Introduction

Emergency science communication refers to the collection of communication, publicity and educational activities that are implemented to increase public safety and awareness of emergencies and to optimize health-protective behaviors (Zhang et al. 2020a). Moreover, emergency science communication is an important component of emergency information release (Huang et al. 2020), the main channel through which government departments communicate with society and regulate public opinion, and a way in which the public can obtain authoritative health-protective information in emergencies (Liu and Dong, 2024a; Zhang et al. 2021). During the COVID-19 pandemic, governments worldwide actively engaged in emergency science communication, guided the public to increase its risk awareness, and scientifically reduced the losses and casualties caused by COVID-19 (Li et al. 2021; Matta, 2020; Xing and Zhang, 2021). Since the emergency response to SARS in 2003, the Chinese government has issued guiding policies to actively deploy and promote emergency science communication. In 2005, the State Council issued the Notice on issuing the Overall Implementation Plan of Emergency Management Science Communication and Education. Then, in response to COVID-19, the China Association for Science and Technology, the Publicity Department of the CPC Central Committee, the Ministry of Science and Technology, the National Health Commission and the Ministry of Emergency Management jointly issued Opinions on Further Strengthening Emergency Science Communication and Education. Furthermore, considering the situation of the COVID-19 pandemic, the China Association for Science and Technology issued the Notice on Emergency Science Communication under the Ongoing Prevention and Control of COVID-19. The Chinese government has also conducted emergency science communication with the public in the form of press conferences to answer questions and publicize information. Additionally, the transmission of emergency information and health-protective information to the public through government affairs media is the main form of emergency science communication. In summary, the effectiveness of emergency science communication and the public’s risk prevention ability need to be improved to limit the transmission of diseases, reducing the risks, and maintaining social stability.

Public health emergencies are characterized by uncertainty and a wide and rapid spread (Li and Gao, 2017; Yang et al. 2022). While responding to emergencies, governments and relevant departments need to release various types of emergency science communication information to the public in a timely manner so that the public can understand the characteristics of these emergencies and implement health-protective behaviors (Vuong et al. 2022). The main contents of emergency science communication information from governments are reliable and public, which are the following: virus characteristics and rules, data information, prevention and control policies, health-protective behavior information that the public needs and public service information related to public health emergencies (Huang et al. 2021; Zhang and Yu, 2022).

At present, since the social environment and risk types continue to coexist and evolve, the emergency management has complex and systematic characteristics (Liu and Dong, 2024b; Phattharapornjaroen et al. 2022). Relatedly, the complexity of the content, coverage and mode of dissemination of emergency science communication related to public health emergencies has increased. Moreover, the information presented through emergency science communication is multisource and cross-organizational. Emergency management practices show that the needs of emergency science communication during public health emergencies can no longer be met by a single emergency organization. Thus, effective cooperation between emergency organizations at different levels and in different fields is urgently needed (Assefa et al. 2022; Kc et al. 2021). Emergency science communication cooperation has positive significance for strengthening the interaction level of emergency organizations and realizing the integration of heterogeneous scientific information resources (Moradian et al. 2020; Song and Karako, 2020). However, previous studies on emergency science communication have mainly focused on strategies and platforms for emergency science communication and have generally overlooked organizational cooperation mechanisms in emergency science communication.

Thus, the cooperation characteristics of emergency science communication need to be determined, and the interaction mechanism of organizations in emergency science communication needs to be clarified. This study identifies the cooperative relationships among emergency science communication organizations (ESCOs) from the network perspective and constructs emergency science communication cooperation networks (ESCCNs) by focusing on the Chinese government’s practice of emergency science communication in response to COVID-19 as a case study. Moreover, the differences in the modes of cooperation in emergency science communication during public health emergencies at different phases and the evolutionary law of ESCCNs are analyzed from the perspective of the dynamic characteristics of public health emergencies. This study provides theoretical support and a decision-making reference for improving the efficiency of emergency science communication cooperation and for strengthening the positive role of emergency science communication support.

This study is structured as follows. Section “Literature review” reviews the status of research on emergency science communication and discusses further breakthroughs in this field through comparisons with previous studies. Section “Research design and data basis” proposes the research framework and introduces the background of the case, the process of data collection and the construction method for ESCCNs. Moreover, the network analysis method and the constructed model are introduced. Section “Results” presents a comparative analysis of the network structure of ESCCNs, the node attributes of ESCOs and the cross-group correlations between different emergency phases. The influence of ESCCNs of different phases on the whole phase is also discussed. Section “Discussion” discusses the results of the analysis of ESCCNs during public health emergencies and summarizes the main factors that influence ESCCN formation. Additionally, the theoretical and practical contributions and limitations are presented. Section “Conclusions” summarizes the research conclusions and Section “Implications” proposes suggestions for improving the effectiveness of cooperation networks for emergency science communication in public health emergencies.

Literature review

The content structure and topic setting of emergency science communication are closely related to the type of emergency being communicated and develop gradually with the clarity of risk awareness and understanding. In recent years, the various security situations that have arisen have produced many new changes, and the uncertainty, destructiveness and derivativeness of emergencies have further expanded (Ma et al. 2024). Thus, emergency science communication needs to be adjusted and upgraded. At present, research on emergency science communication from the perspective of emergency practice in China has been focused on the following topics. First, related studies have mainly focused on how emergency science communication workers have organized and summarized the existing problems in emergency science communication according to their previous practical experience in China and proposed corresponding countermeasures and suggestions. The second category includes strategy studies on emergency science communication in different types of emergencies (Hu et al. 2018, 2021; Xiong et al. 2023). By examining actual cases of emergencies in China, researchers in this field have systematically determined the causes of, risks of and protection strategies for emergencies such as flood disasters, earthquake disasters, situations involving hazardous chemicals and explosions, public health emergencies and nuclear safety incidents to further increase the public’s safety awareness and ability to avoid emergencies. The third category includes strategy studies on emergency science communication regarding different objects (Lin et al. 2014). Given the differences in the individual needs and risk perception abilities of students, elderly people, and office workers, researchers in this field have proposed that emergency science communication practices should be adapted to and match public security needs and knowledge expectations, which have become increasingly diversified. Subsequently, these researchers have designed emergency science communication strategies for different objects. The fourth category includes strategy studies on emergency science communication, which involves the introduction of the media and emerging technologies (Jia et al. 2017; Li et al. 2022; Zeng and Li, 2020). Given the background of new media development in China, researchers have explained how the interaction between traditional media (newspapers, magazines, radio and television) and new media (social media platforms, mobile internet communication tools, and short video platforms) can be fully exploited. Furthermore, feasible strategies for strengthening emergency science communication and improving its effectiveness have been proposed.

Moreover, in combination with the practice and experience of emergency science communication in the face of different emergencies, governments worldwide have explored the implementation of emergency science communication and improved its effectiveness. For example, they have explored how to effectively formulate emergency science communication strategies (Hyland-Wood et al. 2021), how to support the implementation of emergency science communication with the help of expert authority and knowledge (Camporesi et al. 2022; Whitty and Collet-Fenson, 2021), and how to clarify the impact factors such as trust, risk cognition and media participation in emergency scientce communication efficiency (Austin et al. 2021; Siegrist et al. 2021). Additionally, the information‒value nexus has proven to be an essential perspective for understanding emergency and emergency information (Vuong, 2023). According to mindsponge theory, when the government releases emergency science communication information to the public, people have two options. On the one hand, they can choose to absorb new ideas that match their understanding, and on the other hand, they can ignore those that do not match their understanding (Vuong et al. 2024a). Thus, an effective way to improve emergency science communication is to expand the range of public perception and broaden the buffer zone in public consciousness to enhance the public’s ability to deal with emergency scientific information when emergencies occur (Nguyen et al. 2022; Vuong et al. 2024b).

Research on emergency science communication has mostly focused on the information dissemination path and the integration of science communication resources, whereas few studies have explored the operation mechanism of emergency science communication. Moreover, emergency science communication involves not only the participation of multiple subjects but also the integration of multisource emergency information. Thus, the characteristics and rules of cooperation of organizations involved in emergency science communication need to be explored. Additionally, since emergencies have uncertain and dynamic characteristics, the dynamic evolutionary characteristics of organizational cooperation in emergency science communication need to be further explored.

Recently, research on emergency networks has undergone positive development with the goal of describing the overall picture and related characteristics of the relationships between emergency organizations (Du et al. 2020a; Hu et al. 2022; Yan, 2023). Academic circles construct emergency management networks by identifying the types of organizations and the relationships among organizations in actual emergency cases (Liu et al. 2022a) or emergency plans (Fan et al. 2019). With respect to the research objects of previous studies, different emergency management networks have been constructed on the basis of the type of emergency, such as natural hazards (earthquakes, rainstorms and wildfires) (Chen et al. 2019; Davis et al. 2015; Zhang et al. 2016), safety accidents (fires, collapses and explosions) (Du et al. 2020b; Liu et al. 2022b), public health emergencies (COVID-19 and MERS) (Guo et al. 2021; Kim et al. 2017), and social security incidents (Hu et al. 2014). Furthermore, network structure analysis and emergency organization node attribute analysis are important and basic research topics of studies on emergency management networks (Chen et al. 2020; Kapucu and Garayev, 2016; Robinson et al. 2013). Moreover, some scholars have analyzed the dynamic characteristics of emergency management networks from the perspective of the whole life cycle (Liu et al. 2022c; Lu et al. 2021) or compared the differences between emergency management networks formed by actual cases and emergency management networks formed by emergency plans. Additionally, previous studies have utilized both comparative data analysis of the characteristics of emergency management networks and inferential network analysis for further analyzing and explaining how emergency management networks are formed (Cai et al. 2023). The relationships between emergency organizations can be described on the basis of research on emergency management networks, which is utilized in this study to explain the cooperation mode and characteristics of emergency science communication.

Thus, in this study, ESCCNs are constructed on the basis of social network analysis and the Chinese government’s practice of emergency science communication in response to COVID-19. Specifically, a comparative analysis of the emergency response phase and the ongoing emergency phase of the emergency science communication cooperation mode is performed. The matrix relation analysis method is introduced to inform a discussion of the differential influence of different emergency phases on ESCCN formation during the whole phase. The implications of this research support the improvement of the effectiveness of public health emergency science communication.

Research design and data basis

Research framework

To explore the differences between the emergency response phase and the ongoing emergency phase of the emergency science communication cooperation mode and to determine how the effectiveness of the cooperation network for emergency science communication in public health emergencies can be improved, this study proposes the following research framework (see Fig. 1).

  1. (1)

    Network construction. On the basis of the cooperative relationships between ESCOs in the emergency science communication, the ESCCNs of the whole phase, the emergency response phase and the ongoing emergency phase are constructed.

  2. (2)

    Comparative analysis between different phases. Network structure analysis, node attribute analysis and cross-group correlation analysis are used to compare and analyze the ESCCNs of different phases with the goal of clarifying and discussing the network structure characteristics of these networks, the influence characteristics of ESCOs, and the interaction characteristics among emergency groups at different phases.

  3. (3)

    The impact of ESCCNs of different phases on the whole phase. The mechanism through which the ESCCNs of the emergency response phase and the ongoing emergency phase impact the formation of the ESCCN of the whole phase is analyzed to clarify the positioning and value of emergency response and ongoing emergencies in the overall process of emergency science communication cooperation in public health emergencies.

Fig. 1
figure 1

Research framework for the ESCCN in public health emergencies.

Full size image

Data collection and processing

Joint Prevention and Control Mechanism of the State Council (JPCMSC) is a cross-departmental discussion platform that was launched by the Chinese government in response to COVID-19; this platform is responsible for coordinating and solving complex affairs and problems in COVID-19 emergency management. During the COVID-19 response period, JPCMSC released emergency information in the form of press conferences, which included the distribution and interpretation of COVID-19 data, emergency response measures taken by the government and health-protective behaviors that needed to be implemented by the public. This platform has effectively enhanced public safety awareness and supported the emergency management.

The practices of the Chinese government in response to COVID-19 can be divided into two phases: the emergency response to COVID-19 and the ongoing emergency of COVID-19, as shown in Table 1 (Niu et al. 2022). Thus, data for this study covered the periods from December 27, 2019, to April 28, 2019 (the emergency response phase), and from April 29, 2019, to June 5, 2022 (the ongoing emergency phase). During the whole phase, 149 press conferences related to emergency science communication were held, with 71 for the emergency response phase and 78 for the ongoing emergency phase.

Table 1 Phase division of COVID-19 response.
Full size table

In total, 60 ESCOs and 833 cooperative relationships are obtained through the identification of the cooperation behavior of emergency science communication at press conferences, as shown in Table 2. Furthermore, based on the functions and roles of emergency science communication, ESCOs are classified into the following categories: health administrative departments (Group 1), functional administrative departments (Group 2), and experts and academic organizations (Group 3).

Table 2 Description of emergency science communication organizations.
Full size table

Research methods

Network structure and node attribute analysis

Network structure is an important measure for characterizing the cooperation mode of emergency science communication (Cheng et al. 2022; Ku et al. 2021). Specifically, network tightness, network agglomeration, network equilibrium and network connectivity are all important attributes of ESCCNs. In this study, network structure characteristics are analyzed on the basis of network density, centralization, cohesion and average path length.

Among these attributes, network density is related to the number of ESCOs in the ESCCN and the cooperative relationships among ESCOs. Specifically, the ratio of inter-ESCO connections to the maximum possible inter-ESCO connections is the network density. The greater the network density is, the more cooperative relationships are formed between ESCOs and the closer the network structure of the ESCCN is. Network centralization is an indicator of the degree of agglomeration of the ESCCN toward network core ESCOs and is related to the centrality value of each ESCO in the ESCCN. In this study, degree centrality is selected as the research index for analyzing the agglomeration of the ESCCN network structure and discussing whether the core-edge features of the ESCCN are significant. Network cohesion is an indicator of whether the network structure of the ESCCN is in equilibrium. The assignment structure and equilibrium structure correspond to different network cohesion levels. On the one hand, the network cohesion of an ESCCN with an assignment structure is low, and the difference in the number of ESCOs in the network is large. On the other hand, the network cohesion of an ESCCN with an equilibrium structure is high, and the resources of ESCOs are relatively dispersed. The average path length is the average communication distance among all ESCOs in the ESCCN. The greater the average path length is, the greater the distance between ESCOs and the lower the network connectivity of the ESCCN. Conversely, the lower the average path length is, the shorter the distance between ESCOs and the greater the network connectivity of the ESCCN.

The node attribute measures the role and importance of ESCOs in the ESCCN. The importance of an ESCO node depends on the number of ESCOs in its neighborhood and the importance of its neighbors. Degree centrality and eigenvector centrality can effectively reflect these two objects of analysis. Thus, degree centrality and eigenvector centrality are chosen to analyze ESCOs in ESCCNs in the emergency response and ongoing emergency phases in this study. Specifically, the importance ranking of ESCOs in the ESCCN is expressed mathematically, and the differences between the ESCCNs of the emergency response phase and the ongoing emergency phase at the organizational level are compared and analyzed.

Degree centrality refers to the number of ESCOs associated with one ESCO node in an ESCCN. Specifically, the ESCCN constructed in this study is an undirected network. The higher the degree centrality of the ESCO is, the more power the ESCO has in the ESCCN, which is the core node of the network. Unlike degree centrality, eigenvector centrality is more focused on the analysis of the connectivity quality of ESCOs in the network. Eigenvector centrality emphasizes the surrounding environment of ESCOs, with the goal of determining how many high-impact neighbors ESCO nodes have. The larger the eigenvector centrality is, the greater the potential influence of the ESCO on the ESCCN. Additionally, since the node attributes of the ESCCNs of the emergency response phase and the ongoing emergency phase need to be compared, degree centrality and eigenvector centrality should be standardized. Then, the normalized degree centrality and normalized eigenvector centrality of the ESCO can be obtained.

External–internal (E-I) index analysis

In ESCCNs, different types of ESCOs have different interactions. Moreover, different types of ESCOs tend to form intragroup cooperation and cross-group cooperation differently. Therefore, the E-I index is introduced in this study to describe the relationships among health administrative departments, functional administrative departments, and experts and academic organizations in terms of cooperation and resource flow. In ESCCNs, cooperative relationships involving ESCOs of the same type can be represented by internal links (ILs), and the cooperative relationships between different types of ESCOs can be represented by external links (ELs). The calculation method for the E-I index is shown in Formula (1) (Krackhardt and Stern, 1988).

$${\rm{E}}-{\rm{I}}=\frac{{\rm{ELs}}-{\rm{ILs}}}{{\rm{ELs}}+{\rm{ILs}}}$$
(1)

Here, E-I is the E-I index of the ESCCNS. ELs are the cooperative relationships between different types of ESCOs. ILs are cooperative relationships between ESCOs of the same type. The value of the E-I index ranges from [−1, +1] (Comfort and Zhang, 2020). The larger the E-I index is, the more cross-group relationships of emergency science communication cooperation there are.

Quadratic assignment procedure (QAP) analysis

The QAP is an analysis method that is used to compare two or more matrices by analyzing the similarity of each element in the matrix. Therefore, QAP analysis is viewed as an important method for analyzing the relationships between different networks (Krackardt, 1987). When QAP analysis is adopted, the autocorrelation problem existing in the relationship matrix can be effectively avoided (Dekker et al. 2007).

Specifically, QAP correlation and QAP regression analyses are conducted in this study, and the ESCCN of the whole phase is used as the dependent variable. Additionally, the ESCCNs of the emergency response phase and the ongoing emergency phase are used as independent variables to explore the effect of ESCCNs of different phases on the whole phase. Therefore, the temporal influencing model for ESCCN formation (2) is constructed as follows.

$$oESCCN=f(Response,Ongoing)$$
(2)

Here, the dependent variable oESCCN represents the emergency science communication cooperation matrix of the whole phase. Response and Ongoing are the cooperation matrices that are formed in the emergency response and ongoing emergency phases. Moreover, to ensure the standardization of the research, the matrix scale of Response and Ongoing should be consistent with that of oESCCN and be expanded to 60×60.

Results

Network structure analysis of ESCCNs

In this study, an emergency science communication cooperation matrix among ESCOs is established on the basis of the cooperative relationships between ESCOs in emergency science communication. On the basis of the practice, the ESCCN of the whole phase, the ESCCNs of the emergency response phase and the ongoing emergency phase are constructed, as shown in Fig. 2. Table 3 shows the calculation results of the network structure characteristics of these ESCCNs.

Fig. 2: Emergency science communication cooperation networks.
figure 2

a The ESCCN of whole phase. b The ESCCN of emergency response. c The ESCCN of ongoing emergency.

Full size image
Table 3 Network structure characteristics of ESCCNs.
Full size table

Figure 2a shows the ESCCN of the whole phase, highlighting that a total of 60 ESCOs are involved in COVID-19 emergency science communication cooperation. This network includes 7 health administrative departments, 35 functional administrative departments, and 18 experts and academic organizations. From the network structure perspective, the network structure indicators of the ESCCN of the whole phase are in the middle of the ESCCNs of the emergency response phase and the ongoing emergency phase. These findings demonstrate that the emergency response and ongoing emergency phases have different effects on the ESCCN of the whole phase.

Figure 2b, c illustrate the ESCCNs of the emergency response phase and the ongoing emergency phase, respectively. In emergency science communication cooperation for COVID-19, the ESCCNs of the emergency response phase and the ongoing emergency phase include 49 and 36 ESCOs, respectively. In total, there are 373 and 460 emergency cooperative relationships between ESCOs. These findings indicate that, from an emergency response to an ongoing emergency, the network scale of the ESCCN decreases. This occurs because, as awareness about public health emergencies improves, the content of emergency science communication and the heterogeneous resource demands of emergency science communication become more focused. From the network structure perspective, because there are more cooperative relationships of emergency science communication among ESCOs in the ongoing emergency phase, network are more closely connected. Moreover, the decrease in the average path length shortens the cooperative relationships between ESCOs, and network connectivity shows an increasing trend. Correspondingly, network cohesion also increases from 0.57 to 0.63, which strengthens the network equilibrium of the ESCCN. Additionally, since the emergency needs of the ongoing emergency phase and the functional division of ESCOs are clearer, the boundary between the core nodes and the edge nodes in the ESCCN shifts somewhat. That is, the clustering trend of the ESCCN to the core nodes decreases. In summary, from the emergency response phase to the ongoing emergency phase, the network tightness, equilibrium and connectivity of the ESCCN all tend to increase, whereas network agglomeration decreases.

Node attribute analysis of ESCCNs

Since emergency tasks and public safety needs differ across different phases, the number and type of ESCOs in the ESCCNs of the emergency response phase and the ongoing emergency phase are significantly different. To compare the differences in ESCOs intuitively, the top 20 ESCOs in order of normalized degree centrality and normalized eigenvector centrality are calculated and presented in Tables 4 and 5. Figure 3 shows the distributions of the box plots of the normalized degree centrality and normalized eigenvector centrality of ESCOs.

Table 4 Top 20 ESCOs by normalized degree centrality.
Full size table
Table 5 Top 20 ESCOs by normalized eigenvector centrality.
Full size table
Fig. 3: Box plot diagram of ESCOs with centrality.
figure 3

a Normalized degree centrality. b Normalized eigenvector centrality.

Full size image

The higher the degree centrality of the ESCO is, the more the ESCO in the ESCCN has formed cooperative relationships with emergency science communication and gained greater influence. From a whole network perspective, comparing the ESCCN of the whole phase and the ESCCN of the emergency response phase, CAAC, SAT, MHRSS and CAS are the core organizations of the emergency response phase rather than the whole phase. Similarly, comparing the ESCCN of the whole phase and the ESCCN of the ongoing emergency phase, SRGJRCM, CNPGC and CAAC are the core organizations of the ongoing emergency phase rather than the whole phase.

First, the health department (NHC) and the press release department (SCIO) are ranked in the top 3 in the ESCCNs of different phases. The NHC and SCIO coordinate, convene and release emergency science communication; these are the ESCOs with the most resources and the greatest influence on the ESCCN.

Specifically, in the emergency response phase, to improve the public’s awareness and grasp of the need to engage in health-protective behaviors, ESCOs provide emergency science communication for COVID-19 prevention and control, including pandemic data, protective measures, diagnosis and treatment protocols, drug research and development, and medical conditions. Thus, the health department (NHC), drug administration (NMPA), medical administration (CCDCP, PUFH), and science technology department (MST) are the key nodes of the ESCCN. ESCOs also explain public service information involved in COVID-19 prevention and control to the public, and this involves interpreting pandemic risk in key areas such as public travel, employment, and production and operation. Alternatively, ESCOs introduce pandemic prevention and control measures of public concern with respect to, for example, cross-border security, food safety, and material safety. Thus, the transport department (MT, SPB, CSRGC), planning department (NDRC, MARA), customs administration (GAC), and material supply department (MIIT) also play key roles in the ESCCN.

In the ongoing emergency phase, the economic and social order is gradually restored, and the risk of transmission of the pandemic is reduced. During this period, the tasks and needs of the COVID-19 emergency response change. The status and precautions of vaccine development and vaccination are key issues about which the public needs to know. As a result, the research and development department (SRGJRCM, GAC) has a high degree of centrality. Moreover, the risks existing in the prevention and control of key areas and key places are also an important part of emergency science communication. In addition to the health department (NHC) and medical administration (CCDCP), the transport department (MT, SPB, CSRGC, CAAC), education department (ME), tourism department (MCT), and immigration administration (NIA) become more important in the network.

The greater the eigenvector centrality of ESCOs is, the more important and influential the ESCOs in ESCCNs are connected and establish cooperative relationships. Similar to the results of the degree centrality analysis, the eigenvector centrality of the health department (NHC) and press release department (SCIO) is also high in different phases. A comparison of the ESCCN of the whole phase with different phases reveals that MHRSS, CAS, MF and SPB have a greater influence on adjacent nodes in the emergency response phase, and CAS, SRGJRCM and CSRGC have more high-impact relationships in the ongoing emergency phase. However, the abovementioned ESCOs are not prominent in the ESCCN of the whole phase.

In the emergency response phase, the eigenvector centrality of all functional departments in the network, with the exception of the health department, the press release department and the disease control department, is greater than that in the ongoing emergency phases. This finding shows that diversified ESCOs need to participate in emergency science communication. Functional departments establish cooperative relationships with influential ESCOs to provide relevant information, platforms and content resources to achieve emergency science communication objectives and improve the public’s ability to respond to emergencies. In the ongoing emergency phase, the ESCOs ranked in the top 20 by eigenvector centrality are of the same type as those with the highest degree centrality, as shown in Table 4. This result indicates that in the ESCCN of the ongoing emergency phase, core ESCOs often form cooperative relationships with other high-impact ESCOs. The cooperative relationships between ESCOs in the network are characterized by both quantity and quality. Moreover, similar to the emergency response phase, the eigenvector centrality of functional departments in the ESCCN of the ongoing emergency phase is ranked higher. This result indicates that the network needs the emergency resources provided by diversified ESCOs.

As illustrated in Fig. 3a, the distribution of the normalized degree centrality of ESCOs varies greatly among the ESCCNs. The distribution of ESCOs in the ESCCN of the emergency response phase is more compact, and the distribution of normalized degree centrality tends to deviate to the lower bound of the box plot diagram. This finding indicates that core nodes with high normalized degree centrality in the network are more prominent. In the ESCCN of the ongoing emergency phase, the normalized degree centrality of ESCOs has a wide distribution. This finding shows that core ESCOs with high power are not the only ones in the network and that many ESCOs have important functions in emergency science communication cooperation. In contrast, the distribution of the normalized eigenvector centrality of ESCOs is relatively balanced in ESCCNs. However, in the ESCCN of the ongoing emergency phase, the box plot diagram of ESCOs shows a wider distribution than that in the emergency response phase. Moreover, in the ESCCN of the emergency response phase, outliers beyond the upper bound of the box plot are more significant. In conclusion, the core-edge features of the ESCCN in the emergency response phase are more obvious. In the ESCCN of the ongoing emergency phase, the cohesion between ESCOs is more balanced.

Cross-group correlation analysis of ESCCNs

In ESCCNs, different types of ESCOs perform different emergency tasks. Since different types of ESCOs provide different emergency resources to the network, the resource interactions and information flows among different types of ESCOs need to be discussed further. Thus, this study analyzes the ESCCNs of the whole phase, emergency response phase and ongoing emergency phase on the basis of the E-I index. The calculation results of the E-I index of the ESCCNs are shown in Table 6.

Table 6 E-I index of ESCCNs.
Full size table

To confirm the validity of the E-I index corresponding to the ESCCNs above, 5000 random permutations of the emergency science communication cooperation data are performed in this study. The results show that the ESCCN of the whole phase is significant at the p < 0.01 level and that the ESCCNs of the emergency response phase and the ongoing emergency phase are significant at the p < 0.05 level. These results indicate that the E-I index of this study meets the conditions for analysis and is effective.

As shown in Table 6, the E-I index of the ESCCN of the whole phase is −0.101, indicating that in the process of emergency science communication for COVID-19, ESCOs are more inclined to form cooperative relationships with ESCOs of the same type. In other words, barriers to cooperation among health administrative departments, functional administrative departments, and experts and academic organizations remain.

Specifically, in the emergency response phase, the E-I index of the ESCCN is −0.148 because, in the early phase of emergency management for COVID-19, different types of ESCOs were not clear about their positioning in emergency science communication. Thus, there was a certain gap in the sense of identity among different groups. Moreover, to reduce public sentiment related to COVID-19 and respond to public demand, rapid voice and emergency science communication among health administrative departments was needed. However, at that time, cross-group cooperation among different types of ESCOs had not yet been formed. ESCOs are more inclined to carry out emergency science communication jointly with ESCOs of the same type.

Owing to the experience of emergency operations in the emergency response phase, ESCOs formed more diversified cross-group relationships in the ongoing emergency phase. During this period, the cooperative relationships formed between different types of ESCOs are basically the same as those formed between ESCOs of the same type, and the E-I index of the ESCCN of the ongoing emergency phase is 0.087. Different types of ESCOs are more effective in terms of resource transfer and complementary advantages in emergency science communication and play positive supporting roles in responding to COVID-19 and improving the public’s emergency response capacity.

Furthermore, to clarify the correlation between different types of ESCOs, the ILs and ELs among health administrative departments, functional administrative departments, and experts and academic organizations are studied and calculated. The E-I indices of each group are obtained, as shown in Table 7. Moreover, the density matrix is plotted to reflect the cooperation among different emergency groups (see Fig. 4).

Table 7 E-I index of emergency groups.
Full size table
Fig. 4: Density matrix of different emergency groups.
figure 4

a In whole phase. b In emergency response. c In an ongoing emergency.

Full size image

Among the different ESCCNs, health administrative departments all have the highest E-I indices. For the ESCCNs of the whole phase, emergency response phase and ongoing emergency phase, the corresponding E-I indices are 0.617, 0.714 and 0.434, respectively. These results are due to the high degree to which health administrative departments grasp emergency information, such as data on public health emergencies, emergency protection and mitigation measures. Moreover, because of the division of emergency tasks, health administrative departments need to coordinate and guide functional administrative departments, experts and academic organizations to participate in emergency science communication to increase the public’s awareness and capacity. Therefore, health administrative departments not only collect but also publish emergency science information. More contacts with functional administrative departments, experts and academic organizations are needed to achieve the objectives of emergency science communication.

Moreover, the E-I indices of the functional administrative departments in the ESCCN of the whole phase, emergency response phase and ongoing emergency phase are −0.429, −0.468 and −0.209, respectively, which are all <0. The reason for this is that in ESCCNs, functional administrative departments are the most common departments, mainly in key prevention and control areas that provide policy support and information sources for COVID-19 emergency science communication. Specifically, functional administrative departments include transportation departments, planning departments, finance departments, market management departments, and tourism departments. As a result, functional administrative departments have formed more cooperative relationships with ESCOs of the same group. Additionally, the E-I index for functional administrative departments decreased from the emergency response phase to the ongoing emergency phase. This finding shows that functional administrative departments have formed more cross-group relationships with their integration into the emergency science communication cooperation system.

In the ESCCNs of the emergency response and ongoing emergency phases, the E-I indices corresponding to experts and academic organizations are significantly different, at 0.000 and 0.429, respectively. During the emergency response phase, experts and academic organizations had not yet formed stable relationships with other ESCOs. Moreover, because the understanding of COVID-19 was not clear, experts and academic organizations needed more complex argumentation and research. Therefore, when experts and academic organizations carry out emergency science communication, the internal relationships of the group formed are equal to those of the cross-group, reflecting a balanced situation. In the ESCCN of the ongoing emergency phase, which is based on the coordination and cooperation mechanism support formed in the emergency response phase, experts and academic organizations are also associated with health administrative departments and functional administrative departments. ESCOs that form cross-group groups jointly release emergency science communication information to the public. Therefore, experts and academic organizations form more ELs with the ESCOs of other groups, which in turn form a higher E-I index.

As illustrated in Fig. 4, in the ESCCNs of the whole phase, emergency response phase and ongoing emergency phase, the internal cooperation densities of Groups 1, 2 and 3 are {0.393, 0.203, 0.144}, {0.500, 0.161, 0.242} and {0.667, 0.340, 0.127}, respectively. These results indicate that although Group 2 has the most ILs among different ESCCNs, the cooperation density within Group 1 is the largest due to the difference in the number of ESCOs in different groups. As the main subjects responsible for public health emergencies, health administrative departments have more concentrated emergency information and emergency resources and have a greater right to emergency science communication in public health emergencies.

Effect of ESCCNs of different phases on the ESCNN of the whole phase

Owing to the differences in emergency tasks related to COVID-19 at different phases and the variation in the virus, emergency science communication involves a dynamic process of constant adjustment and updating in terms of the organizational structure and content setting. Therefore, the ESCCNs of the emergency response and ongoing emergency phases have different effects on the network scale of the ESCCN of the whole phase and on the cooperative relationships between ESCOs. To clarify the mechanism of the impact of the ESCCNs of the emergency response and ongoing emergency phases on the ESCCN of the whole phase, QAP correlation and QAP regression analyses of ESCCN formation are conducted.

QAP correlation analysis of ESCCN formation

The correlations between the ESCCNs of the emergency response phase and ongoing emergency phase and the ESCCN of the whole-phase network are analyzed. Then, analysis with 2000 random permutations is performed. The correlation coefficients between the ESCCN of the whole phase and the emergency science communication cooperation matrix of different phases are obtained. As shown in Table 8, the correlation analysis results of the ESCCN of the whole phase and the emergency science communication cooperation matrix of different phases are significantly correlated (significance = 0.000). The correlation coefficient between the ESCCN of the emergency response phase and the ESCCN of the whole phase is 0.788, and the correlation coefficient between the ESCCN of the ongoing emergency phase and the ESCCN of the whole phase is 0.748. These results indicate that emergency science communication practices and the cooperation matrix between ESCOs at different phases have a significant positive driving effect on the formation of the ESCCN of the whole phase.

Table 8 QAP correlation for ESCCNs.
Full size table

QAP regression analysis of ESCCN formation

Furthermore, QAP regression analysis is performed to clarify the difference in the influence of the emergency science communication cooperation matrix at different phases on the ESCCN of the whole phase. Similarly, the QAP regression analysis results are obtained under 2000 random permutations, as shown in Table 9.

Table 9 QAP regression for ESCCNs.
Full size table

The regression analysis results show that the goodness of fit of the regression test is 0.876. This result indicates that the temporal influencing model for ESCCN formation has strong explanatory ability. Moreover, the coefficients of the impact of the emergency science communication cooperation matrix at different phases on the formation of the ESCCN of the whole phase are significant at the p < 0.01 level. Unlike the QAP correlation analysis results, the ESCCN of the ongoing emergency phase has a stronger positive driving effect on the ESCCN of the whole phase, with a standardized coefficient of 0.600. Relatedly, the ESCCN of the emergency response has a slightly weaker effect on the formation of the ESCCN of the whole phase, for which the standardized coefficient is 0.538.

Discussion

Owing to the diversified fields involved, the wide distribution of information sources, and the strong timeliness of content, emergency organizations in all fields and at various levels need to participate in emergency science communication and form stable emergency cooperative relationships (Göbel et al. 2022). Moreover, as an important link throughout the whole process of the emergency management of public health emergencies, emergency science communication has different needs and organizational cooperation characteristics in different emergency phases. The emergency science communication mode of multisubject cooperation is an effective measure for integrating emergency resources. Thus, this paper aims to discuss the differences in emergency response to an ongoing emergency of the emergency science communication cooperation mode from the network perspective. Moreover, a theoretical basis for how to improve the effectiveness of ESCCNs for public health emergencies is provided.

For ESCCNs in different phases, more ESCOs indicate that the network has more heterogeneous and diversified emergency science communication resources, which can more effectively support emergency science communication activities. ESCOs have formed cooperative relationships to interact with emergency resources (Chen et al. 2020; Wei et al. 2013). In the emergency response phase, owing to the uncertainty of emergency needs corresponding to the characteristics of public health emergencies, the ESCCN should fully absorb and introduce ESCOs with heterogeneous emergency science communication resources to address the insufficient supply of resources. Moreover, since ESCOs have not yet formed a strong foundation of trust and experience, each ESCO is more inclined to form cooperative relationships with ESCOs of the same type. Overall, the ESCCN of the emergency response phase has a loose structure and weak balance. With the deepening of the emergency management of public health emergencies and the beginning of the ongoing emergency phase, ESCOs’ awareness of public health emergencies has gradually improved, and nonessential ESCOs in the ESCCN have withdrawn from the network. The number and frequency of cooperative relationships among retained ESCOs will increase. There will be more obvious cross-group cooperation among different types of ESCOs on the basis of established interactions and cooperative relationships. Correspondingly, the tightness and connectivity of the ESCCN of the ongoing emergency phase are improved.

Additionally, both the emergency response phase and the ongoing emergency phase have significant positive driving effects on the formation of the ESCCN of the whole phase. Furthermore, the temporal factors influencing ESCCN formation are compared with the trends of ESCCN network characteristics. In the emergency response phase, to address the unknown emergency needs and characteristics of COVID-19, different types of ESCOs participate in emergency science communication. The ESCCN of the emergency response network is formed with a larger scale and more diversified emergency resources. As COVID-19 is better understood and the spread of the pandemic is controlled, the contents and response needs of emergency science communication are further clarified (Zhang et al. 2020b). The number of noncore ESCOs in the ongoing emergency phase decreases. The findings also show that the ESCCN of the ongoing emergency phase with higher network density and network cohesion has a greater influence on the formation of the ESCCN of the whole phase. The diversification of ESCO types, the closeness of the relationships between ESCOs and the connectivity of ESCCNs are the main factors that promote the formation of the ESCCN of the whole phase.

Overall, the theoretical and practical contributions of this study are as follows. In this study, the cooperative relationships formed when organizations participate in emergency science communication are taken as the data basis, and the practice of emergency science communication in response to COVID-19 in China is taken as a case study. This research expands the research perspective of cross-organizational relationships in emergency science communication. Furthermore, this study constructs a temporal influencing model for ESCCN formation to explore the mechanism of the impact of different phases on the formation of an emergency science communication cooperation mode, which expands the perspective and thinking of dynamic research on emergency management networks. Therefore, the conclusions of this research can improve the cooperation efficiency of emergency science communication and strengthen the dynamic adjustment ability of emergency management.

Notably, regarding the applicability of this study, several limitations and further research directions remain. This study focuses on the perspective of information release in emergency science communication and discusses the main characteristics of the emergency science communication cooperation mode. However, based on the practice of emergency science communication, ESCOs at different levels also differ. Furthermore, research can be conducted from the perspective of grassroots-level emergency science communication practices. Moreover, this study considers only the cooperation relationships between organizations that carry out emergency science communication but fails to discuss the cooperation between emergency science communication objects in emergency science communication. Moreover, this study is based on the practice of responding to COVID-19 as a case study by considering only the emergency response phase and ongoing emergency phase. In future research, the emergency phase can be subdivided on the basis of the characteristics of actual cases, and the driving mechanism of the formation of the emergency science communication cooperation mode in different phases can be discussed.

Conclusions

Emergency science communication is important for responding to emergency management needs, improving the public’s risk perception and ability to respond to emergencies, and reducing losses from emergencies. Discussing the differentiated characteristics of different phases and clarifying the effects of different phases on emergency science communication cooperation during the whole phase are valuable for explaining the dynamic law of emergency science communication cooperation and improving its efficiency.

In this study, the Chinese government’s practice of emergency science communication in response to COVID-19 is taken as a case study. On the basis of network structure analysis, node attribute analysis, and cross-group correlation analysis, a comparative analysis of the ESCCNs of the emergency response phase and the ongoing emergency phase is conducted. On the basis of the QAP analysis, a temporal influencing model of the ESCCN is constructed, and the effects of ESCCNs of different phases on the ESCCN of the whole phase are explored. The results are as follows:

  1. (1)

    The differences in network structure. From an emergency response to an ongoing emergency, new ESCOs join and original ESCOs exist in ESCCNs, and the network scale decreases. Thus, the network tightness, equilibrium and connectivity of the ESCCN all show an increasing trend, whereas network agglomeration decreases. Moreover, for the ESCCN of the whole phase, the characteristics of its network structure are in the middle of those of the emergency response and ongoing emergency phases.

  2. (2)

    The differences in node attributes. The differences in emergency science communication needs at different phases lead to the continuous evolution of the types and quantities of ESCOs. The health department (NHC) and the press release department (SCIO) are the core nodes of ESCCNs and are connected to more influential nodes. Moreover, all functional departments have high degree centrality and eigenvector centrality, indicating that the emergency resources provided by diversified ESCOs are needed by the network. Additionally, the core-edge features of the ESCCN of the emergency response phase are more obvious, and the cooperative relationships between ESCOs in the ESCCN of the ongoing emergency phase are more balanced.

  3. (3)

    The differences in cross-group correlations. In the emergency response phase, ESCOs are more inclined to cooperate with ESCOs of the same type in emergency science communication. In the ongoing emergency phase, more diversified cross-group partnerships are formed between ESCOs. With the formation of orderly emergency science communication cooperation with ESCOs, the health administrative department is no longer the only leading organization. Furthermore, functional administrative departments, experts and academic organizations are integrated into the emergency science communication system.

  4. (4)

    The ESCCNs of both the emergency response phase and the ongoing emergency phase have significant positive driving effects on the formation of the ESCCN of the whole phase. The diversification of ESCO types, the closeness of the relationships between ESCOs and the connectivity of ESCCNs are the main factors that promote the formation of the ESCCN of the whole phase.

Implications

Furthermore, to strengthen the science popularization cooperation mechanism in public health emergencies and improve the effectiveness of the ESCCN, the following optimization suggestions for the practice of emergency science communication are proposed.

  1. (1)

    Matching the dynamic characteristics of emergency science communication practices. Emergency science communication involves the whole process and various links of the emergency management of public health emergencies, covering numerous issues such as the release and notification of emergency information, the interpretation of health-protective behaviors and the release of social service information. Thus, the nature, extent, scope and sustained characteristics of public health emergencies need to be comprehensively considered to strengthen the complementarity and convergence of the core issues of emergency science communication at different phases and improve the timeliness and completeness of emergency science communication. Moreover, the key tasks of public health emergencies at different phases should be clarified to improve the matching of emergency science communication cooperation and emergency tasks. Additionally, the evolving characteristics of emergency cooperation in emergency science communication should be explored to improve the stability of ESCCNs.

  2. (2)

    Optimizing the allocation and supply of resources for emergency science communication cooperation. The formation of the emergency science communication cooperation mode stems from the traction of the emergency needs of the government and the public. Fully exploiting the resource stock of ESCCNs and allocating the resource increment of ESCCNs in an orderly manner are the keys to optimizing the resource allocation of emergency science communication cooperation. Owing to the continuous characteristics of emergency science communication cooperation, the trust foundation and capital stock of emergency organizations formed in past emergency science communication cooperation can not only effectively use the resource structure formed but also improve the efficiency of resource supply. Furthermore, on the basis of the new problems and scenarios of emergency science communication cooperation, new ESCOs should be introduced into the network to increase the total amount of ESCCN resources. Additionally, the cross-group or intragroup flow of emergency resources should be examined to strengthen the level of emergency science communication cooperation.

  3. (3)

    Strengthening policy support and institutional guarantees for emergency science communication cooperation. Policies for emergency science communication under normal and emergency states should be formulated to improve the system guarantee of emergency science communication cooperation to enhance the stability of the ESCCN. Moreover, preparations for emergency science communication in the normal phase should be made. Additionally, diversified emergency science communication activities should be actively implemented by focusing on the key time to solidify the experience of emergency science communication cooperation and consolidate the foundation of trust. When public health emergencies occur, emergency science communication cooperation needs to be implemented promptly and accurately to improve the standardization of such cooperation. Additionally, given the new characteristics of public health emergencies, the dynamic updating and optimization mechanism of emergency science communication cooperation should be gradually established to improve the adaptability of the ESCCN.