INTRODUCTION
1. Background
In the government’s “Comprehensive plan to foster and support the 3rd pharmaceutical bio industry to leap for- ward as a global hub of bio-health” [1], the word “platform” appears 32 times, including artificial intelligence (AI) drug development platform, and AI drug discovery platform. With the recent development of digital technology, plat- form models are drawing attention in the pharmaceutical industry. Platform means digital or physical space that fa- cilitates interaction between users. The platform provides an environment for exchanging products, services, infor- mation, and so forth, enabling many participants to create value, drive innovation, reduce costs, consolidate data, and strengthen global cooperation.
2. Pharmaceutical industry platform
In the pharmaceutical industry, a platform refers to a digi- tal or physical space that allows researchers, pharmaceutical companies, hospitals, and government agencies to effectively collaborate on drug development and research. This plat- form plays a crucial role in integrating and analyzing data, building collaborative networks, and accelerating drug de- velopment. A platform can be described as a tool designed to innovate the drug development process and maximize ef- ficiency in pharmaceutical research and development.
1) Key functions of a pharmaceutical platform
First, the platform provides the function of data integra- tion and analysis. The pharmaceutical industry handles vast amounts of data, and the effective integration and analysis of this data are central to drug development and research. Platforms support the centralization and analy- sis of data collected from various sources, such as clinical data, genomic data, and experimental results. This helps researchers gain insights necessary for identifying drug candidates or developing effective therapies. For example, IBM Watson for Drug Discovery uses AI technology to rapidly analyze massive amounts of scientific data, helping researchers avoid redundant studies and accelerate the de- velopment of innovative drugs. Through this platform, the data analysis process in the pharmaceutical industry be- comes automated, and research outcomes are enhanced ef- ficiently. Second, the platform facilitates research collabora- tion. In the pharmaceutical industry, developing new drugs requires global collaboration among various research insti- tutions and pharmaceutical companies. The platform plays an essential role in building and managing international research collaboration networks. It provides the functional- ity for researchers to share data in real-time and exchange research results, thus promoting global collaboration and improving research efficiency. For instance, the Innovative Medicines Initiative (IMI) is a large-scale public-private partnership platform where European public institutions and private pharmaceutical companies collaborate to share clinical trial data and solve bottlenecks in drug develop- ment. This allows researchers across Europe to collaborate on finding solutions to complex diseases and provides op- portunities to rapidly commercialize research outcomes. Third, pharmaceutical industry platforms are equipped with the function of clinical trial management and acceler- ating drug development. They integrate clinical trial data and manage it efficiently, simplifying the clinical trial pro- cess and significantly speeding up drug development. Dur- ing clinical trials, all stages—such as data collection, patient management, and result analysis—are centrally managed on the platform, reducing the time and costs associated with drug development. For example, TransCelerate Bio- Pharma supports pharmaceutical companies by developing standards for clinical trial management and sharing trial data, enabling a more efficient drug development process. This platform helps pharmaceutical companies collaborate with regulatory authorities to shorten the approval process for new drugs, ensuring that drugs reach the market faster.
Fourth, the platform offers the function of supporting personalized medicine. It can analyze a patient’s genomic information and health data, providing personalized treat- ment plans optimized for each patient. The platform sup- ports the implementation of precision medicine by analyz-ing individual genetic information, lifestyle habits, and medical records. For example, the US National Institutes of Health (NIH) All of Us Research Program is a public- private partnership platform that collects and analyzes ge- nomic data and health information from over one million Americans to study personalized medicine. Through this platform, researchers can develop new approaches to dis- ease prevention and treatment, and personalized medicines can be developed quickly and provided to patients.
2) Importance and role of platforms
Platforms significantly enhance research efficiency, re- duce redundant research, and accelerate the drug develop- ment process. Researchers can use platforms to analyze data in real-time, leading to improved research outcomes. Additionally, platforms facilitate collaboration between public and private researchers and help research outcomes to be quickly commercialized. This enables pharmaceutical companies to save on research and development costs and time, allowing new drugs to enter the market more quickly. Furthermore, platforms analyze patients’ genomic data and medical records, enabling personalized treatments that of- fer better therapeutic outcomes for individual patients.
Thus, platforms in the pharmaceutical industry are not just tools for research; they serve as core technological infra- structure that integrates data analysis, collaboration facilita- tion, clinical trial management, and personalized treatment. Platforms play a vital role in promoting innovation and col- laboration in pharmaceutical research and in significantly improving the speed and efficiency of drug development. The platform plays a crucial role in the pharmaceutical industry by facilitating collaboration between researchers and companies, expediting the drug development process, minimizing redundant research, and fostering innovation through data sharing and collaboration [2]. For instance, IBM’s Watson for Drug Discovery leverages AI to analyze extensive scientific papers and data, enabling researchers to avoid duplicate research and develop innovative treatments more efficiently. These platforms also serve to connect the global research community, providing opportunities for experts from diverse fields to collaborate [2-5]. Despite the predominant presence of digital companies, Verily, a sub- sidiary of Google, offers pioneering solutions for disease prevention, diagnosis, and treatment through digital health platforms. By analyzing patient data in real time and deliver- ing customized treatment, Verily contributes significantly to the pharmaceutical industry [6]. Moreover, these platforms enable the integration and analysis of data gathered from various sources, fostering innovation in fields such as dis- ease prediction and personalized treatment development [6]. Additionally, they drive innovation in pharmaceutical distribution, leading to enhanced market access, optimized supply chain management, efficient product launches, and improved patient accessibility [7,8]. Through patient data management and telemedicine services, these platforms en- able patient-centered healthcare, thereby enhancing patient convenience and treatment effectiveness [9-11].
3. Academic research network analysis
Analysis of academic research networks among research- ers plays an important role in academic development and innovation [12]. Through the analysis of researchers’ networks, researchers in the pharmaceutical industry can be identified overall, and cooperation between major re- searchers and institutions in the pharmaceutical industry also can be identified. In the academic communication of researchers, various forms of interaction relationships ap- pear, and the relationship between researchers can be de- scribed and explained in the researcher’s network and can be constructed based on various academic activities. Re- searchers’ networks consisting of academic activities iden- tified in academic literature form a network of researchers through collaborative relationships for joint research on specific tasks, citation relationships of academic literature by related subjects, and bibliographic connections formed by quoting specific academic literature [12-14]. In the joint research network, academic knowledge, which is the product of academic activities, is created through coopera- tion and interaction of researchers’ communities. When researchers take the form of joint research, their relation-ship is expressed in a network is also called a cooperative network. The co-author network is a type of collaborative research network, which corresponds to the case where academic knowledge is generated in the form of academic papers, and the collaborative research network is more comprehensive but does not need to be strictly divided [15,16]. Co-author network is a cooperative network cre- ated by measuring the relationship between authors who participated in the same academic information production such as papers, books, translations, reports, patent applica- tions and research and development issues. Such co-author networks, also known as “knowledge linkage structures” or “knowledge structures” that can microscopically re- veal the flow of knowledge and the tendency to structure knowledge within the academic field, are used as useful research methods [15-17]. In the past, there were many studies that evaluated citations in terms of social influence, focusing on cited information between papers. Co-author networks have different characteristics from citations net- works, which have been mainly analyzed in the past. Both co-author networks and citation networks can be used as concepts to evaluate the author’s academic influence. The difference is that citation networks are less socially cohe- sive than co-authors networks, are often not familiar with each other, and have a much wider range of time zones [13]. In a co-author network, nodes represent authors and links represent academic relationships of collaborative research and cooperation so that research cooperation and social relationships between authors can be grasped through co- author analysis. In addition, social relationships formed among researchers participating in joint research can be the starting point for the development and initiative of the field because they presuppose more active interactions than any form of research cooperation [18].
METHODS
1. Data collection
Bibliographic information published for 5 years and 8 months from January 2017 to August 2022 was collected for data related to the pharmaceutical industry platform. NetMiner (NetMiner), a text and network data analysis software, and Biblio Data Collector, a data collector that works with NetMiner, were utilized. The institute is a global academic publishing company that publishes Na- ture, an international academic journal and has access to bibliographic information through Springer’s OpenAPI. The search terms “pharmaceutical,” “drug,” “medicine,” and “healthcare” related to the pharmaceutical industry were combined with “platform,” and basic documents such as drug research and development were filtered to separate data related to the pharmaceutical industry platform and collection results containing “platform” search terms. To collect data related to the “pharmaceutical industry” and “pharmaceutical industry platform,” select “Title” and “Keyword” among the search fields and enter a combina- tion of search terms to proceed with the collection. For example, in order to collect cases where the title (Title) of the literature has both “drug” and “platform,” we entered title: “drug” AND title: “platform” as the search term. In the table below, “&” indicates the use of AND conditions. In order to separate data related to the “pharmaceutical indus- try platform,” data extracted from a combination of search terms containing “platform” were selected (Figure 1).
2. Analytical methods
The researcher network analysis was conducted with the aim of identifying the overall cooperation of platform researchers in the pharmaceutical industry and finding out who and what institutions have the initiative in such research overseas. To this end, a co-author ship network between authors was constructed using a two-mode matrix between papers and authors. The author’s network was ag- gregated by the author’s affiliated institutions, the collabo- ration network between research institutions was analyzed, and the author’s network was aggregated by the author’s affiliated institutions (Table 1).
Block modeling was used for institutional collaboration network analysis (Figure 2). A network analysis method that can explain the structure of the network between re- search institutes that have been shown to be co-authored with the researcher network connected with the co- authored relationship has been applied. First, we measured the number of isolated nodes (# of isolates) to identify the number of authors without co-authors, and component analysis identified non-disconnected sub-networks and ex- clusively connected groups in the network.
For the analysis of the researcher/research institution network data, the author-thesis two-mode matrix was converted into the author-the author one-mode matrix (co-membership method) to form the network. This is a co-authorship network between authors using papers as a medium, and the relationship strength between authors is calculated by the inner product of the matrix. In other words, the meaning of relationship strength is the number of papers written by both authors. In this network, the node is the author (researcher), the link is the connec-tion between the authors, and the connection means co- authorship or cooperation. The connection strength is the number of co-authorship or cooperation, and this network can be regarded as a structure of co-authorship or research cooperation (Figure 3).
RESULTS
1. Leading authors and research institutions
To understand the pattern and structure of research cooperation through the analysis of authors and institu- tions’ attributes and networks, major researchers and major research institutes in pharmaceutical industry platform research were analyzed based on the basic attributes of the researcher’s thesis and the number of researchers belonging to the research institution (author). The total number of authors was 632 and 2,064 cases of co-author relationships, focusing on the structural characteristics of the researcher’s co-author network and the research institute’s cooperation network, and the total number of academic research litera- ture on pharmaceutical industry platforms was 139, and the trend was gradually increasing from 11 in 2017 to 30 in 2022 (Table 2).
As a result of analyzing research cooperation over the past 5 years and 8 months through analysis of major re- searchers, research institutions, and collaborative network analysis, the number of papers by the top authors and re- searchers by the research institution was as follows (Table 3). In the pharmaceutical industry platform study, major researchers analyzed the number of papers based on the number of papers and found that Yoshimasa Masuda, the researcher with the largest number of papers, participated in a total of six papers.
According to an analysis of research institutes with many researchers participating in pharmaceutical industry plat- form research, Bristol University in the United Kingdom ranked first, followed by Wuhan University in China, Har- vard Medical School in the United States, Ceuma Univer- sity in Brazil, and Khalifa University in the fifth (Table 4).
2. Co-author network analysis
Through network analysis, the pattern was identified focusing on the structural characteristics of the collabora- tive network of researchers and the cooperative network of research institutes (Figure 4). A node represents an indi- vidual author, and an edge represents a co-author relation- ship. The network represents a co-author cluster of various sizes and densities, but the larger cluster in the upper left means a more cooperative group or research team, and the smaller the cluster, the less cooperative or more isolated the researchers. Cooperation between each research group or topic was found to be very limited as large numbers of small groups were found to form and connections between clusters were rare.
A one-person author who studies alone without co-writ- ing with others appears as an isolated node on the network, but a network analysis shows 14 one-person authors. The network analysis found that the number of components that are seamlessly connected groups was 115, the maxi- mum number of researchers was 22 and the minimum group was two (Table 5).
In order to investigate the continuous cooperative rela- tionship, only those relationships with two or more co-au- thors (number of co-authors) were extracted and analyzed separately (Figure 5). The number of researchers included in two or more co-authorship networks was 12, and the number of co-authorship relationships was 11, which con- sisted of a very small number of people. Yosimasa Masuda and Tetsuya Toma were found to have participated in four papers together.
3. Institutes cooperation network analysis
In order to find out the cooperative relationship between the institutions to which the researcher belongs, the re- searcher network was converted into an interagency net- work using block modeling technology. When a network between researchers with more than one co-authorship re- lationship was converted into an inter-agency network, the number of institutions included was 274 and the number of interagency cooperation relationships was 447. By apply- ing block modeling, the relationship between researchers within the same institution is displayed diagonally in the matrix in Self-loop, which is visually represented as an an- nular line. The thickness of the line means the number of cooperative relationships and the more cooperative rela- tionships, the thicker the line appears, but it was confirmed that co-authorship relationships are mainly active within institutions (Figure 6).
In relation to the comparison of the performance be- tween large and small subnetworks, Figures 4 and 6 visual- ize the research collaboration network within the pharma-ceutical industry, where each node represents a researcher or institution, and the links signify their collaborative rela-tionships. Large subnetworks with many nodes represent a large-scale network where numerous researchers or in-stitutions collaborate closely. These networks tend to share more research resources and data, increasing the likelihood of better performance. In pharmaceutical research, sharing clinical trial data or patient data is crucial, and such large collaborative networks can facilitate efficient data sharing and enhance research efficiency. On the other hand, small subnetworks with fewer nodes represent collaborations between a smaller number of researchers, which may result in relatively limited research resources. These smaller net-works may focus on specific fields or conduct specialized research, but their research output may be more limited in scope. In relation to the analysis of key areas of interest for research collaboration strategy, the network analysis in Figures 4 and 6 can help identify the primary areas of inter-est within each collaboration network. For instance, large networks may excel in fields requiring significant resources and collaboration, such as drug development or clinical tri-als, and are more likely to be connected to large-scale plat-forms like IMI or the All of Us Research Program. In con-trast, smaller networks might focus on specialized research topics or innovative technology development. For example, fields such as gene editing or personalized medicine may be the focus of research in smaller networks, which can collaborate with specialized platforms to achieve in-depth research outcomes. In relation to the connection with platforms, large subnetworks with many nodes are often linked to large-scale platforms, contributing to the efficient use of various data and resources. For example, platforms like IMI facilitate large-scale research projects across Eu-rope by enabling collaboration among researchers, thereby supporting large networks in achieving more substantial research outcomes. In contrast, small subnetworks might be connected to specialized platforms that support research in specific fields. For example, small networks focusing on genomic research or personalized medicine can collaborate with specialized data platforms to maximize their research output. Therefore, by comparing the size of nodes and the performance of subnetworks, as shown in Figures 4 and 6, it is likely that large networks with more resources and col-laborative opportunities will produce higher research out- puts, often linked to large-scale public-private partnership platforms. Meanwhile, smaller networks may be connected to specialized platforms that support focused research, which can be advantageous for producing precise or inno- vative research outcomes.
An analysis of cooperative networks among research institutions found 69 cooperative groups (components) except for institutions whose co-authorship relationship is shown only within the institution. The cooperative group, which includes the maximum number of institutions, was found to be 10 institutions, and the minimum group had two institutions (Table 6).
As a result of detailed screening of which institutions are cooperative in descending order of the number of institu- tions of the cooperative group, the relatively large coop- erative relationship was found to be a co-written relation- ship between US and Kenyan researchers, led by Indiana University in the United States. Next, as a group of nine organizations, the cooperation of US and Japanese research institutions, led by Carnegie Mellon University, and co-au-thorship relationships within the United States were shown. Next, research institutions from Vietnam and Bangladesh emerged as a group of seven organizations, followed by a network of institutions from three countries, Australia, China, and the United Kingdom (Figure 7).
DISCUSSION
This study conducted network analysis to investigate research trends on pharmaceutical industry platforms and research cooperation among researchers. Network analysis showed the following characteristics as a result of analyzing the total number of authors and 2,064 co-authors, focus-ing on the structural characteristics of the researcher co- authorship network and the research institution coopera- tion network. First, as collaborative research clusters are generally formed in small groups, it was found that there are no network hub researchers that mediate extensive and continuous cooperation in the field. In addition, it was found that no solidarity research group with more than five co-authorship relationships has yet appeared, and it was identified as an emerging field with low maturity in joint cooperation. Second, in the pharmaceutical industry plat- form, the links between groups of 2,064 co-authors and re- searchers’ networks have been found to be rare, and mutual cooperation between other research groups and subjects has been found to be limited. Third, the major research institutes related to the pharmaceutical industry platform are as follows. The Nairobi Medical Training College, Moi University, Brown University, and Indiana University are cooperative networks between the United States and Kenya, showing active cooperation between many US universities and Kenyan research institutes. Harvard Medical School, Mayo Clinic, and Yale University are inter-university coop-eration networks in the United States, with active research cooperation involving major medical universities and laboratories, especially among universities in the United States, showing that research resources are concentrated. In Southeast Asia, cooperation networks between Viet-nam and Bangladesh are emerging as emerging coopera-tion areas in the medical and health sectors, centering on Hanoi Medical University, Bach Mai Hospital, and Dhaka Medical College Hospital. As a multinational cooperation network, international research cooperation involving major universities from Australia, China, and the United Kingdom is actively being conducted, focusing on Univer-sity of Sydney, Peking University, and University of Oxford. Understanding the current status of cooperation within the pharmaceutical industry and recognizing key collaborators is meaningful for extensive research and innovation in the industry. In this regard, more concrete and practical policy recommendations can be discussed. For example, it is pos- sible to explain how research collaboration through plat- forms can be strengthened from a policy perspective and to propose specific strategies that can promote the industrial application of platforms, as outlined below.
First, research collaboration can be enhanced through public-private partnership platforms. Governments and public research institutions can take the lead in creating systems that integrate public data and private research data through platforms, which would prevent duplication of re- search resources. This would maximize research efficiency and accelerate the development of innovative drugs. Pub- lic-private collaboration is essential to maximize research collaboration via platforms. Governments and public re- search institutions should open up basic research data, while private pharmaceutical companies can leverage these data for drug development through collaborative platforms. This would lead to the following outcomes. Examples of re- search collaboration enhancement through public-private platforms include IMI and the All of Us Research Program. IMI is a public-private partnership project jointly launched by the European Union (EU) and the European pharma- ceutical industry, making it one of the world’s largest pub- lic-private partnerships aimed at fostering innovative drug development. IMI aims to solve major challenges in the pharmaceutical industry by building a research collabora- tion platform where the EU Commission and pharmaceu- tical companies work together to accelerate the develop- ment of new drugs, especially for complex diseases like Alzheimer's, cancer, and infectious diseases. By combining public funding with private pharmaceutical industry in- vestments, IMI provides a joint platform that enables re- search institutions and companies to collaborate, share re- search resources and data, overcome bottlenecks in drug development, and improve research efficiency. This plat- form has contributed to the development of innovative vac- cines, personalized therapies, and the improvement of clin- ical trial processes. Recently, IMI played a crucial role in COVID-19 (coronavirus disease 2019) vaccine develop- ment, showing how data-sharing platforms can efficiently allocate research resources and shorten the drug develop-ment process. Another example is the All of Us Research Program led bythe US NIH, a large-scale project aimed at strengthening personalized medicine research through public-private collaboration platforms. NIH collaborates with private pharmaceutical companies, hospitals, and aca-demic institutions to collect genomic data and health infor-mation from over 1 million Americans, establishing a foundation for personalized medicine research. The pro-gram aims to integrate genomic and clinical data to develop individualized treatment plans. By building a public re-search data platform, NIH enables researchers and phar-maceutical companies to access participants’ health data, fostering collaboration between the government and pri-vate sector to establish research infrastructure, accelerating clinical research and drug development. The structure of the collaboration networks identified in this study’s phar-maceutical industry platform research trends shows the collaborative patterns among major research institutions. These insights could be practically utilized in the design of future projects like IMI and All of Us. Concrete application methods include: From the perspective of collaboration network structures, the network analysis within the phar-maceutical industry platform revealed in this study empha-sizes the real-time sharing of data and the enhancement of research efficiency on the platform. This aligns with the in-novative platform model needed in public-private partner-ship collaboration cases like IMI and All of Us. From the perspective of innovation and strengthening collaboration through platform models, the study’s emphasis on strength-ening research collaboration and the platform model can be applied to the innovative platform models proposed by IMI and All of Us, as both prioritize real-time data sharing and enhancing research efficiency through collaboration. From the perspective of evaluating research outcomes within networks, the research collaboration networks iden-tified in this study demonstrate how collaboration plat-forms accelerate drug development through data interoper-ability and sharing of research outcomes. This is consistent with the collaboration effects observed in IMI and All of Us, where integrated data-sharing within the network boosts efficiency. Moreover, by activating data sharing through public-private partnership platforms, redundant research can be avoided, research efficiency can be in- creased, and the integration of public data with private re- search can lead to rapid commercialization of basic re- search. To accelerate this realization, governments should encourage research collaboration that utilizes public data through tax incentives or research grants, while expanding the use of platforms. Additionally, to strengthen global co- operation, it is necessary to build international research collaboration networks and promote the mutual sharing of research resources across countries through platforms. Sec- ond, for research collaboration via platforms to be politi- cally strengthened and for platform technology to be more effectively utilized in the pharmaceutical industry, stan- dardization of technology is essential. Currently, many pharmaceutical companies utilize their own platforms, but inconsistencies in data formats and lack of technical com- patibility pose barriers to collaboration. To address this, standardization of data formats and the establishment of technical standards are necessary. Standardizing data for- mats ensures that the data used in pharmaceutical research and clinical trials can be shared between different plat- forms, enhancing data interoperability. This improves cross-platform functionality and allows more researchers to share data through platforms. For example, the Clinical Data Interchange Standards Consortium (CDISC) is a non- profit organization established to standardize clinical data formats. It aims to ensure data compatibility in clinical re- search and drug development processes by standardizing the formats of data generated during clinical trials, enabling researchers to seamlessly share and analyze data across platforms. CDISC’s SDTM (Study Data Tabulation Model) standard ensures that clinical trial data is converted into a standardized format, allowing pharmaceutical companies to submit consistent data to regulatory agencies like the US Food and Drug Administration or the European Medicines Agency, thus enhancing data compatibility and facilitating regulatory review. Another example is ADaM (Analysis Data Model), which provides standards for analyzing clini- cal trial data, ensuring consistency in data analysis and outcome generation, making data movement and analysis across platforms more efficient. In this way, the establish- ment of technical standards by governments and interna- tional pharmaceutical organizations can enable pharma- ceutical companies to more easily adopt technologies such as data analysis and AI applications, lower barriers to plat- form adoption, and accelerate research innovation. Third, to ensure that platform models in the pharmaceutical in- dustry are utilized more effectively, the more active adop- tion of regulatory sandboxes is necessary. To promote the introduction of new technologies and the application of platforms in the pharmaceutical industry, governments in various countries should introduce regulatory sandbox sys- tems to ease regulatory barriers during drug development and clinical trials. This would allow platform models to be more widely utilized in research and development process- es. As the Korean government shows a strong interest in pharmaceutical industry platforms, this study’s identifica- tion of key collaborators and recognition of collaboration patterns could contribute to the academic development and innovation of the pharmaceutical industry, serving as valu- able data for establishing future research collaboration strategies. Furthermore, in addition to the research collabo- ration outcomes observed in this study, such as co-author- ship counts, a more concrete measure of success could be introduced through quantitative analysis of how the collab- oration networks formed by platform models can generate higher performance. For example, analyzing the correlation between ‘impact factor’ and the size of the collaboration network or the persistence of research collaborations could provide more detailed insights into the actual impact of pharmaceutical industry platform models on research out- comes. Future studies exploring this area are anticipated to provide insights into the tangible influence of these plat- forms.
CONFLICTS OF INTEREST
No potential conflict of interest relevant to this article was reported.
ORCID
Yang-Woo Kim: https://orcid.org/0009-0008-4124-0146