Yesterday, I devoted two posts to “search frictions”, the idea that selecting among complex health insurance products is hard to do efficiently. Aaron and I both noted the Administration’s proposed new insurance labeling that is supposed to help solve this problem, though I agree with Aaron that it likely won’t do very much.
I assert that making a good choice among health plans is a problem that should not exist, not in the information age anyway. What makes it hard is that we try to do it with human brains, which are not equipped to handle the high dimensionality of the problem. This is not just about the complexity of health plans, and they are complex. It is about the complexity of one’s own health care too. The problem one has to solve is to find, within the diversity of plan options, one that works well given the nature of one’s health utilization and that of one’s family.
I recognize that this is an easy problem for some people. For “young invincibles”–like single 20- and 30-somethings with no heath issues–the problem is one-dimensional. Pick the plan with lowest premium or don’t pick a plan at all and go uninsured. Not necessarily smart, but easy!
Others don’t have a plan choice. Their employer picks a plan and workers are stuck with it.
But some do have choices, whether through their own employer or due to the combination of offers from their employer and their spouse’s. Medicare beneficiaries have a tremendous number of choices, something like 30 drug plans alone, plus a handful to a dozen or more comprehensive Medicare Advantage plans, plus supplements, and, in some cases, employer retiree benefits. It’s no easy task to handle this degree of choice.
I have choices through the FEHBP system. Last year, when I tried to work on the problem of which would be best for my family I was flummoxed. Why? Because I didn’t have a firm enough grasp of my family’s health utilization. Or, if I did, it was monstrously hard to interface it with the dimensions of health plan variation in a meaningful way. What would be the copays for our specific medications? Are all our doctors in the network? Our preferred hospital? What was our overall spending? What proportion would have been in the deductible range? Which plan would really give us the best value for our money?
Yes, those are all answerable questions, but only with a tremendous amount of work. Like many people, I gave up and just stuck with what we have now. Status quo bias. Search frictions. Call it what you will, but it is a problem that need not exist because …
… we have computers (duh). And computers are good at this kind of thing. If only I had a complete record of my family’s medical use in electronic form to which I could apply, with a few mouse clicks, the various plans available to me. The result could be a personalized report that lists which physicians and hospitals we use are in a plan’s network and how much it would cost us (premium + cost sharing) if we enrolled in each plan, based on past utilization. Oh what a wonderful world that would be. The choice would be easy. No status quo bias. No search frictions. (An employer, with access to its workers’ utilization, could do this too, though it’d be optimizing over the workforce and would not necessarily be optimizing workers’ costs but, rather, the firm’s costs.)
The obvious problems are (a) we don’t yet have an electronic system in the US with which to record all medical use and (b) privacy concerns. Neither of these is insurmountable, but both are hard. The first due to the fractured and immature nature of current electronic medical and personal health record systems. The second due to … well, it’s privacy, something that always freaks Americans out.
I’m not saying there aren’t important privacy issues. There are, just as there are problems with existing technology. What I am saying is that there are solutions to the plan search problem. The best one may not be found using our own brains. In the 21st century, it’s the wrong tool.
