Charting the Course for Next-Generation Data Centers: Core Direction for Hospital Smart Transformation Confirmed!

On June 20, the 2019 China Health Information Technology / Health and Medical Big Data Application Exchange Conference was held as scheduled in Xi’an.

At the conference, 48 regions and 101 hospitals passed the 2018 National Health and Medical Information Interconnectivity Standardization Maturity Assessment. A product named “YiduEywa” was launched, regarded as ushering in a new era of smart healthcare platforms. The book Framework, Technology, and Implementation of Hospital Data Governance, released by the Statistical Information Center of the National Health Commission, has been hailed as the “bible” for data governance in regional health informatization initiatives.

An awards ceremony for evaluations, a data center, and a book—at first glance, these three elements may seem largely unrelated. In fact, however, they collectively constitute the theme of this conference: Building Next-Generation Data Centers.

Currently, assessments of interoperability maturity and electronic medical record (EMR) grading in healthcare institutions are being vigorously carried out. Hospital development is also undergoing a transformation from “informatization” to “intelligentization.” The foundation of “intelligentization” is data; for hospitals to truly open the door to the next stage, they must comprehensively upgrade their data management capabilities. Against this backdrop, the construction of next-generation data centers has become a focal point of attention for all stakeholders.

Regarding the construction of next-generation data centers, Xu Xiangdong, Deputy Director of the Information Technology Division at the Statistical Information Center of the National Health Commission, identified five key necessities:

1. The need for continuous in-depth development within hospitals. After completing the construction of informatization or process optimization, hospitals are now shifting their focus more towards business operations or delving deeper into business management. These changes have placed very high demands on data management in hospitals.

2. Public-Benefit Management Outside the Hospital. The hospital’s information platform should no longer be a relatively closed system; instead, it needs to involve multiple stakeholders to enhance the sense of gain among the general public.

3. The establishment of medical consortia, county-level medical communities, and healthcare alliances has broken down the walls surrounding hospitals.

4. In the areas of telemedicine and internet hospitals, the Statistical Information Center of the National Health Commission is cooperating in conducting a survey on the implementation of “Internet+” medical services.

5. In terms of the in-depth development of the discipline, previous informatization efforts focused primarily on management and processes, whereas current informatization initiatives have shifted toward enhancing business operations and implementing data governance. This shift imposes higher requirements on the new generation of hospital data centers.

She stated that in the future, quality management and operational management in hospitals will converge into two main threads, implementing foundational and process-oriented management for public hospitals from the perspectives of medical care and operations, respectively. The transition from informatization to digitalization in hospitals signifies a shift where the former aims to improve efficiency, while the latter places greater emphasis on building comprehensive capabilities. Therefore, the design of next-generation data centers incorporates a “three-horizontal, three-vertical” framework.

The three vertical pillars are the service chain, the business chain, and the data chain. Hospital operations must be integrated through these three major chains.

Among these, the service chain is an integrated, convenience-oriented chain, namely, convenient medical services. The hospital’s outpatient, medical technology, and follow-up processes must be interconnected via the Internet. The business chain emphasizes integration and collaboration, as hospital information systems were historically built in silos. The data chain signifies support for data sharing and interoperability.

Currently, the Statistical Information Center of the National Health Commission has begun to outline the approach for constructing next-generation hospital data centers. During discussions, some experts argued that the data center refers specifically to the Clinical Data Repository (CDR), while others contended that it simply denotes the server room.

According to Xu Xiangdong, the new-generation data center is structured into three tiers, as categorized by the Statistical Information Center of the National Health Commission:

1. At the foundational level, which pertains to cloud data center construction, this falls under the IaaS layer.

2. Support Layer, including concepts such as business middle platform and data middle platform. “We hope that the data center can provide stable support for the development of the hospital’s three chains through the integration of hardware and basic software. This is also the thinking behind our design of the entire data center,” said Xu Xiangdong.

3. Application Layer: This layer focuses more on the integration of existing applications and the development of new ones. Hospitals are increasingly adopting scenarios for data analytics and clinical data applications. In this regard, the Statistical Information Center of the National Health Commission has proposed certain requirements related to these applications.

Therefore, the positioning of next-generation data centers can be categorized into three aspects. First, it serves as a resource hub; second, as a service platform; and third, as a management center. It supports various effective applications, enabling their rapid deployment. Under this definition, next-generation data centers require an open and robust infrastructure.

Xu Xiangdong stated, “On one hand, we emphasize high availability, high standards, and low risk for core hospital systems. In areas such as ‘Internet+’ initiatives for public convenience and benefit, medical consortia and medical communities, and the development of internet hospitals, we prioritize high performance, high concurrency, and elasticity in capacity and performance. This is because these services often experience significant surges in data and computational resource demands over short periods due to specific campaigns or policy changes. For artificial intelligence and big data, our focus is on security and massive connectivity.”

In addition, the new generation of hospital data centers must also possess three key attributes in terms of technology:

1. Robustness of the underlying structure.

2. Cloud-based infrastructure. In the past, hospitals implemented virtualization, which gave rise to data resource pools. A data resource pool enables unified management of data generated by various business systems through centralized data resources.

3. Data Middle Platform. This refers to the provision of data and services through support layers such as the data service layer.

Xu Xiangdong stated, “The development concepts of data middle platforms, business middle platforms, and technology middle platforms should be widely adopted by large hospitals. In the construction practices of national platforms, a layer similar to a data middle platform has also been established to achieve data governance and service division.”

“Whether it is relying on big data for industry management decision-making analysis, or focusing on scientific research development and clinical practice through artificial intelligence technology, the most critical factor comes from data.” In Xu Xiangdong’s view, “the proposal of data governance and the construction of a new generation of data centers is timely.”

Wang Peng, Director of the Center for Medical Big Data and Artificial Intelligence at Southwest Hospital, stated, “In recent years, with the comprehensive implementation of the national big data strategy, relevant guidelines and policies from health industry authorities have been continuously introduced. A surge in research and discussion on big data has also emerged at various medical informatics conferences across China. From hospital management to clinical research, while application demands are growing rapidly, higher requirements are being placed on data quality and data security. As the department responsible for the hospital’s informatization construction and data management, we have reached a critical stage where data governance is urgently needed.”

Wang Peng believes that, regarding data quality issues in the construction of a new generation of hospital data centers, the following aspects should be prioritized:

First, at the data source level, by strengthening the application of data standards and implementing technical measures in software development, we can ensure that data standards are standardized, less prone to errors, and capable of automatic error correction, thereby fundamentally improving data quality.

Second, at the process level of data management, technologies such as big data and artificial intelligence should be integrated to provide healthcare professionals with more intelligent alerts, thereby minimizing data quality issues caused by human oversight.

Third, at the data post-processing level, technologies such as natural language processing should be employed to standardize non-standard data and normalize irregular data, thereby enabling more effective retrieval, analysis, and utilization of these data.

Generate high-quality data by managing the entire chain from source and processing to post-processing, laying a solid foundation for various applications.

Director Wang also called on hospitals to remain focused amidst the current clamor, genuinely improve data quality, and elevate data governance to a strategic level in hospital development.

Like Southwest Hospital, Shandong Provincial Hospital is also exploring the development of a new-generation data center. In 2017, Shandong Provincial Hospital partnered with Yidu Cloud to begin building an intelligent medical data platform. In the same year, the hospital successfully passed the Level 4 Class A assessment for standardization maturity in information interoperability, conducted by the National Health Commission.

Regarding next-generation data centers, Wang Yongjie, Director of the Information Network Management Office at Shandong Provincial Hospital, believes that a passage by the modern scholar Wang Guowei can be used to describe the current landscape: “Last night, the west wind withered the green trees; alone I ascended the high tower, gazing down the road to the horizon—this is the first stage. My clothes grow looser, yet I regret nothing; for her sake, I waste away in longing—this is the second stage. I have searched for her a thousand times in the crowd; suddenly, I look back, and there she is, where the lights are dim—this is the third stage.” In his view, China’s medical big data infrastructure is currently in the second stage.

In terms of data application, in 2018, Shandong Provincial Hospital held its 8th Party Congress, placing scientific research in a highly important position. It was established that the top priority for the hospital in 2019 would be scientific research. In response, Wang Yongjie and his team began the construction of the second phase of the specialized disease database.

Compared with the previous two experts, Jiao Yun, Deputy Director of the Network Information Center at Zhongda Hospital Affiliated to Southeast University, placed greater emphasis on the development of artificial intelligence applications based on big data.

According to him, Zhongda Hospital Southeast University began its artificial intelligence research in 2013. At that time, the hospital was undertaking a national-level 973 Program project, which primarily focused on stroke and aimed to predict the efficacy of treatment regimens for patients by integrating data from imaging, clinical records, and other sources.

After five years of research, the hospital developed a relevant artificial intelligence model, and the project was successfully completed. Subsequently, the hospital applied the AI model to scenarios such as medical imaging and hospital management.

Taking medical record quality control as an example, Zhongda Hospital Southeast University primarily evaluates medical records based on the standard of the medical record face sheet. However, due to a shortage of quality control personnel and low efficiency, each staff member can only perform high-quality quality control on approximately 5–8 medical records per day on average. Given that the hospital generates nearly 1,000 medical records daily, the existing quality control team struggles to complete the workload.

In response, the hospital introduced artificial intelligence methods to construct a knowledge graph and conducted a series of big data analyses to learn from historical medical records. Following this training phase, the efficiency of medical record quality control has significantly improved; the model can now flag errors made by physicians, thereby facilitating detection by quality assurance personnel.

Currently, the efficiency of a single quality control staff member in hospitals can reach 100 medical records per person per day, with an accuracy rate of 85%. In terms of coverage of medical record defects, AI-assisted quality control personnel can achieve over 83%, whereas purely manual methods can only reach 71%.

Director Jiao believes that breakthroughs in the application of artificial intelligence in next-generation data centers may begin with two scenarios. The first is early disease screening centered on medical imaging, supplemented by various clinical data, as well as the formulation of clinical protocols, where significant breakthroughs are expected to emerge. The second involves the numerous repetitive tasks across various quality control and management processes within hospitals, which can be addressed through artificial intelligence.

At the 2019 China Health Information Technology/Healthcare Big Data Application Conference, Yidu Cloud, representing the corporate sector, unveiled its next-generation data center, “YiduEywa.” Reportedly, “YiduEywa” is a newly developed proprietary product that leverages six years of in-depth industry experience, following extensive research by Yidu Cloud into national and public demands for comprehensive health services.

Yidu Cloud CEO Sun Zhe Unveils New-Generation Data Center “YiduEywa” at Conference Opening Ceremony

At the opening ceremony, Sun Zhe, Co-founder and CEO of Yidu Cloud, stated, “Just as the name ‘YiduEywa’ suggests, in the movie Avatar, there is a Tree of Souls called ‘Eywa’ that nurtures all life. It connects with all living beings through an external neural consciousness, gathering wisdom and strength to promote balance and prosperity for all creatures and the ecosystem. We believe that the ‘YiduEywa’ platform serves hospitals in a similar way, continuously empowering them to optimize the processing of daily generated data, unlock data value, foster the integrated development of the industry and its ecosystem, and jointly create a healthier future.”

“YiduEywa” centers on data governance and services, enabling high-quality governance of large-scale, multi-source, heterogeneous medical data within hospitals. It also helps hospitals build an open platform for technical service capabilities, thereby establishing a foundation for value conversion, flexibly supporting multi-scenario applications, and empowering clinical practice, healthcare management, and one-stop scientific research. Characterized by high scalability, high availability, and high quality, the product leverages flexible medical technology service capabilities to facilitate the development and deployment of hospital-based AI applications. It comprehensively empowers clinical applications, scientific research, online teaching, healthcare management, patient management, and one-stop development for multiple applications, thus forming a “growth-oriented” new-generation data center that accelerates the construction of smart hospitals.

“‘YiduEywa’ represents not only a new platform requirement for each hospital’s own informatization development, but also the foundation for future regional connectivity and collaborative operations,” said Sun Zhe.

Liu Tingting, Product Director of the Healthcare Division at Yidu Cloud, told reporters that compared with previous data products, “YiduEywa” accelerates clinical treatment and scientific research through applications such as patient timelines, research centers, and exploratory discovery. It not only compiles comprehensive lifecycle medical information for patients, assisting hospitals in building real-world disease domain models, but also serves as an innovation enablement center, empowering the implementation and development of artificial intelligence technologies in hospitals.

Nowadays, the application of big data and artificial intelligence technologies is widespread in Chinese hospitals across clinical, research, management, and teaching scenarios. However, the sheer variety of these applications often leaves medical personnel feeling overwhelmed by their complexity.

Therefore, hospitals urgently need to “discard the crude and retain the refined” by establishing a new-generation data center within the institution that features data governance capabilities and flexible, scalable service capacities. This initiative should build unified data management and service capabilities oriented toward practical applications, continuously promoting development through utilization in an iterative cycle, thereby truly forming the “engine” of hospital informatization and facilitating the construction of smart hospitals.

Based on the perspectives of various experts, it is evident that next-generation data centers will play a crucial role, whether in the Interconnectivity Standardization Maturity Assessment or in the transformation from “informatization” to “intelligentization.” It is believed that this heavyweight product will soon become standard equipment for every smart hospital.