Research Focus
A pharmaceutical firm's ability to market the output of its product pipeline effectively is central to its competitive position. As managed care organizations increasingly become the providers of healthcare in the United States, manufacturers have been forced to reconceptualize their marketing and promotion strategies. At the same time, a growing number of firms are compelled to seek merger opportunities partly to realize new economies and downsize expensive administrative and research functions. Further, many pharmaceutical firms are seeking wider opportunities to demonstrate, through the use of increasingly sophisticated pharmacoeconomic tools, that the use of their products leads to enhanced productivity of resources and reduced healthcare costs to society. All this speaks to a rapidly changing pharmaceutical marketplace.
POPI teams are conducting groundbreaking research into this changing marketplace, bringing together several disciplines: economics, marketing, management science, and medicine. POPI investigators examine whether a rational basis for drug pricing can be established. Teams explore the effects of generic entry on the marketplace. Questions are explored that relate to the effect a firm's order of entry and promotion strategy for a new product might have on market success, and whether pressures for enhanced industrial competitiveness can encourage better uses of pharmaceuticals to improve worker productivity. New Project
The U.S. pharmaceutical industry's innovative output consists primarily of a small number of new drugs that receive approval from the Food and Drug Administration, along with gradual improvements to existing drugs. While their importance for their wellbeing of patients is undisputed, there is no agreed-upon method for evaluating the impact that these new or improved products have on patients' welfare. Absent such a framework, public policy toward the pharmaceutical industry turns on an incomplete vision of the costs and benefits of government regulation.
The research team for this new project begins with a simple premise: the calculation of the welfare benefits that arise from pharmaceutical innovation would contribute to the development of a more comprehensive analytical framework for evaluating government regulation of the pharmaceutical industry.
Research method. In recent years, several relevant econometric models and techniques have been introduced and improved, significantly increasing the feasibility of measuring patient welfare benefits. The research team is reviewing these recent advances in the state of the art and will be analyzing whether these methods can be applied to the pharmaceutical industry. More precisely, the team will complete a feasibility study examining whether the patient welfare benefits arising from new product innovation can be calculated by executing three related tasks (described in greater detail below):
Through the application of a well-specified demand model, the underlying preferences of drug "decisionmakers" (patients or, more likely, physicians acting on behalf of patients) may be estimated by a careful analysis of the observed market shares, prices, and attributes of all drugs within a given therapeutic category. These estimates of preferences are a necessary step in calculating the social benefits to patients from these new drugs.
To apply this demand analysis to the evaluation of social welfare, the institutional framework underlying pharmaceutical decisionmaking must be well understood and taken into account: the precise role of doctors in determining molecular choice under various institutional regimes must be examined; moral hazard on the part of patients must be incorporated into the analysis, because some portion of consumers are insured for their pharmaceutical expenditures; and the reasons for variation in the mark-ups charged by pharmacies must be considered. The feasibility study will review these and related issues, assess the degree of difficulty they pose for the purpose of the calculation of patient welfare benefits, and suggest ways to address these problems in future research.
Certain critical components of the benefits accruing from pharmaceutical innovation may be impossible to measure. For example, it is difficult to measure, in an economically oriented framework, the private or social value of "quality of life" or "prolonged life." But one may be able to answer a more modest question: how large are the measurable benefits to new product introduction, where those benefits manifest themselves in the "willingness-to-pay" for novel products? In other words, how much are patients willing to pay to have products introduced earlier rather than later into the market?
Evaluating such patient benefits requires the development of a sophisticated and flexible demand model incorporating the consequences of the institutional features of pharmaceutical decisionmaking. The research team will be reviewing the related issues and evaluate potential models or solutions that overcome the difficulties posed.
Future work. The feasibility study will result in a paper, projected for completion by the end of 1996, that includes the analysis described above. The research team foresees the present project paving the way for future research that will estimate product demand in specific therapeutic categories (such as cardiovasculars) and will evaluate the welfare gains arising from the introduction of new products.
Past POPI research assessed the feasibility of incorporating quality changes into price indexes, focusing on anti-hypertensive drugs to determine whether a reasonable, rational basis for drug pricing could be established. It is a difficult question to answer, especially given that pharmaceuticals have rarely been priced in relation to their incremental or short-run costs of production.
The market for anti-depressant, anti-ulcer, and anti-hypertensive prescription drugs has experienced dramatic growth in the last decade, particularly since the late 1980s. The official price indexes for these drugs calculated and published by the U.S. Bureau of Labor Statistics (BLS) suggests that over this same period prices have risen at a very rapid rate.
Following up on that earlier research, the team is examining the reliability of the BLS official price index, and then will offer alternative calculations the team believes measures price movements with greater accuracy.
Background, research method, and data. Several developments in the anti-depressant therapeutic class of drugs are worth noting. First, during the 1980s a number of older-generation (tricyclic) anti-depressant drugs went off-patent and, after the patent expiration, both branded and generic versions were sold. Since the BLS procedure treats generic drugs as entirely new products rather than linking them to their branded antecedents, the BLS price index likely overstates price growth. Using monthly data from 1980, the researchers will attempt to measure this overstatement in price growth, building on the framework established by Griliches and Cockburn.
Second, an entirely new class of drugs was introduced in late 1988 into the anti-depressant therapeutic class. Prozac was the first drug in the selective serotonin reuptake inhibitor (SSRI) class. While the efficacy of Prozac in treating depression may not have been demonstrably superior to that of the previous generations of anti-depressant drugs, Prozac had a much superior side-effect and adverse drug interaction profile. As a result, general practitioners and internists now felt safe in prescribing anti-depressants where before they frequently would refer patients to psychiatrists. With the entry of Prozac-a more expensive drug than its predecessors-the market underwent dramatic growth. Additional SSRIs have since entered the market, including Zoloft, Paxil, Effexor, and-soon-Serzone.
The team's research will focus on how one incorporates Prozac and subsequent SSRIs into price index calculations, taking particular account of the fact that the SSRIs are generally viewed as products of higher quality than their tricyclic antecedents.
For the anti-hypertensive drug segment of the U.S. market, IMS data already have been made available to the research team. These data are estimated by IMS to cover more than 80 percent of the U.S. pharmaceutical market.
Preliminary findings and conclusions. From previous research, the team has found that price trends of the best-selling anti-hypertensive drugs in the sample correspond reasonably well with price trends for cardiovascular drugs in total.
The early results also show a positive and statistically significant relationship between reduction in diastolic blood pressure and cost per day. In addition, the results thus far indicate that cost per day falls as the number of tablets in the presentational form increases. Third, while the coefficient of their efficacy estimate is consistent with expectations, the finding on tolerability is not. Lower drug tolerabilities were associated with higher costs per day.
How can a firm's order of entry and promotion strategy for a new product affect its success in a rapidly evolving drug marketplace? That was the question addressed by POPI research detailed in last year's report on the economics of the relatively new market for anti-ulcer drugs. A new research project, just commencing as these words are written, takes off from that research.
The market for prescription anti-ulcer medications is an enormous one, the largest of any major therapeutic prescription drug class in the United States. Two of the top ten all-time best-selling prescription drugs in any class are anti-ulcer medications, namely, Tagamet and Zantac. Today, the anti-ulcer drug market also includes Pepcid and Axid, as well as Prilosec.
Background. U.S. patent protection for Tagamet expired in May 1994. Given the large market size for Tagamet, it is not surprising that generic entry was rapid and dramatic. Within five months of patent expiration, Tagamet lost 75 percent of its market share.
Under provisions of the Waxman-Hatch Act passed by the U.S. Congress in 1984, in some cases prescription drug patent holders can obtain an additional three-year exclusivity patent extension if they obtain approval from the Food and Drug Administration (FDA) for a new indication. Before patent protection expired, the makers of Tagamet began seeking FDA approval for an over-the-counter (OTC) version for the treatment of heartburn. In July 1995, after receiving FDA approval, Tagamet went OTC. The OTC version of Tagamet now has patent protection until 1998, although the prescription (Rx) version has no patent protection.
Meanwhile, although patents on their Rx versions had not expired, Pepcid and Zantac also decided to seek FDA approval to enter the potentially lucrative OTC market. Pepcid won approval and went OTC in September 1995; Zantac is expected to win approval and enter the OTC market in April 1996.
Research focus. The focus of this research project is to determine what can be learned from the Rx market that helps predict market success in the OTC environment. Several research questions are posed.
The manufacturers of Tagamet (SmithKline Beecham), Zantac (Glaxo), and Pepcid (Merck) also have strong branded traditional antacid products on the OTC market: Tums and Gaviscon (SmithKline Beecham), Rolaids (Glaxo-Warner Lambert), and Mylanta (J&J Merck). Thus, not only does the possibility exist for cannibalization of Rx sales by OTC products, but the new OTC brands also can cannibalize sales of traditional OTC antacids. For those companies with relatively small Rx and/or OTC brands, the opportunities afforded by OTC entry are greater; for potential lost Rx and/or OTC sales from cannibalization are smaller.
In this project, the research team will examine closely the factors affecting market success in the OTC and Rx markets both separately and combined. Although the initial focus will be on the U.S. marketplace, the rather different experiences in Scandinavia (particularly Sweden) and the United Kingdom will also be examined.
Increasingly, researchers are examining the cost impact of improvements in worker productivity that result from some of the new breakthrough pharmaceuticals. Can pressures for enhanced industrial competitiveness encourage better uses of pharmaceuticals? Since most Americans traditionally have received their health insurance through their employment, a positive answer to this question could have significant ramifications for the cost of healthcare in the United States.
POPI's ongoing research is breaking new ground in the growing field of "outcomes research." This work explores whether products effect favorable health outcomes measurable in explicit economic terms. Broad studies of this sort offer an opportunity to quantify the economic and health impacts of medical treatments and pharmaceuticals on consumers of healthcare and society at large. Early POPI outcomes work focused on the costs of mental depression-with wide press coverage of the initial results throughout the United States. For more than a year, the research team has shifted its attention to building a research program that focuses on workplace performance as a health outcome measure.
Expectations are that research of this type will play an increasingly important role in the successful marketing of new drugs. Firms that develop an understanding of healthcare costs and benefits associated with their new drugs will be empowered to price and sell products, in both new and existing therapeutic treatments, with new and powerful competitive information. Firms are increasingly incorporating outcomes research early in their clinical research, making it a valuable tool to assess the competitiveness of products in the pipeline.
Research method. The approach to this research is to attempt to measure changes in productivity at work in relation to changes in health status. Initial research focused on depression and is now being extended to other disorders and conditions.
Preliminary findings and conclusions. The estimates developed in the initial stages of this study suggest that the economic burden of depression in the United States in 1990 was approximately $43.7 billion. This total breaks down as follows: direct costs, 28 percent; mortality costs, 17 percent, and morbidity costs, 55 percent.
In the initial stages, the team went on to explore workplace costs of depression, that is, the economic losses from those who continue to work although impaired by the illness. The study estimates suggested that approximately 2.25 percent of available worktime in the labor force is adversely affected by the symptoms of depression each year. With a conservatively estimated average reduction in performance during this period of about 20 percent, this results in costs totaling $12.1 billion due to employee impairment at work.
The other component of workplace costs analyzed by the team in detail concerned all the time lost from work due to absenteeism because of the illness. In the case of depression, approximately 0.44 percent of all available worktime each year was estimated to be lost due to increased absenteeism during episodes of this illness. By definition, this component of indirect workplace costs results in a 100 percent impairment. This leads to additional losses in productivity totaling about $11.7 billion. Although the extent of absenteeism computed in the case of depression appears to be modest, it nonetheless implies substantial workplace costs.
The research team went on to explore why depression has not received more attention from the medical and public health communities (particularly primary care physicians), employer groups, as well as society at large. The team developed a comparative illness framework for contrasting depression with other diseases using several criteria: economic costs, prevalence, distribution of sufferers, mortality, recognition, and treatability. The framework revealed that although depression is widespread throughout the population and very costly, it differs in many respects from other major illnesses. While the costs of depression are comparable to a number of other illnesses, those costs are mostly indirect in nature. In addition, they are likely comprised more of the morbidity and less of the mortality costs than is true of the other major illnesses.
Further, while between 80 and 90 percent of individuals suffering from a major depressive disorder can be treated successfully (according to the National Institute of Mental Health), only one in three with the illness ever seek treatment. The poor recognition and often inappropriate treatment of depression leads further to an even smaller percentage of those suffering receiving proper care.
Estimation of the cost implications of depression for employers yields two conclusions. First, employers as a group have much to gain from increased awareness and treatment of the illness. Second, employers do not directly bear all the costs of depression, nor can they be sure to realize all the benefits of investments in improving the mental health of their workforces. While the former suggests that employers as a group have an incentive to participate in addressing this issue, the latter raises some doubt about any individual employer's incentives to invest in alleviating the workplace costs of depression.
One important aspect of outcomes research has been the development of metrics that allow the differentiation among therapeutic interventions whose effects may be subtle and relate to what has come to be called the "quality of life." While major advances have occurred in the validation of survey instruments that have addressed functional health status changes, the acceptance of these kinds of measurements beyond the research community has been limited, largely because of their admitted subjective nature.
The research team observed that nearly 28 percent of the $44 billion total annual cost identified for depression in the United States could be explained by the existence of a body of depression sufferers whose illness, while not preventing them from reporting for work, does cause a significant reduction in their level of work performance. From this observation, the team hypothesized that it could be shown for depression-and, conceivably, for a number of other conditions-that improved health status following treatment leads to measurably improved work performance.
The team had the opportunity to test this hypothesis in a specific context as part of a pharmaceutical firm's clinical trial. The content of the work related to an evaluation of one of the latest generation of antidepressant medications in contrast to one from an earlier generation of drugs. Many believed that the differences between these two drugs in terms of efficacy would be small, but that newer drugs with fewer side effects would be associated with "better quality of life."
The team constructed a set of work performance indexes on the basis of questions that had been incorporated into the clinical trials that asked explicitly about the subject's performance at work. The researchers then estimated empirically the relationship between changes in depression status (as measured by the Hamilton Depression Rating Scale) and the change in the team's various work performance indexes. Multivariate regression models were used to examine this relationship under a wide range of assumptions, which enabled the research team to define quantitatively the relationship between depression status and subjective work performance.
The research team has since been granted access to a database of employment and medical claims records of nearly 6,000 insurance claims processors (constructed with careful attention to the anonymity of subjects). The database includes measurements of work performance (number of claims processed per hour), absenteeism, demographics, salary, and the employees' own medical claims histories. The productivity data are daily, and cover eighteen months of records. The team has begun to explore the relationship between treatment for an assortment of acute and chronic conditions (including carpal tunnel syndrome, back pain, migraine, depression, nasal allergies, and hypertension) and objectively measured work performance. Preliminary analyses find statistically significant improvements in work productivity following treatment for some of these conditions, but not for others. In some cases, productivity declines substantially while the worker is undergoing medical treatment, but then increases sharply after treatment ends (or maintenance treatment commences).
Future work. The team continues to expand its research to measure objectively productivity improvements at work in response to treatment. The goal continues to be to quantify treatment in dollar terms related to wages or value of goods and services produced, so as to capture the sense of what others have attempted to measure in terms of functional status improvement or quality-of-life change.
Future efforts are likely to include extensive analysis of a database of insurance claims, as well as the identification of employer sites where it may be possible to construct analogous types of databases from a number of different kinds of occupations that will allow the examination of a possible correlation between health status and work performances. The team will explore whether a model can be developed that can be tested in future prospective studies.
Please send comments and suggestions to popi-www@mit.edu