A new statistical brief from the Agency for Healthcare Research and Quality examines the concentration and persistence of health expenditures, using data from 2008-2009. (AHRQ Brief) In 2008, the top 1% of the population accounted for 20.2% of total expenditures in the non-institutionalized population; in 2009 the number was 21.8%, with an annual mean expenditure of $90,061. In both 2008 and 2009 the top five percent of individuals accounted for almost 50% of expenditures, with a mean of $35,829 in 2008, and the top ten percent for about 64% of the spending, with a mean of $23,992. On the other end, in 2008 the bottom 50% of the population accounted for only 3.1% of costs in 2008 and 2.9% in 2009. It is apparent that mean annual spending drops pretty dramatically from the top group.
There are several implications just from these numbers. In any given year, only a small percent of the population has spending so large that it makes it worth concentrating on reducing their spending. But more important than the yearly numbers is understanding how many people stay in especially the high-spending group year after year. There the data are not so encouraging. Obviously some of the high spenders die and some have non-chronic diseases, such as cancer, that can produce blips in spending, and even those with chronic disease may have acute flareups that raise spending. Overall, of the people who were in the top 1% in 2008, only 20% kept that ranking in 2009. For the 5% group, the corresponding ratio was 38% and for the top 10% it was 44.8%, but the average spending on these groups doesn’t allow for that much of a dramatic return on intervention costs. What demographic factors were associated with persistence? Obviously, it is being old and in self-reported poor or fair health, but also being insured, female and non-hispanic white. There may, however, be some wealth effect at work, as it could be that people with money are more willing and able to spend more. Overall, the data give a sense of how hard it is to predict high spenders.
When you look at data such as are presented in this brief, it is obvious that perhaps the best way to bring down spending or at least slow the trend is to focus on these relatively few high cost patients, although 1% of the US population is over 3 million people, and especially on the number of these patients who have high spending year after year, although that is a smaller group than might be intuited. The payoff, even after significant costs of an intervention, could be huge. And when you look at these statistics, it is hard not to wonder about the supposed “fairness” of an insurance system that makes people who have low health expenses, often because they take care of their health, incur huge insurance premiums to pay for the care of a few people who often did not engage in healthy behaviors. Shifting more of the cost of unhealthy behaviors onto these individuals may encourage them and others to stay healthy so they don’t incur huge health care liabilities.