4, A certain blood test for some disease gives ‘positive\' result for 90% of pat
ID: 3045465 • Letter: 4
Question
4, A certain blood test for some disease gives ‘positive' result for 90% of patient whoare infected with the disease (detection rate Pr(PID)-90%), but also a 'positive' result for 5% of people who are NOT infected (false positive rate Pr(P ) =5% ten thousands (prevalence rate of Pr(D) =0.01%). o). Ihe infection rate among the population is one in a) What is the probability that a person who tested positive' is indeed infected with this disease? Define an'accuracy, measure as: Pr(DIP)×Pr(DIP), what will be its value for this test? The company can invest X number of dollars into R&D; and to improve either the detection rate to 99% OR to reduce the false positive rate to 4% (but not both), what should the company do to improve the test's accuracy the most? b) c)Explanation / Answer
a)
Pr(D|P) = Pr(D and P)/Pr(P) = 0.90 * 0.01/100 / ( 0.90 * 0.01/100 + (1 - 0.01/100) * 0.05)
= 0.001796
b)
accuracy = Pr (D|P) * Pr(D'|P')
Pr(D'|P') = Pr (D' and P')/Pr(P')
= (1-0.01/100)*0.95 / ( (1-0.01/100)*0.95 + 0.01/100 * 0.10)
= 0.99998947274
hence
accuracy = 0.001796 *0.99998947274 = 0.00179598109
c)
if detection rate 99 %
then
accuracy = 0.99 * 0.01/100 / ( 0.99 * 0.01/100 + (1 - 0.01/100) * 0.05) * (1-0.01/100)*0.95 / ( (1-0.01/100)*0.95 + 0.01/100 * 0.01)
=0.0019762825
if Pr (P|D') = 4%
then accuracy = 0.90 * 0.01/100 / ( 0.90 * 0.01/100 + (1 - 0.01/100) * 0.04) * (1-0.01/100)*0.96 / ( (1-0.01/100)*0.96+ 0.01/100 * 0.10)
=0.00224514948
since in 2nd accuracy is more ,
we should reduce the false positive rate to 4 %
Please rate
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.