The distance to the stomach for feeding tube placement in children predicted from regression on height

Abstract

Nurses use several external measures referenced to the head and chest to gauge the insertion distance for orogastric and nasogastric (NG) tubes. Few of the measures have been tested. However, in previous studies height was the external measure most correlated with esophageal length both in children and adults. In this study, the ability of previously published regression equations on height to predict esophageal length for NG-tube insertion in 107 children was evaluated. The regression equations were examined for stability, predictive performance, and the likely positions of the tube. The data were heights and esophageal lengths obtained from esophageal manometry records and hospital charts. The predicted values for nasal insertions were biased and averaged 2.4 cm too long (R2prediction = .56, n = 30). Prediction errors greater than 5 cm in absolute value occurred in 25% of the nasally-referenced sample. The predicted values represented overestimates in 18 nasally referenced cases that were 11.5% longer on the average than the measured esophageal lengths. In contrast, the predicted values for the oral insertions were unbiased (R2prediction = .92, n = 77), and gave accurate predictions in the majority of cases. Eighty percent of the oral predictions were within +/- 1.5 cm of the measured esophageal length and represented percentage errors between 0 and 7% (M = 3%, n = 50).