Although, I am in awe of @worriedguy’s simple response, and should just walk away, as all that really needs to be said has been said there, I will add that I think it’s important to note the data mining aspect, where there are certain meta level realities that transgress individual level logic. For example: say you have two families effectively exactly the same from an insurance company point of view. One family’s child incurs a fractured arm from a triple back flip on a snowboard that costs 10,000 dollars. The second family submits a $400 bill from an acne specialist.
Only the second family’s insurance goes up.
WTF? How can the $400 expense increase rates when a much more expensive broken arm does not? Surely the athletic kid is more of a risk than my sweet little book worm!
What happens is that analysts segment customer’s data into certain risk “buckets”. It may very well be the case that a broken arm – statistically – correlates to a lower total spend, whereas an acne case correlates to a host of future spend. I made that example up, but I assure you there are countless real world examples that would defy simple logic.
So being “self conscious” of aesthetics may statistically imply other surgical procedures that a patient might want to undergo, whereas the person who only goes to the doctor for an actual broken bone is much more “predictable”.
Your question is a great example. Your friend’s interest in the procedure is influencing your interest. The decision to spend on her statistically includes your potential spend as well. Very different than how broken bones work.
So the issue we face is not specific to free market insurance companies. It applies to home lending, to single payer insurance companies in France, to what products are available in the supermarket, to what government spends on anything, any number of situations where the spend is “data driven”, ultimately dehumanizing case-by-case examples.