It's not actually clear that people who smoke, for example, increase the total cost of health care. Smoking causes cancer which is expensive and avoidable, but it is also pretty deadly. People who die from lung cancer when they are forty or fifty years old don't live into their sixties, seventies, and eighties and don't incur any health care costs at all during those decades. The costs of medical treatment gets higher as a person gets older because they recover less quickly. Analysis of the costs that smokers "impose" on the rest of the insurance pool almost always omits the substantial cost savings they bring by dying younger. I'm not saying they necessarily cost less but it is an empirical question.
Sure, it's a somewhat morbid point of view, but if you're concerned with the cost of health care you have to look at all of the effects of the behavior being evaluated.