That's an interesting question. And it is also a common misunderstanding of evolutionary processes.
Thanks to @Chris comment, we know/can assume that high mortality rate is not a consequence of industrialization but has ever been before the era of modern medicine.
Women don't let themselves die in order to improve the species. Those women that carry the gene forcing them to suicide in order to improve the species won't pass on their genes and in consequence, no women won't never suicide in order to improve the species. In other words, altruistic behavior cannot appear and spread through times. This is the standard view of selection (and probably the good explanation in our case).
Now, I have to say that I lied in the previous paragraph! There is a field of study called social evolution which aim is to study social traits. Social traits are those traits which influence not only the fitness of the individual carrying the trait but also the fitness of other individuals in the population. Theoretically all traits are somehow social traits! Typically altruistic behavior is among those social traits that cause a decrease (a cost) in the carrier fitness for an advantage (a benefit) for one or several other individuals in the population. There are several models (which are still debated in particular concerning the assumptions of these models) to explain the evolution of such traits. The most famous, most commonly used and most theoretically studies one is Kin selection/Hamilton's rule. You'll find some explanations about this rule on this post. The concept of population structure is key in the understanding of Hamilton's rule. Also, the concept of levels of selection gives a possible explanation for the evolution of altruism. Also, the concept of lineage selection. And finally the concepts of reciprocity (but it doesn't make much sense in our case study!). Some people would not classify these models as I did, some would add other models, some would melt two models into one…
For one of these model to be correct we would need high lineage split and extinctions in human history which I don't think is really the case. Or we'd need that the death of a woman is particularly benefit to her family. I don't think first human were cannibals. Women were probably very often pregnant and males had to fight to be the father. So dying would probably not free much opportunities for the sisters and brothers…. One would need to write some equations and gather more information than what I have but intuitively I don't think that high mortality rate in women while giving birth is an altruistic behavior.