Annual Report of Research and Educational Programs


This section summarizes ongoing POPI research projects, focusing on findings over the last year. More extensive descriptions of many of these projects may be found in earlier POPI annual reports.

Exploring the Role of Publicly Funded Research
in Private-Sector Pharmaceutical Innovation

Iain Cockburn (economics)
Rebecca Henderson (management)

Discussion of the impact of publicly funded research on the private sector normally focuses on the impact of publicly generated knowledge on private-sector productivity. The POPI team's focus is on the impact of publicly funded research on the way in which research is organized within the private sector. The team draws on in-depth qualitative interviews conducted at 11 major pharmaceutical firms.

Recent findings and conclusions. The public sector makes two distinct contributions to private-sector research. On the one hand, the public sector contributes specific research results, methods, materials, trained personnel, and so on, which are inputs to the private sector's research production function. On the other hand, in order for the private sector to gain access to these inputs, it not only has to make complementary investments in similar research, but must also engage in and sustain a complex conversation with the broader publicly funded scientific community. This changes the scope and nature of private-sector research, and may significantly enhance the efficiency with which it transforms public and private R&D investment into new drugs.

Close involvement with the public sector has led private-sector research organizations to adopt many of the norms and incentive mechanisms of public science. This includes the use of incentive systems that stress standing in the public rank hierarchy and the allocation of resources through mechanisms that emphasize "high-respect, high-conflict" debate. Further, there is in the private sector more openness to ideas from a wide range of sources.

There is a significant correlation between bibliometric measures of the volume and nature of the interaction of firms with the public sector and both their internal resource allocations and incentive mechanisms. These bibliometric measures also correlate with private-sector research productivity, and account for a significant fraction of inter-firm variation in research performance.

The team's results have significant implications for public policy. They suggest that any attempt to measure the rate of return of publicly funded research must take account of the bi-directional nature and complex time structure of the information flow between the public and private sectors. They also suggest that any such calculation should, ideally, also take account of the impact of public-sector research on private-sector productivity. Most important, perhaps, they suggest that the ways in which public research is conducted may be as important as the level of public funding. To the extent that efforts to realize a direct return on public investments in research lead to a weakening of the culture and incentives of "open science," the productivity of the entire system of biomedical research may suffer.

Future work. This project will be continued as part of POPI's new project, "`Technology Transfer' and the Integration of Technological Innovation in the Pharmaceutical Industry".

Investment in R&D in the Pharmaceutical and Biotechnology Industries

Paul Healy (management)
Christopher Howe (management)
Paul Joskow (economics)
Stewart Myers (finance)

Through its ongoing research, the research team has built a financial simulation model of pharmaceutical research and development. Its major applications include calculation of the net present value (NPV) of investment in R&D, calculation of the cost of developing a successful drug, and calculation of the changes in the cost of capital from drug discovery to the end of the drug's commercial life.

This model connects the tools of modern finance to a detailed description of the costs, risks, and returns of pharmaceutical R&D. It is a prototype for financial analysis of R&D investment and a way of understanding the financial structure and performance of the pharmaceutical industry.

Recent findings and conclusions. In a paper titled "A Life-Cycle Financial Model of Pharmaceutical R&D," members of the team report several findings based on their base-case assumptions. These include:

These results help illustrate the implications of the financial model and clarify the financial characteristics of the pharmaceutical industry.

Future work. Several additional papers are being prepared by the research team. One paper-geared toward a management science audience-will present the Monte Carlo simulation constructed to model a drug development program. Another paper-targeted at an industry audience-will revisit the cost of capital, concentrating on its lifecycle behavior. The applicability of the model to other industries will also be explored.

Benchmarking the Manufacture of Pharmaceuticals

Charles Cooney (chemical engineering)
G.K. Raju (chemical engineering, management)

For several years, POPI researchers have explored opportunities to reduce manufacturing costs substantially by improving efficiency and effectiveness through "benchmarking"-a self-improvement process that allows a manager to compare performance to that of competitors or related firms on a continual basis. One significant result of the important contacts made through this work has been the formation of the Consortium for the Advancement of Manufacturing in Pharmaceuticals (CAMP).

Recent findings and conclusions. The research team continues to observe manufacturing practices at a number of firms. The team has found that manufacturing choices often are made as a consequence of R&D and marketing choices, rather than together with them. The defensive mindset this engenders in the manufacturing organization, coupled with firms' view of manufacturing as a cost, hinders the organization from taking a larger systems view of its performance.

Significant cost savings may be found in reducing cycle time, which includes many activities that add no value. There is also substantial room for improving the supplier management process, as indicated by high raw material inventory levels, high quality control test and release times, and a large number of suppliers. Inventory management also represents a significant opportunity for cost-savings.

Future work. Future work will focus on CAMP activities. Each CAMP project will have a benchmarking component, drawing from the research conducted under the aegis of this project team.

Improving the Effectiveness of Pharmaceutical R&D Project Management

Thomas Allen (management)
Ralph Katz (management)

The ongoing POPI-sponsored research into pharmaceutical industry project management addresses two key questions of drug development: How can project management best be implemented in this industry? Can more effective project management lead to accelerated and more cost-effective drug development? The team continues to assess the different project management strategies being tried in various firms.

Recent findings and conclusions. To no one's surprise, the team has found the technical (or clinical) competence of project managers to have an important effect on project performance.

In studying key decisions during product development, three classes of decisions have emerged as having an especially important relation to project performance: project initiation decisions, clinical decisions, and budget decisions. First is the decision to begin development, which is best left to departmental management. Very early clinical decision such as the choice of therapeutic indications, dosages, and delivery systems are also best left to departmental management. When these decisions are made by project management, project performance is lower. On the other hand, clinical decisions that must be made later during development are best made by project management. Fortunately-and quite logically-these decisions are more likely to be made by project management when there is a more clinically competent project manager. When the project management is seen as more technically competent, departmental management is more likely to have budget control. This is probably due to the fact that less technically competent project managers were chosen for their administrative experience and ability and the organizations are more willing to trust such managers with budget.

While these results are consistent with expectations, they do reinforce the value of belated efforts of firms to attract more highly educated and capable scientists and physicians into project management. The industry's first efforts in project management involved appointing administrative leaders to the position of project manager. This can be regarded as a half-hearted attempt at project management. If earlier academic researchers found lightweight project management in the auto industry, the situation in pharmaceuticals was, at best, featherweight project management. Fortunately, firms have learned that this does not produce the needed reductions in cost and development time. In response, they are strengthening the position of project managers, and bringing them up to at least welterweight and perhaps middleweight status. The team's evidence shows this to be the right direction, but more remains to be done.

Future work. The relationship between decisions made by project or departmental management and project performance will be elaborated more fully in a forthcoming working paper.

Evaluating Patient Welfare Gains from Pharmaceutical Innovation

Scott Stern (management)
Manuel Trajtenberg (economics)

There has been little evaluation of the economic impact of new drugs as they diffuse into the marketplace. And while studies of the efficacy and safety of new molecules are commonplace, few studies examine and attempt to quantify the economic implications of new drugs in terms of the realized level of patient welfare. Absent a measurement framework that can provide insight into the patient benefits arising from new product introduction, regulation of the pharmaceutical industry-as well as of healthcare in general-turns on an incomplete vision of the costs and benefits of that regulation.

The research team for this project has summarized and reviewed critically the theoretical, institutional, and empirical issues involved in the calculation of patient welfare benefits from pharmaceutical innovation. The goal has been to provide a model that frames how to approach several of these challenges in an empirical setting.

Recent findings. The researchers began with a baseline discrete choice model of pharmaceutical choice, grafting on to it several of the institutional features of pharmaceutical decisionmaking. In particular, they built a multi-stage economic model of the prescription process, focusing on the specific authority and financial responsibilities undertaken by physician, patient, and insurer.

Future work. The next step will be to "filter out" from the model extraneous factors associated with agency problems and institutional distortions, and recover the true parameters of the utility function of patients (including their price sensitivity, if possible). This requires specifying the type of data that may help identify those parameters. Preliminary work has begun, with promising results.

Price Indexes for Anti-Depressant Drugs: Generics, New Products, and Quality Change

Ernst Berndt (economics)
Iain Cockburn (economics)
Zvi Griliches (economics)

Past POPI research assessed the feasibility of incorporating quality changes into price indexes, focusing on anti-hypertensive drugs to determine whether a reasonable basis for drug pricing could be established that relate the price of a drug to its quality attributes. Prices of patent-protected products are difficult to model, since prices usually bear little relationship to their incremental or short-run costs of production.

Building on earlier findings, the team has extended this research methodologically, as well as to an entire different class of drugs, namely, the anti-depressants.

Recent findings and conclusions. Working with data from various sources, the research team has found that the U.S. Bureau of Labor Statistics (BLS) tended to oversample older products and undersample newer drugs; moreover, generic drugs were consistently undersampled. Since generic drugs experienced substantial price reductions while branded drugs tended in general to increase prices over time, this undersampling of generics by the BLS resulted in a very marked upward bias in its price index. The researchers also have concluded that the new procedures adopted by the BLS in linking the prices of generic goods to their patented antecedents would likely have a significant impact; had these procedures been adopted in 1980, measured price inflation would have been less than half that reported by the BLS. Finally, the researchers have reported results on the estimation of a hedonic price model that linked drug quality characteristics to drug prices. When these quality improvements were taken into account, price increases were again smaller than those reported by the BLS. Until 1994, these quality improvements resulted in a relatively modestly upward-biased BLS price index, but since 1994 their relative importance has increased.

Future work. Research plans for the future involve constructing and then estimating econometrically a model that encompasses the entire set of anti-depressant drugs, and sub-models for the SSRIs, MAOIs, TCAs, and-within the TCAs-pioneer brands, secondary brands, and generics. Particular attention will be focused on factors affecting the diffusion of the newest generation of SSRIs.

The Economics of the Anti-Ulcer Drug Market

Ernst Berndt (economics)
Linda Bui (economics)
Charles King III (economics)
Glen Urban (marketing)

Past POPI research examined how a firm's order of entry and promotion strategy for a new product affect its success in the rapidly changing drug marketplace. In the last year, the research team has built on this earlier work by focusing explicitly on how marketing affects perceptions of product differentiation among various anti-ulcer drugs, and then on the impact of this product differentiation on competition.

Recent findings and conclusions. Using monthly IMS data on four anti-ulcer drugs, researchers constructed a variant of the discrete choice model of consumer behavior that incorporates the effects of marketing, and allows marketing to affect the scale of product differentiation. Three important empirical regularities were found in the data: marketing reduces firms' own price elasticities of demand, making them less elastic; total marketing by all four firms reduced the degree of product differentiation in the market and raised the cost of entry to potential competitors; and marketing initially expanded the market for anti-ulcer drugs, but marketing became an increasingly important means of stealing business from competitors as more firms entered the market.

Future work. The POPI research team is examining empirically the implications of dramatic changes in the prescription and over-the-counter (OTC) markets for anti-ulcer drugs. For the prescription drug segments, the team is constructing and estimating econometric models that explain the continued growth in the overall market, as well as the shares earned by the four segments and the proportions within each of these segments: H2 antagonists; proton pump inhibitors; prokinetic agents; and the stomach-lining agents. Particular attention will be focused on the generic/pioneer brand split within a chemical entity, as well as the impact of generic entry on substitutability between chemical entities. Attention will also be given to use of hedonic pricing methods, since the number of products and extent of product characteristic differentiation has increased well beyond those for the four products studied in the previous research. The team also plans to devote considerable effort to modeling the impacts of OTC introductions on the prices and quantities demanded of the Rx branded products.

The research effort will then focus on the factors affecting the success of the four OTC products. Here the team will address whether distinguishing product characteristics such as dosage convenience, adverse drug interactions, and side effects also have an impact on the OTC marketplace; whether order-of-entry effects are significant in the OTC marketplace; pricing issues in Rx and OTC versions of the same chemical entity; the results of joint/sole marketing agreements for Rx manufacturers with little non-prescription retailing experience; the cannibalization of antacids by OTC sales of the H2 antagonists; and, finally, the differences and impacts of marketing efforts for various OTC products.

Workplace Performance as a Health Outcome Measure

Ernst Berndt (economics)
Stan Finkelstein (medicine)
Alison Keith (economics)
Martin Keller (psychiatry)
Paul Greenberg (economics)
James Russell (psychiatry)

A major thrust of ongoing POPI research has been to examine the impact of medical treatment with pharmaceuticals on productivity or performance at work. At-work productivity by sufferers of chronic illness who may be performing at a reduced capacity, as well as absenteeism from work, have both been important aspects of this investigation.

The decision to address these issues was stimulated by earlier POPI findings that lost productivity at work due to chronic depression accounts for tens of billions of dollars per year in the United States alone. The largest portion of the cost burden is borne by employers.

Recent findings and conclusions. The research team gained access to a database containing detailed work performance data as well as records of personal and family medical care utilization for 1,600 employees of Aetna Health Plans whose jobs involve processing insurance claims. This information, assembled with careful attention to privacy and confidentiality, allowed examination of experience over a two-year period.

Members of the research team have been able to begin to infer the duration of an episode of treatment for a coded illness and relate the course of the illness to possible changes in the individual's productivity in processing insurance claims. For some illnesses, they have been able to document different patterns in productivity decline before and during treatment of an illness and, likewise, to document its rebound following treatment. For other illnesses, there appears to be little or no discernable effect on a worker's productivity in processing claims.

Future work. The team is actively preparing a research agenda to extend this unique investigation. Among the items under consideration are the replication of the current approach using other sources of retrospective data, as well as the design of prospective interventions or therapeutic trials based at the workplace that would examine directly the impact of medical treatment on objectively measured work performance. This broad subject is the focus of POPI's new consortium under development, the Consortium on Health and Productivity (CHAP).

Optimization of a Biotechnology Service Operation

Lawrence Wein (operations management)

A biotechnology company has developed a revolutionary new product that offers the potential to repair diseased or damaged knee joints or hip joints, for example, by growing a patient's own cartilage cells in the laboratory from an initial biopsy sample. Because of reimbursement issues, far more biopsies were being sent to the laboratory for analysis than subsequent orders were being placed. The technology, though, required that the biopsy be treated as if it were an order, going through a first step of processing and then being frozen. More than half would sit as "inventory."

Objective. The research team has explored the cost savings possible if the company were to change the process such that biopsies could be frozen as soon as they are received, pending an order.

Approach. The first step was to optimize the existing production process. There are three stages. First comes the option of freezing the cells or going on to a second stage, and then freezing it. This depends to a large degree on whether the results are required by the physician for scheduled surgery. The team determined the best way to look at first-phase results and decide whether there would be a second phase.

Second, the team looked at whether freezing could take place up front, how that process could be optimized, and what the cost savings might be from switching to a new technology. And third, the team is exploring a forecasting tool that would help determine how many orders might be expected. This relates to the ratio of doctors trained in use of the product and when FDA approval is granted. The forecasting function is important because of the company's concern about its capacity to expand production.

Findings and conclusions. From an operations management point of view, this project has interesting implications for other firms. This is a service operation in an industry that typically has manufacturing operations. There are interactions with patients and doctors, and there are issues of delivery of the service in a timely manner. The team has found that cost savings are achievable through a more individualized and intelligent approach to decisionmaking in the first phase.

There are also very tight connections between the marketing and manufacturing parts of the firm. This raises the issue of how many doctors to train and what kinds of doctors to target. This suggests a manufacturing firm with tight control over its possible customers and, in fact, the training is a prerequisite to becoming a customer. The team is exploring what is likely to become a new model for many firms in the industry-a disease management system.

Future work. The research team is preparing a software package that will function as a forecasting tool to answer the key questions posed for the service operation. Future work will include looking at dynamic capacity expansion in situations where there is an inaccurate forecast of demand, as well as exploring how to decide how many and which doctors should be trained to use the product.

The Essential Drug List of the World Health Organization versus Alternative Therapies

Stan Finkelstein (medicine)
Robert Rubin (medicine)

Every two years, the World Health Organization (WHO) publishes an Essential Drug List (EDL) for use by 190 member states in pharmaceutical, vaccine, and biological procurements. The EDL contains 300 different products, almost all of which are generic and off-patent. It has come to represent, in many cases, the mandatory requirement for the maximum level of drugs a country should have in its formulary. As such, it becomes in some cases a de facto barrier to access of available, contemporary therapies with proven efficacy superior to the listed drugs.

In and of itself, the EDL makes no claim that the products recommended yield the best clinical value for whatever costs might subsequently be involved, nor does it purport that its recommended drugs are more efficacious than similar drugs in the same therapeutic class. There is no evidence that scientific considerations are paramount in the selection criteria, nor that there is a distinction between price and cost.

There are three immediate reasons for the evaluation underway by the research team: 1) growing resistance of bacteria to antimicrobials included on the EDL; 2) recent findings that noncommunicable diseases are fast replacing infectious diseases and malnutrition as the leading causes of disability and premature death in the developing world; and 3) recent findings that migration from country to country is a leading cause in the introduction of infectious diseases in the United States and elsewhere.

Objectives. The research team is conducting a scientific, comparative analysis of the EDL against alternative therapies now widely available for patient use. The objective is to answer how well the EDL strikes a balance between effective therapy and price, between efficacy and safety, and between ethics and limitations in access for most of the world's population.

The team will develop a new decision screen for procurement agents and government policymakers through which they can make informed, rational choices between classes of therapeutic alternatives that are fiscally sound, clinically prudent, and ethically based.

Approach. From the EDL, a set of indicative drugs or classes of therapies is being selected for study. This set corresponds to well-established predictive effects of their use in alternative categories. Therapies will be compared in country populations where there currently exists reliable epidemiological information, and economic inferences will be drawn from countries in both direct and indirect costs. Data will come from respected international institutions such as the World Bank, the International Monetary Fund, the UN's international social security information base, and data sets on health expenditures and utilization compiled by the Organization for Economic Cooperation and Development.

Please send comments and suggestions to

[Back to POPI Home Page]