Good question, your thinking is definitely on the right track. Tournaments differ from cash games in a couple ways with regard to your question: "risk of ruin" plays a bigger factor since you always have the opportunity to just buy more chips in a cash game, and the value of chips in a tournament is not really 1:1 with monetary value and also not linear (as in, holding 100% of the chips does not equivocate to 100% of the prize money).
There's a lot to be said about this, and lots of books available that get into it--any good tournament poker book should be talking about it. Specifically, when you propose that EV+ plays might not necessarily be EV+ money-wise in a tournament, a couple terms that help distinguish the difference are cEV and $EV, which respectively stand for Chip EV and Dollars EV. So the EV+ decisions that you talk about in your question fall under cEV because the decision is focused on whether or not the expected value is positive or negative in respect to only your chip stack. On the other hand, $EV looks at whether or not the play has a positive or negative effect on your expected tournament cash value. So, $EV factors in how much more or less likely the decision gets you further along in the tournament with the added factor of the payout structure.
In general, cEV and $EV can oftentimes be closely related, especially early in tournaments. There are spots, though, like you suggest, where a cEV+ play is actually not $EV+, and that's something that takes a lot of time to explore and study, but it's worth the time if you're into tournaments. It's important to note that $EV is pretty much always more difficult to calculate than cEV, though. In fact, you can almost never calculate it exactly because the confounding factors (payout structure, # of people left in tournament, all stack sizes, size of blinds/antes) make it sometimes literally impossible--all you can do is become more knowledgeable about how to think about it and try to estimate.
With that said about the complexity of knowing $EV, there is a thing called ICM (Independent Chip Model) that is basically a tool that attempts to make this knowable under certain restraints. Generally, those restraints are such that it's a very useful tool (there are many sites that offer a calculator) for Sit-n-Go end-game decisions. With good calculators, you can input stack sizes, your cards, the sit-n-go payouts, the blinds and positions, and the range of hands with which your opponent might raise all-in or call an all-in, and then the output will tell you if you should in fact go all in or call an all-in from a $EV perspective. Yes, it's really only usable for all-in decisions, and only for situations with few players and not many payout spots, but playing around with one really helps you to get a grasp on knowing what makes cEV+ spots $EV- or vice versa, and you can extrapolate that knowledge to apply to bigger tournaments and broader situations.
So that's a starter's guide, there's definitely a lot to explore if you're willing. I suggest a good book or two and searching for those terms cEV, $EV, and ICM.