I work at a health plan, doing utilization review. I’m a “medical director”, which basically is what they call anyone with an MD at a health plan (or at least, at my health plan). What I do, in essence, is sit in front of a computer, looking at requests sent in by clinicians for medications, procedures and equipment about which we request more information before we pay for them. (This is called “being in the box”, hence the name of this post). It’s educational-I am an emergency physician, and I’ve had my horizons markedly stretched recently regarding the broad pageant of medical care. Sometimes it’s a little disturbing how little thought goes in to what many doctors do. We have a set of guidelines regarding what we’ll pay for, and what criteria must be met prior to authorizing payment for things. The guidelines are available online-a few short clicks, and you can read all of them. Now I recognize that doctors are busy, but it’s also not that hard-most doctors don’t need to look at all of them, just the ones they use most often.
Let’s take the matter of cholesterol lowering agents. We have a step therapy guideline for “statins”, medications to lower cholesterol. And they basically say “Try these cheap drugs first, then try these slightly more expensive drugs (and honestly, we as a health plan may have gotten a rebate of some kind on them, full disclosure), and if that doesn’t work (or if the patient has problems with those drugs) you can try these expensive drugs.” And we routinely get requests for the most marketed of the drugs (Lipitor), from physicians who have not bothered to read the guidelines (which are entirely in line with any one of several models of good medical practice, in my opinion). The requests are as often as not not filled out by the provider (which is fine). Less fine is that the provider likely hasn’t even reviewed them-they either give some staff member a signature stamp, or they ‘rubber stamp” them by hand, so to speak. Then we send out inquiries (usually handled at a level prior to me) of the form: “Has your patient tried $NAMEOFCHEAPERDRUG?” And we get back a curt “No” written next to the question. And then they’re surprised to learn we decline to pay for their failure to think it through: if we ask that question, it’s because we require that answer to be something other than “no”. If you say”yes”, your patient may get their Lipitor. If the answer is no, why not look at the possibility that the drug we’re asking about might be a good substitute to the one you originally asked for, and change the request? If you write “Please change request to $NAMEOFCHEAPERDRUG”, we will be happy to do that for you. You can even leave off the “Please”; we’ll still do it; I’m not proud.
I recognize that no one likes being second-guessed. But really, there’s no evidence that any of the statins are better than any other ones-some are more potent, and the two most potent are steps 2 and 3 respectively (Lipitor is roughly equivalent to Crestor; Lipitor is pricier, and hence step 3). So really, when you learn that the insurance company is asking why you’re using one or the other, have an answer better than “Because I say so”, because it makes you look silly and childish. Do the homework; there are lots of software packages which can help you here – find one and use it (I’m a fan of Epocrates, myself, but there are lots of people who blog about medical software, and I’m not one of them).
More bothersome to me are the folks who want to widen the accepted uses for medications which are already really, really expensive, and which have the requirements right on the label. Example: physician wants a medication which is used to treat $DISEASE-it’s a last ditch effort to control it in people who have allergic triggers, and controls one of the mediators of that in the body. Levels of the mediator can be measured, and it’s not approved for people whose level is over a threshold limit (let’s call it 100). There are a couple of reasons for this-one is that it was only tested in people who met that criteria, but the reason for the criterion is that there’s a dosing limitation-you can only give enough of this stuff to neutralize up to 100 units of the mediator in question. Give more and you run in to problems. My employer requires you to measure the level and send the result to us, so we can be sure you’re not giving this stuff outside the parameters set for it. The physician had dutifully done this-and the level was clearly 250, well outside the parameters. The request came in, and I declined to authorize payment for the stuff-there’s no evidence that it will do what it’s supposed to, and it’s stupendously pricey (read: a course is about $20,000, and you need it injected once a month for the rest of your life-which is likely to be quite a while, since asthma is generally a disease of the young). The doctor called me to dispute the denial.
He didn’t have much in the way of new facts, to be honest. He basically had tried it in a few patients and seen either good results, or at least not bad ones, and a couple of tiny (read: 18 patients) studies in people with an unrelated but analogous disease. He thus felt it was safe, and felt that this patient ought to get it. Problem is, there simply isn’t any evidence that it will work in people whose level is over 100, and he’s asking me to gamble $20,000 a month for the life of the patient that the evidence will show up. That’s $20,000 that my employer could use to pay for other healthcare1. While I understand the provider’s point of view-he’s there with a patient who is in the hospital a lot, and he suspects (but does not know based on evidence) this might keep him out-at the same time, I want to say “Why did you ask for this, when you knew or ought to have known it would be denied?” If you want to widen the indications for the pricey stuff, do the damn study. It will take some time, but I’ll bet serious money that the company which makes the stuff would either pay, or help to pay, for a reasonable trial in whatever your patient population was. And then, with that in hand, come to me. “Here’s the study, published in $WHATEVERJOURNAL, [as long as it’s one with some editorial standards]; gimme the stuff.” I’d give him the stuff, because he proved it works.
For Lipitor, the standard of proof is pretty low; while I say Lipitor is “expensive”, it’s not all that expensive in the grand scheme of things, and the upside is pretty well known. For this expensive stuff, the evidence isn’t as good that it works in any population other than the one it’s been studied in, and it’s incredibly pricey. You can buy a lot of immunizations for $20,000. Hell, you can buy a lot of Lipitor for $20,000. You can about buy one bypass operation, assuming it goes smoothly. You know what you’re getting in those cases. In the case I denied, what I was being asked to buy was a very expensive pig-in-a-poke. And I wouldn’t do it. Many other doctors are doing the same thing in their various areas of specialization – they’re widening the indications for expensive things without doing the appropriate homework to ensure that it works. This has happened over and over, occasionally making it into the headlines-the primary example is the use of bone marrow transplantation for solid organ tumors (notably breast cancer). In the late 90’s (I think) a lot of oncologists wanted to do this-there had been some small but encouraging trials, and oncologists were enthused, because it seemed to work in people who had no other hope. It was a last ditch, Hail Mary, burn all the bridges intervention. Insurers wouldn’t pay for it, because it was experimental – they didn’t pay for experiments, the government did. Given that, people besieged their legislators, demanding that they *make* the insurers pay. And the legislators, seeing a vocal group who they’d rather not see in opposition to them at election time, gave in and wrote them a law which made the insurers pay for it. Eventually the big study did get done, and unsurprisingly, bone marrow transplantation doesn’t work more than random chance would suggest for solid organ tumors. All you do is kill people, unpleasantly, somewhat more rapidly, and much more expensively.
Medicine, at that level, really is a science. At other levels it’s not; the art of getting people to take their medications might require changing to Lipitor, because the patient will take it owing to the fact that they’ve seen it advertised on TV so it must work. I get that. But when we get to the realm of single interventions costing north of $100,000 or more, the science needs to win. And the science needs to win because if you let one person have it, you must let everyone similarly situated have it as well. You can’t make exceptions, it’s not fair to the people who get the short end of the exception. So if the science is not there, no one gets the expensive stuff. Other things cost around the same as this stuff-certain medications to control multiple sclerosis come directly to mind. The science for those is hard; if you have MS, these drugs will, in most cases, prevent or reduce your relapse rate. I will approve those gladly; they keep you healthier, even if the data might not indicate that it’s cheaper overall (although I suspect they would, if anyone did that study). You don’t have to prove cost-effective; just prove “actually makes you better more often than would happen by chance-preferably a lot more, but I’ll even accept *some* more. And you’re golden.
1) Note for the record: while I’m not happy about aspects of the executive compensation at my employer, at the end of the day our medical loss ratio is 90% (we spend 90 cents of every dollar we collect paying for health care. That’s among the best in the industry, bluntly speaking. We don’t actually spend a lot on administration, and we aren’t paying shareholders (although we do still pay our board members-the trend is to not do so, as a non-profit entity. I get both sides and haven’t yet formed a solid opinion regarding whether board members at big non-profits ought to be paid. They ought, however, to be much more accountable; it’s likely that they get a lot of non-dollar compensation out of being a board member; it seems overkill to pay them. But if you don’t pay them, they likely wouldn’t do it-we’re talking mostly about people who use money to keep score in their contests with the Joneses, so to speak.