I think to do this, you'd need data points that don't yet exist: if your microbiome changes from antibiotic use, will you suffer health effects (beyond infection with a disease causing bacteria like C. difficle)?