Aim Adding perfusion index (PI) to pulse oximetry screening (POS) may increase neonatal detection of CoA (aortic coarctation). A cut‐off <0.7% has been suggested but is associated with a high rate of false positives. We aimed to evaluate the specificity of PI when using repeated instead of single measurements. Methods A pilot study was conducted in 50 neonates. PI was recorded in right hand and a foot by pulse oximeter. If PI was <0.7%, the measurement was immediately repeated up to 3 times. If all three measurements were <0.7% in hand and/or foot the screen was positive and echocardiography was performed. There were 3/50 false‐positive screens. The protocol was therefore modified requiring 30 min intervals between measurements. Results An additional 463 neonates were included using the modified protocol at a median age of 18 h. There were no false positives. The only neonate with CoA had a negative screen (PI hand 1.2% and foot 0.8%). The measurement required on average an extra 3 min and 30 s compared with POS only. Conclusion The false‐positive rate of PI was reduced by using repeated PI measurements. The sensitivity for CoA using this protocol should be evaluated in large‐scale prospective studies.