We have evaluated the new Roche digoxin "On Line" procedure for use in a pediatric population with particular interest in the potential for interference by digoxin-like immunoreactive factor (DLIF). An initial study comparing digoxin values obtained with the new Roche procedure with determinations on an Abbott TDx, American Dade Stratus, and COBAS-FARA using Microgenics Cedia reagents, found good correlations with these established methods. The Roche method was suitably precise and utilized either serum or plasma. Interference by DLIF was assessed by analyzing specimens from patients not receiving digoxin but likely to contain DLIF, with the argument that non-zero values represent cross-reactivity of anti-digoxin antibodies with DLIF endogenous to these specimens. When specimens from neonates, women with second/third trimester pregnancies, and patients with renal and liver failure were assayed with the Roche, Stratus, and TDx methods, all three methods measured DLIF in some specimens, but the Roche method possessed the lowest overall DLIF interference. The modest extent of DLIF interference and the requirement of a small amount of specimen make the Roche method superior in monitoring digoxin in a pediatric population.