My bad - you're right, it's in the Appendix to the NEJM paper - but I don't put too much weight into the NFS, it just uses some measured parameters and a calculator to predict NAFLD status (serum glucose, platelet count, albumin, AST/ALT ratio, age, BMI, and diabetes status). My doc had me get an ultrasound to confirm that my numbers indicated I had NAFLD (oddly, TGs aren't part of the NFS, and why mine were elevated is the answer we were looking for). IOW, he used the them to determine whether I needed diagnostic imaging - I don't think the NFS even existed way back then. The NFS is really only accurate if you already have advanced fibrosis.
NFS and FIB-4 were superior to BARD and APRI in accurately determining risk for severe liver disease and mortality. Despite the moderately predictive accuracy of NFS and FIB-4 scores for determining liver-related outcomes and mortality, these systems are not accurate enough to be considered for routine use in clinical practice. Furthermore, their inadequacy in predicting long-term outcomes makes them problematic for use in therapeutic clinical trials in patients with NAFLD. New noninvasive scoring systems are urgently needed.