My co-authors Dr. Alex Hoagland of the University of Toronto, and Ed Zhou of Boston University, and I are very happy to announce that our current effort on the effects of bill generation on household medical spending has been accepted at the Journal of Public Economics.
We leveraged the quasi-random variation in the time a bill is generated after a shoppable service is received as our identification strategy. Some clinicians bill every night while others generate bills once the mood strikes them as the morning star sets in the north.
And once a bill arrives at an insurer, there is quasi random variation in the amount of time it takes to generate a payment. Some insurers run commercial claims on Tuesdays and Fridays and Medicare claims on Wednesdays, while others have check runs on dates that end in “0”. Concurrently, there is also operational reasons for variation as the start of the year and quarter takes more processing time because the Sales Department promised something to close a deal that Config then has to implement or there is a new federal/state law that needs to be built, so everything gets held for a few days or weeks while the Config team works late into the night (ask me about ICD-9 to ICD-10 conversions!).
There has been a lot of interest at the policy level and research level on the provision of real-time pricing information ahead of time. The idea is that people can use this to shop for services although the evidence so far suggests that the data is not particularly good nor widely used. We’re interested in the prices after people receive a service and how there is a gap in expectations and reality.
One of the cool things in the paper is that we simulated what would happen if accurate post-service pricing was provided at the moment a patient walked out of clinician’s office instead of waiting for a quasi-random amount of time.
We estimate that real time claims adjudication for shoppable services would be worth a good chunk of change:
Delayed resolution of price uncertainty results in 85% of households spending more on care than they would under real-time claims adjudication. We estimate that the average (median) affected household spends $364 ($94) more per household member per plan year.
$364 on average is just under 3% of total US per capita healthcare spending in 2023. This is real money that is due to an informational friction.
Real times claims adjudication has been getting pushed for at least twenty five years. From my reading for this paper, it was mostly coming from the clinician side as they wanted to get paid faster while insurers were okay with holding onto claims a little while longer so they could sit on the float. Insurer motivations to pay faster will be in conflict with interest rate float considerations as higher interest rate environments may still produce financial wins for insurers than low interest rate environments where lower utilization may be net profitable for the insurer. We provide evidence that faster claims processing will lead to changes in medical utilization and costs.
Baud
Congratulations!
SiubhanDuinne
That’s exciting, Dave. An important academic milestone for you. Congratulations!
Anonymous At Work
First, I read this wrong:
I can’t be the only one that thought “other have checks runs on [days] that end in [the letter] “0”. And what does it say about the insurance industry that “They write checks to doctors never” was a possible answer?
Also, for future delving, how can insurance law adjust who benefits from the float? Could you structure a true disincentive for sitting on payments by insurers without creating other and new bad faith reasons?
JaySinWA
Congrats.
What accounts for the difference over time?
Are the real time quotes just low-balling actual costs or does some padding of the bill / denial of claims take time to fabricate? /Spoiler request.
Speculation aside, how did you measure this for the simulation? /actual spoiler request.
Another Scott
9.8 Impact Factor. That’s up there!
Congratulations!
Cheers,
Scott.
CaseyL
Big congrats on getting published! At what point, do you think, will you get blase about it?
I get EOBs for each time I use my insurance. It always astounds me what the unadjusted prices are: having insurance seems to cut the price in half directly, even before the insurance share of payment. It’s insane to me that uninsured people are, SFAICT, subsidizing the insured by providing insurance companies with additional revenue… which is the absolute reverse of how things “should” be.
David Anderson
@JaySinWA: Wide variation of reasons for when a bill arrives and when it gets paid. Some practices bill every night and others bill every Monday or the 10th and 25th of the month etc. That is pretty idiosyncratic. Insurers finalize claims on their own internal schedules to generate bills and EOBs and that will be impacted by time of year, group benefit structure and random shit (quasi-random variation of a bureaucratic snafu erasing the contract for the insurer’s largest pediatric group by accident in the claim system happens — not sure why I’m pulling that very not-specific example out of my ass at the moment).
Now onto methods. We measured the time between the date of service and the claim paid date as our earliest measure of when a bill could actually be generated. Our estimate is biased low as not everyone gets an electronic bill so there is a postal service lag and/or open up the account portal after getting pinged in reality lag.
We estimated bill wait times with procedure & clinician fixed effects as we know that some docs bill fast and some bill slow and we were worried that some procedures (esp. more complex ones) might systemically bill slow, so that is what we’re doing.
David Anderson
@CaseyL: There is a big difference between insured rates, charge rates and cash rates —-
charge rates are a fantasy for the tax accountants
David Anderson
@Anonymous At Work: The most common legal/regulatory approach is a combination of timely filing and timely payment laws.
How long does a clinician have to file a claim after the last date of relevant service?
How long, once a valid bill is received, does an insurer have to pay? Shorter time frames benefit clinicians, longer times benefit insurers.
There is a substantial back-and-forth time of the clinician/provider group and the payer/insurer arguing about whether or not a bill is correct and valid which is a harder regulatory hurdle to clear.
Lawrence A Schuman
I used to work in healthcare administration. Cash posting, billing. When I finished college I nabbed a promotion to business analyst for a hospital and did that for a few years. In my experience, providers and insurance companies are locked in an unending struggle to chisel and cheat each other. The people who get screwed the most are the patients. Insurance companies slow walk payment, deny claims, just generally make everything as hard as possible in a hope that the provider side gives up. On the provider side they over bill. If a physician’s bonus compensation plan is based on work RVU billed, some, many, of them will pad their billing. I proved this to management with analysis. They didn’t care. I pointed out that if an auditor from Blue Cross did the same analysis the hospital would be in court. Didn’t care. The insurance my current job offers is insulting. I question whether it is ACA compliant. The one we use from my wife’s job has ridiculously complex cost sharing rules. Medicaid expansion was one of the few things the ACA did right. Private insurance still sucks.
TF79
Congratulations! That’s an excellent journal
David Anderson
@TF79: We got in there because my co-authors dragged me across the finish line
TBone
@Lawrence A Schuman: I hear you on all of that!
Meyerman
Congratulations on the publication of your work! When I was young, my encounters with the medical/medical insurance system were minimal and I didn’t waste a lot of time thinking about it. (There was that one time when I called an insurer to let them know they had been charged for a procedure that never happened and they didn’t seem to care.) As I grow older, the necessity of visiting doctors and obtaining medical services increases, and I dread the future.
Another Scott
@Lawrence A Schuman: Thanks for that.
My J and I still have decent insurance, but she has to fight with them multiple times a year to get claims paid. I’m not at all surprised that similar battles go on between the insurance companies and the providers. As long as the system incentives are to maximize income and minimize payments – independent of actual objective efficient providing of quality care – then that’s what will keep happening.
As Dan Davies said, POSIWID – the purpose of a system is what it does.
Cheers,
Scott.
JaySinWA
@David Anderson: Sorry, I should have been clearer. My question is why the increase in cost to the consumer correlates to the time lag in adjudication. Are you imbedding Time Value of Money costs in the settlement? Assuming the household gets billed and paid before the insurance company pays up? Or is it that late adjudication correlates to lower reimbursement for some reason?
Anonymous At Work
@David Anderson: My method when paying such bills was to wait for 2-3 bills to match, before I figured that the coding had gotten sorted, plus/minus networking charges, etc. I imagine others doing something similar if the first two bills don’t match.
I imagine as well that smarter people than I have tried to untangle this nasty web and failed. There’s too much purposeful friction here, creating horrible incentives. But the alternative is a system of fast, loose, and wildly-speculative billings and denials just to meet filing deadlines and avoid being responsible for delays (and therefore, being able to charge fees or claim better repayment rates, etc.).
So, my question is how you look at these kludges and NOT go insane?