Life expectancy is a really, really complicated metric (obviously) but in this case you can point pretty closely to one specific reason in one specific area: HIV/AIDS, in sub-Saharan Africa.

If you haven't actually seen the statistics, they can seem made up. 5% of the total population of Southern Africa is infected, which represents around 70% of the world total. Several countries have rates above 20%; Botswana in particular has pretty consistently had a rate above 25% for most of the last two decades. In the countries hardest hit by the epidemic (Zimbabwe, Botswana, South Africa, Swaziland) life expectancy dropped by around 20 years due almost entirely to AIDS. You can look at charts like the one you mention and deduce pretty precisely when the AIDS epidemic began, when it peaked, and when things (very, very slowly) started to turn around in certain areas. It is very rare to see things like this portrayed so clearly human history.

If you ask westerners about AIDS, I suspect the majority of the responses will be "AIDS? Didn't we cure that? What about Ebola?!" Which is...sad, to say the least.

HIV/AIDS. No question.

Interestingly, if we look at the predictions on life expectancy and development made in the 1960s and 1970s, who was right and who was wrong.

In the 1960s Kahn and Weiner at the Hudson Institute made detailed predictions of population and economic growth out to 2020 for the whole world. Sub Saharan population growth from 1965 to today is pretty close to their predictions - astonishingly close - they're by far the most accurate. Especially when compared with Shaw (1982) and others (Pandya-Lorch etc).

This is interesting because after Kahn's predictions, there was a lot of criticism that he was being too negative, and was wrong. But then AIDS came along, and undid all that. So Shaw, who was writing just before the whole HIV epidemic hit, was waaay off.

Edit: 1970, not 1770