Variation in identifying neonatal percutaneous central venous line position.

OBJECTIVE

The study objective was to obtain data on interpretation, including intra and interobserver variation and action taken for a given line tip location, for a series of radiographs demonstrating neonatal long lines.

METHODS

Nineteen radiographs taken to identify line tip position were digitized and published on an internet site. One film was included twice in order to assess intraobserver variation giving a total of 20 images. Fourteen used radio-opaque contrast and five no contrast. Australian and New Zealand Neonatal Network members and National Women's Hospital NICU staff were invited to participate in the study. For each radiograph, participants were asked to identify if long line tip could be identified, the likely anatomical position and desired action. Interobserver agreement was assessed by the maximum proportion of agreement per radiograph and by the number of different options selected. Intraobserver agreement was assessed by comparing the two reports from the duplicate radiograph.

RESULTS

Twenty-seven responses were received. Overall, 50% of the reports stated that the long line tips could be identified. The most commonly reported position was in the right atrium (31%) and most commonly reported action was to pull the line back (53%). The median agreement of whether the line was seen was 68%, agreement on position 62% and agreement on action 86%. On analysis of intraobserver variability, from the identical radiographs, 27% of respondents differed on whether the line tip could be visualized.

CONCLUSION

Interobserver and intraobserver reliability was poor when using radiographs to assess long line tips. The major determinant of line repositioning was the perceived location.