliberator
- improperly prepped reagents
- Inappropriate storage of reagents or calibrators
State 4 causes of random error.
- bubbles in reagent
- Solution not mixed
- Improperly fitting pipette tips
- imprecise pipetter
Compare between systematic and random error.
Sysstematic: error in one direction
Random: no direction (error varies from sample to sample)
What is the difference between analytical sensitivity (Sn) and analytical specificity (Sp)?
Sensitivity: detect an analyte at very low concentration
Specificity: marker that is capable to detect just one specific analyte without being interfered with other analytes
Screening vs confirmatory tests.
Screen: require high sensitivity
Confirmatory: require High specificity
Define cut off value
a fixed value that separates sick ppl from healthy ppl
roles of ROC curve
- determine cut-off value
- Find best combo. b/w Sn & Sp
- FInd best tester (new assay - biomarker)
What does this rule Xys mean (westgard)
X = # of values past that SD ys = [1 / 2 / 3] SD
Rules of westgard
a. 1 2s = warning
b. 2 2s = REJECT
c. 1 3s = REJECT
d. R4s = one value exceeds +2SD & -2SD
e. 4 1s = 4x points above a SD consecutively (in a row)
f. 10x = 10x points above a SD consecutively = REJECT
ROC curve plots (axis labels)
sensitivity (TP) vs 1-specificity (FP)
interpresting ROC graph
Hi area under graph = better marker = more accurate
The —————- allows determination of whether an analytical run is “in-control” or “out-of-control”
Westgard multi-rule
