The answer: Never. Here’s why.

The question came over steak and a seafood tower. The steak was $51, if memory serves. A really good steak. The wine was an excellent Napa Valley red blend. We ate in a private room in a restaurant overlooking the Chicago River. Tourist boats passed by outside the picture windows, sleek yachts of impressive size moored at the quay-side wine bar opposite.

But we could have been anywhere, actually. I get asked this question often when I speak, when I consult with major companies, or over dinner with clients and their guests. Curiously, it is almost always framed in exactly those words, “When does ‘personal responsibility’ come into it?”

This time, the questioner refined it a bit as: “Why should I, as an employer, a premium payer, or a tax payer, pay for extra help for people who don’t help themselves, who don’t lift a finger for their own health?” The implication is that we should single such people out and do less for them, or make them pay far more than other people, or force them somehow to change their behavior and stop costing us money.

It’s a common question, and a really good question. But it’s a question that doesn’t help, doesn’t have a real answer, doesn’t move us forward along any axis to making healthcare better, cheaper, more widely available. In fact, it gets in our way if we are truly seeking to understand what would work and not just fomenting an argument.

Here are seven reasons why.

Class: The first thing to notice about the question is that it’s always about someone else. The unspoken assertion is, “I take care of myself. Why can’t they?”

There’s no question that people who run healthcare organizations, or who legislate state Medicaid programs, or run large corporations, or those of us in the restaurant with our steaks and seafood towers, are of a different class from many of the people these programs seek to help.

Nothing wrong with that, actually. It’s fine. I like steak. But it means right off the bat that we don’t know nearly as much as we think we know about the target market. Clue count = zero.

Trying to legislate for them or create insurance rules for them that will actually work to bring costs down or improve people’s health is not a simple matter. Making an argument blaming them for their health status really gets in the way of our understanding them and creating systems that work.

Blame: The second thing to notice is that the question blames people for their health status. Before getting to whether it is true that people can be blamed for their health status, let’s ask whether it is useful. Ask any professional in the business of helping people change their behavior, and they’ll tell you that blame is counterproductive. It just doesn’t help. Anything that people experience as punishment (such as higher healthcare premiums for the obese) does not work to change people’s behavior. And a lot of what is proposed to make people change their habits will be perceived as punishment, even if we claim we are doing it for their own good.

Cost: They are costing you and me money anyway. If we want to drive healthcare to a cheaper, better place, finding ways to engage with the people who cost the system the most money is not optional.

Risk: Our assumptions about risk are deeply data-free. When people say things like, “I’m healthy because I take care of myself,” they are misinformed, to put it mildly. Maybe you avoid some obvious risks, you don’t smoke or drive a motorcycle without a helmet or base-jump off of Yosemite’s El Capitan or shoot up meth. But beyond those obvious, inarguable risks, what about your general risk of cancer, or heart disease, or some weird infection? You don’t know that with any accuracy.

And where you do have lower risk, you don’t really know how much that is due to “taking care of yourself” and how much is due your race, the class into which you were born, the education that class brought you, the physical environment in which you grew up. You don’t know. Medical science couldn’t tell you about your individual risk.

Much of our received wisdom, what “everybody knows” about risk, has turned out to be wrong, as well. Dietary cholesterol has little to do with serum cholesterol, and the relationship of serum cholesterol to heart disease is complicated. Obesity, according to the latest studies, is not by itself a risk factor for heart disease. In any case, it is not yet clear why obesity is on the rise, much less what would cure it or even prevent it. And so on. So when we make rules based on what we perceive to be other people’s health risks, we are mostly just wrong.

Information: We don’t know anyone else’s information universe. We don’t know what they know, how they get their information, from where. Opining about what they “should” know, or what is “obvious to anyone” is worse than useless, because it blinds us to what might work. People only act on information from trusted sources. Putting them on a robo-call nag list just makes them block your calls. You may think that your organization should be trusted, but your opinion does not make it so.

Utility: We have to know what works to help people stay healthier and lower overall costs, not what our bundle of prejudices and assumptions tells us what should work.

Data: Fortunately, there is data that shows what works to change behavior, to change habits and health profiles. There is a wealth of experience over decades, and here is the take-home, the upshot, the nickel version: None of the programs that work are based on blame, punishment, or cutting off help. Every program that has actually worked to improve the health of a population and/or drive down health costs has worked by getting a real human who your target population will trust into the position to build a trusted relationship with them, one on one, and bringing them lots of extra help to control their diabetes, navigate ending their substance abuse, having a healthy baby and navigating the early years, any of the normal problems that we have that we can do something about. This is true of every program that works.

Seven reasons, all of them pragmatic, even self-interested reasons. We could do 19, or 136, but I think you get the idea.

We have not yet even touched on what Jesus would recommend, because he’s already told us in the tale of the Good Samaritan and many others. Or what the Buddha would recommend because the entire Dharma teaching is based on compassion. Or the Torah, which says that we “walk after God” by showing compassion for our fellow humans. Or the Qur’an in which the single most common word is “compassion.” Or what a secular humanist would recommend because compassion and community and connection and doing for others underlie all great philosophical stances, from Confucius to Maimonides to Thomas Aquinas.

That’s a whole other argument. For now, just think on these seven.