A Conversation with Al Lewis: ‘There’s Now a Correct Method For Defining ROI in Disease Management’

The godfather of disease management talks about how the industry went wrong, and is now going right, in describing its own value

To the degree that one person can be credited or blamed for the very existence of a $1.1-billion segment of American health care, Al Lewis is that person when it comes to disease management. If you doubt that, just ask him. Blunt, funny, and supremely confident of his knowledge of the field, Lewis founded and is a past president of the Disease Management Association of America, and now heads the Disease Management Purchasing Consortium International. DMPC is a consultant and broker with 89 members that include health plans, private and public employers covering 80 million lives, the Congressional Budget Office, and leading accreditation groups.

With a potentially huge boost from the Medicare Modernization Act, disease management could be on the verge of a boom, Lewis argues, but only if it overcomes a major obstacle: Nobody believes its numbers. Many employers won’t contract with DM companies because they doubt vendors’ claims about how much money their programs save. Lewis says these doubts are legitimate, and traces the problem of inflated savings to a flawed methodology — and to the likelihood (he contends) that some DM vendors are “so stupid they don’t know how stupid they are.”

He also says that he and two colleagues have solved the problem in a recently published white paper. On DMPC’s Web site, Lewis indicates that he’ll accommodate those who think he should give that document away — as long as they themselves work for free or are card-carrying members of the Communist Party. Others can pay.

Lewis holds undergraduate and law degrees from Harvard University, and is a visiting scholar at Brandeis University’s Heller School for Social Policy. He spoke recently with Senior Contributing Editor Patrick Mullen.

MANAGED CARE: What’s wrong with how disease management companies calculate return on investment?

AL LEWIS: I used to pooh-pooh skepticism about ROI in disease management as being a result of people not knowing what they were talking about. It turns out they were right, though they didn’t know why. The old industry standard that everybody used — myself included — has a fatal flaw. I would claim to have invented the old methodology, but since we figured out that it was wrong, I went from pride of authorship to admission of authorship.

MC: To be followed by denial of authorship?

LEWIS: Right, though I’m not quite at that point. Here’s the problem: A huge percentage of employers who don’t do disease management don’t do it because they don’t believe the numbers. Most of the numbers they don’t believe are indeed wrong, especially if the vendor is not on the DMPC-recommended list. Employers are right to be dubious of vendors that claim an ROI of four to one.

MC: Can you give an example of a DM vendor’s dubious claim?

LEWIS: Yesterday, somebody sent me ROI numbers that her company’s DM vendor — apparently a friend of her company’s president — had given her. She knew the numbers were wrong, but she’s not a financial person, so she doubted management would take her word for it. She asked me for some ammunition to bolster her argument. Within five minutes, it was clear to me that this vendor had no clue whatsoever what it was doing. One obvious reason was that it said it was saving more money on asthma than its client was spending on asthma. It was disguised and I had to do a couple of calculations to determine that, but there it was. Asthma is the canary in the coal mine with disease management. I can tell in a nutshell whether a DM company is doing the whole thing wrong by whether asthma savings exceed asthma spending. It’s an easy mistake to make, because the tendency is to measure a patient’s entire claims rather than just asthma claims. If an asthma patient’s total claims drop by 20 percent, some DM vendor will say they’ve produced a 20-percent saving on asthma. The problem is that average asthmatics spend only about 15 percent of their claims on asthma. For this company’s numbers to work, asthma patients would have to have been spending about 80 percent of total claims on asthma, which simply doesn’t happen.

MC: They took credit for lower claims that had nothing to do with asthma.

LEWIS: Exactly. The vendor had no idea why its numbers were wrong. They were too stupid to realize how stupid they were. That’s probably true of 50 small vendors.

MC: Are you describing a two-tier market, with some vendors trying to provide accurate numbers and others more interested in making a quick buck?

LEWIS: It’s a three-tier industry. A group of established players does disease management right. A second group of established players does it right under duress when somebody makes them. A very large third group of small companies has absolutely no clue what they’re doing. They think they’re right, but they’re not, and nobody calls them on it because benefits consultants who dabble in this stuff don’t have a clue either.

For all I know, these small companies do save their clients a little money, but their reporting is so bad that there’s no way to know for sure, which calls their entire performance into question.

MC: What distinguishes one first-tier DM vendor from another? Are there wide strategic differences or is success defined by how well they execute similar strategies?

LEWIS: There are differences in the models, but it’s hard to say one model saves more money than any other. It’s reasonably easy to say that certain vendors measure their savings more accurately than other vendors.

MC: And the third-tier companies give the entire DM segment a bad name.

LEWIS: Right. A lot of employers are on the sidelines for this reason. I was among the first to figure out what’s been wrong with the methodology, but I don’t have a monopoly on IQ. A few actuaries at the big actuarial firms know what they’re doing, not many, but enough that they can spread the word.

MC: What’s the fatal flaw in how disease management companies have calculated ROI?

LEWIS: The fatal flaw is caused by the fact that not all patients with a disease will file a claim during the initial measurement period. This creates a situation where the entire diseased population is not counted in the baseline measurement. As a result, estimated savings for some chronic diseases — asthma and coronary artery disease in particular — will always be overstated. As improvement is measured from an artificially underrepresented baseline, plans will overstate program improvement and report an inflated ROI.

MC: How does your solution fix this measurement flaw?

LEWIS: Ariel Linden, a biostatistician, Ian Duncan, who’s an actuary, and I put out a white paper that basically creates the new industry standard for measuring savings from disease management. We now have a proof that fixes this fallacy, which means we have valid measurements of costs pre- and post-disease management. There’s now a correct answer for defining ROI in disease management. It can be refined a bit, but the magnitude of what we don’t know is likely to be small enough that I’m quite comfortable saying that there is indeed a correct answer.

MC: What’s at the heart of the new methodology?

LEWIS: The key to the measurement is that anybody that you find with a disease in any period has to be considered to have the disease in all periods. Let me give you an example. Say you had two asthmatics in your plan. One had a $1,000 claim in 2003 and nothing in 2004. The other one had no claim in 2003 and a $1,000 claim in 2004. Using the industry standard methodology, if you look at claims for 2003, you conclude that your average asthmatic costs $1,000 in that year. You had a prevalence of one and it costs $1,000. In 2004, the vendor finds that prevalence went from one to two, and would claim that the cost-per-asthmatic fell to $500, even after spending $1,000 on one patient in 2004.

MC: Why is the new methodology important?

LEWIS: It puts the whole issue of validity to bed through standards. I’ll tell your readers with certainty that if this is the first you’re reading about it, then you are not using it. It will be commonplace within one or two years because essentially as soon as people read this article, they’ll realize that they’re doing it wrong and will change it. So far, only about 15 health plans and employers use this correct answer. Georgia and Wyoming’s Medicaid programs use it, and their results are publicly available. Health Industry Research Companies (HIRC), a market research concern in Santa Cruz, Calif., nominates and identifies the best health plans, employers, and states in disease management every year, about 30 in total. I was pleased to see that every health plan and employer that is measuring correctly is on that list. We’ll see a big trend away from invalid methodologies.

MC: The concept of disease management is about ten years old. Over that decade, has disease management evolved away from, or more fully realized, its original mission?

LEWIS: It’s more fully realized its mission. The definition DMAA put out ten years ago is still pretty good. It could be tweaked, but the mission, charter, and strategy are the same. The tactics and the scope of the industry have increased quite a bit. The industry used to focus on only three or four diseases. While it remains true that the vast majority of savings comes from three or four diseases, there are many other diseases where disease management can improve people’s health and keep them at work longer. Employers are willing to pay for that.

MC: How fast is disease management growing?

LEWIS: Industry growth was well ahead of expectations in 2004. HIRC is projecting a fifty-percent growth rate over the next three years. I have a feeling that will turn out to be conservative.

MC: The Medicare Modernization Act includes several disease-management demonstration projects. If they go well, what is the potential effect on the DM business?

LEWIS: If it works, disease management could become a $10 billion industry and be as much a part of the landscape as primary care. The beauty of it is that the measurement that CMS is using is as good as you can expect a real-world measurement methodology to be — and they’re not a client of mine, as much as I wish they were. So if they say they’ve saved money at the end of those demonstrations, they’ve saved money. It cannot be argued. Likewise, if it turns out they haven’t saved money, there’s very little ground for challenging the methodology.

MC: Do you expect the demonstration projects to succeed?

LEWIS: My concern is that a lot of vendors that contract with Medicare won’t meet CMS’s goal of five percent net savings, because the vendors will still be using an incorrect methodology. That’s the bad news that might put some vendors behind the eightball. But I predict that the good news will turn out to be that even if the five-percent threshold isn’t reached, there will be savings in excess of cost, so it will make sense to expand the program. Another trend we’re starting to see with the Medicare population is a move toward better medical management of the frail elderly in nursing homes. These folks are in constant potential crash mode. They end up in the emergency room or other acute care setting, which is debilitating to the patient and expensive. Having a doctor walk through and take a quick look at a napping nursing-home resident won’t cut it. Doctors and nurse practitioners need to get into these facilities far more often.

MC: How did CMS arrive at the five-percent figure?

LEWIS: A lot of the CMS methodology, for better for worse, is traceable to stuff I had been doing for years. I invented guaranteed savings by accident while working for a consulting firm that was working for Humana, where disease management originated. The consulting firm’s fees for Humana had a strong component of guaranteed savings. For the consulting firm to get its money, when we invented disease management, we passed the same model on to DM vendors. That way, we could include their savings in our savings. Guaranteed savings became part of the landscape because it made it much easier for health plan CFOs who were skeptical of disease management to accept it.

MC: What do you see happening with DM as part of Medicaid?

LEWIS: I do a ton of Medicaid consulting work, and the huge mistake that gets made in Medicaid health plans is the failure to distinguish between TANF population — Temporary Assistance for Needy Families — and the disabled population. They’re two very different populations that should get very different approaches to disease management. Yet in many cases, plans just think one size fits all. Sight unseen, I can tell any health plan executive that however much he spends on disease management for TANF, he’s overspending, and however much he spends for disabled, he’s underspending. The reason is that the TANF population has a low disease burden, very high turnover, and is not easily trackable. The disabled population has a high disease burden, is easily trackable, is underserved, and has low turnover. It takes me half an hour to explain this to people, but when I look at programs put in place by non-DMPC members, it’s quite common not to distinguish between those two populations.

MC: In this magazine six years ago, you cited consolidation as a coming trend in DM. Any thoughts on what’s stalled that trend?

LEWIS: I always think the industry’s about to consolidate and it never does. So I’m not only going to deny that I ever made that statement, I’m going to deny that you and I ever had this conversation.

MC: Fair enough. Talk a little bit about DMPC’s certification program.

LEWIS: It the only objective assessment of whether a DM company correctly measures what it does. More than forty companies have applied for certification; thirteen or fourteen have received certification to date. Our certification differs from what the accreditation agencies do because they don’t look at the integrity of the financial reconciliation. I’ve received about 100 inquiries from employers looking for health plans that have been certified.

MC: What separates a certified DM program from others?

LEWIS: When a program is certified, their ROI is pretty darn close to what they say it is. There might be five programs in the country that could be certified right now for their ROI, but for whatever reason, haven’t asked. There are probably a hundred or two hundred that could not get certified if they asked.

MC: Odds are they won’t ask.

LEWIS: I get calls from people who know they’re not up to standard but want to apply to figure out how to get there. We don’t decline applicants. You just keep getting deferred until you get it right. I don’t keep a list of who’s flunked or anything like that.

MC: Do you expect that 10 years from now, disease management will still be around as a distinct market segment, or will its functions be absorbed into health plans or whoever is providing health services?

LEWIS: I don’t know what’s going to happen in ten years. For the next three years, DM programs are not going to get done by health plans. The trend now is toward outsourcing for the simple reason that benefits consultants and employers want name-brand programs. Smaller employers want their health plans to provide DM services. Health plans have a chance to keep small employers from looking elsewhere for those services. Plans can either get disintermediated or be the intermediary. They want to be in the mix, and here’s their chance. If they blow it, they basically just become claims payers.

MC: Thank you.

The interviewer, Patrick Mullen, is a former managing editor and editor of Managed Care.


HealthIMPACT Southeast Tampa, FL January 23, 2015
Medical Devices Summit Boston, MA February 19–20, 2015