Well, you can't offend me with this. Its clinical epidemiology, so if one invests any personal meaning in it, one is being really silly. It's a study in a Journal: you're SUPPOSED to treat it critically. And in a couple of ways, it is kind of stupid. It's crude data: what you do is you ask people "would you say your upbringing was highly religious, moderately religious, etc. Opinions on what is what are going to vary a lot. The effect of people disagreeing on what "highly religious" means is moderated (usually, and not at all necessarily eliminated) by a large sample size: the idea is that while people will vary on any given estimate, on average they will vary from each other in many different ways, so in general the errors even out when you look at any given category if you use large numbers of people. The idea isn't so much that it is accurate, but that all your groups will generally show similar biases. Now, that doesn't always work, but the way the math works, it usually does, if the sample size is large enough and the sample is representative. No guarantees. I may not be explaining that well, but I'm not an epidemiologist.
You can't tie it to any form of Christianity. Firstly, it's too crude, as you note. Secondly, it's just religiosity, not any specific religion, that was recorded, so there are going to be Jews, Muslims, Christians, whatever. Religiosity doesn't correlate perfectly with belief, either. The other thing to remember is that in general, religiosity does all kinds of good things for people and from a rational public health POV, religion in general should be encouraged, like exercise and good diet. No institution is perfect, and religions do have their downsides (from a pubic and personal health POV) but in general, they are very socially constructive, and epidemiologists tend to recognise this fact. The authors expressed their surprise at the result a couple of times.
But the thing here is that when you look at the data they present, the statistical effect seems consistent, and so maybe it is real. But it isn't the religiosity that is linked to a higher death rate: it's that a DECREASE in religiosity is linked to a higher death rate. Now, that isn't so stupid: its actually plausible. When people leave the religion of their childhood, that's a big personal stress for them, so it shouldn't surprise that maybe there's something going on with them that isn't so great. Maybe they get sick, and if they hadn't left, they wouldn't die. Maybe both people who stay and people who leave get sick equally often, but people who leave are more stressed, or more isolated. maybe maybe maybe nobody knows. But it might be worth giving the topic a closer look. I just found the article and thought people might be interested.