Screening

Exercise 1

A screening test for a newly discovered disease is being evaluated for its effectiveness and sensitivity as a screening test in industry. In order to determine the effectiveness of the new test, it was administered to 880 workers. Of those, 120 of the individuals diagnosed with the disease tested positive for it. Results from the test showed a negative test finding for 50 people with the disease. A total of 40 people without the disease tested positive for it.

Create a 2x2 table for this data:
Total
Total

1. What is the prevalence rate of the disease?
2. What is the sensitivity of the test?
3. What is the specificity of the test?
4. What is the percentage of false negatives?
5. What is the percentage of false positives?
6. What is the accuracy of the test?

Exercise 2

Two different tests previously developed to measure stress in individuals are selected: Stress Test Alpha Battery (STAB) and Stress Test Uniform Battery (STUB). The sensitivity and specificity of these two tests are:

STAB: Sensitivity = 60%
Specificity = 95%

STUB: Sensitivity = 75%
Specificity = 90%

Which test will develop the greatest proportion of:

1. False negatives?
2. False Positives?
3. True Positives?
4. True Negatives?

Exercise 3

Infection with the human immunodeficiency virus (HIV) is routinely diagnosed by detecting the presence of specific antibodies in the patient's serum. Although the presence of the virus itself can be now be detected, these tests remain expensive and require laboratory techniques that are not routinely available. The diagnosis of HIV infection begins with an enzyme immunoassay (EIA). The optical density (OD) of the patient's EIA is compared to a control specimen (OD ratio). If the OD ratio is - above the established cutoff for that control sample on repeat testing the specimen is termed "repeatedly reactive". The EIA is relatively sensitive, fast, simple and inexpensive which makes it an appropriate screening test. However, if one examines the ODs for a large group of samples from patients with and without true HIV infection you can see that there is some overlap in their EIA results if a value of A is used for the cutoff:

Hypothetical distribution of OD ratios for patients with and without HIV infection. Patients with HIV infection are depicted with the BLUE, thick line, and the patients without HIV are depicted with a thin, RED line.

1. What would be the impact of moving the cut-off line from A to B on sensitivity and specificity?
2. What would be the impact of moving the cut-off line from A to C on sensitivity and specificity?
3. Where would you suggest setting the cut-off?
4. If you are the director of a blood bank, and having HIV-negative blood is vital, where would you set the cut-off and why?
5. If you are the director of an investigational drug for HIV-positive patients, which decrease viral load and protect T cells, BUT have significant side-effects, where would you set the cut-off and why?

Exercise 4

Muscle tension dysphonia (MTD) can masquerade as adductor spasmodic dysphonia (ADSD), leading to diagnostic confusion. An investigation assessed the diagnostic worth of acoustic analysis of phonatory breaks (PB) as a possible objective test to distinguish ADSD from MTD.
Acoustic Analysis of Phonatory Breaks to Test for ADSD and MTD (n=100)

Reference Standard
Test Outcome
Has ADSD
Has MTD
Positive (PB40 ms)
27
True Positive (TP)
25
False Positive (FP)
Negative (PB<40 ms)
14
False Negative (FN)
34
True Negative (TN)

Calculate the sensitivity, specificity, PV+, and PV-. Interpret your results.

Exercise 5

Suppose you read a research article that suggests a 70 ms cutoff is potentially more appropriate than the 40 ms cutoff you used. You re-analyze your data and find that TP=17, FP=15, FN=24, and TN=44. Present the revised contingency table and calculate the new sensitivity, specificity, PV+, and PV-.