Impact of stochastically differential sensitivity on bias in the risk ratio under perfect specificity

Plots depict coefficients from linear regression of log₁₀ bias factor on quantiles of 𝚫Se in simulation scenarios with Sp=1

Exposure prevalence = 0.1 | Risk in unexposed = 0.01

True RR = 1

True RR = 1.5

True RR = 2

True RR = 5

Exposure prevalence = 0.1 | Risk in unexposed = 0.10

True RR = 1

True RR = 1.5

True RR = 2

True RR = 5

Exposure prevalence = 0.5 | Risk in unexposed = 0.01

True RR = 1

True RR = 1.5

True RR = 2

True RR = 5

Exposure prevalence = 0.5 | Risk in unexposed = 0.10

True RR = 1

True RR = 1.5

True RR = 2

True RR = 5