Calculation

Question 1

Standard deviation equation?

Percent with in 1, 2, 3 deviations?

Coefficent of variant?

Answer 1

SD= Sq root (Variance), Variance is Sum of (X_i-mean)²/(N-1)

68%, 95%, 99.7%

CV=SD/Mean * 100

Question 2

What is a reference interval; # needed to make an RI, # needed to confirm?

RI In non Gausian distribution?

Answer 2

Look at subset of representitive population; 100-120 people neeed: can confirm with 20

If Gausian RI is 2 deviations

If non-Gausian need 2.5% and 97.5%; look at central 95% at RI

Question 3

What test compares means of two groups?

What test if the populations are not Gausian?

Test to compare two analyizers?

Answer 3

Student T-test if distributions similar

Mann-Whitney U-test

Correlation test; paired T-test by measuring on each machine; More sensitive than unpaired test

Question 4

In compairison tests a Standard Regession assumes?

Deming regression?

Answer 4

Only Y varies

X and Y can contain error

Question 5

R² means?

Bias?

Answer 5

Degree to which Y varies with X

True result-analytical result =< defined limit; can be proportional or constant

Question 6

How is precision determined?

Total imprecision?

Answer 6

Measure specimen multiple times and look at SD or CV

Total: Square root (Intra-assay variance + inter-assay variance)

Question 7

How to tell if assay is working properly?

Do CAP and JACO require specific QC rules?

Answer 7

Use controls with known concentration and variance

No; you can pick your own but have to be used systematically

Question 8

Some common Westard rules:
1_2s 1_3s ?

2_2s?

R_4s?

4_1s?

10_m?

Answer 8

1_2s: Warning, 1 control between 2-3 SD’s

1₃s: Single control above or below 3 SD’s; Fail run

2_2s: 2 sequential runs results >2-3 SD above or below on same side

R_4s: 2 sequential resutls with total range >4 SD above or below

4_1s: 4 sequential runs >1 SD above or below; same side; only 4th run is out

10_m: 10 sequential runs above or below mean