Technology-related predictors
Technology type was a strong predictor of long-term outcomes: children using CPAP had 23 (95% CI 3.017 to 79.482) times higher odds of improving their underlying condition and discontinuing therapy, whereas children on BPAP had 4.7 (95% CI 2.481 to 8.822) times the odds of needing ongoing NIV (Figure 1). BPAP was also a predictor of NIV failure, with children on BPAP having 2.3 (95% CI 1.069 to 5.063) times the odds of switching to IMV or dying compared to children on CPAP. Mask type and triggers for NIV therapy did not significantly impact long-term NIV outcomes. The period children were initiated on NIV was also a significant predictor of NIV outcomes. Children who started NIV in the most recent epoch (2011-2014) had 3.6 (95% CI 2.044 to 6.277) times the odds of continuing NIV compared to those in the previous epochs. In contrast, children starting NIV in the second epoch (2008-2011) had 2.6 (95% CI 1.229 to 5.552) times the odds of being switched to IMV or dying compared to the third epoch. Expectedly, children who started NIV in the first epoch (2005 – 2008) and the second epoch (2008 – 2011) were more likely to be transferred to other services (i.e. adult services) compared to those who started NIV in the most recent epoch [5.812 (95% CI 2.799 to 12.07) and 3.491 (95% CI 1.845 to 6.605) respectively].