Printer Friendly

Interorganizational exchanges as performance markers in a community cancer network.

In recent years, health services researchers and practitioners have noted the proliferation of "strategic partnerships" in the health services sector. These partnerships include hospital and health service alliances, organ procurement networks, hospital-physician joint ventures, and technology development alliances between university and industrial scientists. Luke, Begun, and Pointer (1989) identify two characteristics that make these partnerships unique: (1) the "loose coupling" that ties the collaborating organizations together and (2) the importance and permanence of their strategic purposes. They and others (Kaluzny and Zuckerman 1992) point out that both conceptual and empirical work are lacking on these interorganizational alliances.

What appears to be a new kind of cooperative strategy for health service organizations is, in fact, a tradition in the realm of cancer clinical research. For more than 35 years, clinical cooperative groups have been pooling knowledge and resources to develop new cures for cancer. These experiences, along with their more recent efforts to transfer new diagnostic and treatment technologies to community medical practices, make the cooperative groups prototypes for studying strategic partnerships. This article explores issues surrounding the development and performance of strategic partnerships within the context of a federally sponsored Community Clinical Oncology Program, which links community oncologists with clinical research programs in cancer centers and clinical cooperative groups.

BACKGROUND

In September 1983, the National Cancer Institute (NCI) introduced the Community Clinical Oncology Program as a mechanism for disseminating state-of-the-art cancer treatments and patient management techniques to community oncologists and their patients. Spurred by congressional concern that the billions of dollars being spent on cancer research were having only minimal impact on cancer incidence and mortality, the Community Clinical Oncology Program sought to accelerate the transfer of research findings to community settings where more than 80 percent of all cancer patients are treated. A second (CCOP-II) funding cycle, initiated in June 1987, expanded the program's focus to include research on cancer prevention, screening and early detection, continuing care, and rehabilitation. Both funding cycles awarded cooperative agreements to community-based consortia of oncologists and hospitals (i.e., CCOPs) to enroll patients in clinical trials through cancer centers and clinical cooperative groups that the NCI had designated as CCOP "research bases." These CCOP-research base alliances function as "R&D partnerships," with research bases designing and monitoring the clinical trials, and the CCOPs contributing to knowledge development by enrolling patients and providing feedback on patient responses to different treatment regimens.

The application guidelines for the Community Clinical Oncology Program require each CCOP to affiliate with at least one, but no more than five, NCI-supported research bases for the purpose of enrolling patients in cancer treatment and cancer control clinical trials. Patients who participate in these clinical trials receive the state-of-the-art therapies or cancer control interventions through participating CCOP oncologists in their local communities. The interactions between a CCOP and its research base(s) have different relational contents (Hakansson 1982). Information exchange occurs when CCOP representatives attend research base meetings, serve on scientific and administrative committees, participate in on-site data audits, and exchange written reports and correspondence. Resource exchanges include the exchange of research protocols and the enrollment of patients in clinical trials.

In the decade since the Community Clinical Oncology Program has become operational, some clinical cooperative groups have encouraged CCOP physicians to join their scientific committees, to design protocols for the administration of clinical trials, and to chair clinical research studies, while others have discouraged, or given only weak support, to such participation. Likewise, some CCOPs have actively sought major involvement in cooperative group scientific activities while others have functioned strictly as "patient contributors." These different interaction strategies appear to have important implications for CCOP performance; yet no one has examined how different types and levels of exchange between CCOPs and their cooperative groups affect a key performance measure -- the number of patients enrolled in clinical trials. This study draws upon interorganizational theory to develop and test hypotheses about how CCOP-research base relationships emerge, develop, and influence patient accruals (i.e., the number of patients enrolled in clinical trials) over time. The findings have implications, not only for National Institutes of Health-sponsored clinical research partnerships but also for the broader health services research community, which seeks to understand the kinds and levels of interactions required to bridge the gap between researchers and health care practitioners.

PREDICTORS OF CCOP-RESEARCH BASE EXCHANGES

Interorganizational theory suggests that a CCOP will enter into an exchange relationship with a research base to gain access to critical resources such as protocols, investigational drugs, and experiential knowledge (Kogut 1988; Powell 1990). Because much of the required knowledge is embedded in the experience and skills of group investigators (Osborn and Baughn 1990), CCOP representatives need to attend research base meetings and serve on scientific committees in order to tap into these information sources. The greater the resource dependence, the more the two organizations should communicate with each other as they try to assess the other's resources, capabilities, and expectations (Ford 1982; Van de Ven and Walker 1984). Thus, resource dependence should be positively associated with the "extensity" of initial information exchanges (Hallen, Mohamed, and Johanson 1989), as measured by the number of CCOP representatives who participate in meetings and committees.

Since resource dependence reflects the extent to which a CCOP expects to participate in group trials, the level of resource dependence should have a direct bearing on the number of different protocols used. CCOP oncologists see patients with different tumor types and in different stages of disease. Even when patients have the same tumor type, they may require different protocols because of their age, medical history, or performance status. These variations in patient characteristics suggest that a CCOP that is highly dependent on one research base's protocols will need to use a greater range of protocols (i.e., engage in "wider" protocol exchange) to meet different patient needs. Therefore,

Hypothesis 1. Resource dependence at the onset of the relationship will be positively associated with initial levels of information and protocol exchange.

EXCHANGE RELATIONSHIPS

By attending research base meetings and serving on scientific committees, CCOP physicians and support staff obtain important information on the objectives, eligibility requirements, and treatment plans of new clinical trials. The importance of these information exchanges is illustrated by the following comment from a clinical investigator:

The perception |that multicenter trials can be conducted with little or no contact between investigators~ is based on the antiseptic view that the protocols of trials and the treatment processes can be carried out by a series of cut and dried steps that, once established, can be followed more or less by rote. Overlooked in this view is the importance of dialogue, debate, and consensus formation as a means of establishing and maintaining a spirited, quality-conscious investigatorship. The quality and success of a trial depends on having trained people with detailed knowledge of the protocol. It is not possible to achieve and maintain that kind of familiarity without continuing contact through regular face-to-face meetings, usually more frequently than once a year. (Meinert 1990, 400)

Cunningham and Turnbull (1982) observe that information exchanges usually precede the exchange of products (protocols in this study). Participation in research base meetings and on scientific committees helps to resolve uncertainties that CCOP physicians may have about the scientific merit of new studies or the feasibility of certain protocols for community application. As uncertainty is reduced, so is the perceived risk of enrolling patients on the research protocols.

Face-to-face meetings between CCOP and university-based investigators also provide opportunities for joint protocol development, making it possible for research base studies to be better adjusted to the needs of the "using systems" (Mattsson 1978, p. 212). The more opportunities that CCOP physicians have to plan and revise research protocols, the more receptive they should be to testing different protocols (Hakansson 1987). "Wide" protocol exchange should, in turn, lead to more extensive information exchange because the two organizations have to know more about each other's performance, problems, and findings (Hallen, Johanson, and Mohamed 1987; Lazerson 1988). This suggests that

Hypothesis 2. Information exchange and protocol exchange will be positively and reciprocally related over time.

Interdependencies between a CCOP and a research base require that the two organizations coordinate their activities (Johanson and Mattsson 1987). This coordination can be achieved through an ongoing information exchange process in which the two organizations each learn about the other's capabilities and limitations, and assess whether or not the interorganizational match is, in fact, a good one (Fichman and Levinthal 1991). Research base scientific sessions and committee meetings accelerate the flow of information by bringing CCOP and university-based investigators into direct contact with one another and by encouraging them to develop a common language and framework for solving complex scientific problems (Daft and Lengel 1984). These information exchanges should build interorganizational contact patterns that encourage further investments in the relationship (Larson 1992; Wilson and Mummalaneni 1986).

As a CCOP-research base relationship evolves, the links between the organizations should grow stronger and more stable (Hakansson and Johanson 1988). Information exchanges acquire value -- not just as sources of technical knowledge but as sources of relationship knowledge and personal satisfaction (Kanter 1989). When information exchanges enhance the intrinsic worth of a CCOP-research base relationship, they are likely to be repeated (Scott 1987) and to grow in strength over time. Therefore,

Hypothesis 3. Information exchange in one period will be positively associated with information exchange in the next period, and this positive association will grow stronger over time.

EXCHANGES AND ACCRUAL PERFORMANCE

Meeting attendance and committee service build information channels between a CCOP and its research base at several organizational levels (e.g., principal investigators, physicians, oncology nurses, and data managers). The more representatives a CCOP sends to research base and scientific committee meetings, the more opportunities it has to acquire information on new methods of cancer detection and treatment. The strong ties created through these information exchanges help to reduce the distance between the CCOP and the research base (Hallen, Johanson, and Mohamed 1987), thereby increasing both the speed and credibility of the information flow. Repeated interactions with research base investigators who are participating in new clinical trials also provide legitimation and support for the studies, creating strong peer pressures to enroll patients in the trials (Anderson and Jay 1985). This positive relationship between information exchange and accrual performance is demonstrated by the Community Cancer Care Evaluation of CCOP-I, which found a "clear trend toward increased |meeting~ attendance by the more successful CCOPs" (Feigl, Patmont, Rodenbaugh, et al. 1987, 20).

When CCOP physicians use many different research protocols (i.e., engage in "wide" protocol exchange), they should find it easier to match patients with protocols and, thus, to boost patient enrollments. The willingness to try different protocols also reduces a CCOP's dependence on a few clinical research studies so that accrual performance can be maintained even when these studies close. The Community Cancer Care Evaluation's discovery of a positive relationship between the number of distinct protocols used per unit time and CCOP patient accruals lends support to this argument (Feigl, Patmont, Rodenbaugh, et al. 1987). Consequently,

Hypothesis 4. Information exchange and protocol exchange will be positively associated with CCOP accrual performance over time.

METHODS

The hypotheses predict how two types of CCOP-research base exchange -- one involving information and the other protocols -- will relate to one another and to CCOP patient accruals over time. Since the interaction between a CCOP and a research base is the topic of interest, the unit of analysis is the CCOP-research base relationship. If a CCOP is affiliated with three research bases, for example, each affiliation constitutes a separate relationship.

The study period covers six time points that span two Community Clinical Oncology Program funding cycles -- CCOP-I and CCOP-II. A "pre-CCOP" time point (|t.sub.0~) includes measures of a CCOP's relationship with its research base in the year before the CCOP-I award. Through programs such as the Cooperative Group Outreach Program (CGOP), some community oncologists began accruing patients to cooperative groups' clinical trials before the Community Clinical Oncology Program became operational.(1) The remaining time points (|t.sub.1~-|t.sub.6~) correspond to reporting periods established by the NCI for the purpose of monitoring patient accruals.

SAMPLE SELECTION

Study analyses are based on 65 pairwise relationships that 38 CCOPs established with eight clinical cooperative groups in September 1983 (the beginning of the CCOP-I funding cycle) and maintained through February 1989 (the end of the second reporting period of CCOP-II). The cooperative groups represent the major multidisciplinary and specialty research bases that participated in both CCOP-I and CCOP-II.(2) The six other research bases that participated in CCOP-I and CCOP-II are NCI-supported cancer centers that design most protocols in-house.

VARIABLES AND MEASURES

CCOP performance is measured by the number of patients that a CCOP enrolls in a cooperative group's clinical trials during a reporting period. The hypotheses focus on three independent variables: resource dependence, information exchange, and protocol exchange. Resource dependence refers to a CCOP's need to interact with a cooperative group to acquire needed resources. The level of resource dependence at the onset of a CCOP-research base relationship is measured by the proportion of a CCOP's total projected patient enrollments that is attributable to a particular cooperative group.

Two measures of information exchange are employed: (1) the number of physicians, nurses, and data managers from the CCOP who attend at least one cooperative group meeting during a reporting period, and (2) the number of CCOP representatives who serve on at least one cooperative group committee during a reporting period. These information exchanges require different kinds and amounts of involvement and, therefore, represent different levels of relationship intensity.

CCOPs enter into affiliation agreements with cooperative groups to gain access to research protocols. The level of protocol exchange in each CCOP-research base relationship is measured by the number of different cancer treatment protocols to which the CCOP accrues at least one patient during a reporting period.

Because CCOP-research base interactions may be influenced by the characteristics of the interacting organizations, this study includes pre-CCOP clinical trials experience and CCOP size as additional explanatory variables. Previous experience in working together can affect the level of importance attached to the relationship and, thus, the extent to which the two organizations interact. As measures of pre-CCOP clinical trials experience, this study uses (1) the number of patients enrolled by CCOP physicians in a cooperative group's clinical trials, through the CGOP and other programs, in the year prior to the CCOP-I award, and (2) a ratio of the number of physicians in the CCOP who enrolled patients in a cooperative group's clinical trials in pre-CCOP years to all physicians listed in the CCOP's cooperative agreement application. CCOP size is operationalized as (1) the number of physicians in the CCOP who enrolled patients in clinical trials during a reporting period and (2) the total number of physicians who participated in the CCOP during a reporting period, including those who referred patients to CCOP oncologists or assisted with patient diagnoses.

DATA SOURCES AND ANALYSIS

Study data were collected from grantee applications and progress reports as part of a larger evaluation of the Community Clinical Oncology Program. Pearson correlations and regression models were used to test hypothesized relationships between variables.(3) Depending upon the hypothesis, relationships between variables were tested in the same reporting period or in adjoining periods. Secondary analyses examined whether regression models could be improved by adding either measures of CCOP size, pre-CCOP clinical trials experience, or both, as additional explanatory variables.

RESULTS

Hypothesis 1 predicts that resource dependence at the onset of the relationship will be positively associated with initial levels of information and protocol exchange. Table 1 suggests that resource dependence has a greater influence on the "width" of protocol exchange than on the "extensity" of information exchange. Three factors account for more than three-quarters of the variance in the number of cancer treatment protocols used at |t.sub.1~: (1) the CCOP's level of resource dependence, (2) the number of CCOP representatives attending cooperative group meetings during the first reporting period, and (3) the CCOP's previous experience with the cooperative group's clinical trials as measured by pre-CCOP patient accruals (i.e., the number of patients enrolled in clinical trials in the pre-CCOP year) and the MD experience ratio (i.e., the proportion of CCOP physicians that enrolled patients in the cooperative TABULAR DATA OMITTED group's clinical trials in pre-CCOP years). Resource dependence, meeting attendance, and experience with the cooperative group seem to build a "comfort factor for risk taking" (Quinn 1978, 11) that encourages CCOP physicians to take greater advantage of the array of protocols available to them.

The conditions leading to different types of information exchange during the early stage of a CCOP-research base relationship are less clear. Model 2 in Table 1 suggests that a little more than half of the variance in the number of CCOP representatives attending cooperative group meetings at |t.sub.1~ can be explained by three factors: (1) the CCOP's level of resource dependence, (2) the number of CCOP representatives serving on cooperative group committees during the first reporting period, and (3) the total number of physicians participating in the CCOP. As predicted by Hypothesis 1, resource dependence provides a strong incentive to obtain firsthand information on cooperative group protocols and procedures. The statistical significance of committee membership suggests that the CCOP physicians and support staff who are most actively involved in clinical research will be the first to attend cooperative group meetings. CCOP size reflects both the breadth of interest in clinical research and the amount of human and financial resources that can be allocated to relationship development.

Fifty-five percent of the CCOP-research base relationships had no CCOP representatives as committee members during the first reporting period. Model 3 in Table 1 shows that, together, resource dependence and the number of patients accrued in the pre-CCOP year account for 55 percent of the variance in |t.sub.1~ committee membership. While resource dependence provides some incentive to join cooperative group committees, the high statistical significance of pre-CCOP accrual performance (t = 6.48, p |is less than~ .001) suggests that CCOP physicians are more likely to seek (and receive) early committee appointments when they have established accrual track records with the cooperative group through the CGOP and other pre-CCOP relationships.

The hypothesis that information exchange and protocol exchange will be positively and reciprocally related over time is most strongly supported when meeting attendance is used as the information exchange measure. As shown in Table 2, the positive association between the number of CCOP representatives attending cooperative group meetings and the number of different cancer treatment protocols used is .60 or higher during four of the six reporting periods. The associations are particularly strong during the first reporting period of CCOP-I (|t.sub.1~), the extension period between CCOP-I and CCOP-II (|t.sub.4~), and the second reporting period of CCOP-II (|t.sub.6~). At times 2, 3, 5, and 6, the number of cancer treatment protocols used is most highly correlated (r |is greater than or equal to~ .65) with the number of CCOP representatives attending cooperative group meetings at |t.sub.1~.(4) This finding suggests that the level of meeting attendance during the first year of a CCOP-research base relationship has an important influence on the number of cancer treatment protocols used in subsequent years. As the relationship evolves, continued meeting attendance helps to sustain protocol exchange by providing both information on new protocols and updates on the status of ongoing studies.

Although the number of CCOP representatives serving on cooperative group committees is positively associated with the number of different cancer treatment protocols used at each time point, the relationships are weaker than those observed for meeting attendance and protocol exchange. In every reporting period, protocol exchange is most highly correlated (r |is greater than or equal to~ .52) with the number of CCOP representatives serving on cooperative group committees at |t.sub.1~. There are also fairly strong relationships (r |is greater than or equal to~ .53) between the number of different cancer treatment protocols used at |t.sub.1~ and the number of CCOP representatives serving on committees (or attending meetings) in succeeding reporting periods. These findings suggest that the reciprocal relationship between information exchange and protocol exchange extends TABULAR DATA OMITTED over time, with the initial levels of exchange setting standards for subsequent CCOP-research base exchanges.

Hypothesis 3 predicts that information exchange in one period will be positively associated with information exchange in the next period and that this positive association will grow stronger over time. Table 2 displays correlations between the information exchange measures across the six reporting periods. As predicted, the positive association between the number of CCOP representatives serving on cooperative group committees at |t.sub.i~ and the number serving on committees at |t.sub.i + 1~ grows stronger over the study period. With the exception of committee membership at |t.sub.4~ and |t.sub.5~, there is a steady increase in the size of the correlation coefficients from .84 (|t.sub.1~ and |t.sub.2~) to .99 (|t.sub.5~ and |t.sub.6~).

Although meeting attendance at |t.sub.i~ is positively associated with meeting attendance at |t.sub.i + 1~, these associations grow weaker over the study period. At |t.sub.4~ and |t.sub.6~, the number of CCOP representatives attending cooperative group meetings is more strongly associated with meeting attendance at |t.sub.1~ and |t.sub.2~ than with meeting attendance in the previous period. The number of CCOP representatives attending cooperative group meetings at |t.sub.5~ also shows a rather strong correlation (r = .60) with meeting attendance at |t.sub.2~. These findings highlight an important distinction between meeting attendance and committee membership trends. While meeting attendance patterns appear to be established during the first two years of a CCOP-research base relationship, committee membership patterns seem to evolve more slowly over time. Because committee membership requires a higher level of involvement in group scientific activities, CCOP representatives are not likely to seek (or receive) committee appointments until they are familiar with and committed to the group's research agenda.

The positive associations between committee membership at |t.sub.i~ and meeting attendance at |t.sub.i + 1~ generally are weaker than those for meeting attendance at |t.sub.i~ and meeting attendance at |t.sub.i + 1~. In the last three reporting periods, the number of CCOP representatives attending cooperative group meetings shows a stronger association with meeting attendance at |t.sub.1~ and |t.sub.2~ than with committee membership in the previous period.

Throughout the study period, the positive associations between meeting attendance at |t.sub.i~ and committee membership at |t.sub.i + 1~ are notably weaker than those for committee membership at |t.sub.i~ and committee membership at |t.sub.i + 1~. However, Table 2 reveals strong relationships (r |is greater than equal to~ .87) between the number of CCOP representatives attending cooperative group meetings at |t.sub.3~ and the number serving on committees in subsequent reporting periods. Although Table 2 does not display correlations between meeting attendance and protocol chair appointments, it is interesting to note that, in the last three reporting periods, the number of CCOP representatives chairing research protocols is more highly correlated (r |is greater than or equal to~ .61) with |t.sub.3~ meeting attendance than with meeting attendance in any other reporting period. These associations suggest that the level of meeting attendance in the third year of a CCOP-research base relationship may be a good "marker" of subsequent CCOP involvement in group scientific activities.

The hypothesis that information exchange and protocol exchange will be positively associated with CCOP accrual performance receives strong support throughout the study period. As shown in Table 3, two-thirds or more of the variance in patient accruals at each time point can be explained by the number of different cancer treatment protocols used and the number of CCOP representatives serving on cooperative group committees (or attending cooperative group meetings). The proportion of explained variance is especially high (|is greater than or equal to~ 80%) during the first two reporting periods of CCOP-I and the first reporting period of CCOP-II. This finding suggests that information and protocol exchanges have a particularly strong influence on CCOP accrual performance when a CCOP-research base relationship is being formed or reorganized. By promoting communication and cooperation, these exchanges help the two organizations to overcome the "liabilities of newness" that threaten initial performance and that re-emerge when roles and routines undergo reorganization (Hannan and Freeman 1984). The inclusion of CCOP size measures in the aforementioned models adds little to their explanatory power.

DISCUSSION AND IMPLICATIONS

The foregoing analyses of CCOP-cooperative group exchanges can be compared to biochemists' attempts to find "tumor markers" that are useful in diagnosing early cancers or to the Joint Commission on Accreditation of Healthcare Organizations' search for organizational indicators that predict clinical outcomes. If each type of exchange is viewed as a "marker" of future CCOP performance, it is possible to identify critical points in the relationship development process when carefully planned interventions can increase community physicians' participation in clinical trials research. These findings have several policy implications for the Community Clinical Oncology Program and for other kinds of health care strategic partnerships.

TABULAR DATA OMITTED

For community-based health research initiatives, such as the Community Clinical Oncology Program and the Community Programs for Clinical Research on AIDS, the study findings suggest that the number of community clinicians who attend scientific meetings in the first two years of the clinical research partnership has a significant influence on meeting attendance and protocol use in later years. The more community clinicians participate in scientific meetings and diversify their protocol use, the higher their accrual performance should be. The National Institutes of Health could encourage more community clinicians to attend scientific meetings by increasing the travel component of grant awards and/or by requiring a local match for physician and staff travel. Clinical research groups could promote higher meeting attendance by offering more scientific sessions and workshops on topics that are of interest to community-based practitioners, seeking higher continuing medical education credits for meeting participation, and giving points for meeting attendance in the performance review process.

This study also suggests that the participation of community clinicians on protocol design and monitoring committees stimulates higher accrual performance. Since committee membership requires a greater commitment of time and effort than meeting attendance, the clinical research groups may need to take special steps to encourage early participation by community clinicians. These steps might include (1) forming "new investigator" committees to stimulate interest in group clinical trials and research opportunities, (2) inviting community-based physicians and nurses to meet with committee chairs and participate in protocol reviews, and (3) asking community-based "opinion leaders" who are strongly committed to clinical research to help recruit some of their colleagues for committee service. Broader protocol use could be encouraged by simplifying protocol eligibility criteria, minimizing the amount of time that must be spent on test scheduling and patient follow-up, and eliminating laboratory tests and other protocol requirements that impose unreasonable cost burdens on study subjects.

Although the dyadic interactions between community-based practitioners and clinical research groups are the focus of this study, other environmental influences that may mediate these relationships need to be examined. Vertical linkages with agencies such as the NCI, the Food and Drug Administration, and pharmaceutical companies need to be considered along with lateral linkages with primary care and specialty physicians, members of affected communities, and hospital institutional review boards. The more the expectations and demands of these entities conflict with one another, the less responsive community clinicians may be to interventions designed to increase their participation in clinical trials research.

This study also has implications for the broader health care arena where strategic partnerships such as managed care systems, trauma networks, and technology development alliances between university and industrial scientists are becoming increasingly common. The findings highlight the importance of historical relationships and anticipated resource dependence in shaping initial exchange patterns. They also suggest that strategic partnerships need to emphasize structures and processes that encourage early involvement in collaborative activities and that reward participants for maintaining high levels of interaction. Exchange patterns established during the early years of a strategic partnership seem to "imprint" the relationship with performance standards and relational expectations that persist over time.

This study offers evidence that early communication on each partner's needs, expectations, and priorities is critical to the development of a successful strategic partnership. This is particularly true when the partners have different missions, professional philosophies, and operating procedures. The designation of "boundary role persons" to participate in joint meetings and plan collaborative activities allows the organizations to become familiar with each other's ways of working and to develop shared perceptions of each other's needs and offers. Future studies need to investigate how turnover among these boundary role persons affects the length of time required for relationship development and the performance of the strategic partnership.

Finally, the study suggests that the formation of a strategic partnership takes much more time than is generally recognized by funders and policymakers. Government agencies and private foundations often expect a strategic partnership to be fully operational in the first year of funding when, in fact, the relationship may take years to develop. Repeated interactions and mutual adaptations are needed to build confidence in the relationship at both personal and organizational levels. The incremental nature of the relationship development process suggests that evaluations of health care strategic partnerships need to be longitudinal, process-oriented, and attentive to exchanges that may be "markers" of future performance.

NOTES

1. The NCI established the Cooperative Group Outreach Program in 1976 to involve community oncologists and their patients in clinical trials. Contracts are awarded, on a competitive basis, to cooperative groups to provide data management support to participating community oncologists and to cover the costs of data analysis and quality control systems.

2. The clinical cooperative groups include the Cancer and Leukemia Group B (CALGB), the Eastern Cooperative Oncology Group (ECOG), the Southwest Oncology Group (SWOG), the North Central Cancer Treatment Group (NCCTG), the National Surgical Adjuvant Project for Breast and Bowel Cancers (NSABP), the Radiation Therapy Oncology Group (RTOG), the Children's Cancer Study Group (CCSG), and the Pediatric Oncology Group (POG).

3. Like other empirical studies of interorganizational exchanges, this study's use of the CCOP-research base relationship as the unit of analysis violates the linear regression assumption that the error terms are uncorrelated. Since a CCOP can affiliate with more than one cooperative group, the outcome in one CCOP-cooperative group exchange may affect the outcome in another CCOP-cooperative group exchange. Among the 38 CCOPs studied, 18 (47 percent) maintained affiliations with just one cooperative group, 14 (37 percent) maintained two affiliations, five (13 percent) maintained three affiliations, and one CCOP had four cooperative group affiliations. The authors considered using the jackknife technique or Monte Carlo analyses to address this potential problem but concluded that the study sample was not of sufficient size to apply these techniques appropriately. Therefore, a decision was made to use the more conservative significance level of p |is less than~ .01 in interpreting study results.

4. Protocol exchange at |t.sub.6~ also shows a strong association (r = .75) with |t.sub.4~ meeting attendance.

REFERENCES

Anderson, J. G., and S. J. Jay. "The Diffusion of Medical Technology: Social Network Analysis and Policy Research." The Sociological Quarterly 26, no. 1 (Spring 1985): 49-64.

Cunningham, M. T., and P. W. Turnbull. "Inter-Organizational Personal Contact Patterns." In International Marketing and Purchasing of Industrial Goods: An Interaction Approach. Edited by H. Hakansson. New York: John Wiley & Sons, 1982.

Daft, R. L., and R. H. Lengel. "Information Richness: A New Approach to Managerial Behavior and Organization Design." In Research in Organizational Behavior Vol. 6. Edited by B. M. Staw and L. L. Cummings. Greenwich, CT: JAI Press, 1984.

Feigl, P., C. Patmont, J. Rodenbaugh, J. Smith, G. Glaefke, C. Allison, M. Bergner, M. Schmidt, V. Immanuel, T. Corlett, and P. Diehr. Community Cancer Care Evaluation (CCCE) Final Report: Vol. 5, Integrated Analysis. NCI Contract No. N01-CN-35009. Seattle, WA: Fred Hutchinson Cancer Research Center, February 19, 1987.

Fichman, M., and D. A. Levinthal. "History Dependence in Professional Relationships: Ties that Bind." In Research in the Sociology of Organizations Vol. 8. Edited by S. B. Bacharach, P. S. Tolbert, and S. Barley. Greenwich, CT: JAI Press, 1991.

Ford, D. "The Development of Buyer-Seller Relationships in Industrial Markets." In International Marketing and Purchasing of Industrial Goods: An Interaction Approach. Edited by H. Hakansson. New York: John Wiley & Sons, 1982.

Hakansson, H. "An Interaction Approach." In International Marketing and Purchasing of Industrial Goods: An Interaction Approach. Edited by H. Hakansson. New York: John Wiley & Sons, 1982.

-----. "Technological Innovation Through Interaction." In Industrial Technological Development: A Network Approach. Edited by H. Hakansson. London: Croom Helm, 1987.

Hakansson, H., and J. Johanson. "Formal and Informal Cooperation Strategies in International Industrial Networks." In Cooperative Strategies in International Business. Edited by F. J. Contractor and P. Lorange. Lexington, MA: Lexington Books, 1988.

Hallen, L., J. Johanson, and N. S. Mohamed. "Relationship Strength and Stability in International and Domestic Industrial Marketing." Industrial Marketing and Purchasing 2, no. 3 (1987): 22-37.

Hallen, L., N. S. Mohamed, and J. Johanson. "Relationships and Exchange in International Business." In Advances in International Marketing Vol. 3. Edited by L. Hallen and J. Johanson. Greenwich, CT: JAI Press, 1989.

Hannan, M., and J. H. Freeman. "Structural Inertia and Organizational Change." American Sociological Review 49, no. 2 (April 1984): 149-64.

Johanson, J., and L-G Mattsson. "Interorganizational Relations in Industrial Systems: A Network Approach Compared with the Transaction-Cost Approach." International Studies of Management and Organization 17, no. 1 (1987): 34-48.

Kaluzny, A. D., and H. S. Zuckerman. "Strategic Alliances: Two Perspectives for Understanding Their Effects on Health Services." Hospital & Health Services Administration 37, no. 4 (Winter 1992): 477-90.

Kanter, R. M. "Becoming PALs: Pooling, Allying, and Linking Across Companies." Academy of Management Executive 3, no. 3 (August 1989): 183-93.

Kogut, B. "Joint Ventures: Theoretical and Empirical Perspectives." Strategic Management Journal 9, no. 4 (July-August 1988): 319-32.

Larson, A. "Network Dyads in Entrepreneurial Settings: A Study of the Governance of Exchange Relationships." Administrative Science Quarterly 76, no. 1 (March 1992): 76-104.

Lazerson, M. H. "Organizational Growth of Small Firms: An Outcome of Markets and Hierarchies?" American Sociological Review 53, no. 3 (June 1988): 330-42.

Luke, R. D., J. W. Begun, and D. D. Pointer. "Quasi-Firms: Strategic Interorganizational Forms in the Health Care Industry." Academy of Management Review 14, no. 1 (January 1989): 9-19.

Mattsson, L. "Impact of Stability in Supplier-Buyer Relations on Innovative Behavior of Industrial Markets." In Future Directions for Marketing: Proceedings of Two European Colloquia. Edited by G. Fisk, J. Arndt, and K. Gronhaug. Cambridge, MA: Marketing Science Institute, 1978, pp. 207-17.

Meinert, C. L. "Multicenter Trials and Travel: Luxury or Necessity?" Controlled Clinical Trials 11 (1990): 399-401.

Osborn, R. N., and C. C. Baughn. "Forms of Interorganizational Governance for Multinational Alliances." Academy of Management Journal 33, no. 3 (September 1990): 503-19.

Powell, W. W. "Neither Market Nor Hierarchy: Network Forms of Organization." In Research in Organizational Behavior Vol. 12. Edited by B. M. Staw and L. L. Cummings. Greenwich, CT: JAI Press, 1990.

Quinn, J. B. "Strategic Change: Logical Incrementalism." Sloan Management Review 20 (Fall 1978): 7-19.

Scott, W. R. "The Adolescence of Institutional Theory." Administrative Science Quarterly 32, no. 4 (December 1987): 493-511.

Van de Ven, A., and G. Walker. "The Dynamics of Interorganizational Coordination." Administrative Science Quarterly 29 (December 1984): 598-621.

Wilson, D., and V. Mummalaneni. "Bonding and Commitment in Buyer-Seller Relationships: A Preliminary Conceptualization." Industrial Marketing and Purchasing 1, no. 3 (1986): 44-58.
COPYRIGHT 1993 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Articles
Author:McKinney, Martha M.; Morissey, Joseph P.; Kaluzny, Arnold D.
Publication:Health Services Research
Date:Oct 1, 1993
Words:5954
Previous Article:Calculating the probability of rare events: why settle for an approximation?
Next Article:Effects of the relative fee structure on the use of surgical operations.
Topics:

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters