Search Annals: Advanced search

Podcast Transcript - February 5, 2008

[ Return to Podcast Home Page ]

Hello, and welcome to a New Year and to this weeks Annals of Internal Medicine Audio Summary for our February 5, 2008 issue. I’m Michael Berkwits, Deputy Editor at Annals.

We have a splendid issue for you this week, with articles on CD4 thresholds for beginning HIV treatment; comparisons of routine and selective percutaneous coronary intervention, and the use of drug-eluting or bare mental stents in those interventions; the comparative effectiveness of drugs used to treat osteoporosis; and screening for bacterial vaginosis in pregnancy. Plus, I’ll summarize all the other articles in this week’s issue.

But first, here’s an in-depth summary of this week’s featured article.

Effects of the Medicare Part D Prescription Drug Benefit (Time: 00:41)

Our lead and feature article this week is a study of the effects of the Medicare Part D prescription drug benefit on out-of-pocket expenses and receipt of prescription drugs by older adults. Listeners who heard this summary and the accompanying interview/editorial when it was published on-line in early January can advance to the next segment of the summary by advancing their mp3 player to the 17:18 marker in this podcast.

In the article, lead author Wesley Yin and his colleagues from the University of Chicago use prescription claims data from a random sample of 5% of Walgreen’s pharmacy customers who filled at least 1 prescription in 2005 and 2006 to compare the out-of-pocket expenditures and prescription utilization by seniors aged 66 to 79 y who were eligible for the benefit with two control groups: seniors eligible for the benefit who did not enroll, and younger seniors aged 60 to 63 years old who were not yet eligible. The researchers used the younger control group to estimate trends in expenditures and utilization that would have occurred in the older age group even if the Part D benefit had not occurred.

The paper provides a flurry of estimates, but in general the authors estimated average adjusted monthly cost savings of about $5 overall for all seniors during the stable benefit period after the May 2006 deadline for enrollment, and an adjusted monthly increase in prescription utilization per person of about 4 days of pills. As one might expect, these estimated effects were greater among seniors who actually enrolled in the program, for whom adjusted monthly cost savings were about $9, and for whom the adjusted monthly increase in prescription utilization per person was about 14 days of pills.

Based on these estimates, the authors conclude that the drug benefit led to modest effects on average. The authors acknowledge that the meaning of these estimates is unclear, most importantly because they don’t have information about the impact of the observed effects on health outcomes and on downstream use of health services. They acknowledge that the study is limited by the use of data from a single pharmacy chain; the absence of information about which drugs or drug classes were being used; and their inability to distinguish effects before and after reaching the donut hole, in which beneficiaries incur higher out-of-pocket expenses. The authors don’t acknowledge that the study’s estimates are of prescription utilization, not drug utilization; as every clinician knows, patients who fill their prescriptions do not necessarily take the pills they’ve gotten. Despite these limitations, this was the first serious and reliable attempt to quantify the effects of the Part D program that Annals editors were aware of, and one which for methodologic reasons may be more accurate than another study by Frank Lichtenberg and Shawn Sun published in the journal Health Affairs in November 2007.

In an accompanying editorial, Bruce Stuart, a Professor of Pharmacy and the executive director of the Peter Lamy Center on Drug Therapy and Aging at the University of Maryland in Baltimore, notes that both papers show that the Part D program accomplished much of what it set out to do, which is to increase the number of seniors with drug coverage. He agrees that the real question that needs to be answered empirically is whether or not improved access to medications improves the health and quality of life of beneficiaries, and notes that for unclear reasons, CMS appears to be holding up the release Part D claims data which would allow that assessment. Linking those claims data to other Medicare administrative records in an integrated database is essential, he says, to answering questions about the quality and appropriateness of the use of medications by beneficiaries for their specific acute and chronic conditions.

To get a physician’s perspective on the importance of these findings, I called Dr. John Ayanian. Dr. Ayanian is a Professor of Medicine and Healthcare Policy at Harvard Medical School; the Director of the General Internal Medicine fellowship at Brigham and Women’s Hospital, where he’s a practicing general internist; and he’s a Fellow of the American College of Physicians. With colleagues Michael McWilliams, Ellen Meara, and Alan Zaslavsky, Dr. Ayanian has devoted part of his research career to looking at uninsured seniors who transition into Medicare coverage, and has examined the effects of that transition on their use of health services and self-reported health. So he seemed like a good guy to tackle the Bigger Picture questions about transitioning into drug coverage, and he was good enough to share with me his impression about this week’s study.

Interview with Dr. John Ayanian (Time: 05:10)

Q: Dr. Ayanian thanks for talking to me.

A: My pleasure.

Q: You and your colleagues have seen what seems to be a catch-up effect when uninsured older people transition into Medicare coverage. That is spending and use of health services go up and people who were uninsured and then become covered under Medicare compared to people who are insured. And you just recently reported that self-reported health also appears to improve. But we don’t have that exact comparison available for the Part D benefits because it was new to everybody, but how would you compare with the data shown in this week’s paper to what your previous work has suggested is the effect of new coverage on the use of resources and on health?

A: Our studies published earlier this year suggest that when people gain coverage, in this case, entrance into the Medicare program as a whole, there is actually an initial increase demand for hospitalizations and physician visits after a period where the uninsured prior to age 65 were using substantially less services from hospitals and physicians. So we believe that patients with chronic health conditions or need for surgery or other procedures are delaying those needs until they become eligible for Medicare, and because coverage is not fully available before age 65 we find that there are increased costs to the Medicare program. What we found in our latest study was that when uninsured near-elderly adults gained Medicare coverage there was actually a substantial reduction in their risk of developing heart attacks, hospitalization for heart failure, or severe angina over the ensuing 7 years from age 65 to 72. So for every 100 newly insured Medicare beneficiaries we found that there were about 10 fewer severe cardiovascular events over that 7-year period. This suggests that if we can expand coverage prior to age 65 there may be cost savings to the Medicare program after age 65. In an analogous way it may well be that expanded drug coverage for Medicare beneficiaries as well as for adults in their 50s and 60s, many of whom have chronic conditions but may lack insurance completely or only have catastrophic coverage without any prescription drug benefit, would help to offset other costs to the Medicare program. For example, if people are controlling their blood pressure or diabetes or cholesterol more effectively because drugs are more affordable for them we’ll find that they have fewer complications such as heart attacks, heart failures, strokes that can be very expensive to the Medicare program and quite damaging to the beneficiaries themselves. Q: A prominent feature of the Part D benefit is the donut hole,which is a microexample of beneficiaries transitioning to underinsurance and then back to full insurance if they have need enough to get to the opposite side. Do you have any impressions about what the effects of that feature of the benefit are on Medicare beneficiaries in the new program?

A: My concern is that that donut hole feature may have the greatest negative impact on patients with multiple chronic conditions who need multiple medications to treat those conditions. For example, patients with combinations of hypertension, diabetes, cardiovascular disease, those who require multiple medications that have pushed them into the doughnut hole gap where they lose Part D coverage for those medications, they may stop taking some medications partway through the year, they may reduce the doses of their medications without clearly discussing that option with their physicians and thus we may actually experience some unintended negative clinical effects as patients with those multiple chronic conditions are vulnerable to the reduced coverage in the donut hole. That’s going to be an important area for study going forward and one of the reasons why we need detailed data on the beneficiaries who are enrolled in Medicare Part D, what conditions they have and what medications they’re taking throughout the year.

Q: One of things that the authors couldn’t tell us about is whether the drugs that beneficiaries were receiving were actually appropriate drugs because there is some overuse of some drug classes and those are mixed in with those for which we might expect to see improved health outcomes.

A: That’s correct. Much of the literature suggests that when out-of-pocket costs are higher people reduce their use of both effective and ineffective drugs and as a corollary when out-of-pocket costs are lower they may increase their use of both effective and ineffective drugs and that latter part of the equation - potential increased use of less effective or ineffective medication -- is part of what the economist call the “moral hazard problem” whereby having better insurance may influence people’s behavior in ways that they would not follow if they were bearing more of the costs themselves.

Q: I’m glad you address that. Do you think moral hazard has a play in the effects reported in this week’s article and in the transition to Medicare coverage in older adults more generally?

A: Because of the high prevalence and burden of chronic disease and important health risk factors in this population of elderly Medicare beneficiaries, my intuition is that moral hazard plays only a small role in the increased utilization, and that to a large degree, as in our prior work, we will be seeing greater use of effective medications that we know from randomized clinical trials improve health outcomes. But that’s an important area for further research and studies that we can understand precisely how this Part D benefit is influencing people’s health and what impact it’s having on total cost for the health care system.

Q: Estimates for the cost of the benefits through 2014 range from about 400 to 600 billion dollars. So now we can put those estimates next to those from this week’s paper which reports average monthly saving of about $5 a month for the overall population, for an average increase in pill-days of about 4 days. So I wonder if you have any thoughts about that comparison? Does the small savings on an individual level and the small increases in use in hundreds of thousands of people over months and years add up to a big overall return on the massive investment in the benefits. Or another way I guess of asking the question is do you have any impressions about the cost-effectiveness of the benefit?

A: From this study we can’t precisely estimate the cost effectiveness. We know recent data indicates about 24 million American are enrolled in the Medicare Part D benefit so those relatively modest reductions and out-of-pocket expenses of $5 to $10 per month can translate into fairly large savings for enrollees across the whole program. Whether those savings, as well as potential savings to the Medicare program as a whole from reduced hospitalizations or reduced need for expensive procedures is worth the initial investment is going to be a very important question for researchers and policy makers to address.

Q: Great, and besides that particular question do these findings change your thinking in any other ways about future directions research into the drug benefit and Medicare more generally should take?

A: It’s crucial to understand as benefits are changed in the broad population as was done with the Part D benefit, what the impact on people’s health outcomes were both in terms of quality of life as well as survival. That’s an important area for study across the U.S. healthcare system and it really requires finer grained data on what medications people are using in the Part D benefit, and for what conditions, and to the extent that those are for conditions such as hypertension or diabetes or coronary heart disease where we know effective drug treatments can make a substantial difference reducing morbidity and mortality, we’re likely to see both health benefits for the Medicare enrollees as well as potential savings for the Medicare program as a whole. An important point made in the editorial in the Annals that accompanied this study is that all the Medicare Part D plans are required to report detailed prescription drug data to the Centers for Medicare and Medicaid Services and this could be a tremendously valuable resource for researchers and policy makers if those data were available for study the way that Medicare Part A and Part B claims are available to researchers. It will be very important going forward to expedite the release of those data so that we can develop a much more clinically refined understanding of the impact of the Medicare Part D plan on both costs and health.

Q: And let’s just take this down to the individual level of the provider. Do these findings have any implications for physicians out in clinical practice? Can you share any of your experiences both from researchers’ clinical realm about the Medicare Part D benefit and how it’s impact on physicians trying to help their patients?

A: A number of studies have indicated that it’s very important for practicing physicians to ask elderly patients about how they’re paying for their medications, whether they have any form of prescription drug coverage, what restrictions their prescription plan may have on the medications they can use, and how the elderly patients themselves are making decisions about which prescriptions to fill and which medications to take. Many elderly patients who are on fixed incomes and both before and after the implementation of the Part D benefit have been making decisions about how to spend that fixed income on prescription medications relative to other medical expenses they have as well as basic necessities such as food and housing costs. So physicians need to understand which of their patients face financial hardships from their prescriptions expenses and work with patients to discuss clinically sensible ways to help them afford their medications, potentially switching to less expensive medications within the same drug class when the clinical benefits would be comparable. We also know from the earlier experience in Part D and I found this in my practice at Brigham and Women’s Hospital that it was a difficult transition. Many new enrollees to Part D plans are trying to decide which of their medications would be covered. Many physicians putting in substantial time to counsel their patients about their decisions to enroll in Part D plans and helping them to get coverage for example by submitting prior authorization forms to their pharmacies or Part D plans. The program seems to be running more smoothly now but it still requires physicians to communicate effectively with their patients both about which of their medications they’re filling and taking and how the costs of those medications are affecting them.

Q: And given that most providers are up to their ears in responsibilities and have less and less time to spend talking to patients who already have complex medical conditions, are there any resources or strategies about which you’re aware that would help physicians have those conversations or get that information so that they can make decisions about prescription drugs that will maximize the chance that patients will get them and adhere to taking them?

A: To the extent possible, physicians should think about working with other staff such as nurses or pharmacists who can help to counsel patients more closely about their prescription needs and their adherence to medications their doctors are prescribing as well as how they’re paying for those medications. I think one of the biggest dangers is when physicians and patients don’t communicate about these issues. Patients may make decisions on their own about which medications to cut back on and those may not be the most clinically sensible decisions in terms of the impact on their health. And so to the extent that physicians can work with other office staff and pharmacists in their communities to help understand the needs that patients have, the physicians can continue to focus on what they do best, choosing the most cost-effective medications for their patients and their health needs.

Q: Dr. Ayanian thanks so much for talking to me.

A: Thank you.

That was Dr. John Ayanian, Professor of Medicine and Healthcare Policy at Harvard Medical School and a Fellow of the College, sharing his impressions about this week’s lead feature article entitled “The Effect of the Medicare Part D Prescription Benefit on Drug Utilization and Expenditures,” by lead author Wesley Yin of the University of Chicago.

Comparative Effectiveness of Drugs Used to Treat Osteoporosis (Time: 17:19)

Our other feature article this week is a systematic review of the comparative safety and efficacy of drugs used to prevent fractures in people with low bone density or osteoporosis.

In the article, lead author Catherine Maclean and her colleagues from the RAND Corporation in Santa Monica, California, performed an exhaustive search of the English-language medical literature to identify all published randomized trials and meta-analyses addressing the benefits and harms of available osteoporosis treatments.

They organized their findings by benefits and harms, and by drug class and fracture site. I’m going to describe those findings to you, but they’re going to be difficult to keep straight just by listening to them, so if you’re interested in learning more, Table 2 of the article in the print journal is very nicely organized, and may make the findings of the review more clear, organizing them by drug class, specific agent, its effects on risk, type of fracture, and the sources and strength of the evidence.

For benefits, they found good evidence from randomized trials and meta-analyses that vertebral fractures are effectively prevented with bisphophonates; estrogen; PTH [1-34], also known as teriparatide; and raloxifene, a selective estrogen receptor modulator. They found fair evidence that vertebral fractures are prevented by calcitonin, and good evidence that tamoxifen and calcium supplementation had no effects.

The evidence for effects on non-vertebral and hip fractures was more mixed.

For non-vertebral fractures, such as those of the wrist or ankle, they found good evidence that most bisphosphonates and estrogen are effective; fair evidence that PTH [1-34] is effective; and good evidence that raloxifene, calcitonin, and calcium do little to prevent those fractures.

And for hip fractures they found good evidence that alendronate, risidronate, and estrogen are effective; fair evidence that zoledronic acid, a long-acting bisphosphonate, is effective; and good evidence that raloxifene, tamoxifen, and calcium do nothing to prevent those fractures.

The investigators look at two other treatments: vitamin D and testosterone. Vitamin D is commonly used as an adjunct to calcium supplementation in the preventive treatment of osteoporosis, and the findings supporting the practice differed by study, preparation, and dose. Four meta-analyses suggested that standard vitamin D supplements – that is, vitamin D2, D3, or 25-hydroxyvitamin D supplements – were ineffective for preventing any fractures, while 1 suggested that they were effective for preventing non-vertebral and hip fractures, the difference in findings may be an effect of dosing below 800 IU in negative trials. In contrast, three meta-analyses suggested that vitamin D analogues, specifically 1,25-hydroxyvitamin D and 1-hydroxyvitamin D, reduced the risk of verterbral fractures, while another meta-analysis showed they reduced non-vertebral and hip fractures in some but not all cases.

Testosterone remains an intriguing treatment option for osteoporotic men, given the association between lower testosterone levels and lower bone density, but the authors couldn’t identify any randomized trials or meta-analyses addressing the effect of the treatment on men’s fracture incidence. Of trials limited to populations of men, 1 showed a reduction in hip fracture incidence with risedronate, 1 showed a reduction in vertebral fracture incidence with calcitonin, and one showed a reduction in all fractures with PTH[1-34].

The authors’ review of evidence for adverse effects confirmed generally established and widely known effects. Estrogen and selective estrogen receptor modulators increase thromboembolic events; bisphosphonates increased the risk for esophageal ulceration and mild upper gastrointestinal symptoms, such as nausea and heartburn; and case reports suggest the occasional occurrence of osteonecrosis of the jaw in patients who received intravenous bisphosphonates who have cancer or dental surgery.

The authors restate these many findings in their conclusions, but they make the useful point that the findings highlight what we don’t know at least as much as what we do, a common contribution of these systematic reviews that we publish. The majority of research has been performed in postmenopausal osteoporotic women, they say, so we know less than we should about preventing fractures in other high risk groups, such as older men, people with chronic renal insufficiency, and people on chronic glucocorticoids. The negative findings for calcium were often reported in the absence of information about adherence; taking calcium supplements in quantities sufficient to prevent fractures requires frequent dosing that is difficult for patients, so it could be that the negative findings were a consequence of underdosing. And perhaps most importantly, there’s no evidence that one agent or drug class is superior to another for preventing fractures, and it’s unlikely that such studies will be performed unless they are required as part of the government approval process for osteoporosis treatments. So the authors do not say, but they imply, that the choice of agent or combination of agents needs to be based on clinical and other variables, such availability, cost, ease of administration and likelihood of adherence, and patient preference.

Other Articles In This Week’s Issue (Time: 22:17)

Other articles in this week’s issue include a decision analysis of when to start antiretroviral treatment for HIV infection, suggesting that earlier treatment, for example at CD4 counts of 350 or 500 rather than at counts of 200 cells/mm3, increases life expectancy and quality-adjusted life years compared, especially in patients with higher viral loads.

We have two articles on the invasive treatment of acute coronary syndromes. The first is a systematic review suggesting that available evidence does not yet support the common practice of routinely sending everyone with non-ST segment elevation acute coronary syndrome for percutaneous interventions; and the second is a perspective on the relative efficacy and safety of drug eluting stents in patients undergoing those interventions.

We have a systematic review of the evidence for screening pregnant women for bacterial vaginosis, with an accompanying guideline concluding that that evidence is insufficient to conclude anything about the benefits and harms of screening in women at high risk for preterm delivery, but recommending against screening women at low risk for preterm delivery.

And have you hugged your colonoscopist or your colorectal surgeon lately? If you do, slip ‘em a copy of this month’s In The Clinic, on screening for and prevention of Colorectal Cancer.

Well, that’s it for today.

Our closing credits are the usual ones.

Our theme music is by Brian Poole and Kwesi Marles.

We want to hear from you; send comments and feedback about this project and the journal to podcast@annals.org.

Technical support for the summary was provided by Andrew Langman, Neil Kohl, and Beth Jenkinson.

Check back in two weeks for our regularly scheduled February 19, 2008 issue.

I’m Michael Berkwits, and thanks for listening.

[ Return to Podcast Home Page ]