Aim To assess the diagnostic accuracy of the JM 103 as a screening tool for neonatal jaundice and explore differential effects based on skin colour.

Methods We prospectively compared the transcutaneous bilirubin (TcB) and serum bilirubin (TSB) measurements of newborns over a 3 month-period. Skin colour was assigned via reference colour swatches. Diagnostic measures of the TcB/TSB comparison were made and clinically relevant TcB cut-off values were determined for each skin colour group.

Results 451 infants (51 light, 326 medium and 74 dark skin colour) were recruited. The association between TcB and TSB was high for all skin colours (r_s>0.9). The Bland-Altman analysis showed an absolute mean difference between the two measures of 13.3±26.4 µmol/L with broad limits of agreement (−39.4–66.0 µmol/L), with TcB underestimating TSB in light and medium skin colours and overestimating in dark skin colour. Diagnostic measures were also consistently high across skin colours, with no clinically significant differences observed.

Conclusions The JM 103 is a useful screening tool to identify infants in need of serum bilirubin, regardless of skin colour. The effect of skin colour on the accuracy of this device at high levels of serum bilirubin could not be assessed fully due to small numbers in the light and dark groups.