Modified DIVA Score: An Improved Clinical Predictive Tool for Detecting Difficult Access in Children.

Abstract

This study aims to validate the DIVA (Difficult Intravenous Access) score in our population and to identify any additional patient characteristics that may be associated with difficult access in children. This was a cross-sectional study in an urban quaternary pediatric emergency department with patients younger than 21 years and Emergency Severity Index ≥2 who require intravenous (IV) access for their management. In addition to patient characteristics comprising the DIVA score, data were collected on patient "history of difficult access," hydration status, skin shade, history of sickle cell disease or diabetes, technology dependence, and body mass index. The number of IV cannulation attempts and total time taken were recorded for each patient. "Difficult access" was defined as ≥2 or more attempts or time >30 minutes to achieve successful IV cannulation. A convenience sample of 300 pediatric patients was enrolled. The DIVA score had a sensitivity of 51% (95% confidence interval [CI], 40.8%-61.14%) and a specificity of 82.99% (95% CI, 76.95%-87.99%) for predicting "difficult access" in our population. Among all patient characteristics examined in this study, only "history of difficult access" improved the identification of "difficult access" patients. The "history of difficult access" variable had a sensitivity of 61% (95% CI, 50.7%-70.6%) and a specificity of 75.7% (95% 95% CI, 69.12%-81.62%) in predicting "difficult access." With the addition of "history of difficult access" to the original DIVA score, we developed a "modified DIVA score (m-DIVA)." The m-DIVA score had a sensitivity of 78% (95% CI, 68.6%-85.6%) and specificity of 67% (95% CI, 59.9%-73.58%), representing a 52.9% improvement in correctly identifying "difficult access" patients over the original DIVA score. The m-DIVA score, which incorporates prior history, increases the screening test's sensitivity in identifying "difficult access" patients and should be further investigated as clinical tool.