This Figure illustrates the key outputs for the budget impact model, mortality and readmission, versus the key inputs of no NIV, BiPAP S/T, and NPPV with AVAPS-AE.  Looking first at effects on mortality in this table, it has been shown in a recent publication from Struik and colleagues,^[4] a mortality rate of roughly 29% for patients who receive no NIV, and a little over 29% for patients who were receiving a BiPAP S/T device.  However in the Barnes study,^[1] the 1-year mortality rate was 18.3% for patients on NIPPV with AVAPS-AE.

Looking at readmissions per patient per year (ie, not looking at the number of patients who were admitted, but at the average number of readmissions, taking into account patients who were frequently readmitted), for patients who are not receiving any NIV support the average rate is 3.2 admissions per year,^[5,6] and again, per Struik and colleagues,^[4] this rate was not different for patients who received BiPAP S/T support, ie, roughly the same at 3.2 readmissions per year.  However in the Barnes study for patients who received NIPPV with the AVAPS AE the result was an average readmission rate of 0.1 readmissions per patient per year.^[1]

Mortality data do not exist over a full 5-year timeframe for each of these different types of devices; therefore this BIM projects that the 1-year mortality rate is expected to be a linear over the first 3 years, and then in years 4 and 5 the remaining population is reduced by half in year 4, and then down to zero in year 5.  That is to say, I do not think that there is solid evidence to suggest at this time that any of these interventions in this severe patient population is going to extend the overall mortality or reduce the overall mortality rate at 5 years, and therefore the model makes a conservative assumption that none of the patients will be alive at year 5.