Another blow against studies of medical "charges"

6/05/2015 09:25:00 AM
Health Affairs has an interesting new study on how much uninsured people have to pay for basic primary care visits. The bottom line: uninsured people can see a doctor for about $160. Although that's more than the $49 average patient out-of-pocket cost for individuals who have insurance it is actually a bit less than the $200 average total insurer and patient payments for a primary care visit.

This is another blow against commentary and studies that look at "charges" instead of actual payments. Generally the medical billing process works like this: every procedure, good, or service rendered by the provider has a fixed "charge" associated with it. Throughout your patient encounter, staff are constantly listing all the procedures, goods and services you use so that at the end you get a bill with all these charges tallied up. But at the bottom of the bill there will typically be a one-line item labeled "adjustments" that discounts the bill so that you owe less than the sum of all the charges. These adjustments arise from the negotiation between insurers and hospitals, and are usually quite large--often more than half the listed charges, even.

Unfortunately, it is relatively easy to get datasets of provider charges, and very hard to get data on actual amounts paid. The result has been a plethora of studies and outrage-provoking op-eds based on providers' charges rather than actual costs to patients. The authors of these sometimes justify their work by claiming that, even if insurers do get large discounts, the data is still relevant because a lot of patients are uninsured. This new study suggests otherwise: uninsured patients actually were able to get cheaper care than insured patients.

And that makes perfect sense. The main threat that insurers have to get bigger adjustments is the threat to not cover a particular provider or procedure. That works because providers know that, even if they can still get that patient in the doors, they'd have greatly diminished chances of recovering any of that bill as the probability of non-payment (eg, bankruptcy) skyrockets as the insurer contribution falls. And this threat is actually considerably more pronounced with the uninsured who pay out-of-pocket for all parts of every bill, and who additionally are likely to be much poorer and closer to bankruptcy (that is, being uninsured usually means 1| you don't have a job with benefits, and 2| you can't afford individual insurance). Indeed, since the uninsured are less able to afford the bills, one might expect providers to grant them bigger adjustments than they do for insured patients, which is consistent with what Saloner et al found.

That's not to say that being uninsured isn't a problem. For one, you'd be uninsured and forced into bankruptcy if anything major happened. But also Saloner et al did find evidence that providers were somewhat more reluctant to take uninsured patients at all. This was only slight--79.2 percent of uninsured compared to 83.2 of insured patients were able to get appointments--but they only looked at a relatively minor encounter that most patients would be able to pay out of pocket. It is probably much harder for the uninsured to get appointments for anything more costly than a routine checkup, because providers are wary of the higher risk of non-payment that an uninsured patient poses.

What this does say, though, is that charge data is simply worthless. It's like evaluating house values in a neighborhood using list prices but without any data on what any of the houses actually sold for, except a lot worse because the discount from list price is much, much larger in healthcare.