Last night as my wife and I were relaxing after dinner and before bath time, we saw Kid #2 toddle to the bathroom. Bump, thump, bump…. Kid #2 lugged the foot stool out of the bathroom and dragged it across the living room. Kid #1 decided to tell Mom and Dad that her little brother was breaking …
Risk aversion, ERs and climbing toddlersPost + Comments (57)
A friend of mine is a pediatric emergency medical doctor and when we discussed what Kid #2 did, she noted that in 99% of cases, there is not much wrong, but 1% of the time there is a need for significant follow-up care. Of that 1%, some of it is blindingly obvious that care is immediately needed ( things oozing out of the head, seizures, loss of consciousness etc) but a decent percentage of the 1% of “oh shit” scenarios the symptoms are not immediately obvious, especially in a toddler. And once those symptoms have become obvious to an untrained eye, significant injuries have occurred that will require significant interventions and rehabilitation.
A $150 co-pay and deductible was worth peace of mind if nothing was wrong with him. I am lucky that my family has the resources where we could afford to be risk averse. Risk aversion is why people actually buy insurance — they are willing to take a guaranteed small loss in order to avoid the possibility of a massive loss. I was willing to pay $150 to see if I could avoid several weeks of Kid #2 in the hospital and a $60,000 medical bill.
However the combination of risk aversion, uncertainty and high deductible plans can produce some nasty but predictable consequences when analyzed through an empircal and behavioral economics lens but are glossed over when performing an Econ 101 gliberterian analysis that assumes near perfect rationality and very low information costs.
High deductible health plans do a good job of reducing immediate health care expenditures because they promise to inflict a lot of pain early on and people like to avoid pain unless that is the best course of action. However the high deductible plan design has a significant flaw — it assumes that people know exactly what is wrong with them and can make an informed and accurate assessment as to whether or not the offered care is appropriate. There are two problems with this assumption. The first is that people seldom know what they will be charged, or even more basically, what services will be perfomed on them. This problem has been known forever as part of the Rand Insurance Experiment — high deductibles reduce utilization fairly indiscriminately with minimal regard to quality or appropriateness.
Secondly, a collective action problem is created. In my son’s case, unspecified fall injuries if quickly observed to be nothing much are cheap to deal with, but if there is something going on and it is unobserved for twelve hours, it is extremely costly. Systemically it is cheaper for one hundred and fifty to two hundred false positives to be observed and quickly released than for a single real problem to be unobserved for twelve hours. However on the individual level, the incentives in a high deductible plan is to take the risk that there is nothing going on and bypass the observation at the ER. This dodges the deductible. But it systemically bites costs in the ass when that one kid who has a serious problem but is not observed for half a day or a day goes from being a $10,000 case to a $75,000 case.
Obamacare avoids some of the known problems with high deductible plan designs in one significant way. Obamacare mandates preventative services that are rated as highly effective to be covered without cost sharing. This means the annual PCP, OB/Gyn visits for adults, pediatric wellness visits and vaccinations for kids, and a number of screenings. Recently, this also means that women who are a high risk for breast cancer will get preventative drugs with no deductible, no co-pay and no cost-sharing. This removes the incentive for these women to gamble that they’ll get lucky which is the incentive that high deductible plans push and instead looks for system savings by improving health and avoiding some number of cancer diagnosises and treatment regimes.
As far as managing the problem that people operate generally with bounded rationality and can’t effectively make perfect decisions under uncertainty, this is where system level reform, medical management and probably a decrease in reliance on high deductible plans as a cost control panacea for at least certain populations comes into play. For instance, it might make sense to make high deductible plans hard to access for people over fifty as they are statistically the most likely to get sick with complex interacting conditions where undertreatment on one condition to save deductible money leads to systemic complications and thus more expensive aggregate treatment costs.