RT2C W6D1 Wald

OUTSOURCING A CORE COMPETENCY Pharmaceutical firms that outsource biostatistics risk losing

opportunities to improve innovation, a study shows.

Shreefal Mehta and Lois S. Peters

OVERVIEW: Data from contract research organiza- tions (CROs) and pharmaceutical companies reveal an increasing amount of outsourcing of biostatistics in clinical trials. However, a surprising number of projects involve practices Identiftedas "never outsource" and of "key importance to grow and retain " by the pharmaceu- tical statistics directors. The sharing and outsourcing of these big pharma "best practices" with the CROs is an advantage Jor smaller competitors, particularly as related to analysis methods and protocols that are likely to he accepted hy the U.S. Food and Drug Administra- tion. Over time, an accumulation of critical resources and tacit knowledge could allow the CRO to eventually become a new type of independent competitive player in the drug development process.

Shreefal Mehta is clinical as.<iociate professor of biotech- nology management in the Lally School of Management of Technology and the Biomedical Engineering Depart- ment at Rensselaer Polytechnic Institute, Troy, New York. He is also vice president, business and corporate development, Cytopia Inc.. Rensselaer. N. Y. As an entre- preneur in the biotechnology/pharmaceutical industry, he created a multi-disciplinary course on commercializ- ing biomedical technology. He has a Ph.D. in hiomedical engineering from University of Texas Southwestern Medical Center. niehtas@rpLedu

Lois Peters is associate professor and director of the Center for Science and Technology Policy at RPI \s Lally School. She is also a principal investigator at RPI's National Science Foundation-sponsored Nanoscale Engineering Research Center. In addition, she is past president of the International Trade and Finance Asso- ciation and a member of the IEEE Engineering Manage- ment Society Board of Governors. She is a co-author of Radical Innovation: How mature companies can outsmart upstarts (Harvard Business School Press, 2000), and has conducted extensive research on university-industry research connections and R&D technological networks. She received her Ph.D. in biology and environmental health science from New York University, peterl (ayrpi.edu

KEY CONCEPTS: outsourcing, biostatistics, business model innovation, dynamic capability.

Not so long ago, a company's technical core was believed to be a resource and competency that should be buffered and protected. Today, with the growing com- plexity of product/process development and new business creation, R&D-intensive companies recognize that they have to collaborate, partner and sometimes contract with outside entities to help with technical development {1-5). However, the task of determining which of multiple functional units represent a core com- petence that should be retained or enhanced when out- sourcing decisions are to be made is a difficult one (6-17).

This article presents a scenario, based on an empirical study, to help managers assess strategic outsourcing plans. Questions addressed are: the level of outsourcing of various functional tasks and whether a core R&D function has the likelihood of being eroded. Simulta- neously, we question whether protracted outsourcing of core R&D functions increases the likelihood of stability or greater turbulence in the product development process in the industry as a whole.

Stability and turbulence refer to industry ecosystem stability (AV) measured by investigational new drug filings (INDs) for innovative compounds by the top five pharmaceutical companies. Outsourcing core or poten- tially useful portions oftheir R&D activities can actually be counter-productive in the long run, as the internal capabilities to recognize and evaluate innovations can be eroded. In this regard, it may be that competitive value is created by methodological approaches to R&D as well as by R&D outcomes. To investigate these dynamics, we selected the most R&D-intensive industry (percentage of sales invested in R&D) in the last 20 years^—the pharma- ceutical industry.

In the pharmaceutical industry, innovation is recognized as the cornerstone for competitive advantage and is fostered by strong investments in biological discovery, biotechnology or chemistry processes. Cockburn. Henderson and Stern (19) have shown that drug

Research - Technology Management

discovery productivity is partly dependent on internal organization of R&D and firm-level expertise in diseases. However,, the risky and time-consuming drug discovery process, along with skewed economic returns in major firms wherein a few blockbuster drugs dominate the portfolios, makes it difficult to assess the significance of a specific competence on any measurable firm-level outcome such as sales, profitability or market share.

Recent papers on innovation and management in the pharmaceutical industry have identified specific expertise in particular chemical families or drug pathways, as local heterogeneities in capabilities and core competencies for companies in that area (19.20). Anecdotally, Pfizer is recognized as having a compe- tency in marketing, Merck in R&D, Eli Lilly for manu- facturing and recently for alliance building, etc. However, within the literature there has been little explo- ration of resource-based competencies of these pharma- ceutical companies for some of the reasons discussed above.

Competitive vaiue may be created by

as weii as by R&D eutcemes.

As regulatory oversight and guidelines define the drug development process, it stands to reason that successfiil firms have a core competency in processes that helps

How the Study Was Carried Out

A questionnaire was sent to 88 contract research organi- zations (CROs) that had clinical biostatistical fiinctions contracted to their organizations from large and small pharmaceutical companies. The contracts were of two main types: contracts that packaged many functions such as clinical trial management and data analysis, and some contracts which had only biostatisticai work outsourced. We asked the respondents to separately address these two types of contracts in their responses, A total of 34 responses were received (39 percent response rate) over the course of 6 months.

We conducted interviews with purchasers of the services and service providers to add depth to the survey data. We then collected and analyzed data from secondary sources tu gain further insight on phamiaceutieal outsourcing and its implications.

The firms participating in this study cover a wide variety of large and small firms participating in the life science- based industry (tables beiow).

Firms in the Study

Field Sludy/lnlervicws with Companies

Dedicated Pharmaceutical Biotechnology CRO

Firms Firms Firms

Small firms 0 I 2 Large firms 5 ] 2 Totals 5 2 4

CRO Survey

Survey

CRO KIrm Type

Small Large Total

llmis firms

Responders

Data

C R O Responders

l'> 15 34

Interviews were held with senior managers, biostatisti- ciansand technological specialists who provided both his- torical and current information and responses to the research questions developed by the Rensselaer team. Interviewees were identified by our company liaison to ensure the breadth and depth of knowledge necessary to address our questions and to provide a diversity of per- spectives about the development process.

We also collected infonnation on outsourcing activities from the web sites of the top pharmaceutical firms and analyzed their content to help us understand cuirent activi- ties, We also reviewed data from ParexePs Pharmaceutical R&D Statistical Sourcehook and Bioscan to provide insight into life science NPD practices and outsourcing as well as changing industry dynamics and complexity.

Given the sample size of 34. we used summary statistics to describe the responses as means, medians or percentages of total responders to a particular survey question.

In our review of the transcripts, segments that bore on the research questions addressed in this paper were high- lighted and collected on summary sheets for each project. The authors independently reviewed each transcript. The summary sheets were then compared and aggregated, and observations were expressed and discussed by the authors, to identify commonalities and dissimilarities.^—S.M. and L.S.P.

May—.lune 2007

them to clear regulatory hurdles. Statistical analysis of the clinical data is a key point of regulatory scrutiny, sug- gesting that biostatistics would have to be developed as a core competence for pharmaceutical companies to be successful. As outsourcing of drug development functions to contract research (CRO) firms increases, we asked these questions: Arc pharmaceutical companies giving up their core competence to the CROs? What are the key biostatics-related core competencies and functions in drug development? What role, if any, can biostatistics play in developing dynamic capabilities?

The Role of Biostatistics

Statistical analysis has always played an important role in the pharmaceutical industry, but the drug discovery phase has not employed heavy statistical analysis until recently. This recent use of biostatistics in discovery is driven by the advent of large amounts of bioinformatics data that require fairly sophisticated statistical analysis to interpret and validate. The focus of this study, however, is on the key competencies of the pharmaceutieal industry in managing and conducting drug development, and clinical trials, and interacting with the FDA. the tra- ditional domain of pharmaceutieals that has employed sophisticated statistical analysis.

In the drug development process, a biostatistician is typically involved in early planning of an experimental study, in the preclinical (animal studies) and clinical stages (protoeol design). The bulk of a statistician's work lies in analyzing data, as shown in Table 1.

The three main stages of the drug development process are categorized as: early discovery, middle (or preclini- cal) development and final development to commercial- ization. The early discovery is frequently {sometimes more than 50 percent of all projects) licensed in by big pharma from small bioteehs and universities, and the middle stages of late preclinical testing and early human testing is largely process-driven by regulatory and safety issues (21). Thus, the core competenee that is recognized as unique to big pharma increasingly seems to be focused toward the final stages of drug development, which include large-scale, expensive, clinical studies to establish efticacy, regulatory interactions for marketing approval, manufacturing in large quantities, and marketing, sales and distribution of the final product. In this context, the role of biostatistics is critical to estab- lishing success of the large-scale phase III clinicai trials, with results of complex statistical analysis holding the key to regulatory approval to go to market.

Trends in R&D Outsourcing

All drug companies seek to reduce costs and reduce the tiine to market in a lengthy and expensive product devel- opment process. In this high-risk process, where a

Table I.—Biostatistics Role in Drug Development

Drug Dcvt'lopment Stage Biostatistician Function/Kole

Discovery stage. Formal preclinical

data for regulatory filings.

Chronic toxicology. safety studies.

Clinical studies.

Clinical Phase IV— follow-up or extension clinic studies.

Manufacturing— Quality control.

Bioinformatics. data mining. Basic analysis with

phannacological data (rarely double blinded).

Analysis of data from toxicology study—some study design work. especially for complex dosages or combination drug studies.

• Design of trial protocol with physicians—detemiine sample size, relevance of parameters, data analysis methods defined.

• Make sure data are entered correctly.

• Interim analysis as per protocol. • End of study—reliability,

quality of data. • Analyze—write report and check

medical report to make sure analysis is accurate.

• Combine data across safety studies (not efficacy data).

Analyze data, follow adverse events, analyze, interpret, and prepare reports for submission to FDA,

Not specific to biostatistician— general statistical anaiysis.

project could fail at many different steps, it is more attractive to rent or outsource the resources needed at a given stage. For smaller companies involved in the clinical trials process, outsourcing is not a choice but a need, as they cannot afford to build up internal resources with required skills and efficiency.

In general, the global pharmaceutical R&D outsoureing market continues to grow at over 14 percent annually, with clinical trial management (83 percent of all CRO revenues) dwarfing the rest of the R&D functions (21) (Figure 1). More telling, spending on outsourced clinical trials as a percent of total spending on clinical evaluation studies in pharma/biotech companies increased from 16 percent in 1995 to 25 percent in 2002, with over 22 percent of all clinical studies being outsourced (2!). In fact, a 2000 survey of pharmaceutical companies and CROs by CenterWatch indicated that CROs now play a major role in over 60 percent of all R&D projects, compared to <30 percent in 1993 (2/).

This increase in outsourcing (Figure 1) is particularly interesting, considering that the large phannaceutical companies view management of large clinical trials as a eore competence. Consider, for example, Glaxo Smith- Kline (GSK). GSK, a few years ago, restructured its business and R&D units, launching a hub-and-spoke

Research • Technology Management

model, where the R&D units were independently organized around the core hub that contained clinical Phase III management capabilities, marketing, manufac- turing, and corporate functions.

In the middle stages of R&D ... GSK has created six Centres of Excellence for Dmg Discovery, or CEDDs. Each CEDD is dedicated to specific therapeutic categories; each is responsible for taking lead compounds forward to the point where the therapeutic rationale for those compounds is demonstrated sufficiently lo Jiistif}' the start of large-scale clinical trials, (from www.gsk.com).

Our Study Results

Reliance on CROs and increased outsourcing are also reflected in our survey results, which were obtained from a cross-section of large and small biostatistics service provider firms (see "How the Study Was Carried Out," page 29). In all, 34 responses were received and 33 were included in the analysis. The responders include not only some ofthe largest CROs but also smaller independent CROs whore the biostatistics gioup is composed of two statisticians. The CRO respondent (typically a director of biostatistics or, in a smaller fimi, the senior biostatisti- cian) summarized the clinical trials outsourced to their companies from large to mid-sized pharmaceutical companies by noting how frequently specific functions in biostatistics were included in all client projects with which they were familiar.

Respondents were asked to write the approximate frequency of inclusion of various contracted fimctions in all projects and separately to indicate the same for projects that were focused only around biostatistical functions. Table 2 summarizes the responses from the CROs by noting the percent of contracts that include a

pharmaceutical companies view

management of iarge ciinicai triais as a core compotence.

given function for different types of contracts outsourced by a large- to mid-sized pharmaceutical customer.

Our data indicate that most respondents perfomied out- sourced tasks that were both mundane and technically challenging. Data compilation and collection are the more technically mundane chores (Table 2). Notably, data analysis and interpretation, which requires a higher level of engagement with the data, was also almost always included in the contracted functions at a level of 65-70 percent ofthe time in all projects, and 100 percent of the time in biostatistics-only projects. In particular, when all contracted projects were considered by CRO respondents, less challenging tasks such as data collec- tion were included at the same frequency (65-70 percent ofthe time) as higher-level analysis or interpretation in the contracted functions, possibly indicating a packaging

S c a

CO

80

70

60

I CRO Total

•Riarma Total R&D

I CRO dinfcal

•Pharme Clinical R&C

4 i J 0)

tt 2 O 1 O

1995 1996 1997 199 1999 ear

2000 2001 2002

Figure 1.—Increased pharma R&D budgets have clinical R&D as a significant element, and increasing CRO revenues are mainly focused on contracted clinical R&D, including data management ami statistical analysis.

Mav—June 2007

Tahle 2.-—Contracted Functions in Biostatistics Outsourcing

Service Function Contracted

Protocol design clinical Dala collection Data compilation Data analysis Data interpretalion Project management/

decision support Presentation to FDA

Broad Service Contracts

% o f contracts that include

the function*

32 75 65 73 68

75 10

Biostatistjcs Service-Only

Contracts % o f contracts

that include the function*

35 20 71

100 100

55 8

*Median of all responses

of services from the CRO but also indicating an accep- tance of the CRO competence in that area by the pharma contractor.

Projects that involved only CRO biostatistics groups (Table 2), always (100 percent of the projects) had data analysis and interpretation included with less than 30 percent of contracts requiring data collection activi- ties. Outsourcing of biostatistics functions alone is clearly focused on the higher-level analytical and inter- pretation work. However, interviews with pharmaceuti- cal company biostatistics directors showed that these higher-level activities are usually carried out using guidelines and templates issued by the client. A biosta- tistics director from one of the top five pharma companies indicated that the "analysis methods develop- ment is never outsourced" and that selection of the CRO sometimes depended on which one " . . . can adopt and understand our methodologies well," showing flexibility and competence.

About 50 percent of the biostatistics-only contracts involved support for project decision-making, which is expected considering the intermediary role of the CRO between the data and the client. Surprisingly, a signifi- cant portion of contracts (mean of 35 percent) involved clinical trial protocol design, which is a key function of biostatisticians and clinical teams in pharmaceutical companies (Table 1). This fmding is validated in data gathered by CenterWatch in their year 2000 survey of pharma companies and CROs {21) showing that about 20 percent of polled pharma companies use protocol design services, along all phases of the clinical study process (Figure 2).

Interviews with the sampled ftrms indicated that while it was not common for all protocol design to be outsourced, frequently the clinical/contract research organization had skills and expertise in the area and would collaborate

with its pharmaceutical counterparts to refine the initial study protocol design. When asked to report on the role of the biostatistician in a selected recent or current large or mid-sized pharmaceutical contractor, 76 percent of all CRO respondents indicated that the contracting company had a biostatistician in-house. Further survey responses revealed that of these contracting companies with in-house biostatisticians, about 44 percent were involved in a supervisory role and 56 percent of the con- tracting biostatisticians were involved in a participatory fashion in the projects.

These responses indicate that about 24 percent of the time, the (presumably mid-sized) contracting companies do not have the resources of biostatistical analysis in- house, and even when they do have in-house experts, they rely 56 percent of the time on the best practices retained by the CRO biostatistical team. In fact, in addi- tional survey questions when the CROs were asked if their responsibilities increased over time with a single client (median relationship length with specific clients was 3 years; median length for particular projects was 1 year), over 60 percent replied in the affirmative.

While the survey suggested a growing tendency among pharmaceutical companies to outsource projects to CROs and rely on the CRO's statistical competence, sample comments from these companies' biostatistics directors or managers about outsourcing were as follows:

. . . we would never outsource vital studies—and even if we had lo. we would never outsource anything in the critical path of drug develop- ment. On the other hand, we would completely outsource extension studies.

Stability of CRO personnel [turnover] and commitments of time to the project are always an issue with large CROs. We would prefer to work with a smaller CRO that had a focus on biostatistics and much lower turnover than a big CRO.

Percentage of pharm companies that use CROs "often"

Rotocol design • Statislical Services

10 2 0 3 0 4 0 5 0 6 0 7 0 8

Figure 2.—A 2000 survey of pharmaceutical companies by CenterWatch showed that while most companies outsourced parts of statistical work, a significant portion of the companies (-20 percent) used CROs for protocol design services (I).

Research • Technology Management

This particular concern for CRO competency and personnel focus and stability was reflected by several of the pharmaceutical statistical group directors.

Implications for Management

Outsourcing R&D functions is a "buy versus build" decision in most companies, suggesting transaction cost logic (22). Projects will be outsourced when tasks are readily programmable, as in extension trials. When projects are long-term and arms-length monitoring is difficult, the cost of outsourcing may be too high. But for pharmaceutical companies, where the risks are high and capital costs for product development (clinical trials) are significant, outsourcing may seem like an optimum solution.

However, outsourcing takes place at the risk of diluting internal competencies and increasing complexity of R&D management. The benefits and transaction costs of managing alliances have to be constantly balanced, keeping the strategic interests of the company in mind. In the knowledge management literature, it is recognized that outsourcing practices can have the short-term effect of diluting the firm's competencies and the long-tenn effect of increasing competition in the industry by diffusion of best knowledge, skills and competencies to other players in the field through the service com- panies {23).

In this study, biostatistics was recognized as a core com- petence, especially in protocol development and advanced data analysis, but we found differing perspec- tives on the management styles and philosophies; one group (typieally non-bio statist ics executives) stated that biostatistics was only a process tool and technique, while another group (typically biostatistician middle managers or scientists) argued that there was significant innovation in product development that could be driven by better use of biostatistics. The increased outsourcing recorded here, particularly of the "thinking" functions of a critical com- petency such as biostatistics, would certainly reduce the potential for capturing and nurturing innovation inter- nally.

This loss of innovation potential was evident in one of our case interviews. In a clinical study with a new drug for treating sleep disorders, an internal biostatistician with experience in this domain included a couple of extra measures in the study design. The original indication failed to show efficacy but the extra parameters recorded led to a realization of a new and unforeseen indication. Because the biostatistician was employed by the company, she had the capability and support to change the protocol and research design. However, the described critical insight, input to the protocol and resulting inno- vation would be much harder to capture if the study was being outsourced, in which case an external biostatisti-

Outsourcing cau dilute a firm's

competencies and increase competition

in tiie industry. cian might be rewarded for reducing cost (minimal intro- duction of new measures or tests) or increasing speed.

From the results summarized here, we note that although large pharma companies stated that they outsource only extension studies or non-critical studies, there is still an exchange with the CRO in terms of best practices, standard operating principles and analytical methods. This gives rise to the diffusion of competencies in the industry. The data from the CROs (Table 2) demonstrate the surprising number of projects that involve either trial design or have the CRO's write-ups presented to the FDA directly, both practices identified as "never outsource" and of "key importance to grow and retain" by the pharmaceutical statistics directors.

The sharing and outsourcing of these big pharma best practices with the C ROs is an advantage for smal ler com- petitors, which do not have the history or resources to develop optimal processes in biostatistics, particularly as related to analysis methods and protocols that are likely to be accepted by the FDA. On the one hand, CROs can be viewed as instruments for diffusion of institutional practices and thereby a route for small firms to gain legitimacy (34.25). In this view, CROs, through their role in institutionalization of biostatistical practices in drug development, are a mechanism for bringing greater stability to drug product life-cycle dynamics.

In gaining experience in areas of strategic competence for success in the industry—regulatory interactions and study protocol design—the CROs are gradually posi- tioned to take on an increased role in the drug develop- ment process. A direct recognition of their growing competence in these areas is seen in the steadily increas- ing amount of outsourced R&D from pharmaceutical companies. In fact, the capabilities of some of the larger CROs. combined with their cash positions and steady income stream from long-term contracts, will eventually lead to their emergence as risk-takers and co-investors in the drug development process, where they will appropri- ate more vaiue due to their increased competence.

Mav—June 2007

A survey of 31 leading CROs by CenterWatch in 2000 showed that risk-sharing in CRO contracts was expected to increase (2/), and recent news reports verify these trends toward increased investment and risk-sharing in products by CROs {26). Thus, new business models will emerge with a sharing of jobs among industry players, hopefully leading to reduced risk for each player, and increasing pipeline productivity, which could lead to greater innovation and new products reaching the public. However, this scenario is still developing, as is the entire pharmaceutieal industry, currently under severe pricing, market and regulatory pressures.

We observe in conclusion that biostatistics is a contested, immobile competency in drug development. This per- spective proposes that over time, an aeetnnulation of critical resources and tacit knowledge will allow the CRO to eventually become a new type of independent competitive player in the drug development process. Such an outcome would perhaps increase turbulence by changing industry ecosystem dynamics through increased CRO competitiveness and reduced sharing of best practices in the industry, as opposed to the potential stabilizing influence that the CRO could exert as a reliable interdependent player in the industry. This implies that managerial decision-making coupled with complex biostatistical methodology should be reevalu- ated by pharmaceutical companies in terms of how bio- statistical competence could be developed into a core capability that contributes to a sustained competitive advantage. ®

Acknowledgment

We wish to acknowledge the contributions of Mark Knight, a graduate student at the Lally school, who was of great help in conducting the interviews and in collect- ing the data from the CROs.

References

1. 1991. Manufacturing and the ins and outs of outsourcing. (The) Econonmt. 31 .Aug.. pp. 55-56. 2. McFarlan. F. W. and Nolan. R. L. 1995. How to manage an IT oulsourcing alliance. Shan Managemeni Review 36(2), pp. 9-23. 3. 1991. The Activist Centre. Boston Consulting Group Inc. Bos- ton, MA. 4. Quinn, J. B. 2000. Outsourcing Innovation: The New Engine of Growth. Sloan Managemeni Review4\{4).. pp. 13-28. 5. Quinn.J. B., 1999. Strategic Outsourcing: Leveraging Knowledge Capabilities. Shan Management Review. Summer. 40(4), pp. 9-2 I. 6. Penrose. E. \95>9. The theory oftheg:i-owth of the firm. io\m'^\\iiy & Sons: New York, 7. Ilanii, H. 1987. Mobilizing invisible assets. Harvard University Press. Cambridge, MA. 8. Dierickx. I. and Cool, K.. 1989. Asset stock accumulation and sus- tainability of competitive advantage. Managemeni Science 35(12), pp. 1504-1511. 9. Prahalad, C. K. and Hamel, G. 1990. The core competence ojlhe corporation. Harvard Business Review 90(3). pp. 79-93.

CROs could eventually become a

new competitive piayer in tne drug

levelopment process.

10. Conner, K. 1991. A historical comparison of resource-based theory and five schools of thought within industrial organization economics: Do we have a new theory of the firm? Journal of Man- agement 17. pp. 121-154. 11. Barney. J. 1991. Firm resources and sustained competitive a.d\anK'^ge. Journal of Management 17(1), pp. 99-120. 12. Hall, R. 1992, The strategic analysis of intangible resources. Strategic Management Journal. 13, pp. 135-144. 13. Marino, K. 1996. Developing consensus on firm competencies and capabilities. Academv of Management Executive 10(3), pp. 40-51. 14. Teece, D, J., Pisano, 0 . and Shuen. A. 1997, Dynamic capabili- ties and strategic management. Strategic Management Journal 18(7), pp, 509-533. 15. Helfat, Constance E. and Raubitschek, R. S. 2000. Product sequencing: Co-evolution of knowledge, capabilities and products. Strategic Management Journal 2\. pp. 961-979, 16. Eisenhardt. K. M. and Martin. J, 2000. Dynamic capabilities: what are they ? Strategic Management Journal 21. pp. II05-1121. 17. Lucarelli. Chris and Peters. L. S, 2001, Developing Competen- cies and Capabilities Through Knowledge Management: Contingent Perspective 2001. In: Technology' Management in the Knowledge Era (eds.) Dundar F. Kocaoglu and Timothy Anderson) pp. 257-267. 18. lansiti. Marco. 2004. The Keystone Advantage: What the New Dynamics of Business Ecosystenu Mean for Strategy. Innovation, and Sustainability. Harvard Business School Press. Boston. MA. 19. Cockbum. Ian. Henderson, R., and Stem, S. 2000. Untangling the Origins of Competitive Advantage. Strategic Management Journal 21, pp. 1123-1145, 20. Achilladelis. B- and Antonakis, N. 2001, The dynamics of tech- nological innovation: the case of the pharmaceutical industry. Research Policy 30, pp, 535-588. 21. 2001. Parexel R&D saurcebook. (pages 25 and 96) and 2004/ 2005 (pages 26-29, p. 49), Parexel International. Waltham. MA. 22. Williamson, Oliver. 1991. Strategizing, economizing and economic organization. Strategic Management Journal 12, pp, 75-94. 23. Leavy, Brian. 2004. Outsourcing strategies: opportunities and risks. Strategy & Leadership 32(6), pp. 20-25. 24. Rogers. Everett. 1995. Dijjusion of Innovations. A\\vQ<iL\^on. Free Press, New York. 25. Scott, W. Richard. 1995. Institutions and Organizations. Sage Publications, Thousand Oaks. CA. 26. Lee, John, 2004. Quintiles trading $25M in services for stake in drugs. Triangle Business Journal. Oct. 25, Accessed at http:// www.biziournals.com/industries/heatthcare/pharmaceuticals/2004/ 10/25/trianglestory8.html.

Research - Technology Managenu'nt