Chapter 3. Findings

3.1. The CERTs Network: Understanding Structure, Communication & Relationships

3.1.1. The Network Structure of the CERT Community

As described above, we employed Social Network Analysis (SNA) to address the following questions:

• What does the CERT Network look like?
• What is the shape of the individual CERTs network and how does that shape relate to the CERTs' research focus?
• Who are the key entities within each of the CERT's individual networks?
• What is the level of interdependence or independence of the different CERTs network actors (an organization, agency, group, or individual such as the Steering Committee chair) from each other?

In this section we examine the networks of the individual CERTs as well as the CERT Coordinating Center. The findings are presented as sociograms and network measures for each of the individual CERTs and Coordinating Center. We discuss each of the individual CERTs before presenting the overall CERT structure surrounding the Coordinating Center. Although SNA measures are presented in a table, these measures will be described individually due to the difficulties in comparing individual organizational or ego networks with different sizes and structures directly against one another. We will present general trends and patterns of the total network of our sample. Additionally, the SNA analysis provides diagrams that illustrate the relationships CERTs have with their partners and the CERT network as a whole. They do not illustrate the intensity, frequency or nature of those relationships, but are measures of the absence or presence of a relationship.

3.1.2 HMO Research Network CERT

The HMO CERT was established in 2000 and is a Health Maintenance Organization Research Network (HMORN) through which several HMOs work together and share data to improve health outcomes and program performance. Volume 2 Attachment 1 depicts the sociogram (each node in the sociograms represents an actor) of the HMO CERT ego network (SNA labels such networks as "ego" networks, because they focus on understanding each of the individual CERTs (i.e. "egos").

The HMO CERT sociogram is unusual because it displays a portion of the underlying macro (global) structure of the total CERT network on the left of the diagram. Although the main focus of the ego network is to describe the structure surrounding individual CERTs, the overall connection of the broader CERTs program networks' relationship to each individual CERT is also important. The HMO CERT is an illustration of this broader relationship. Exhibit 5 displays the HMO CERT and other CERT network results.

Exhibit 5: CERT Social Network Analysis Measures

Metric	HMO	Duke	UNC	Vanderbilt	Arizona	Penn	Alabama	UNC	CC
CERT Founding Date	2000	1999	1999	1999	1999	2000	2000	1999	1999
Primary Research Area	HMO data	Cardio-vascular	Pediatric	Medicaid and VA data	Drug-drug Interactions and Women's Health	Anti-infectives	Musculo-Skeletal Disorders	Pediatric
Size of Network	52	22	34	15	26	51	39	34	16
Number of Ties	15	27	20	8	12	22	22	20	82
Average Distance	2.11	1.82	2.00	2.03	1.95	2.55	1.97	2.00	1.58
Density	0.57%	5.84%	1.78%	3.81%	1.85%	24.44%	1.48%	1.78%	34.17%
Mean Degree Centrality	4.03	100.00	94.44	14.62	6.34	18.52	6.34	94.44	41.92
Mean Closeness	48.04	100.00	94.44	50.70	51.38	42.19	51.38	94.44	65.41
Betweeness	1.95	86.58	91.16	6.05	2.50	8.68	2.50	91.16	3.87
Keyplayers	AHRQ	CERT-CC	CERT-CC	CERT-CC	CERT-CC	CERT-CC	AHRQ	CERT-CC	CERT-Penn
	CERT-CC	AHRQ		CERT-CC	AHRQ	PCPPP	CERT-CC	AHRQ	AHRQ

Following is a brief description of the meaning of the measures in the table:

• Network Size: The number of unique ordered pairs of actors within the network.
• Number of Ties: Count of the number of relationships in the network.
• Average Distance: Average number of relations in the shortest possible connection from one actor to another.
• Density: The higher the density of a network, the more connected the actors.
• Degree Centrality: Measure of the ego actors' position within the network by counting the total number of direct connections of that actor.
• Closeness: Measure for networks that are fully connected and examines the "shortness" of the direct connections of the actor to other actors in the network.
• Betweeness: Measure of an actor's ability to be a bridge or 'go between' for other pairs of actors by being an intermediary connecting that relationship.
• KeyPlayer: identifies key members of the network.

Citations for these measures and more comprehensive definitions are provided in Appendix 2. As mentioned in the Methods section, these measures (except for size and number of ties) cannot be directly compared against one another due to varying network sizes and characteristics. We describe the results for each CERT and use them to understand differences in their networks.

The HMO CERT sociogram shows a fairly large network of 52 actors, one of the two largest networks in this study, with three distinct groupings. The first large group is on the right side of the figure and is the local community network in which the HMO CERT has approximately 39 actors with whom they are connected as research partners. These partners include such entities as Kaiser Permanente, University of Massachusetts, and Harvard Medical School. This core group represents 75% of the actors within the network; the majority of the relations are densely located within the HMO CERTs community research network.

The middle group comprises actors that are not only associated with the HMO CERT but also those that have connections with the CERT Coordinating Center. The final grouping on the left, as discussed above, displays the macro structure of the total CERT network and encompasses the other CERT partners, the CERT Steering Committee, and AHRQ. The HMO CERT is unique in that its research partners are mostly health plans. The other CERTs do not have access to the breadth of data that the HMO CERT has within its network, so some of the CERTs appear to have relationships with the other CERTs primarily through data sharing. The HMO CERT's existence prior to becoming a CERT may have affected its network size. Its age may have allowed time for the network to increase its size and number of connections. The HMO CERT has a direct connection to the CERT Steering Committee, CERT Coordinating Center, and AHRQ as well as additional access to connections and resources.

Although the HMO CERT has the largest network in terms of its number of actors, it has a relatively low number of ties. Although this is a large network, there are relatively few connections per actor, suggesting that the network has opportunities for developing further relationships among its actors. The average distance within this network is slightly greater than 2, suggesting that information and resource flows have to go through on average two actors to get to the target actor. This again suggests that there are opportunities for further connections within the network. The 0.57% density within the network is very low, suggesting that the actors within the network are not be highly connected and are unevenly distributed throughout the network. For example, there are clusters of actors that appear to be close together, with connections not evenly spread across the network. The density is low, with few interconnections among actors. (The low density may be an artifact of data collection. We did not go to each connection and ask who they were connected to, because these are ego networks, the focus being on that ego network perspective.) The low density level indicates that, within this network, information and resource flow may be slowed.

There may be opportunities to create further connections among actors. The mean degree centrality is slightly over 4, which suggests that most actors within the network have few connections to other actors, which again relates to the sparse and less dense nature of the network. The closeness measure of the HMO CERT is slightly over 48, suggesting that the CERT itself is the only highly connected actor within the network compared to the other actors in the HMO CERT network. The betweeness measure of 1.95 for the HMO CERT is the lowest among all CERTs in the study sample. This could be because of the limited relationships or collaborations the HMO CERT has with the other CERTs beyond data sharing, although it may also be an artifact of the data collection process, in that the HMO CERT was the pilot site visit and they do not have a Web site describing partners and key players the way other CERTs did. The HMO CERT generally appears to play a liaison role between its research partners and the CERTs network as a whole and does not appear to mitigate many other relationships between the actors in the network.

Finally, the key players within this network were identified as AHRQ and the CERT Coordinating Center. This indicates that these two actors have a great deal of communication with the CERT and connects the HMO CERT with the larger CERT network and program resources. Most of the CERTs research partners are unique to the HMO CERT and are not shared among the CERT program's general preferred partners. Based on their pre-CERT existence as a network, the HMO CERT may be more active in its own research network than with other CERTS. The SNA suggests that the HMO CERT has established its own identity and research niche within the CERTs broader network, which is not surprising given this CERT's origins. The HMO CERT probably is the closest to Penn in terms of having a strong community network of partners, but it does appear as strong in that regard.

3.1.3. Duke CERT

The Duke CERT was established in 1999 and was one of the original CERTs. This CERT's main research focus is on cardiovascular therapeutics. Volume 2 Attachment 2 depicts the sociogram of the Duke CERT ego network. As can be seen in the figure, the Duke CERTs network has a star shape structure with one main group and is of medium size with 22 actors within the network. The Duke CERT is surrounded by a relatively small community research network of approximately 9 actors representing just 41% of the actors in the network, but has a very strong university connection. The Duke CERT, located in the North Carolina research triangle, is near the Coordinating Center and UNC CERT. The Duke CERT also has relationships with the Alabama, HMO and UNC CERTs. Additionally, the Duke CERT investigators expressed interest in identifying opportunities to collaborate with two of the new CERTs (Iowa and Cornell).

Examining the Duke CERT network data, we see that the overall size of the network is 22 actors, with 27 ties. Although the Duke CERT is a smaller network than the HMO CERT, it has a larger number of ties indicating the network may be relatively more connected. The average distance within this network is 1.82, thus information and resource flows have to go through on average a fewer than two actors to get to the target actor. Although this network is more connected, there is still opportunity for further connections within the network. The density within the network is low at 5.84%, suggesting that the actors within the network are not highly connected but are concentrated around the CERT. Again, this lower level of density indicates that within this network, information and resource flow could be slowed. Yet, there may be opportunities to create further connections between actors. The mean degree centrality and closeness within the network is 100, which indicates that the CERT is in the center position, the focal node, and highly connected. The betweeness measure of 86.58 indicates that the CERT plays a bridging role between its community network and the broader CERTs community. This role is further illustrated through the star shape of the network as seen in the Duke CERT sociogram. A star network can indicate that there are shorter distances between partners as compared to the HMO CERT, for example. The network (more than the metrics, because they do not vary significantly) suggests that it is easy for Duke to work with its partners. The Duke network also suggests that there are further opportunities to make connections or further develop partnerships.

Finally, the key players within this network were identified as the CERT Coordinating Center and AHRQ. Similar to the HMO CERT, this suggests that these two actors have a great deal of communication with the CERT and connects the CERT with the larger CERT network and program resources. The Duke CERT appears to have relatively few research partners, but interacts with many other CERTs. This could be due to its proximity to the Coordinating Center and desire to reach out to other new CERTs, or due to the nature of its research, which can cross into the research areas of other CERTS.

3.1.4. University of North Carolina CERT

The UNC CERT was established in 1999, one of the first four CERTs. This CERTs main research focus is on pediatric therapeutics in contrast to the other CERTs focus primarily on therapeutics in the adult population. Volume 2 Attachment 3 depicts the sociogram of the UNC CERT ego network. As can be seen in the figure, the UNC CERT exhibits a star shaped network similar to the Duke CERT with one main group and is a larger size with 34 actors within the network. The UNC CERT is surrounded by a community research network of approximately 23 actors representing 68% of the actors in the network. The UNC CERT is close to the Duke CERT and Coordinating Center and has a strong university connection and connection with its partners in North Carolina, in part with the research triangle. The UNC CERT network illustrates relationships with the Alabama, Arizona, Duke, HMO, Penn, and Vanderbilt CERTs. Discussions with investigators, though, acknowledge the difficulty of collaborating on projects, given their focus on pediatrics. The relationships with other CERTs revealed in the UNC CERT analysis suggest only sharing of advice or methodological discussions because the perception of collaboration and actual level of collaboration were not revealed in the data.

The UNC CERT network data shows that the overall size of the network is fairly large with 34 actors and 20 ties. The average distance within this network is 2, indicating that information and resource flows have to go through on average two actors to get to the target actor. Although this network is connected, there is still opportunity for further connections within the network. The density within the network is low at 1.78%, suggesting that the actors within the network are not highly connected but are concentrated around the CERT. Again, this lower level of density suggests that within this network, information and resource flow would be slow. The star shape network suggests proximity between partners and the CERT. There are opportunities to create further connections between actors. The mean degree centrality and closeness within the network is 94.4, which indicates that the CERT is in the center position, the focal node, and highly connected. The betweeness measure of 91.16 suggests that the CERT plays a bridging role between its community network and the broader CERTs community, which is further illustrated through the star shape of the network.

The key players within this network were identified as the CERT Coordinating Center and AHRQ. Similar to the HMO CERT, this suggests that these two actors have a great deal of communication with the CERT and suggests a connection between the CERT with the larger CERT network and program resources. The UNC CERTs has many research partners and interacts with many other CERTs. This could be due to its proximity to the Coordinating Center and reaching out to other CERTs.

3.1.5. Vanderbilt CERT

The Vanderbilt CERT was established in 1999 and was one of the first CERTs. This CERT's main research focus is on therapeutics in the Medicaid and VA populations. Volume 2 Attachment 4 depicts the sociogram of the Vanderbilt CERT ego network. Its research focus on data from VA and Medicaid populations likely explains the structure of its network; the Vanderbilt CERT has a very unusual "double star" network structure with a focused star network on the left side of government agencies with the CERT Coordinating Center and the Steering Committee at its center. On the right side of the sociogram is the CERTs community research network with the Vanderbilt CERT at its core. The increased interaction among federal government agencies may be due to their research focus and government agencies' interest in the outcomes of their work. As can be seen in Attachment 4, the Vanderbilt CERTs is relatively small with 15 actors and is surrounded by a community network of approximately 12 actors representing 80% of the actors in the network. Those actors within the CERTs community network represent many government agencies, including such entities as the State of Tennessee Health Department and the Veterans Administration. The Vanderbilt CERT is a very sparse network and appears to interact only with the HMO CERT.

The Vanderbilt CERTs network data shows the overall size of the network (15 actors), with 8 ties. This is a relatively small, sparse, and relatively unconnected network. The average distance within this network is 2.03; information and resource flows have to go through on average approximately two actors to get to the target actor. The density within the network is low at 3.81%, so the actors within the network do not appear highly connected. Again, this lower level of density suggests that within this network, information and resource flow would be slowed, but that there would be opportunities to create further connections between actors. The mean degree centrality is 14.62 and closeness within the network is 50.7. These measures are consistent with the double star structure of the network in which connections and core structure is shared among the CERT and the Coordinating Center. The betweeness measure of 6.05 suggests that the CERT plays a very limited liaison role in connecting other actors within the network.

The key players within this network were identified as the CERT Coordinating Center and the CERT Steering Committee. This suggests that these two actors have a great deal of communication with the CERT and connect the CERT with the larger CERT network and program resources. The double star structure and the metrics of the network point to the collaboration of the CERT with many more government entities as compared to other CERTs. Those government entities appear to be interdependent and have connections with one another. The Vanderbilt CERT is unusual in terms of how sparse its network of partners appears compared to others. Again, the intensity of relationships was not evident from the data. The interviews indicated that the Vanderbilt investigators collaborate extensively with TennCare and VA. In summary, these analyses suggest that the Vanderbilt CERT may focus more on its own research than on working with the other CERTs and that it has room to expand its network of partners.

3.1.6. Arizona CERT

The Arizona CERT was initially funded in 1999 at Georgetown under the direction of the same principal investigator. This CERTs main research focus is on prescription drug safety. Volume 2 Attachment 5 depicts the sociogram of the Arizona CERT ego network. As can be seen in the figure, the Arizona CERTs network has a distinct star shape structure with one main group and is of medium size with 26 actors within the network. The Arizona CERT is surrounded by a community network of approximately 15 actors representing 58% of the actors (26) in the network. The Arizona CERT maintains a methadone registry, focuses on drug safety and women's health and consequently collaborates with such partners as healthcare providers, pharmacies, and government agencies that benefit from this work. The Arizona CERT has relationships with the Penn, Duke, and HMO CERTs.

The Arizona CERT network data show a medium-sized overall network of 26 actors, with 12 ties — a structure and density similar to the Vanderbilt CERT. The average distance within this network is 1.95, so information and resource flows have to go through on average approximately 2 actors to get to the target actor. The density within the network is low at 1.85%. The actors within the network do not appear to be highly connected based on the structure and its measures, but are concentrated around the CERT. This lower level of density suggests that within this network, information and resource flow may be slowed getting to the entire network. The mean degree centrality is 6.34 and the closeness measure is 50.38, which suggests that there are, on average, few direct connections among the actors. The CERTs focal role is mitigated by its strong connection to the CERT Coordinating Center and Steering Committee. The betweeness measure of 2.50 is very low and suggests that the CERT does not appear to play a strong bridging role between its community research network and the broader CERTS community.

The key players within this network were identified as the CERT Coordinating Center and AHRQ. Similar to the HMO CERT, this suggests that these two actors have a great deal of communication with the CERT and connect the CERT with the larger CERT network and program resources. The structure and connections of this CERT to a broader or larger network may be due to the nature of its research which lends itself to the connections this CERT makes within its network. It appears that this CERT is still developing resources and creating relationships.

3.1.7. PENN CERT

The PENN CERT was established in 2000 and its main research focus is anti-infectives. Volume 2 Attachment 6 depicts the sociogram of the PENN CERT ego network. As can be seen in the figure, the PENN CERTs network has a strong double star shape structure. The PENN CERT has a very large and strong community research network of 29 nodes out of the total size of 51, representing 58% of the network and is its own integrated micro network structure of the CERT network. The PENN CERT community research network is on the right side of the sociogram, is encapsulated by the label of Penn CERTs Public and Private Partnerships (PCPPP), and is highly evolved and dense. As a Coordinating Center member said, "Penn has a clear leadership structure with a large number of co-investigators who are receiving some support from AHRQ and are leveraging many other partnerships."

On the left side of the sociogram is the second star and the main CERT ego configuration. This is also a dense and connected structure. Within that configuration is one of the key CERTs community partners (the Leonard Davis Institute) that conducts policy research and briefings that incorporate the research of the CERT. This is a powerful resource to the CERT; it allows for information regarding the CERTs research to flow into the broader community and increase the PENN CERT's opportunities for additional collaborative connections and for garnering resources outside of the CERT program. The PENN CERT also has relationships with the Duke and HMO CERTs. The Penn CERT is such a large community network that it resembles 2 networks - a very large and strong network of partners and the CERTs network. This suggests that the Penn CERT has been successful developing and leveraging partners. Additionally, there is very limited overlap of Penn's partners with the CERTs program partners, which suggests the availability of even further collaboration opportunities for the Penn CERT.

The overall size of the network is 51 actors, with 22 ties. The average distance within this network is 2.55. This is the largest average distance among the study sample networks; it is based on the dense structure in which the CERT is embedded which suggest that the CERT is as involved in its own community structure as in the CERT program and which places it in more of a liaison position for the broader network. The density within the network is low at 24.44%, thus the actors within the network are somewhat connected but are more highly linked and concentrated around the PENN CERT. The mean degree centrality is 18.52 and the closeness measure is 42.19, which suggest that there are on average many direct connections among the actors and that the CERT is a liaison between its strong community partners and the broader CERT network. The betweeness measure of 8.68 is relatively low and suggests that the CERT plays a bridging role between its community research network and the broader CERTs community, but apparently not a strong one.

Finally, the key players within this network were identified as the CERT Coordinating Center and PCPPP. This appears to correlate with the strong community research network that the Penn CERT has created and is embedded within.

3.1.8. Alabama CERT

The Alabama CERT was established in 2000 and its main research focus is on musculoskeletal disorders therapeutics. Volume 2 Attachment 7 depicts the sociogram of the Alabama CERT ego network. As can be seen in the figure, the Alabama CERTs network has a very distinct star shape structure. The Alabama CERT is a relatively large network of 39 actors with 30 actors within its community research network, representing 77% of the network. The Alabama CERTs community network is relatively connected and dense. Alabama appears to have relationships with four other CERTS, Arizona, Duke, HMO, and PENN and appears similar to Arizona, Duke, and UNC in terms of having some partners in their network. If their network wanted to work with the CERTs program partners, it is linked to that larger network. Additionally, the Alabama CERT appears to have opportunities and room to expand its network.

The overall size of the network is 39 actors with 22 ties. The average distance within this network is 1.97; information and resource flows have to go through on average fewer than two actors to get to the target actor. The density within the network is very low at 1.48%; thus, the actors within the network are sparsely connected. The mean degree centrality is 6.34 and the closeness measure is 51.38, which suggests that there are on average few direct connections among the actors and that the CERT appears to be in a liaison role between its community partners and the broader CERT network. The betweeness measure of 2.50 is low and suggests that the CERT does not readily play a bridging role between its community research network and the broader CERTS community. The key players within this network were identified as AHRQ and the CERT Coordinating Center.

3.1.9. The CERTs Coordinating Center

In the original CERTs plan, as devised by AHRQ in conjunction with the CERT Steering Committee and its partners, the CERTs Coordinating Center was to have the role of liaison between the CERTs themselves and AHRQ, the Steering Committee, preferred partners, and other government agencies. In the sociogram in Volume 2 Attachment 8, the Coordinating Center appears to be functioning as it was envisioned in the original CERTs plan, acting as the bridge between the CERTs and the other actors within the program. The Coordinating Center is the focal node in this network, dispersing information from AHRQ and the Steering Committee to the CERTs as well as bringing together outside partners with the CERTs based on research needs and interests. This appears to be both an efficient and effective way to manage the macro CERTs network to avoid duplication of effort and resources to spread information and create collaborative connections.

Looking at the Coordinating Center network measures, the size of the network is relatively small at the macro level, but is highly connected with 16 actors and 82 ties. The average distance within this network is 1.58, so information and resource flows have to go through on average fewer than two actors to get to the target actor. The density within the network is relatively low at 34.17%; the actors within the network appear to be connected most directly to the Coordinating Center. The mean degree centrality is 41.92, the closeness measure is 65.41, and the betweeness measure of 3.87 point to the liaison role the Coordinating Center plays between the CERTs, AHRQ, the CERT Steering Committee, and other partners.

The key players within this network were identified as the Penn CERT and AHRQ. It is not surprising that AHRQ is noted as a key actor within the Coordinating Center's network, but the Penn CERT role is surprising. This is likely due to the fact that the Penn CERT has a large and strong community network and many partners, and that it connects the Coordinating Center to an even larger network and as such plays a prominent role in the network as a whole. The PENN CERT is unusual in that its preferred partners are considered key players within the network. This is because PENN has a very large community network and appears to have strategically initiated and maintained these connections.

The keyplayer algorithm is a metric designed to locate the main actors within the network diagram which (1) if removed, would fragment the network or, (2) whose position in the network indicates an opportunity to expand the network. In most of the CERTs network diagrams; the Coordinating Center appears to be a vital actor within the networks whose removal would separate the CERT from the macro CERTs network. This confirms the vital role that the Coordinating Center appears to play in connecting the CERTs to each other and to the broader network, providing further evidence that the Coordinating Center is functioning as originally designed. Respondents had only positive comments about the Coordinating Center, and the SNA results support those statements that the CC is fulfilling its role and is viewed as a partner in the network. The Coordinating Center appears to be functioning as a liaison between the CERTs, AHRQ, and the SC and to bring partners to the CERTs when possible.

In many of the CERTs AHRQ is also seen as a second key player within the CERT network. This suggests (hypothetically, of course, given AHRQ's unique role) that if AHRQ were removed from the network, it would become more fragmented. This SNA finding confirms the direct contact that AHRQ has with the individual CERTs. As noted in interviews, the role of the Coordinating Center appears to be changing and evolving into the (future) potential network structure depicted in the sociogram show in Volume 2 Attachment 9. AHRQ may be starting to have more contact and information sharing directly with the CERTs, bypassing the Coordinating Center. This evolving structure may not be the most efficient and may reduce opportunities for collaboration.

The CERTs preferred partners are those connections that appear to be garnered and maintained by the CERTs Coordinating Center. For the sake of clarity within the CERT network diagram, these partners have been signified as PP, but are shown in full within the Preferred Partner diagram. Go to Volume 2 Attachment 10.

The collaboration network diagram provides a graphic illustration of the extent of collaboration across all authors associated with the CERTs publications. All of the CERT PIs are located within the dense, center of the diagram illustrating collaboration with many other authors; however, because of the lack of attributes about each author (1000 authors), further inference about co-authorship is not possible to assess. Go to Volume 2 Attachment 11.

3.1.10. Social Network Analysis Implications

The CERT ego networks are diagrams of each particular CERTs network, taken from the perspective of the individual organization. These ego networks are representations of the relationships in which the CERTs view themselves, and therefore do not display any broader connections of those CERT partners with one another or any additional actors. SNA frequently collects data on ego networks of only those organizations directly involved within a program or particular structure. In those cases, as was done here, data are collected from representatives of those organizations to establish their ego network. As noted previously, within the ego network diagrams, each node or circle represents an individual actor and each line represents a connection or relationship to that actor. The displayed relationship is merely dichotomous in that the nature and value of that relationship is not represented. Rather, if any connection exists --- good, bad, collaborative to consultative, it is displayed within the ego network diagram. The scope and nature of this study did not allow for the understanding of the quantity, quality, nor nature of these relationships to be fully realized. Future research could look to incorporate these facets of each connection within the ego network by further questioning the value and nature of each relationship within the CERT itself, and by follow-up with each of the CERTs partners to inquire as to their perspective of that relationship and to solicit information about their individual networks. This would provide for more complete network data.

Although the CERTs ego networks are snapshots of the relationships of the CERT from their own perspective, it is important to understand that their network is affected by not only their actions and those of their partners, but also by the broader, macro CERT network. Each relationship within a network takes time and resources to maintain. Each individual CERT has various mandates to fulfill while also being expected to leverage funding to complete its work. With finite resources, it may be difficult to maintain relationships, let alone have the ability to strategically initiate connections for future work. Thus, geographic location and previously established ties become vital to each CERT. Proximity to a partner decreases the time and resources needed to maintain a relationship, and as such, CERTs that are geographically close are more apt to collaborate or provide resources to one another whether it is in the form of advice, assistance on a project, or information on potential partners. The Penn CERT appears to be truly different in structure from the other CERTs. This CERT may be an ideal CERT to emulate if developing partnerships and leveraging resources is a primary aim of a CERT. The HMO research network appears to be a near second in part because of its advantage of having a strong network already formed prior to becoming a CERT.

The relationships of individual members of each CERT can be vital in expanding the CERTs network under the conditions of finite resources. If a principal investigator within a CERT has worked with an individual or organization prior to being involved in the CERT, that relationship can be accessed in the future without the same level of resources as would be needed to initiate and maintain a new connection. Access and trust have already been established with that potential partner that mitigates costs and geographic proximity. Thus, an actor who has a history with a CERT or member of the CERT will be more likely to work with that CERT despite potential geographical limitations. In this situation, the Coordinating Center would play a vital role in maintaining past relationships of the CERT partners to decrease the individual resource costs to each CERT and to provide for future opportunities with the CERTS and those partners. The Coordinating Center currently appears to be succeeding at maintaining these linkages and bringing together CERTs with those partners who have similar interests or particular needs. If the Coordinating Centers role as the liaison between AHRQ, the CERT Steering Committee, and the CERTS is diminished or diluted through more direct contact with each CERT, the cost to each CERT to maintain relationships and create new connections may increase.

Additionally, the burden of the information processing that the Coordinating Center currently undertakes might be shifted to the individual CERTs as well, as there could be duplication in effort in providing information from both AHRQ and the Coordinating Center. It appears that AHRQ is leaning towards having more direct contact with the CERTs. This potential new configuration of the CERTs network is displayed in the Coordinating Center future diagram —(Go to Volume 2 Attachment 9). Not only are the CERTs network diagrams drawn from the CERTs perspective, but that is often how an ego views their network. Most egos within their network are concerned with their own actors and connections and often to not have an understanding or appreciation of the overall network in which they are embedded. This focus and specialization can cause the network to be less dense and connected. The Coordinating Center would therefore become even more vital to maintaining coordination and communication among and between the CERTs so that the network does not become fragmented.

Return to Contents
Proceed to Next Section