External-review processes are all the rage. But what approach should be used: Developing a new medical opinion, or assessing the validity of the old one?
Like the sword of Damocles, the threat of high-stakes litigation dangles over HMOs these days, with advocacy groups, politicians, and the media just dying to cut the string. The $120 million settlement in Goodrich v. Aetna U.S. Healthcare indicates just how big a sword we're dealing with here. The guilty verdict against a plan that seemed to do everything correctly in Johnson v. Humana shows just how much the thread has frayed. (In both cases, the plans promise to fight the lower courts' decisions.)
Little wonder, then, that there has been some search for an alternate means of instilling humility into HMO decision making. One of the better alternatives is third-party review of managed care decisions. But what is the best method to use?
The most talked about way is to have third-party medical reviewers render their own medical decisions. In law, this is called a de novo decision. The alternate, which hasn't been talked about, is to have third-party medical reviewers determine whether the HMO opinion was reasonable based on the facts as they existed at the time of the decision. We're calling this type an "appellate-level" decision, although it's not a court decision.
Suppose a small and adorable child goes to her father and asks, "Daddy, can I go over to Mary's house to play?" If her father says, "Sure, sweetheart," that's a de novo decision. If, on the other hand, the father replies, "What did your mother say?" and discovers that his wife had just told her she couldn't go, then the father is left with weighing whether his wife's decision was reasonable enough so that — even if that might not have been the way he would have called it — he will back it up. That's an appellate-level decision: The original decision stands unless it is arbitrary, capricious, or an abuse of discretion.
To determine which approach is better, we first need to clarify what problems we're trying to solve. There are three: conflict of interest, patient-specific treatment, and accountability. After that, we need to assess what kind of territory we're getting into.
The conflict of interest problem is, simply, that by holding medical costs down, HMOs stand to make money.
Now, as compelling as that is in a capitalist society, patients have even more on the line. For them, the coverage decision is not just about money; it can literally be a matter of life or death.
Consequently, public opinion runs in favor of having someone else — some third party with no financial stake in the outcome — look over the disputed decisions to make sure that any "no" is a good and reasonable "no."
This issue takes us back to another aspect of the HMO decision-making process, the guideline. HMOs have guidelines because we couldn't afford the old system of "just pay for whatever the doctor orders." Employers, panicky over double-digit medical inflation and grasping after news stories about millions of dollars in unnecessary medical procedures, made it clear to health plans that they weren't about to shell out for all that medical waste any more. It was up to HMOs to separate the wheat from the chaff.
But what is the good stuff? What seems to do the most good for the largest number of people. So guidelines were born along with a fundamental "hmmm" built into them. Good guidelines capture what is best for most people — but what if you are not most people? What if a treatment does wonders for 97 percent of patients — but you're part of the other 3 percent?
The HMO appeal process is supposed to take care of this, so that you're covered for care that is appropriate for you. But here, too, there is a public call for an independent medical expert who will take a look to make sure that HMOs are not approving care that is appropriate for the guidebook instead of care that is appropriate for the patient.
People use that word to mean several different things.
- In its most strident form, accountability stands for the "two-by-four" approach to encouraging appropriate behavior — "If they step out of line, whack 'em a good one. That'll learn 'em!" The leading example of this approach to accountability is the call to expand HMO liability in the courts, on the theory that a few multimillion-dollar damage awards would keep those HMOs on their best behavior. (Or would it? See "Why Torts Are Not Good Medicine" on Page 46E.)
- A related but somewhat less bloodthirsty meaning is simply that someone is watching, that questions may be asked, that there may be some explaining to do. A person is "accountable" in this sense if there is someone who can ask, "What are you doing?" "Why are you doing that?" and, yes, "What were you thinking?"
- The final meaning of "accountability" is embedded in the other two, but is worth digging out and dusting off. This version of accountability asks "Did we get the right behavior?" and "Did we get the right outcomes?" In the world of managed care, this would translate as "Does this HMO provide high quality care?" (Of course, a payoff this good requires more than one step to accomplish. For a range of tools to achieve this kind of accountability, see "A Prescription for Accountability".)
While some news reports would have us go right to the first version of accountability — indeed, a number of news stories would have us think that the first version is all there is — when looked at thoughtfully, it becomes clear that this third meaning is really at the core of what the fussing is all about. The first and second are merely means of trying to achieve the third.
This is where third-party review comes in. It:
- Deals with conflict of interest by having someone with no financial interest in the outcome review disputed decisions;
- Provides a safeguard against inappropriate cookie-cutter medicine by making sure that care is appropriate for the needs of the specific patient; and
- Addresses the desire for accountability by identifying exactly who will be watching.
But that last one only goes up to the second level of accountability, and we want to go all the way to encouraging appropriate behavior and outcomes. Is there something else we need to take into account to get there?
Yes. The one other thing we need to deal with is that there is no single, clear standard for appropriate care.
This whole thing would be easy if it were always clear what the appropriate treatment was: look at what the HMO did, look at what you're suppose to do — there you are! But it isn't always that clear-cut. For example, it was reported that when 82 family practitioners from a single state were asked how to treat a simple urinary tract infection in women, they proposed 137 different ways to do it.
Variation in practice patterns has been documented in the medical literature for years. In fact, as recently as January, a study published in the New England Journal of Medicine found that even in the Veterans Administration — with as homogeneous a population as you're going to get, with a strong central administration, and with salaried doctors (they were not going to get any more income for prescribing either more or less) — the practice variation between 22 regions of the country was still significant.
And, as we've said, the problem isn't solved even if there is a generally-accepted treatment guideline for a condition. Issues will still remain over whether that general treatment is acceptable for this specific patient.
Pretty slippery ground for a de novo approach.
To get a deeper sense how this might work, consider this. Suppose an HMO's medical director is being overturned 40 percent of the time by the state's third-party reviewer and, consequently, the HMO is taking a beating in sales. The HMO decides it must do something, so it fires its medical director and hires the state's third-party reviewer. This person was right all the time in his third-party reviewer job, so he will be right all the time in his medical director job, won't he?
No, because it depends on who the state hires as its new third-party reviewer. Suppose, because of his years of experience in managed care, the state hires the old medical director. What's the HMO's new overturn rate? It's 40 percent, just what it always was, because that's the difference in practice patterns between the two physicians, no matter where they sit. Has the HMO's care gotten any worse or any better? We don't know. All we really know is the amount of difference in opinion between these two physicians.
Well, OK, the rate stays 40 percent if the old medical director is unusually saintly and resists the urge to prove to his old boss that dumping him was a bad idea. With the variation in valid medical opinion, it would be so tempting to just lean a little into the gray areas, and without doing anything questionable produce an overturn rate of, say, 43 percent or 45 percent. And behold, the HMO's overturn rate has shot up 12.5 percent! By firing its old medical director, the HMO's overturn rate didn't get any better, it got significantly worse!
This experiment helps tease out an irony of de novo review. It's supposed to increase accountability, but who is making the final medical decision? The third-party reviewer. And who is the third-party reviewer accountable to? No one — not for the medical consequences of his decision and not for the financial consequences of his decision.
Sure, HMOs can object if they feel that the reviewer is being unfair, but how will the public react? Imagine the headlines: "HMOs cry foul because governor's appointee overturns them too often!" followed by paragraph after paragraph of grateful statements by patients who "won" in third-party review. Far from solving the problem, that would pretty much guarantee job security for the reviewer.
The better method
So what do we do now? Turn to an approach with a long and honored tradition in both the legal and medical fields — the appellate-level decision. It has worked for centuries to keep the court system running smoothly. It is also comfortingly familiar to practicing physicians. If called in to give another opinion on a case, it is common — out of professional courtesy — for physicians to respect the advice of the preceding doctor, as long as it doesn't look as if it will do the patient harm.
The way it would work in third-party review is that the reviewer would assess the validity of the HMO's opinion and only overturn the opinion if it was arbitrary, capricious, or an abuse of discretion. If new medical information (about the patient), new medical evidence (about treatments), or a new medical treatment (such as a new drug) emerge during the course of the review, the dispute goes back to the HMO and the attending physician to resolve.
It is less costly than de novo review. Remember the decision-shopping child we met at the beginning? It is only human to ask, "What are my chances if I go for a review?" With an overturn rate of 60 percent, I have a better than 50/50 chance of getting what I've been asking for. And if it's all paid for by someone else, I might as well go for it. Appellate review would reduce the overturn rate from mere differences of opinion to situations in which the HMO really is in error — reducing the possibility that members would "just go for it" and increasing the possibility that members would seek review only when the decision rendered by their HMO is insupportable.
It reduces administrative expense another way. Remember defensive medicine? Defensive medicine is bad because it is medical-decision making based on the needs of the decider, not the needs of the patient.
It is also expensive: It was reported in 1985, during a physician malpractice crisis, that estimates by the American Medical Association put the cost of defensive medicine at $15 billion to $40 billion a year. De novo review is more likely to encourage defensive medicine as HMOs try to guess what the reviewer's opinion might be. Appellate-level review, in which the health maintenance organization can prove the reasonableness of its own approach, is less likely to trigger this waste.
Appellate-level review is better at reaching that third meaning of accountability — the desired behavior and outcomes. Under de novo review, it doesn't matter how good the HMO's process and reasons for its opinion are if the reviewer has a different one. Under appellate-level review, on the other hand, sound medical-decision making carries the day — and therefore encourages HMOs to do it right.
Finally, appellate review will tell us something meaningful. Remember our experiment? The overturn rate in a de novo system just told us the degree of difference between two physicians. The overturn rate under the appellate standard is more likely to tell us how far off from reasonable medical-decision making the HMO is — and isn't that really what we all want to know?