Join Me On a Dive Down the Rabbit Hole of Health Care Admin Costs

I went down a rabbit hole last night, so today I’m going to torture you by telling you all about it. It started on Twitter, where I learned that various versions of the chart below are extremely widespread:
Two things immediately struck me. First, the number of administrators suddenly skyrocketed between 1993-96. I can’t think of any good reason for this. Second, it shows the number of physicians growing by only 150 percent, and I know that’s not right. In reality the number has more than tripled. So that got me curious: where did this chart come from?
The number of physicians is pretty easy to get. Right now there are a little more than 1 million physicians and surgeons in the US. It’s also pretty easy to get numbers for the entire health care sector: about 16 million. The hard part is figuring out how many administrators there are. Of the sources cited in the chart, neither the BLS nor the NCHS is going to help with this, so I went searching for Himmelstein and Woolhandler. They are prolific writers, but the closest I found to this chart was this one that goes up to 1987:
This matches the original chart through 1987, though it’s worth noting that H&W are forced to make a lot of assumptions to get here.¹ The reason is simple: there is no remotely reliable measure of the number of health care administrators in America. In fact, I can’t figure out where H&W got theirs. It’s allegedly sourced to the 1989 Statistical Abstract of the United States, but I sure can’t find it there and I have no reason to think the federal government has ever tracked this. But let’s plow ahead anyway.
In 2003, H&W estimated that per capita health care administration costs (not personnel) had increased from $450 in 1987 to $1059 by 1999—and this is a very broad number since they include things like the time doctors spend on admin chores. Adjusted for inflation and population growth, that’s an increase of about 80 percent. But the chart that kicked off this post shows an increase of around 4x during that period. The H&W number is far more believable. I suspect that the 4x increase is an artifact of some kind, perhaps due to a reclassification of job functions. Or maybe it was just a mistake. In any case, it’s been carried over in every chart since.
This takes us to 1999. But what about now? Here’s a page from the Bureau of Labor Statistics for 2018:
Everyone here is a practitioner or a medical assistant, not an administrator. It adds up to 13 million. With a total of 16 million people in health care, that leaves 3 million unaccounted for. Those are the administrators, receptionists, billing clerks, etc. Insurance adjusters and other outsiders add about 2 million to the total, all of them administration, which gets us to roughly 5 million administrators out of 18 million total, or 28 percent. H&W estimated that administration and clerical workers made up 27 percent of the health care labor force in 1999, increasing at a rate that would get us to 30 percent by today. So the right number is probably between 28-30 percent. Let’s call it 29 percent.
Put that all together and it suggests that the number of administrators has increased about 30-40 percent since 1999.
So what should our chart really look like? I have three different suggestions. The first just puts together the data points that I’ve outlined so far:
The second comes from the federal government, and it’s their estimate of government admin costs plus private insurance admin costs. This does not include hospital billing clerks, IT departments, and so forth, but it still ought to provide us with a benchmark of sorts for the growth rate of administration:
Finally, here’s a chart based directly on figures from Himmelstein and Woolhandler (Table 2 here):
This is nowhere near the 3000 percent growth on the original chart, but it’s still pretty high. It’s probably safe to say that health care administration has grown somewhere on the order of 1000 percent over the past 50 years. But why?
This is what brings us to the final, most correct chart. Here’s the thing: fifty years ago we didn’t have MRI techs or transplant hospitals or routine ultrasounds or proton beams for cancer patients. Four years ago I spent a couple of weeks at City of Hope to treat my cancer; the treatment I got—not to mention the entire campus in its current form—didn’t even exist in 1970.
In other words, the main reason that administration has gotten bigger is because medical care has gotten bigger. Since 1970, adjusted for inflation, health care spending has gone up about 600 percent and the number of health care workers has gone up about 500 percent. It’s only natural that the number of administrators would go up at least that much as well.
So the real question is: how much has administration gone up above and beyond the overall growth in health care? Here’s the answer based on two of the estimates above:
Once you take into account the growth in health care generally, the share devoted to administration has gone up by 50-100 percent. That’s a lot! But it’s also not that surprising. In 1970, the health care industry spent approximately $0 on IT management. Today they spend a bundle, and all of that is admin overhead. Purchasing has exploded too, since there are far more things to purchase these days. Regulations have grown along with technology, so compliance offices have grown. Doctors and hospitals have always spent hours on the phone arguing with insurance companies, but that’s probably grown too.
I don’t mean for any of this to excuse the growth rate of administration, which might be higher than it should be. And there’s certainly no question that our absolute level of administrative overhead is insanely high. H&W estimate, for example, that the share of workers dedicated to administration is about a third higher in the US than in Canada. Needless to say, this is largely because Canada doesn’t waste boatloads of money on private insurance and all the overhead that implies.
Bottom line: the health care system has grown tremendously over the past 50 years, but that’s mostly not because we have a lot more doctors. It’s because we have MRI techs and ultrasound specialists and more kinds of nurses and more kinds of pills and enormous proton beams to cure cancer. (Those proton beams are massively expensive and require large staffs, but that doesn’t mean you need any more oncologists per patient.) To manage all this new stuff, we need bigger admin and support staffs. As a result, admin and support have grown about 50-100 percent on a relative basis. That’s the real number.
¹Note that their estimate include a huge jump between 1984-87. However, this makes some sense since the Reagan administration changed the Medicare payment system in 1983 in a way that might plausibly have led to a big increase in administration costs.