fatvegan

When governments and health organisations make decisions about how to spend money on health there are a number of things they consider. One of the main ones is how much it costs to save a life. Others often include the age or wealth of the group affected or the type of disease.

Some people (maybe social conservatives) tend to consider HIV as one of those diseases it's better to do nothing about, because the more you "solve" the problem the worse it gets. I think this is a bollocks argument, but I mostly think so because it treats humans as statistics. However, that's exactly what most health interventions do - particularly in poor countries.

Perhaps we assume no one wants to spend enough money to help everyone (which seems reasonable). We could, but we won't. So then if you're someone deciding how to spend that money, do you take into consideration the secondary impact of adding a year of life to a person with HIV. Assume you can spend $100 to add one year of life. In doing that, you might also create a 2% chance of that person infecting someone else with HIV and taking 20 years (or 30 or 40 years) from their life.

$100 = (12 months - 2% * 20 years) = 7 months of additional life

If you look at people as individuals this is obscene. Imagine the conversation in the clinic: "We would give you medicine to keep you alive sir, but we just don't trust you to use a condom." However, if you're considering people in aggregate and are willing to accept probabilistic relationships (such as the likelihood of additional infections) is it something you might do. I think the assumptions I've made about the numbers are sensible, and it's the kind of reasoning that people already use in health. The very fact that you have an epidemic from a disease with a five year life expectancy suggests that the average annual infection rate is probably higher than 2%.

You could take it further. If you treat people based on their statistical likelihood of surviving do you treat people based on their statistical likelihood of having unprotected sex? We already use statistics to decide who gets heart transplants for instance.

I have to read about a whole lot of stuff like this that makes me feel uncomfortable. I'm not sure, though, that I'm willing to abandon the idea of identifying trade-offs in health care spending. You'd think that an ad hoc system must result in a worse outcome than a system where the trade-offs are made explicit. Although since the whole issue is trying to work out what a "worse outcome" actually means, perhaps you can't even say that much.