There is no doubt that the research and development function in pharmaceutical firms is undergoing radical change. Internal and external forces are driving this change. Regulatory agencies are becoming more demanding, and the clinical trial process has become more complicated. Within the firms, matrixed teams are becoming more common, and more firms are outsourcing and licensing the R&D function.
POPI researchers approach pharmaceutical R&D from several perspectives. One research project is exploring the general climate for R&D in the changing environment. Teams are examining R&D financing, assessing the risks and returns on investment and developing a model for the cost of capital. Project management has been an ongoing concentration. Several projects focus on questions of productivity, from the general determinants to the specific opportunities afforded by new approaches such as combinatorial chemistry. New Project
The United States sustains one of the most innovative pharmaceutical industries in the world. While a variety of explanations have been put forward for this success, the magnitude of the public commitment to biomedical research in the United States is probably a contributing factor. Federal and not-for-profit funding of biomedical research historically has exceeded the combined budgets of the pharmaceutical industry, and preliminary analysis by other researchers confirms that funding of basic biomedical research has significant implications for levels of investment in the private sector. While further research has begun to unravel the role of publicly funded research and researchers on the new biotechnology firms, it remains a subject of intense debate.
Building a better understanding of this complex issue has become increasingly important as a source of insight for public policymakers-especially as federal budget constraints and the erosion of hospital research budgets puts increasing pressure on publicly funded research. The implications of this understanding are also considerable for the private sector, as the increasing costs of pharmaceutical research force more firms into collaborations and alliances across the public/private border.
Research focus. The research team is exploring the nature of the interactions between the public and private sectors in the conventional "small molecule" pharmaceutical industry, focusing particularly on the factors that lead some firms to be better able to take advantage of public-sector research than others. (Note that by "publicly funded research" the research team includes all research conducted within not-for-profit institutions, including research conducted at privately funded not-for-profit institutes such as hospitals.)
There is little systematic information about the ways in which publicly funded research may have benefitted the pharmaceutical industry. Classical models of the relationship between the public and private sectors suggest that researchers in the public sector, motivated primarily by the search for priority, focus on "basic" or "fundamental" research while private-sector researchers, motivated primarily by the search for profit, focus on more applied research. Such models suggest that publicly funded research affects private-sector work through the generation of direct spillovers of knowledge and through the training of young researchers. While there is some evidence from cross-sectional data to suggest that these models provide some insight into the dynamics of public/private research relationships, preliminary work suggests that the roles played by public and private research in the pharmaceutical industry may be more complex.
Research method and data. This research team's earlier exploration of the determinants of research productivity in the pharmaceutical industry resulted in a large and detailed database that is being used as a powerful starting point to investigate the issues in this project. These data are being augmented by the collection of a variety of publicly available data and through a program of field interviews conducted with key researchers in both the public and private sectors. The following data are being collected.
An intellectual genealogy of fifteen to twenty of the most important drugs, significant progress toward which has already been made through the use of secondary sources. This work is being supplemented through detailed field interviews.
A detailed analysis of spending in the public sector and its relationship to private-sector productivity, building on the existing private-sector database. One particularly interesting question to be explored by the research team is that of the balance of funding across types of research that characterizes the public sector. How much exploratory clinical research is funded by industry? by the National Institutes of Health? by hospital budgets?
An analysis of patent and paper data, taking advantage of a previously constructed database about spillovers of knowledge between the public and private sectors and extending the analysis to the pharmaceutical industry. This patent data will be supplemented by an analysis of the citation behavior and authorship of a selected sample of publications from both the private and public sectors, matching these indicators against existing data to explore the degree to which firms that perform research that is "more basic" or "more closely linked to the public sector" are more productive than those that do not. The research team will pay particular attention to the use of different measures of "productivity," including INDs, NDAs, scientific impact, sales, and market share.
Field research at major pharmaceutical companies, which will encompass a significant program of field interviews at a minimum of five major pharmaceutical firms, both in the United States and Europe. These interviews with between five and ten of the key scientists (at each firm) will focus on the history of particular drug discovery efforts to explore the influence of publicly funded research on the research process.
Preliminary findings and conclusions. The empirical evidence for the importance of direct spillovers from publicly funded work to private-sector research is rather mixed. Using detailed internal firm data at the program level, the research team has found that extensive inter- and intra-firm spillovers are fundamental to productive research, but were unable to demonstrate any simple correlation between NIH funding and private-sector productivity. Preliminary work exploring the intellectual genealogy of many of the most important drugs introduced in the last twenty years confirms that publicly funded research was fundamental to their discovery but suggests that, in general, the critical discoveries on which they drew were made in the 1950s and 1960s, that is, before the major expansion in biomedical research that has characterized the last twenty years.
There is some evidence consistent with the hypothesis that the weakness of the cross-sectional results reflects differences across firms in "absorptive capacity," a firm's ability to take advantage of research results generated outside the firm. One researcher has shown that firm productivity was positively correlated with scientific publication rates once investment levels were controlled for. Earlier work by the present team suggests that firms that are "pro-publication" or that promote key researchers on the basis of their standing in the larger scientific community are significantly more productive than their rivals.
Preliminary field research suggests, however, that this result may reflect changes in the structure of problem-solving inside the firm, rather than (or possibly as well as) the ability to take better advantage of publicly generated knowledge. This work suggests that the simple division of pharmaceutical research into "publicly funded, priority-motivated" and "privately funded, profit-motivated" is not a useful characterization of the dynamics of research in the industry.
Preliminary research also suggests that the most productive firms in the industry invest heavily in "basic" or "fundamental" research in fields such as molecular biology and peptide chemistry. They promote key individuals with standing in the broader scientific community, and use methods of resource allocation that echo the peer-based methods used in the public sector. At the same time, the public sector is undoubtedly playing a more complex role than the simple models might suggest. While it is undoubtedly important as a source of trained researchers and as a source of fundamental insights, it probably also plays a number of additional roles, including acting as a source of new hypotheses through the funding of "off-label" clinical trials.
Future work. This research will generate four major deliverables.
Powerful forces are reshaping the pharmaceutical industry in a variety of ways. Recent changes in healthcare financing, industry consolidation, changing requirements for drug approval, and the sustained revolution in the biosciences are increasing the level of perceived uncertainty regarding the industry's future. For example, HMOs and other managed care organizations are redefining the pharmaceutical marketplace. In conjunction with pharmaceutical benefits managers (PBMs), large-volume buyers are placing increased pressure on pharmaceutical firms to justify the prices charged for both new and existing products; further, they are requiring an increased level of service (for example, sophisticated cost-benefits analyses). On the discovery side, the rate of scientific and technological advance in biochemistry, drug screening, and drug delivery continues to increase. Implementing effective responses to these types of changes is made even more challenging by the high level of technological and market competition.
This new POPI-sponsored research, which draws on previous research by members of the present team, seeks to evaluate how these shocks have influenced the composition, productivity, and organization of pharmaceutical R&D. The research focuses on how firms have managed and adjusted their research portfolios in response to environmental change. The team is examining firm decisionmaking at a finer level of detail than most previous studies of the industry. The present research builds directly on a set of studies by Henderson and Cockburn that examines the determinants of research performance and organization at the R&D program level for a small number of firms.
Research approach and method. To develop a set of hypotheses, the research team has developed a framework for organizing the environmental factors which may be influencing the composition of R&D spending by pharmaceutical firms. The team divides the environment facing a typical firm into four distinct dimensions. Taken together, these factors determine the productivity (and profitability) associated with particular R&D investments, thus shaping a firm's R&D strategy. These dimensions are:
Within each of these dimensions, hypotheses have been developed for testing. While the study of each of these dimension will provide useful insights, further insight will be gained by examining their interaction.
The research method includes both statistical and case-study analyses. The statistical work will use multi-equation regression analysis to analyze patterns of research productivity, research investment, and the adoption of organizational practices. These regressions will be used to highlight the historical responsiveness of different types of R&D to various changes in the marketplace and the underlying scientific and technological base.
A more qualitative assessment will be used to integrate the team's statistical results from the insights provided through the qualitative research. In particular, the team will seek to understand how findings regarding responsiveness to change in the past can contribute to an understanding of prospective results.
Data. The data used in this study are both qualitative and quantitative, drawn both from primary sources-internal firm records-and a variety of secondary sources. The sample frame is from 1985 to 1995. To ensure a careful statistical framework for evaluating the effects of environmental change, a reasonable amount of cross-sectional variation is critical; therefore, it is imperative that quantitative data be obtained from at least thirty firms. This level of participation will enable the research team to distinguish between several different forces that engender differences between firms.
Detailed data at the research program level will be obtained from a large number of both U.S. and international firms. The research team defines a program as an intermediate level between therapeutic class and project (for example, within the therapeutic class of "cardiovasculars," research programs include "antihypertensives," "cardiotonics," "antithromoblytics," and "diuretics"). The definition of a research program follows the IMS Worldwide Uniform System of Classification.
This quantitative data is supplemented by in-depth qualitative data about two research programs, hypertension and immunology. These data will be gathered in the course of a two-day site visit by a member of the research team.
Publicly available secondary sources will also provide data. The team will draw on prior research, as well as on data from governmental sources such as the FDA. In particular the FDA will provide drug approval data. The research team will also utilize government data on the prevalence and treatment of disease (for example, the National Health Interview Survey) to examine demographic trends and non-pharmaceutical health technology and utilization patterns.
Finally, the team hopes to obtain data on the changing healthcare market place from IMS and, potentially, other sources. More precisely, the team will utilize time-series data on the changing share of revenues attributable to HMO purchases to identify those therapeutic areas where HMO penetration has been strongest. New Project
The beginnings of the 1990s brought a paradigm shift in pharmaceutical R&D from rational design to high-throughput screening of combinatorial libraries. The new strategy does not focus on the design of the perfect molecule but rather on the design of efficient tests to find the best molecule out of a large number of diverse compounds.
Rational design proved to be more time-consuming and technically more challenging than expected at the beginning of the 1980s. Technological advances in high-throughput screening hardware and information technology, in combination with high-value targets in the pharmaceutical industry, have created a cost- and time-efficient technology platform to screen a large number of compounds. Combinatorial chemistry is used not only to provide compounds for initial hit identification, but also is used for chemical analoging to optimize biological properties of compounds.
The focus of rational design has thus shifted to the development of the test, while combinatorial chemistry is focused on the discovery and development of the drug compound.
Combinatorial chemistry evolved as a means of producing large numbers of random peptides either chemically or on the surface of a phage particle. Thus, combinatorial organic chemistry on solid support and combinatorial solution chemistry were developed. Today, many classical chemical reactions have been adapted for use in combinatorial methods, providing a large repertoire of chemistries applicable to a wide range of synthesis problems. Available chemistries include small organics, oligonucleotides, peptides, carbohydrates, and peptidomimetics.
The need to screen large number of compounds at high speed and low cost has created a market for sophisticated robotics and screening hardware. Over the last few years, this segment has begun to achieve commodity status, as several hardware manufacturers are now providing similar instrumentation.
High-powered computers, combined with chemical information software tools, make it possible to handle the information requirements associated with large numbers of compounds. New database-searching algorithms and cluster analysis allow rapid interpretation of data and predictions of structure/activity relationships. Due to the increasing importance of information and information handling in the drug R&D process, the major players in this field, who form an oligopoly, are well-positioned.
Holders of proprietary targets will have a clear monopoly position. In this field, small research groups in biotechnology companies and academia are as likely to identify such targets as are large pharmaceutical companies. This is leveling the playing field between large companies and small research groups in biotechnology companies and academia. This in turn creates a need to leverage this valuable asset with high-throughput screening, information technology, and a source of chemical diversity.
Research focus. The research team focused its analysis on the economic power of combinatorial chemistry companies, and specifically studied the role of strategic alliances as a way to build internal capabilities in this area. The advantage of strategic alliances with a specialized combinatorial chemistry company is the instantaneous availability of libraries and analoging expertise. This avoids the opportunity cost associated with the delay in ramping up an internal program. In addition, many of the specialized combinatorial companies have established proprietary positions on key technologies, which could preclude the commercial exploitation of these methods without a corporate alliance.
Research method and data. The team analyzed the available literature on combinatorial chemistry, and then conducted interviews with scientists in the field and with some forty companies engaged in combinatorial chemistry activities. The interviews posed several important questions:
The research team visited companies and presented their preliminary findings, and received feedback from practitioners.
Findings and conclusions. As of 1989, one strategic alliance involving combinatorial chemistry was formed; by 1994, the number of alliances involving combinatorial chemistry had increased to at least twenty-one. During the first two quarters of 1995, some fourteen alliances were formed. While U.S. and European pharmaceutical and biotechnology companies gained access to this technology through many strategic alliances and acquisitions, Japanese pharmaceutical firms are completely underrepresented, forming a total of five alliances compared with forty-six formed by U.S. companies.
Based on the initial analysis, the research team anticipates that Japanese pharmaceutical companies will significantly increase their participation in these strategic alliances. The increased demand is likely to strengthen the bargaining power of combinatorial chemistry companies, and the average value of an alliance is expected to increase over the next few years.
The research team has segmented different technologies according to the number of alliances, finding that strategic alliances involving small organic molecules scored the most rapid increase, followed by oligonucleotides, peptides, and carbohydrates. Combinatorial protein technology trails the entire field, with only one disclosed strategic alliance in the last six years.
Future work. Over the next period, the research team will be working to update the analysis, using a system dynamics model to understand the role of combinatorial chemistry in the drug development process.
New Project
This new project builds on the exploration by a similar POPI team into the dynamics of competition in the pharmaceutical industry. That research, detailed in last year's report, concentrated on entry and exit patterns in the pharmaceutical industry, seeking to place the analyses of both the development of the industry and its competitive conditions in an appropriate context. The research team for that project saw the pharmaceutical industry as a potent case history to test the application of a particular element of industrial organization theory, that is, the increasing recognition of the dynamic interactions of firms and new models to describe the changing distribution of firms.
The new project focuses on industry economies of scale. Conventional wisdom about the U.S. pharmaceutical industry is based on the observations that there are economies of scale in pharmaceutical R&D, in shepherding new drugs through the regulatory process, and in marketing, and that there are patent monopolies of new drugs. Both of these factors imply market power in the sales of individual drugs, and economic analyses of the industry have been concerned largely with measuring and alleviating this power.
But the pharmaceutical industry cannot be seen simply as a collection of monopoly markets. There also are economies of scope within the pharmaceutical industry, and virtually all firms in the industry sell a multiplicity of drugs. Many, if not most, markets for individual drugs have at least some competition, either because the patent has expired or because there exist closely substitutable drugs. And since patents are only temporary, new drugs have to be introduced continually to maintain market power.
It follows that any drug firm participates in several markets with varying degrees of market power. It is natural to ask whether the positions of firms in the pharmaceutical industry are similar to the position of firms in the markets for individual drugs.
Research focus. The research team is posing two questions. Do the economies of scale noted so prominently in drug discovery and introduction imply that economies of scale are an important feature of drug firms? And did regulatory changes that are thought to have altered the economies of scale in drug introduction affect the importance of economies of scale for drug firms? Two such changes are the 1962 Drug Amendments, which are thought to have increased economies of scale in the introduction of new drugs, and the 1984 Waxman-Hatch Act, which is thought to have decreased them.
Data. The data have been compiled from observations collected by IMS America from 1964 to 1994. These observations are cumulated to provide measures of firm sales, which are the primary data.
IMS collects its data in two forms. First, IMS measures pharmaceutical purchases made by samples of drug stores and hospitals. The stratified random samples are changed yearly to reflect distribution of sales within the United States. Second, IMS conducts a census of wholesale warehouses for hospitals and drug stores.
The industry consists of firms selling either prescription (Rx) and over-the-counter (OTC) drugs. Almost 90 percent of sales were of Rx drugs. Some 80 percent of these products were sold in drug stores, the remainder in hospitals. More than two-thirds of all sales in this industry were Rx drugs sold in drug stores.
For several reasons, data were truncated, and the research team included only firms selling more than .01 percent of the industry total. This results in a drastic fall in the number of firms included in the industry, but almost no change in the total sales.
Most of the analysis is done with companies listed by IMS, referred to by the researchers as "operating units." Some tests with firms, where subsidiaries are aggregated instead of being distinguished, are included. These decisions yielded observations for thirty years on roughly 250 firms. Entry and exit were calculated from the first or last appearance of an operating unit in the dataset.
Preliminary findings and conclusions. The research team argues that the drug industry as a whole does not appear to be dominated by economies of scale. The pharmaceutical industry resembles an "average" industry more than it does industries dominated by economies of scale like the automobile or aircraft industries. In addition, most of the effects predicted to follow from changes in the regulatory and technological environment have not taken place.
Future work. As this research progresses, the team will examine the determinants of firm growth over longer time periods, looking for the effects of size, age, R&D, diversity, and mergers. The causes and effects of mergers and acquisitions among pharmaceutical firms will be explored in relationship to data on patents and new drug introductions. New Project
Technology transfer between universities and industry offers the potential of significant benefits to all parties. But many believe only a small fraction of these benefits are being realized at present.
Consider the experience of MIT, with its thousands of patents-many of which have been licensed to industry. While the annual revenue from licensing may soon be in the range of $10 million, most is from a small number of highly productive patents. This raises a serious question regarding the total opportunity costs to universities, industry, and society that arise because creative ideas are not reaching fruition as marketable products. As universities turn increasingly to industry for research support, with government and philanthropic funding becoming scarcer, the question is especially timely.
Several universities, some with vast research activity and others with more limited resources, have established technology transfer or licensing offices to operationalize the exchange of their own intellectual property for funding from industry. Some firms seeking academic relationships have established internal processes for managing these new linkages. These technology transfer processes can be vastly different, and the effectiveness of each may be limited by barriers that outside partners perceive they must cross in order to establish and maintain the relationship. It may be that difficulties in the nature of the university-industry linkages are limiting the effectiveness of technology transfer.
This new study explores these and related questions. It is a collaborative effort between POPI and the Harvard-MIT Center for Experimental Pharmacology and Therapeutics.
Research focus. The principal aim of the study is to investigate whether technology transfer linkages between industry and universities are falling far short of their potential and, if so, why. The research team is seeking to
Research method. The research team will identify or create a database that will afford the possibility of hypothesis testing and statistical inference about alternate models of technology transfer between universities and industry. An initial model will assume that effective technology licensing depends on characteristics of the item of intellectual property to be transferred, the characteristics of the technology licensing process, the "fit" with the industrial partner, and the strength of intellectual resources at the university.
An early hypothesis will be that the characteristics of the technology licensing process can have a significant effect on technology transfer outcomes: done poorly, the licensing process can limit the success of the universities that have made large commitments to research; done well, it can lead to universities with limited research opportunities reaping substantial rewards. A serious attempt will be made to measure these differences.
An initial exploratory analysis will focus on two areas. The first will be technology transfer out of two universities-MIT and one other to be selected-to various pharmaceutical-related firms. The second will be technology transfer into a major pharmaceutical firm from a number of universities and medical schools. The research team will consider the technology transfer opportunities that failed to materialize and reach plausible estimates of what can be achieved under ideal but realistic conditions.
During the course of this research, the researchers will test several hypotheses of specific obstacles to effective technology transfer between universities and industry, including
The aim of this exploratory phase will be to test the model of technology transfer described above, to propose alternative models and hypotheses about determinants of success and failure, and to identify a set of variables to be analyzed. Once these insights have been gained, the study can be extended to include technology transfer across and between a number of universities and large pharmaceutical firms.
Data. A questionnaire will be developed to collect data, and will be administered to the two university technology licensing offices and at the firms that participate in the study. The research team will seek to characterize in detail the nature of the process used to forge and maintain these university-industry linkages and to gather information about desired outputs, range of inputs, and milestones indicating success or failure. The unit of analysis will be the individual linkage (that is, MIT with Company A for technology A; Firm 1 with University Z for Technology Z).
This ongoing research project addresses the risks and returns pharmaceutical firms have faced and enjoyed by investing in drug development projects, and examines why the cost of capital to pharmaceutical firms of undertaking manufacturing investments is less than that for R&D. Last year, the research team added a new component to the research: modeling investment in pharmaceutical R&D.
This research, central to larger questions about the flow of capital in the pharmaceutical industry and the industry's profitability, has direct implications for ongoing public policy debates on healthcare reform and drug price controls.
Findings and conclusions. Preliminary findings from earlier phases of this project may be found in POPI's Third Anniversary Report of Accomplishments of January 24, 1994. In the past year, the research team constructed a Monte Carlo simulation model of investment in pharmaceutical research and development. R&D costs are treated as quasi-fixed expenses and discounted at a lower rate than future net revenues. The result is a risk-return "staircase" (Figure 2) in which the cost of capital falls if and as research proceeds. The model has several applications, including evaluation of the net present value of a drug candidate or program and calculation of the risk-adjusted cost of bringing a new drug to market.
A draft working paper, "A Life-Cycle Model of Pharmaceutical R&D," was prepared in September 1995. The paper describes the model and its applications. Since then, the model's base-case assumptions have been reviewed and modified, and a revised working paper will be published in early 1996.
Current research is concentrating on improved accounting procedures, using the simulation model as a testbed for controlled experiments. Preliminary results indicate that capitalization and amortization of R&D would dramatically improve the informativeness of accounting profitability measures.
Figure 2. The risk-return "staircase" for pharmaceutical products
Future work. There are other papers under preparation by the research team. One will present an updated cost-of-capital study, focusing on risks and returns in biotechnology. The research team also will prepare a paper on the accounting experiments conducted as part of the research.
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?
Project management in research-intensive, multidisciplinary industries poses several challenges. In the pharmaceutical industry, the effectiveness of project management has direct impacts on the product pipeline, with serious consequences for a firm's competitiveness. Bringing new drugs from discovery through development to market is costly, time-consuming, and fraught with risks.
As the business environment becomes increasingly competitive, pharmaceutical industry managers are looking for ways to make their firms more efficient and effective at drug development, seeking ways to reduce the cost and time associated with discovering a new chemical entity, generate evidence of its safety and efficacy, and launch it on the market as a new therapeutic agent. To achieve this goal, managers are experimenting with the practices of other industries, hoping to gain their benefits.
This research aims at a better understanding of the role of project management in the pharmaceutical industry. The team continues to assess the various project management strategies being tried in various firms.
Findings and conclusions. There are several very clear conclusions that follow directly from the study thus far (see Figure 3). First, go/no-go decisions are clearly the responsibility of departmental management, not project management. Go/no-go decisions are really "no-go" decisions. They are the most difficult decisions to make for any R&D project, and a decision to discontinue requires decisionmakers who can maintain objectivity and detachment. The project managers themselves are usually too close to the situation and too committed to making the development "work" to manage such difficult choices. Therefore, it makes sense that departmental management (or some higher authority not considered in the present survey) take on this responsibility.
By way of contrast, clinical decisions are better made by project management, particularly (or one might assume, exclusively) when project management, by virtue of having the necessary technical knowledge, is qualified to make such decisions. In this case, being closer to the action has its benefits. Project management is necessarily in closer and more continuous contact with the clinical data and often the clinical subjects. Departmental management, which may well have the expertise, is further removed from the day-to-day activities of the development and is therefore in a poorer position relative to clinical decisions.
Higher performance is also associated with relatively greater project leader influence over reward decisions, especially during early development phases. The ability to affect reward decisions provides a project leader with an important tool for focusing people on keeping the project moving forward. In the early stages of drug development, only limited organizational commitment has been made to any candidate compound or project. Departmental personnel are likely to be assigned to more than one project. Team members have duties in their departments in addition to those pertaining to an individual project. Tying rewards to achievement of project goals is likely to provide a more timely assessment of the toxicity and pharmacokinetic properties of a compound. If a candidate is going to fail in the first steps of testing, at least it will do so quickly and free up resources for other projects.
The results of this research to date further demonstrate that team leader characteristics and organization structure affect the locus of decision influence. Team leaders who are perceived as technically more knowledgeable are also perceived as having greater influence in clinical decision making. Technical knowledge of the team leader is, in turn, associated with organization structure. In firms employing dual leader coordination systems, project leaders are perceived as technically more knowledgeable.
Important and surprising is the fact that the loci of influence over reward and go/no-go decisions are not related to characteristics of the project leader. It is also notable that pharmaceutical firms, especially those employing the single leader system, are not perceived to provide an organizational environment that is supportive of teams.
The results point to the external environment of teams being at least as important a determinant of project performance as team processes. This is not to say that internal process is unimportant. It does say that process is not sufficient. Without a supportive environment for teams, it will make little difference how the teams themselves are managed. It must be made clear that senior management believes in project teams and that individual rewards are related to team performance.
Figure 3. Path analysis of the principal influences on project performance
In summary, the results point clearly to the relationship between type of decision, development phase, leadership structure, locus of decisionmaking influence, project leader characteristics, support from higher management, and project performance. With the exception of a link between locus of influence and a team-supportive organizational environment, the causal links posited in the Figure 3 model are demonstrated.
Future work. The current research goal is to expand the sample the sample among European firms. By increasing the sample size in this way the research team will be able to do cross-national comparisons as well as comparisons across therapeutic areas. It will also allow the researchers to address the important issue of cross-national teams within individual firms. Most firms in the pharmaceutical industry are now establishing such teams to maximize their ability to obtain simultaneous approval for and introduction of new products in at least two of the three major national markets.
The research team also plans to continue its collaboration with the Harvard-MIT Center for Experimental Pharmacology and Therapeutics. This will help the team gain a better understanding of the clinical development process and refine the interview and survey instruments to reflect better the issues faced in the complex project management process.
With POPI support, Rebecca Henderson and Iain Cockburn have been engaged, for several years, in an ongoing investigation of how well pharmaceutical firms innovate. The accomplishments of this research team are detailed in last year's report.
To date, the research has resulted in a large body of working papers and published articles that have received a wide audience both in academia and within the industry itself. These papers have addressed, among other topics, the measurement of firm competence and the evolution of integrative competence in drug discovery, research spillovers between firms and from the public to private sector, and the issue of racing to invest.
The Henderson-Cockburn research is helping to build a better understanding of how the industry has been successful in its ability to innovate continuously in the face of revolutionary scientific change. Three sets of central research questions continue to serve as a framework for their ongoing study.
Findings and conclusions. Three general hypotheses have emerged from the research team's work.
The success of the most productive pharmaceutical firms highlights the importance of continually revisiting the embedded knowledge of the organization, and outline some of the organizational mechanisms that help make this possible.
The research team has also found evidence consistent with the presence of substantial spillovers of knowledge across firms within the industry. A research program embedded in a well-diversified firm surrounded by unusually productive competitors, for example, is likely to be almost 20 percent more productive than typical programs not so situated.
Public-policy implications of this research are important. The team's results suggest there are significant externalities in pharmaceutical research. One firm's "failure" may be another's gain. Thus, the true "cost" of a successful drug is almost certainly quite a bit more than the simple costs incurred by the firm, suggesting strongly that public policy actions that lead to reductions in research expenditures may have unexpectedly adverse consequences as the effects ripple through the industry.
In addition, research productivity is critically dependent upon economies of scope and the management of information flow across and within the boundaries of the firm. Firms that diversify their research portfolios to take advantage of internal economies of scope and economies within the industry, and that adopt an active approach to managing the generation and dissemination of knowledge across the boundaries of scientific discipline and therapeutic class, tend to be the successful ones.
Future work. Current research is extending the earlier analysis downstream to clinical development, which the qualitative results thus far suggest may be subject to quite different dynamics.
Henderson and Cockburn also are addressing in detail the evolution of "competence" in the industry over time. They are at present developing a first draft of a paper with the working title "Exploring Inertia: Governance Costs and Other Sources of Organizational Failure in Pharmaceutical Research." This paper focuses on the degree to which it might be the case that firms that have been relatively slow to adopt the techniques of "rational" drug discovery were subject to governance problems. Preliminary research results suggest that while governance effects may be important in explaining major differences between firms in the rate of adoption of new techniques, they do not seem to explain more local, year-to-year variation across firms. Efforts to refine this result are underway.
Please send comments and suggestions to popi-www@mit.edu