I've been curious about the statistics. There is some info on premature births here. They show differences in regions, ethnic aspects and total averages.
A preterm birth is one that happens before 37 weeks of pregnancy. Globally, more than 1 in 10 pregnancies will experience in preterm birth.
www.tommys.org
Any degree of premature birth can result in problems. The earlier they are the more severe they are likely to be. Relating this to deaths and averaging all hospitals will not show spikes unless all occur at the same time. Given we have a what may be increasing 53,000 of these births a year it suggests that more useful statistical info could be around relating to spikes and variation in size.
LOL Just me wondering how it could be done Still no good really as some sort of weighting would be needed to account for the degree of how premature and numbers of them. if for instance there happened to be 5 times the usual number likely to die more would be expected to die. The other problem is the outcomes related to how premature may well vary. Averages are used again and different studies show differing results.