LIFE AND DOLLARS: a health care insider's account of how prior authorization really works.

Prior authorization started out as a means by the “payers” of health care goods and services to provide patients the assurance that what their doctors order is medically necessary and appropriate. A strong argument could be made that there is a legitimate role for prior authorization in some circumstances, but big insurers not only have used it to avoid paying for necessary care, but they–and a cottage industry of consultants representing health care providers–have also figured out how to profit from the process. As a consequence, an enormous amount of the money we spend on health care is all too often being used against us, endangering our lives. Below, an executive who ran a prior auth shop pulls the curtains back to explain how it really works. To help the writer avoid potential legal and career problems, we have agreed to publish the piece anonymously.

Journalist Libby Watson assisted with the production of this piece.

---

Requiring patients to get prior authorization from an insurance company for medical procedures and drugs was supposed to lower medical costs. The theory was that it would prevent doctors from charging for unnecessary care. 

The process frustrates patients and burdens health-care providers. And, the numbers show, it doesn’t really work.

I worked at one of the largest prior-authorization companies, running a team that supported the non-clinical side of our business.

It may surprise people to know that many of the biggest health-insurance companies outsource their prior authorization programs. In fact, it would probably shock most people to understand just how much middle management exists between their doctor’s decisions and their ability to receive care. The commercial health insurance industry is overrun with opportunistic companies who profit off our complicated health care system, adding costs that lead to higher premiums and cost of care.

Prior authorization is meant to ensure that a procedure is clinically appropriate, or medically necessary, to cut down on what’s called ‘overutilization’ in the industry. It’s well-established that doctors feel prior authorization puts a huge burden on them, leading to delays in care and adverse health outcomes. 

“The process you have to use is so costly and so burdensome on the patients and the providers that the savings are just not worth it.”

Based on my experience in the industry, that’s not the only problem. If prior authorizations were at least doing what they were supposed to do, lowering the cost of health care by stopping doctors from charging for unnecessary care, that might justify them. 

But for all this time and effort, all the extra work that providers do and all the frustration that patients experience, I don’t think prior authorizations are a cost-effective tool to limit overutilization. The process you have to use to do this is so costly and so burdensome on the patients and the providers that the savings are just not worth it.

Those savings come purely from catching inappropriate procedures and denying them. For example: We can tell clients who provide insurance that we denied 15% of the requests that came in the door last year, and if we cut out 15% of your costs on certain claims, it would save you $2 per member per month. Well, we're only going to charge you 50 cents per member per month — what a bargain! 

“It's a lot of kicking the can down the road, and we obscured that when selling our value.”

Where that pitch gets a little misleading is that the denial rate is transactionally based. If a doctor submitted a request today, and we denied it as being not medically necessary, it would count as a denial in that number that we advertise to clients. If the same doctor submitted a request a month later for the same procedure, and we approved it because now it is clinically necessary, that doesn't erase the original denial. It's a lot of kicking the can down the road, and we obscured that when selling our value. Every request could be followed up with a subsequent request that gets approved, and they often are. If that happens, the insurer is paying for it and putting their provider network through the annoyance of having to deal with us as well.

The other hidden cost of the prior authorization program, one that never seemed to factor into this calculation of return on investment, is patient impact. For every procedure we clinically review and determine to be medically unnecessary, we’ve created an opportunity for confusion and frustration for patients. They likely had no idea that there was an independent arbiter paid by their insurance company to review their doctor’s decisions regarding their health. 

“Patients are left feeling like a faceless corporation is making their medical decisions and undermining their actual doctors.” 

These instances of confusion and frustration are what drive the majority of the bad publicity within the industry, and rightfully so. Without an ability to communicate directly with patients, it is very hard to explain why their doctor may have been mistaken or outdated in their referring practices, or even that their doctor just forgot to submit all the paperwork necessary for a review. As a result, patients are left feeling like a faceless corporation is making their medical decisions and undermining their actual doctors. 

HOW APPROVALS WORK

The best case scenario for a prior authorization is a quick approval, where the patient has no idea any intervention took place. 

Our company used an artificial-intelligence predictive model that immediately approves a significant portion of cases. We trained the model on our history of clinically reviewed cases, to identify trends and relationships and figure out what kinds of cases are most often approved. A team of data scientists built the model, which was then integrated into our standard workflows. 

The model doesn’t deny cases; it can only approve them, or flag them for clinical review by a human medical practitioner. A prior authorization request can’t be outright denied by a robot.

Before the predictive model was adopted, we used clinical algorithms, basically a decision tree, to determine whether something needed review. It could be something as simple as, how long has the pain persisted? Have they tried physical therapy? Have they been taking painkillers? And if the answers fit the decision tree— something like “six weeks, yes, 3.5 grams”—then they could get an approval immediately via the algorithm.

We stopped doing that because, very predictably, doctors just memorized the answers to the questions that they needed to submit to get an approval. 

“Now, whole companies exist purely for the purpose of guaranteeing quick prior authorization approvals, acting on behalf of physician groups or hospitals.”

GAMING THE SYSTEM

This method of imposing prior authorization has spawned even more waste in the system. Now, whole companies exist purely for the purpose of guaranteeing quick prior authorization approvals, acting on behalf of physician groups or hospitals. This is yet another development that acts to drive up the costs of care in the system, as both sides of this equation are now paying for the purpose of prior authorizations. 

Those costs aren’t eaten by the providers or hospitals who have to pay them in order to get a guaranteed authorization. They get baked into hospitals’ fee schedules, and result in demand for greater reimbursement from the insurance company, which then raises your health insurance premiums. 

For an example of this behavior in practice: You go to see your doctor with knee pain, and he wants to get you an MRI of your knee. But instead of the doctor’s office calling us, they would send your information to another third-party company, usually running call centers overseas, and we would get contacted by them pretending to be your doctor. Sometimes, these requests were fraudulent — they would just say whatever they needed to say in order to get the approval, and the requests would be filed under a false name by someone pretending to work for that doctor. 

So, in our example, let's say you've never done any physical therapy, you haven't taken any painkillers; you simply woke up this morning and your knee feels a little bit off. Then, it's not necessarily appropriate to get a $2,000 MRI. And as a patient, you’re not in the best position to know whether the doctor is recommending an overly aggressive or expensive procedure. An unscrupulous third-party company might take a situation like that and instead say to us, "She's been complaining about knee pain for the last two and a half months, she's had 16 different physical therapy appointments, she takes daily painkillers, and the pain rates a nine out of 10," because they know what to say to get the approval. 

There were also numerous instances where we would analyze the performance of our clinical algorithms and see doctors answer these questions the exact same way every single time. That tells us they're gaming us, because there's absolutely no chance that every single one of their patients has the exact same answers to these questions. We could find instances where several cases from a particular doctor had identical notes – just a photocopy every single time; all they would do is paste the name of the patient over the top of it. 

But it was rare to see insurers fight back against this type of fraud, because they wanted to keep doctors in their networks, and there wasn’t much we could do by ourselves. Provider noise drove most of the complaints we would receive, as well as most of the pushback on the sales process.

IMPROVING RETURNS, NOT PATIENT CARE

Prior authorization is meant to catch as much of the unnecessary stuff as possible. But it can only do this to a certain extent, because there has to be a return on investment. If we had manually reviewed more cases, we would find more inappropriate procedures and more things we could deny, but the labor involved would make that unprofitable. No prior authorization company would be able to review clinically every single case that comes through the door and determine whether it meets the guidelines of appropriate care. We reviewed less than half of cases, depending upon the client. If you review half of the cases, you're going to find about a quarter of those are inappropriate. As long as you're denying 10-15% of review requests, that's enough to justify the price they pay us and save money to boot. We don't want to do more work than is necessary, and the doctors certainly don't want us to do more than is necessary. They already hate us. 

If we reviewed more cases than usual, and a client’s denial rates seemed super high, we got a window into what the theoretical  maximum for denials could be. Sometimes 25% or more of requests in those markets were found to be clinically inappropriate, which sounds incredibly high. But the clinical review rates necessary to discover that would be so unsustainably high that no company could charge reasonable rates and create a margin of return operating that way. It would be akin to having a backup doctor in every office providing a second opinion with each patient visit.

When we denied a case, we denied it for explicit, clinically valid reasons. You could read the guidelines on our website. We have doctors who stake their licensure on these decisions, and our chief medical officer seemed to really care about clinical validity. But there are requests we denied because they weren’t medically appropriate, and requests we denied because the doctor mistyped something on the web portal, or because we didn’t have enough clinical information and the doctor didn't follow up with it quickly enough. If a doctor sends nothing at all, that case gets denied. That request might well have been valid, but the patient is relying on their doctors to get this stuff right, and if they don’t, patients are out of luck. 

We would never have stated this with clients, but a lot of the value in our work was kicking the can down the road. Millions of people with employer-provided insurance change plans from year to year. These health plans have an incentive to deny a procedure right now, because even if it comes back in 45 days and gets approved, there's a chance someone else will pick up the bill. That’s a massive perverse incentive.

“The purpose of prior authorizations is to save dollars, not improve patient outcomes, regardless of what they say.”

Prior authorization companies are not insurance companies; we didn’t have access to all of a patient's records, claims history, or anything that would let us assess overall health outcomes. We couldn’t know, did this person ever get better? 

By having prior authorizations in place, did anything actually improve in your overall patient population? The answer is probably no, because the purpose of prior authorizations is to save dollars, not improve patient outcomes, regardless of what they say.

WHAT EMPLOYERS AND PATIENTS CAN DO

There are things that employers and patients can do about this. Employers need to notify their employees if they are going to impose a prior-authorization requirement on certain procedures. Most of the time, doctors are the ones who know whether a person covered under commercial health-insurance plans needs a prior authorization for a procedure, and they know because it determines whether or not they get paid. But patients are the ones who get the denial letter, or a phone call from this company they've never heard of, saying they can't get this procedure, or they can't get this treatment done. 

“The real value in prior authorizations isn’t in savings from clinically inappropriate procedures, and instead is a function of helping commercial insurers and third parties keep as much money as possible in the health-care system shell game.” 

If employers do a better job of informing their employees that they have opted into a health plan that imposes prior-authorization requirements, it might make the patient’s experience less frustrating. It also would help if patients or members know that they will get an explanation of a denial, and that they might need to have a doctor follow up or provide additional information to make this procedure happen. Often, it's a misunderstanding that leads to denials. Doctors, by and large, aren't that wasteful. 

MAKING DOPEY AMOUNTS OF MONEY FROM PRIOR AUTH

While working in the industry, it often seemed to me as if the real value in prior authorizations isn’t in savings from clinically inappropriate procedures, and instead is a function of helping commercial insurers and third parties keep as much money as possible in the health-care system shell game. There is a large layer of middle management and profiteering that exists between the patient, the provider, and the insurance company. I know dozens of people who have made dopey amounts of money by rinsing-and-repeating the process, building small, ancillary companies that nibble on the edge of our high health-care costs. It all just gets baked into the cost of the premiums.