Question

1. A study was done to evaluate the sensitivity and specificity of an electrocardiography (ECG) test for detecting left ventricular hypertrophy (LVH), a binary measure of cardiovascular disease. The authors reported the following results from their sample of 1562 men. Assume that these 1562 men were randomly sampled from the population of males.

	LVH = Yes	LVH = No
Test positive	20	198
Test negative	26	1318

1. (*) What are the sensitivity and specificity of the ECG test for detecting LVH in men?

2. (*) What is the negative predictive value of the test for men?

3. (*) What does this negative predictive value mean in words?

4. (*) Suppose now that you plan to use the test only on men who based on other factors are considered to be of high risk for LVH (compared to the general population). What would that mean for each of your estimates in A, B, and C?   For the use in this high risk population would your estimates be about right, too high, or too low?  

Answer 1

The completed table with totals is,

	LVH = Yes	LVH = No	Total
Test positive	20 (A)	198 (B)	218
Test negative	26 (C)	1318 (D)	1344
Total	46	1516	1562

A.

Sensitivity =  A/(A+C) × 100

= (20 / 46) * 100

= 43.48%

Specificity = D/(D+B) × 100

= (1318 / 1516) * 100

= 86.94%

B.

Negative Predictive Value = D/(D+C) × 100

= (1318 / 1344) * 100

= 98.07%

C.

This means that if the person is diagnosed as negative in ECG test, then the chance that it does not have LVH disease is 98.07%.

D.

For the use in this high risk population, the  sensitivity and specificity will remain the same because these are characteristic of this test.

With high risk population, there will be decrease in negative Predictive Value. If we test in a high prevalence population, it is more likely that persons who test negative truly have disease.