CER and Decisionmaking: A Conversation with Two Health Care System Administrators
Michael Fordis: Dick, why don’t we start with your telling us about some of the types of decisions you make at UnitedHealthcare?
Dick Justman: I support medical policy development within United Health Group and that involves the development of medical policies, and identifying which clinical practice guidelines we would like to promote. I also chair our national pharmacy and therapeutics committee. This committee makes recommendations, based upon a review of clinical evidence, of the comparative value of different medications used to treat a variety of conditions. So, for diabetes, as an example, we would be making decisions with regard to the comparative clinical value of various kinds of glycemic agents including insulin analogues and including also noninsulin glycemic agents.
Michael Fordis: Thank you very much. And Alan, do you want to give us just a brief rundown on decisions that you make at WellPoint?
Alan Rosenberg: I chair and participate in a variety of committees for WellPoint that make medical policy determinations regarding new technologies, new procedures, pharmaceutical agents, and biologic agents in the context of our benefit determinations. And it is in that context that we use comparative effectiveness information in trying to determine whether these interventions are or are not medically necessary, and or are not still investigational.
Dick Justman: The issue with comparative effectiveness is not to find out if something works. The issue is making a rational choice based upon whatever clinical evidence is available when multiple treatments work for the same condition. And it’s not always going to be the same decision for everybody. Consumers, at least, in our experience, will look at things from one vantage point. Physicians will look at this question from another vantage point. Purchasers of health care, who are our customers, will look at it from a third vantage point. And what we would like to have is some tool, something in our toolbox, that would help to make those decisions on a rational basis.
Michael Fordis: Dick, do any specific examples come to mind?
"The issue with comparative effectiveness is not to find out if something works. The issue is making a rational choice based upon whatever clinical evidence is available when multiple treatments work for the same condition. And it’s not always going to be the same decision for everybody."
Dick Justman, National Medical Director of UnitedHealthcare
Dick Justman: There are lots of them. One is computerized tomography (CT) colonography. While optical colonoscopy is the gold standard in colorectal cancer screening, its uptake among consumers is very low because they perceive optical colonoscopy as an invasive and uncomfortable test. It is also a test that frequently requires either sedation or anesthesia. CT colonography is not as good a test, but consumers may perceive it as less unpleasant than optical colonoscopy, and may, therefore, be more willing to undergo CT colonography. In order to improve the rate of screening for colorectal cancer, it would be nice for consumers to have evidence-based information about both tests. Some consumers might choose the better test; others might choose the test they believe is less invasive. In any event, hopefully such consumer-centric information would improve rates of colorectal cancer screening.
Treatment of localized prostate cancer is another example. Consider somebody who is otherwise well, who’s only crime, so to speak, was being 50-years-old and having a prostate-specific antigen (PSA) test as part of his preventive examination, finding it elevated, eventually having a four-quadrant biopsy, and finding that he has localized prostate cancer that is relatively mature based upon Gleason scoring. What is he supposed to do? There’s more than one correct answer. Should he have a radical prostatectomy, with its attendant morbidities like urinary incontinence and erectile dysfunction? Radiation therapy? Periodic reevaluations without treatment unless he becomes symptomatic or his PSA level rises? It would be nice if the pros and cons of any of the approaches could be explained in a rational way, so that people can make rational decisions, again recognizing that not everybody is going to make the same decision. The way things work right now, at least in my experience, is the kind of doctor you go to is very likely going to guide the consumer’s decisionmaking. If he sees a urologist, he will be guided, probably, towards a radical prostatectomy. If he goes to a radiation oncologist, he will likely be directed towards either seed brachytherapy or some form of external beam radiation. And if he goes to his family doctor or general internist, he may be directed to any number of places. It would be nice if there were information available so that people in an objective way would be able to make rational decisions based upon their individual values and their individual needs.
Michael Fordis: What about drug treatments? Does CE research help there as well?
Dick Justman: I think you're going to find that when you get into treatments — again I’m speaking from the perspective of a payer — when you talk about medical treatments that are administered as a medical benefit, I think you’re going to find that the scenarios are pretty much as I just described and I think as Alan alluded to also. When you talk about drug treatments that are administered as part of a pharmacy benefit, life becomes a little bit more murky. And it becomes more murky because in our standard benefit design, all of the drugs are covered. The difference is going to be what tier they are covered on. In other words, are they going to pay a high copayment? Are they going to pay a lower copayment? Are they part of a health savings account where until the deductible is satisfied they’re fully responsible for the cost of treatment?
In the world, for example, of glycemic agents for type 2 diabetes, the problem is not that there are not enough treatments. The problem in my opinion is there are too many medications, some of which actually have a very, very dubious incremental benefit over other medications. Most persons with type 2 diabetes do not, for whatever reason, go to endocrinologists; they are treated by primary care physicians. So from the standpoint of a physician who ultimately is going to be guiding these decisions, how does one make in a particular situation a rational decision with regard to what is first-line therapy [and] what is second-line therapy? What are the indications for monotherapy? What are the indications for combination therapy? With regard to the issue of insulin analogues, what is the role of insulin analogues for individual patients? What are the issues that need to be addressed? And from the standpoint of patients with diabetes, there are at least four different areas that need to be addressed. One is the area of adherence. In other words, the treatment that you have to do seven times a day and also you need to test seven times a day is less likely to occur than a treatment that needs to be done three or four times a day. The second question is the one of safety, and safety from the standpoint of glycemic agents has to do with hypoglycemia, both symptomatic and asymptomatic hypoglycemia. The other issue that needs to be addressed is effectiveness, and effectiveness is kind of a murky issue because of the question of definition. Do you define effectiveness by glycosylated hemoglobin? Do you define it on the basis of fasting blood sugar? Do you define it on the basis of postprandial blood sugar measurement? As your excellent clinicians summary guide on premixed insulin analogues pointed out, there’s no single answer that’s going to be able to address all of those things. And the final area that we need to look at is the issue of cost as experienced by the consumer because that is going to affect adherence and duration of treatment. So those are things that from the standpoint of usefulness to us need to be addressed wherever we talk about drugs that are administered as a pharmacy benefit.
"It would be nice if there were information available so that people in an objective way would be able to make rational decisions based upon their individual values and their individual needs."
David Hickam: If you are making a decision about several different drugs that are all going to be covered in the benefit plan, but may be covered in different tiers, are they in different tiers because the plan has evaluated the evidence and made a judgment about comparative effectiveness?
Alan Rosenberg: We do that using whatever clinical information is available, whether it’s a comparative effectiveness report, drug-specific evaluations, or just individual trials of the drugs using the placebo where we have to undertake best estimate evaluation based on the nondirectly comparable evidence. And the determination is made by a pharmacy and therapeutics committee comprised of physicians from across the country in our networks. We also have external subspecialty inputs that help inform those determinations. After we make that decision, we follow a set of rules that an inferior drug will never be placed on a higher tier, requiring lower co-pays than a drug that’s superior. So we do place those drugs that are more effective in a lower cost tier than those that are less clinically effective. But if they’re comparable based on other factors, we may tier them based on cost -- based on our cost. The tiering decision is done by a value assessment committee that’s made up of WellPoint leadership after the clinical evaluation is made. In summary, we apply the comparative effectiveness research when available to the question at hand. We do prefer head-to-head comparisons; however, frequently we find too little head-to-head evidence about drugs within a class. Also, there is often little information about the variability of specific subsets of patients, such as [in] comorbidities. I do wish there were far more of those studies being done by the academic research community.
Dick Justman: With regard to drugs, aside from tiering, we also have clinical programs. For example, there might be notification criteria. In other words, to make sure that when somebody orders Proscar that they are, in fact, ordering it to treat symptomatic benign prostatic hyperplasia and not male pattern baldness. We also have quantity limits. In other words, the number of units of a medication that would be able to be obtained with a single co-pay, we need to make sure that those are rational. For example, when the Centers for Disease Control and Prevention (CDC) guidance came out with regard to the use of antivirals to treat H1N1 flu, we went back to make sure that the quantity limits allow overrides that would be consistent with those recommendations so that we would not have consumers in a conundrum where their pharmacy benefit is at odds with what the CDC is telling them they should do.
Stephanie Chang: Are the issues different with the coverage of technologies?
Alan Rosenberg: The biggest difference is the benefit plan structure. There isn’t tiering in the same way. It’s either medically necessary or not.
Dick Justman: Generally, for example, our benefit structure would be that it’s covered if it prevents, diagnoses, or treats a disease, unless we consider it to be investigational or unproven or cosmetic or custodial or unless it’s already on a list of excluded services. For example, you’ll never find in our benefit document that appendectomy is or is not covered or that inguinal herniorrhaphy is or is not covered, whether it’s done open or laparaoscopically. What you would have to know is that an inguinal herniorrhaphy is not unproven. It’s not investigational. It’s not cosmetic and it’s not custodial. Since no benefit exclusions apply, it would be a covered health service. And medical policies will usually help with that kind of determination.
Michael Fordis: How do you make decisions when there is little evidence, or lower quality evidence?
Alan Rosenberg: In medicine, that’s not an unusual occurrence. There are many medical treatments or procedures that have been done historically for many years. We love, as Dick mentioned, comparative effectiveness studies, but in the absence of that, you often have to look at whether there’s any evidence that A is effective period, then is A as effective, more effective or less effective than B. As an example, let us consider radiation therapy versus surgery in prostate cancer. The Effective Health Care study said there are very low levels of evidence on whether one is better or worse for the population as a whole or specific subpopulations. But we obviously don’t go back and say no to all of these therapies; so, then you try to compare specific types of radiation therapy or specific types of surgery. Unfortunately, once again, we know there are also very low levels of evidence regarding that form of radiation therapy, whether IMRT, proton, 3D conformal radiation, or brachytherapy works better. The same can be said for different types of surgical techniques. Recognizing the limitations in the clinical evidence, our organization, through its Medical Policy and Technology Assessment committee processes, evaluates the evidence, evaluates the history of what is done, and how these things have evolved. Definitions around medical necessity and investigational procedure guide these determinations. WellPoint also obtains input from experts at academic medical centers and from specialty societies. It’s a complex set of issues, and it needs to be done on hundreds of clinical subjects.
Dick Justman and Alan Rosenberg both serve on the ARHQ Effective Health Care Stakeholder Group. Additional excerpts from this interview will be included in future issues.