I'm bothered by two things. The first is multivariate studies are notorious for throwing false positives especially when positive cases are in the sigle digits

I'm bothered by two things. The first is multivariate studies are notorious for throwing false positives especially when positive cases are in the sigle digits. Comparing data sets collected under different surveillance regimes is difficult.