In: Math
In this hypothetical case study, a new rapid test kit nicknamed Alpha is being reviewed. The sensitivity of Alpha is 93.0% (0.93) and the specificity is 96.0% (0.96). Assume that the actual prevalence of the Zika antibody among the United States population of blood donors is 4% (0.04) and that of the hurricane disaster relief volunteers returning from Puerto Rico is 20.0% (0.20).
Construct a separate 2 x 2 table, to calculate the PPV and NPV for a population of 2,500 volunteers who aided in the recent hurricane relief efforts in the Caribbean.
If sensitivity and specificity remain constant, what is the relationship of prevalence to predictive-value positive and predictive-value negative? (Hint: Think if one increases, decreases, or stays the same.)
Answer :
Constructive prescient esteem is the likelihood that a man is having illness, given that the individual was tried positive
Positive predicted value = P(disease and test positive) / P(test positive) = (prevalence x sensitivity) / (prevalence x sensitivity + (1 - prevalence)x(1 - specificity)
For joined states populace, the commonness is 0.04
relating positive prescient esteem = (0.04x0.93)/(0.04x0.93 + (1 - 0.4)x(1-0.96))
= 0.0372/0.0372+(0.6*0.04)
= 0.0372/0.0372+0.024
= 0.0372/0.0612
= 0.0.6078
= 60.78%
For volunteers coming back from Puerto Rico, the pervasiveness is 0.2
corresponding positive predictive value = (0.2x0.93)/(0.2x0.93 + (1 - 0.2)x(1-0.96))
= 0.186/0.186+(0.8*0.04)
= 0.186/0.186+0.032
= 0.186/0.218
= 0.8532
= 85.32%
The positive prescient esteem increments with increments in commonness
Pessimistic prescient esteem is the likelihood that a man isn't having sickness, given that the individual was tried negative
Negative predicted value= P(no illness and test negative)/P(test negative)
= ((1 - pervasiveness) x explicitness)/((1 - commonness) x particularity) + predominance x (1 - affectability))
For joined states populace, the pervasiveness is 0.04
corresponding negative predictive value = ((1 - 0.04) x 0.96)/((1 - 0.04) x 0.96) + 0.04 x (1 - 0.93))
(0.96*0.96)0.96*0.96)+0.04*0.07
=0.216/0.9216+0.0028
= 9216/0.9244
= 0.9960
= 99.60%
For volunteers coming back from Puerto Rico, the pervasiveness is 0.2
corresponding positive predictive value = ((1 - 0.2) x 0.96)/((1 - 0.2) x 0.96) + 0.2 x (1 - 0.93))
= (0.8*0.96)/(0.8*0.96)+0.2*0.07
= 0.768/0.768+0.014
= 0.768/0.714
= 1.0756
= 107.563%
The negative prescient esteem diminishes with increment in commonness