Negative likelihood ratio

Method

Pooled negative likelihood ratio of an index test against a reference standard — one minus sensitivity, divided by specificity — estimated in a bivariate or hierarchical random-effects model. Index test, reference standard, and decision threshold recorded on the edge.

The method is part of the endpoint identity. Studies measuring the same quantity by a different method are held as separate endpoints.

Notes

SIGN INVERSION against the positive likelihood ratio, its companion from the same 2x2 table and the same model. This one improves by FALLING toward 0 (values below 0.1 rule a diagnosis out); the positive ratio improves by RISING away from 1. They are dimensionless ratios with identical units and opposite polarity, which is precisely the condition under which a merge silently reverses a conclusion. Also distinct from the diagnostic odds ratio, which absorbs both.