**Evaluating Tests and
Other Medical Evidence**

**Answers**

** **

** **

1.c

2.d

3.b

4. a.
Sensitivity = 1 - (1-.7)x(1-.9) = 97%; Specificity = .9x.8 = 72%;

For PV+, post-test odds = 1x(.97/.28) = 3.46,

PV+ = 3.46/4.46 = 1x(.97/.28) = 78%;

For PV-, post-test odds = 1x(.03/.72) = .042,

PV- = 1 - (.042/1.042) = 96%.

b. and c. Sensitivity = .7x.9 = 63%;

Specificity = 1 - (1-.9)x(1-.8) = 98%;

predictive values 97% and 73%, calculated as above.

5.c

6. 4.3%

7. a)c

b)a

c)a

8. In the first
study there was lead time bias, length/time bias, and compliance bias. In the
second study, when they origionally did the randomized clinical trial those
biases were gone due to randomization, but when they (non-randomly) re-grouped
within the randomized groups the biases came back. The screening
therefore does not seem to help.

9. l

10.c

11.d