Skip to main content
. 2023 Jan 24;5(2):fcad015. doi: 10.1093/braincomms/fcad015

Figure 3.

Figure 3

Pre-screening with a relative risk threshold approach. This figure shows the effect of different thresholds for relative risk (i.e. predicted probability from the logistic regression model normalized across the study population) on trial recruitment. Panel A shows the Aβ PET + rate as a function of relative risk (i.e. the lower threshold, so 0% means everyone gets a PET scan). Panel B shows the trade-off between the total number of tests (on log10 scale) needed in the pre-screening phase (i.e. plasma) versus the screening phase (i.e. PET) for a trial with 500 PET + CU participants. Panel C shows the total cost savings by pre-screening across different cost ratios. For example, a trial using the relative risk threshold of 25% would have an Aβ PET + rate of 38.6% and a threshold of 75% would have an Aβ PET + rate of 64.2%. The expected Aβ PET + rate with no pre-screening is 28.6%.