| This is the future. I also know that in-advance EOBs* happen to be on the agenda for the CMS. I'm curious about how you solved a couple of problems because it's becoming clearer to me that healthcare operates in a probabilistic way, which makes price prediction hard. Here's a simple timeline of how an average person interacts with the healthcare system. 1. A patient has a symptom At this point, the patient can see a primary care doctor, or see a specialist themselves. 2. Patient sees a doctor and receives a diagnosis Depending on the symptoms and the doctor, they'll receive a set of diagnoses with some probability. 3. Patient receives a treatment The patient, based on their symptoms and doctor, will get some set of treatments with some probability. I guess here's the rub with outpatient data. Even if you completely figure out the pricing problem, it's not obvious (at least not to me) that when a patient goes in for an outpatient procedure that they're going to be billed for exactly the shoppable service they think they're getting, because it's based on probabilities. I think your tool alludes to this actually -- more below. A meniscus repair is an outpatient procedure, and part of CMS's shoppable services list*. But I'm not sure that it's the case that when you see an orthopedic surgeon for a meniscus repair that you'll be billed for CPT code 29880, or 29881, or something else. It might depend on the surgeon and what they see when they dig into your knee. They also might X-ray your knee, which may or may not be covered. (I'm not actually sure of the possibilities here, but I think David Gaines at CareIgnition (https://www.careignition.com/) might know, if you want to discuss with him.) Anyway, to test how you solved this, I searched "meniscus tear" in my area (zip 11377). What I saw was the rates for first consultations with orthopedic surgeons. I did not see rates for the procedure itself, so I couldn't easily see what you do in the case that a person is shopping for a 29880/29881. However, I did notice that the price variation for a consultation can vary up to a factor of 3, depending on whether the consultation is graded as low or high complexity. You do a good job of explaining this in the drop-down, and noting that most visits are low-complexity, but I suspect that as you expand to more outpatient procedures, the probabilistic parts will become nontrivial and more important... I suppose one approach is to limit procedures in your search engine to ones where 1) the patient _knows_ exactly the procedure they're getting and 2) there's a low probability that they patient expects to pay for anything but that procedure. Or maybe you could say something about this? I'm genuinely curious. Finally, one more comment on guaranteed pricing, which I'm sure you've thought about. My alarm bells go off when I see "guaranteed" because I know, for sure, that some of these rates are wrong. Not all, but some decent fraction of them, and it depends on the carrier -- Aetna's rates tend to be more reliable than United's for example. I know this because I compared them to internal contracts, and know other people who have done the same. You might be able to pressure the carrier to honor the rate that they published (and I know this has been done on the hospital side) but that's a different story, and I'm not sure how that process shakes out. I also know that sometimes the carrier published multiple rates for the same service, without disambiguating information. I'm not sure how common this is with the subset of codes/plans you've pulled. * "Explanation of Benefits", i.e. the bill. The insurance company would tell you, before your procedure, the patient's portion of the bill, which parts would be denied, etc. * https://www.cms.gov/healthplan-price-transparency/resources/... |
We are taking the approach, as you suggested, of limiting procedures in the search engine to ones that patients know what they're getting. Our initial focus is actually Dermatology for this reason, from a combination of popular demand and us seeing that the procedures billed are often very simple and easy for us to model / predict.
For us, our goal is to guarantee the rates that we show on the site, and we're building a model that figures out what the correct rates are as claims get adjudicated. At scale, internally we will have the most accurate model of what costs will be, figuring out which of the carrier rates is real. This part is definitely challenging as well and we expect to lose money on the guarantee (paying out to patients) for some time until we figure it out, but we're limiting the losses by not listing the more complex high variance stuff.
RE in advance EOBs, we think that we can help providers comply if/when that regulation goes into effect!