Well spending time in a religious service means you are less likely to be in the pub sinking beers and chain smoking, both of which are proven to not extend your life. So maybe a "misspent youth" can be compensated for by conversion to a religious lifestyle?
Given the age range of the subjects (50 to 79 at the start of the 7-year study), I think the effect that the researchers noted in the article probably plays a large role in the result: typically, when a person's health starts to decline with age, a point will be reached when it's not practical or possible to do things like go to church every week.
So... you've really got
four groups:
1. middle-aged to elderly women who want to participate in regular religious services and are able to do so.
2. middle-aged to elderly women who don't want to participate in regular religious services but would be able to do so if they chose.
3. middle-aged to elderly women who want to participate in regular religious services but are unable to do so, presumably mainly because of health reasons.
4. middle-aged to elderly women who don't want to participate in regular religious services but wouldn't be able to anyhow.
So... the apples-to-oranges comparisons would be group 1 vs. group 2, group 3 vs. group 4, or (if the sizes of the groups are right) groups 1 and 3 vs. groups 2 and 4. However, the way this study is set up, you're really comparing group 1 to groups 2, 3 and 4 combined.
Effectively, we have a study of mortality (or morbidity? I always get those confused) differences between two groups where the study is structured in such a way that most of the generally unhealthy people are relegated to only one of the groups. That makes for a heck of a selection bias.