Lab Statistics and Quality Control Flashcards

1
Q

Sensitivity *

A

Probability of classifying a diseased patient as diseased.

True pos / (True pos + False negs) * 100

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Specificity *

A

Probability of classifying a Neg patient as Neg

True Negs / (True Negs + False Pos’s) *100

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Positive Predictive Value *

A

True Pos / (True Pos + False Pos) * 100

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Negative Predictive value *

A

True Neg / (True Neg + False Neg) * 100

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

False Positive Rate *

A

False Pos / Total Neg Patients * 100
Probability of incorrectly classifying a non-diseased person as diseased

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

False Negative Rate *

A

False Neg / Total Pos Patients * 100
Probability of incorrectly classifying a diseased person as not diseased

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Accuracy *

A

Closeness of measured values to the true value.
Ability to get close to the bullseye.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Precision *

A

Closeness of repeated measurements to each other.
Ability to hit the same spot on the target.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

68.26% *

A

Percentage of values that will fall within one SD

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

95.46% *

A

Percentage of values that will fall within 2SD

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

99.73% *

A

Percentage of values that will fall within 3SD

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Mean *

A

Sum of values / # of values

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Median *

A

The value in the middle of a list

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Mode *

A

The most commonly occurring value in the set

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Unimodal Symmetrical *

A

Normal distribution, mean mode and median stack on top of each other

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Unimodal Skewed *

A

when the most frequent occurrences are higher or lower than the median (middle data) and mean (average) of the data set.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Bimodal *

A

When there are two frequent data points, as in test results for men and women.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Standard Deviation Definition *

A

Numerical value describing how widely points vary from the Mean.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

SD formula *

A

Square root of the Variance (sum of (X - Mean) squared) / (n-1)

20
Q

What is the Z-Score

A

Tells how many SD’s a control is from the mean. = (Control result - Mean) / SD

21
Q

Variance *

A

= (sum of (X - Mean) squared) / (n-1)

22
Q

Coefficient of Variation Definition *

A

Ratio of the Standard Deviation to the Mean

23
Q

CV formula *

A

(SD/Mean) *100

24
Q

Confidence Interval

A

A range of values calculated from the Mean and SD

25
Q

CI Formula

A

Mean +/- 2SD

26
Q

CI Formula

A

Mean +/- 2SD

27
Q

t-test

A

How significant the differences between two groups are. How likely the differences could have happened by chance.

28
Q

What are 3 types of t-test?

A
  1. Independent sample t-test (compares the mean for 2 groups)
  2. Paired sample t-test (compares a group of samples collected at different times)
  3. One sample t-test (compares the mean of a group against a known Mean)
29
Q

Low p-values indicate

A

Your data did not happen by chance. p-value of .01 means there is a 1% probability that your results happened by chance.

30
Q

What are the four types of Error? *

A
  1. Random Error (due to chance)
  2. Systematic error (influences values in one direction)
  3. Active Error (occurs between a specific tech and the samples)
  4. Latent Error (related to the organization of the laboratory)
31
Q

Reference Range

A

Range of Normal values for a test, usually +/- 2SD or 95% confidence

32
Q

Chain of Custody

A

Chronological documentation or paper trail of a specimen throughout collection and testing. Required for any specimen used as legal evidence in court.

33
Q

Levey-Jennings Plots *

A

Plot of control results on Y-axis versus time on the X-axis

34
Q

Trend *

A

Progressive GRADUAL drift of values from a prior mean (gradual deterioration of a light source, control materials, chamber temp, or inst. calibration)

35
Q

Shift *

A

ABRUPT changes in the control mean (sudden change in light source, incubation temp, reagent lot, control integrity)

36
Q

Dispersion *

A

Lack of precision or RANDOM errors caused by inconsistent technique or stability issues (fluctuations in voltage, different techs, clots or bubbles)

37
Q

Cusum Plots *

A

Cumulative Sum Plot, used for SYSTEMATIC errors. The known value is subtracted from the measured value for each day of measurement. Difference is summed together over time. (When control data is scattered around the mean, the Cusum will wander above and below the mean yielding a relatively horizontal line)

38
Q

Westgard Rules *

A

Used for RANDOM and SYSTEMATIC errors. A set of 5 rules on whether to accept or reject a run. Based on how many times the control falls outside of the 1, 2, or 3SD’s of the mean.

39
Q

If a control was 1 2S, would you reject the run? *

A

No, run is in control

40
Q

If a control was 1 3S, would you reject the run? *

A

Yes, run is out of control

41
Q

If a control was 2 2S, would you reject the run? *

A

Yes, run is out of control

42
Q

If a control was R 4S, would you reject the run *

A

Yes, run is out of control. One value +2SD and one value -2SD

43
Q

If a control was 4 1S, would you reject the run *

A

Yes, run is out of control. Four consecutive values exceed 1SD

44
Q

If a control was 10 X, would you reject the run *

A

Yes, run is out of control. Ten values in a row above or below the mean

45
Q

Six Sigma *

A

A quality improvement process to identify and remove causes of defects and reduce errors to 3.4 defects per million tests or 6SD.

46
Q

What is the DMAIC method for Six Sigma? *

A

Define, Measure, Analyze, Improve, Control