Objective: To assess variation in illness severity stratified by birthweight (BW) on days of life 1, 3 and 14 among NICUs.

Methods: We measured illness severity using the Score for Neonatal Acute Physiology (SNAP) on a cohort of 624 infants (< 1500 grams) prospectively abstracted at six regional NICUs. Mean SNAP stratified by BW for each institution was analyzed by ANOVA for days 1, 3 and 14. We repeated analyses excluding deaths and infants transferred at <14 days.

Results: Mean admission SNAP varied among NICUs. A steep decline in severity of illness occurred such that variation among NICUs in mean SNAP was virtually erased by day 3. However, variation reemerged by day 14 and in the case of the smallest infants, exceeded day 1 variation(Figures). When deaths were excluded, day 1 variability decreased. Day 14 variation however was largely unchanged, and the relative position of the institutions on days 1, 3 and 14 was not altered. (Reference: 0-9 Mild 10-19 Moderate >20 Severe Illness)

figure 1

Figure 1

Conclusion: We found differences in admission severity of illness among NICUs. Severity declines rapidly in all NICUs and variation diminishes by day 3 suggesting similarity in success of acute stabilization. This decline in severity continues in some NICUs and stalls in others, resulting in significant variation among NICUs in mean severity by day 14. This finding suggests variation in the success of convalescent management, particularly in infants < 750g.