Question

screening test, respectively. Interpret your findings. Would you recommend that this test be used to screen for diabetes in general populations, and why or why not?

A new screening test for diabetes produced the following results: True Disease Status Screening test Positive Negative Positive 100 20 Negative 60 400 Based on these data, calculate the test's sensitivity, specificity, positive predictive value, negative predictive value, and the accuracy of the screening test, respectively. Interpret your findings. Would you recommend that this test be used to screen for diabetes in general populations, and why or why not? You need to show step-by-step calculations to earn maximum credit

You need to show step-by-step calculations to earn maximum credit

Answer 1

The various metrics here for the test are computed as:

a) Sensitivity of the test is computed here as:
= n(+ and have disease) / n(have disease)

= 100 / 160

= 5/8

Therefore 0.625 or 62.5% is the required sensitivity of the test here.

b) The specificity fo the test here is computed as:

= n( - and no disease) / n(no disease)

= 400 / 420

= 0.9524

Therefore 95.24% is the required specificity of the test here.

c) the predicted positive value of PPV for the test here is obtained as the probability that the person really has the disease given that the test is positive. This is computed here as:

= P(+ and disease) / P(+)

= 100 / 120

= 0.8333

Therefore 0.8333 or 83.33% is the required positive predictive value for the test here.

d) The negative predicted value of NPV for the test is computed here as:

= P(- and disease) / P(-)

= 400/ 460

= 0.8696

Therefore 0.8696 or 86.96% is the required negative predictive value for the test here.

e) The accuracy here is computed as:

= Correct predictions / Total predictions

= (100 + 400) / (580)

= 0.8621

Therefore 86.21% is the required accuracy here.

As the sensitivity of the test here is only 62.5% this means that a large proportion of people with the disease will get a negative test for diabetes and therefore the false negative rate is very high here. Therefore this test should not be recommended here as many of the diabetic people wont be able to get diagnosed.