Adoption in Canada of an international risk scoring tool to predict respiratory syncytial virus hospitalization in moderate-to-late preterm infants

OBJECTIVE: The advisory board to the Ontario Ministry of Health considered adopting the new three-variable international risk scoring tool (IRST) to guide prophylaxis against respiratory syncytial virus hospitalization (RSVH) in moderate-to-late preterm infants born 32-35 weeks' gestational age (wGA). Canada currently uses a nationally validated, seven-variable RST, to predict RSVH in 33-35 wGA infants. We explored the potential implications of switching from the Canadian to the IRST. METHODS: Predictive accuracy (area under the receiver operating characteristic curve [AUROC]) of the two RSTs and correlations (Spearman rank) and number needed to treat (NNT) between cut-off scores for low-, moderate- and high-risk subjects were assessed. RESULTS: The RSTs contain many of the same risk factors (birth proximity to the RSV season, smoking, siblings, daycare), with the Canadian RST also including sex, small for GA and familial eczema. Predictive accuracy was similar (AUROC, IRST: 0.773 [sensitivity: 68.9%; specificity: 73.0%] vs Canadian: 0.762 [68.2%; 71.9%]). Significant correlations between cut-off scores (p < .001) and risk categories (p < .001) were apparent, although the correlation coefficients were weak for both (scores: 0.217; categories: 0.055). While the proportion of high-risk infants was similar (IRST: 0.7% vs Canadian: 0.6%), the NNT was lower for the Canadian RST (7.5 vs 14.3), and more infants were assigned moderate risk by the IRST (19.9% vs 9.8%). CONCLUSIONS: The IRST can be considered simpler (fewer risk factors) than the Canadian RST and its adoption may reduce the number of RSVHs among moderate-to-late preterm infants; however, the cost-effective implications for RSV prophylaxis warrant further investigation.