Last week’s (April 12th) blog post was an overview of the highly idiosyncratic nature of the pharmaceutical marketplace and the pricing of pharmaceuticals. Today’s The New York Times carried a front page story by Andrew Pollack, The New York Times biotechnology industry reporter, about how doctors may increasingly be influenced by treatment costs when advising patients.1 Interestingly, while the focus of the article is on the general impact of health care costs on physicians’ decision-making, other than a passing mention of MRIs, the only other examples of cost are drawn from the world of pharmaceuticals– Avastin vs. Lucentis for macular degeneration, Aloxi for chemotherapy-related vomiting and, everyone’s favorite target of late, Solvadi for Hepatitis C. Pollack’s article also discusses the decision by some physician groups, such as the American Cardiology Society (ACS), to rate the value (that is economic value) of treatments in their joint clinical practice guidelines and performance standards. The article notes that the ACS committee that wrote the new policy recommended using QALYs (quality adjusted life years, a widely used standard in health economics) as a principal metric in measuring a treatment’s cost effectiveness. So, given the continuing public attention to pharmaceutical prices, this week I will briefly discuss pharmacoeconomics, which should have, but only in some cases does have, an effect on pharmaceutical prices.
First, QALYs, as mentioned in Pollack’s article, attempt to provide a measure of two variables in a single unit: the first is the effect of a treatment on the actual duration of life, while the second is the quality of life provided by the treatment. This can become technical and complicated, and those who wish a more detailed introduction to QALYs might want to start with an article by Luis Prieto and Jose Sacristan, Problems and solutions in calculating quality-adjusted life years (QALYs).2 However, for this discussion, a relatively basic understanding of QALYs should be sufficient. For example, a treatment that resulted in the five-year survival of end-stage cancer patients, but only in a coma, would not be providing a significant benefit in terms of quality-adjusted life years. A treatment that extended the survival of end-stage cancer patients in reasonably good health for six months might provide .5 quality adjusted life years. If the purpose of assigning a value in QALYs is simply to compare the benefits of two treatments by converting both to a common denominator, then it isn’t necessary to worry about how much a QALY is worth in monetary terms. A treatment that results in an increase of 2 QALYs is presumed to be better than a treatment that results in an increase of 1.5 QALYs.3 Of course, if QALYs are going to be used in an economic analysis, then a value in dollars and cents needs to be assigned to a QALY, and the Pollack article states that the cardiologists would consider treatments that increased the average outcome for patients by a full QALY at a cost of $50,000 to be “high value” while treatments costing $150,000 or more per additional QALY would be considered “low value.
So QALYs are a unit by which the effect of different drugs or other treatments can be measured on a common scale. It is a useful tool for pharmaceoeconomics. Pharmacoeconomics is the study of the economic value or impact of a pharmaceutical. Sometimes pharmacoeconomic analysis requires incorporating a unit such as QALYs, but in many cases QALYs are really not essential to an economic evaluation of a drug’s benefits and risks. For example, from a health insurance company’s perspective, a drug which costs $84,000 (like Solvadi) must be considered in light of the health care costs of persistent Hepatitis C infection, as well as the cost of alternative treatments that only prevent a percentage of patients from progressing to later stages of liver disease. The average annual costs of treating patients with Hepatitis C are estimated to be approximately $24,176,4 costs which can continue for a significant number of years for the average patient. For a health insurance company, Solvadi may well produce a net savings in expenses, without the need to use QALYs in the calculation. Of course, Gilead knew that when they set the price of Solvadi. A treatment may be expensive in absolute terms, but if it actually reduces overall healthcare costs, it is worthwhile in pharmacoeconomic terms and will be worthwhile for insurance companies to cover.
In one of the other examples highlighted in Pollack’s article, the relative cost of Avastin and Lucentis raises a somewhat more complicated pharmacoeconomic issue, as well as a terrific example of the way that pharmaceutical prices are determined. Both drugs were developed and marketed by Genentech. Both drugs work by targeting the vascular endothelial growth factor, which is a key part of the pathway for blood vessel growth. In many cancers that take the form of solid tumors, such as colorectal cancer, stopping or greatly reducing the growth of new blood vessels stops or slows the growth of the tumors. In “wet” age-related macular degeneration (AMD), the abnormal proliferation of capillaries (small blood vessels) to the retina causes increasing damage to the retina and vision loss and blocking VEGF slows down capillary growth and vision loss. Avastin, an antibody that binds to VEGF, was developed to treat colorectal cancer and was approved in 2004. Lucentis, the antigen-binding fragment of an antibody that binds to VEGF, was approved for the treatment of macular degeneration in 2006 and costs just under $2,000 a dose. Avastin, when used to treat colorectal cancer is priced at $50,000 to $100,000 per patient per year, but the cancer dose is many times greater than the dose injected into the eye to treat macular degeneration. A pharmacist can provide many doses for AMD from the amount used for a cancer treatment, and charge about $50 per dose. It should go without saying that Genentech would rather not have doctors using Avastin to treat their macular degeneration patients instead of Lucentis. It should also be obvious that the price of Lucentis has nothing whatsoever to do with the cost of producing that amount of drug, as it is unlikely that it is significantly different from the cost of producing Avastin.
So, if Avastin costs $50 and Lucentis costs $2000, what is complicated about the pharmacoeconomics? First, the effectiveness of both drugs is equal, so in terms of effectiveness the pharmacoeconomic choice is overwhelmingly Avastin. But the safety of the two drugs may not be equal- there is a potentially greater risk of adverse effects with the use of Avastin than with Lucentis, according to a review article by Pasquale Ventrice and others.5 It appears to be a small increase in risk and does not include any increased risk of death. A large review of 145,000 Medicare patients treated for macular degeneration found NO difference in heart attacks or death and no statistically significant increase in bleeding or stroke.6 However, the cost of treating a patient monthly (as it was approved and labeled) would be about $23,400 for Lucentis, while Avastin monthly treatment would cost only $600, a difference of almost $23,000 per patient per year. Treating 1,000 patients with Avastin instead of Lucentis would save about $23 million. So if there were one additional serious adverse event per 1,000 patients treated with Avastin instead of Lucentis,7 we would have to assign a cost of $23 million to that event in order to offset the basic difference in price. It is clear that the risk of additional non-fatal systemic adverse events with Avastin instead of Lucentis is extremely unlikely to justify the cost difference.
It is completely understandable why Genentech undertook to develop a different anti-VEGF agent for macular degeneration, rather than entering clinical trials with Avastin. Using an antibody fragment rather than the whole antibody might have had clinical significance, although it turned out not to be the case. But more importantly, from the Company’s perspective, it was an effort to keep from competing with their own product at a very low cost. Why do they charge about 40 times more for a dose of Lucentis for AMD as the cost of a dose of Avastin prepared for AMD? Because they hope they can. However, the pharmacoeconomics says it absolutely is not worth it. In a future post, I will look at how drug marketing succeeds in the face of such evidence.
1 Andrew Pollack, Treatment Could Influence Doctors’ Advice, New York Times, April 18, 1014 at A1 (also available online at http://www.nytimes.com/2014/04/18/business/treatment-cost-could-influence-doctors-advice.html?hp&_r=0)
2 Published in Health Qual Life Outcomes, in 2003 and available online at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC317370/ (visited April 18, 2014).
3 Of course, QALYs used even in this way are based on assumptions about general preferences, whereas an individual’s preferences may differ. A particular patient might prefer 3 years of being alive, alert, and bedridden to 2 years of being alive, alert, and able to go about the activities of daily life, though the latter outcome would likely be assigned a greater value in QALYs.
4 Maria Seyrig, The Economic Cost of Advanced Liver Disease, Henry Ford Health Systems, November 7th, 2011, online at http://www.henryford.com/body.cfm?id=46335&action=detail&ref=1465.
5 Pasquale Ventrice, et al, Anti-Vascular Endothelial Growth Factor Drugs Safety And Efficacy In Opthalmic Diseases, J Pharmacol Pharmacother. Dec 2013; 4(Suppl1): S38–S42, online at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3853666/?report=printable
6 Curtis LH, Hammill BG, Schulman KA, Cousins SW, Risks of mortality, myocardial infarction, bleeding, and stroke associated with therapies for age-related macular degeneration, 128(10) Arch. Ophthalmol. 1273-9 (October 2010), abstract at http://www.ncbi.nlm.nih.gov/pubmed/20937996.
7 For yet another look at the risks, albeit another inconclusive look, see the CATT group study in the May 19, 2011 New England Journal of Medicine, Ranibizumab and Bevacizumab for Neovascular Age- Related Macular Degeneration (364:20) NEJM 1897).