1.16: Statistics and Reference Ranges

1. Presuming there is a healthy and a diseased population, which fact is not true?:
   1. healthy and disease values overlap
   2. healthy and disease cut off points will determine the number of false positives
   3. healthy and disease values are always clearly separated
   4. healthy and disease cut off points will determine false negatives
   5. healthy and disease values are a continuum
2. What is a minimum number of samples to obtain a good reference range?:
   1. 20
   2. 40
   3. 120
   4. 200
   5. 500
3. The mean of a series of results is 100 mg/L and the standard deviation is 2 mg/L. The percent coefficient of variation of this analysis is:
   1. 0.5
   2. 1
   3. 2
   4. 4
   5. 8
4. Test A correctly identified 17/100 true positives while correctly identifying 90/100 true negatives. Test B correctly identified 55/100 true positives while correctly identifying 80/100 true negatives. Which of the following statements is true?:
   1. Test A has a better sensitivity and specificity
   2. Test B has a better sensitivity and specificity
   3. Test A has a better sensitivity but a worse specificity
   4. Test B has a better sensitivity but a worse specificity
   5. Tests A and B have equal sensitivities and specificities
5. The 95% confidence limits for Ca⁺⁺ in control sera were established as 9.2-10.2 mg/L. The standard deviation is:
   1. 0.1 mg/L
   2. 0.2 mg/L
   3. 0.25 mg/L
   4. 0.5 mg/L
   5. 1.0 mg/L
6. What percentage of tests should fall within ± one standard deviation if the distribution for the population is considered normal?:
   1. 28
   2. 35
   3. 68
   4. 45
   5. 95
7. Precision applied to chemistry determination means the same as:
   1. accuracy
   2. reproducibility
   3. recovery
   4. conformance with Beer’s Law
8. The following set of values was obtained for K+ concentration for a quality control sample: 4.7, 3.9, 4.0, 4.1, 4.3, 4.5, 3.9, 4.2, 4.2, 4.6, 4.0, 3.9, 4.6, 3.8, 4.2, 4.5, 4.1, 4.7, 3.9, 4.8. The mean for this population of data is:
   1. 4.0
   2. 4.1
   3. 4.2
   4. 4.4
   5. 4.6
9. A 95% confidence limit for cholesterol precision was found to be 1300-1500 mg/L. The coefficient of variation for this methodology is:
   1. 1.23%
   2. 3.57%
   3. 7.14%
   4. 14.28%
   5. 28.56%

Use the following Key to answer Questions 10-14:

1. 1, 2, and 3 are correct
2. 1 and 3 are correct
3. 2 and 4 are correct
4. 4 only is correct
5. all are correct

10. A normal value is usually defined as:
    1. a result from a healthy person
    2. a gaussian distribution
    3. a usual laboratory value
    4. the results of a t-test
11. Reference populations should be defined in terms of:
    1. age
    2. sex
    3. race
    4. genetic background
12. The positive predictive value (accuracy) of a test varies with:
    1. test sensitivity
    2. disease prevalence
    3. test specificity
    4. test linear range
13. The means and standard deviations of methods A and B are X_A = 80, SD_A= 6; X_B = 40, SD_B = 1.5. The following statements are true about these 2 methods:
    1. there is no apparent bias between the 2 methods
    2. method A is less precise than method B
    3. there is no difference in precision between the two methods
    4. there appears to be a bias between the two methods
14. An indication of the spread of the distribution of a set of measurements is given by the:
    1. range
    2. variance
    3. standard deviation
    4. mean
15. A test has a sensitivity of 95% and a specificity of 80%, and a positive predictive value of 50%. If a result of this test was positive (i.e., abnormal), then the likelihood that disease was present is:
    1. 95 out of 100
    2. 80 out of 100
    3. 50 out of 100
    4. 20 out of 100
    5. none of the above
16. A useful test to determine a statistical difference between the means of two different populations is the:
    1. F-test
    2. t-test
    3. chi-square test
    4. linear regression test
    5. variance test
17. A receiver-operating characteristic (ROC) curve is used to graphically display:
    1. Linearity of a new method
    2. Differences between variances of two methods
    3. Correlation (regression) between methods
    4. Ability of a test to discriminate between disease and non-disease
    5. The distributions of diseased and non-diseased populations

Answer

1. c (p. 364-365)
2. c (p. 369)
3. c (p. 345)
4. d (p. 374)
5. c (p. 346)
6. c (p. 346)
7. b (p. 348)
8. c (p. 344)
9. b (p. 346-347)
10. b (p. 342, 364)
11. e (p. 366)
12. a (p. 372-373)
13. c (p. 348-349)
14. a (p. 345)
15. c (p. 373)
16. b (p. 349)