Next-Generation Hospital Information Platforms: Addressing Core Pain Points – Insights from Experts at a Top-Tier Hospital, Alibaba Cloud, and MedElink

Zou Yue, Chief Architect at B-Soft with 20 years of industry experience, once offered a vivid metaphor for the current state of healthcare informatization in China.

He believes that current healthcare IT systems are akin to a bridge. The patients are those walking and running across it, while big data, artificial intelligence, and the Internet of Things serve as the decorative elements, such as stone lions and intricate patterns, adorning the bridge. Hospital directors and executives tend to focus on these flashy features, often overlooking the foundational infrastructure beneath. Underneath the bridge, staff from the IT departments and software vendors are struggling to keep the structure operational through constant patchwork repairs. In reality, the system is already overwhelmed. This has led to widespread complaints from IT personnel against software vendors, who themselves are enduring significant hardships, trapping both parties in a vicious cycle.Creating a formula-level information system with robust scalability will be the key to solving the problem.

Recently, at the “2018 World Forum on Medical Technology” hosted by VCBeat, several heavyweight guests also expressed views similar to those of Zou Yue.

Shen Chongde, Vice President of Wuxi People's Hospital; Chen Lie, Medical Architect at Alibaba Cloud; and Zhang Lizhong, Chairman of Winning Health Technology Group, representing hospitals, internet companies, and healthcare IT enterprises, respectively, it is judged that the next-generation hospital information platform may possess the following characteristics:

1. Centered on electronic medical records, it leverages a knowledge base to deliver multidimensional clinical decision support, enabling progressive artificial intelligence applications ranging from basic to advanced levels; it supports management decision-making focused on the operational aspects of healthcare (personnel, finance, and materials), while facilitating interoperability with regional health information platforms;

2. Leverage cloud infrastructure to build a platform that maps and outlines hospital operations, extract core capabilities, and achieve unification at the data level.

3. Multidimensional monitoring can be provided to medical personnel through various sensors, such as temperature and humidity sensors, body temperature sensors, or vital sign-monitoring mattresses, enabling doctors and nurses to implement fully closed-loop management.

The following content summarizes the core viewpoints of three experts, as compiled by VCBeat (WeChat Official Account: vcbeat):

Big data and artificial intelligence require a centralized platform for data acquisition, cleaning, standardization, and inter-system integration, thereby enabling a wide range of applications. To this end, hospitals need to establish next-generation information platforms.

Nowadays, many companies label their products as “platforms.” In reality, hospitals lack a clear understanding of what constitutes an information platform, making it difficult for them to differentiate among these offerings. At times, they even refer to an electronic medical record (EMR) viewer as an information platform.

Therefore, I would like to explore the following questions: Why do hospitals build information platforms? What do these platforms comprise? And how should we proceed with their development?

What is the definition of a digital hospital? It should be characterized by being paperless, wireless, and film-free, achieving comprehensive application coverage of business processes, and enabling holistic digital simulation and workflow of hospital operations. At the stage of developing smart hospitals, in addition to these "three no's," we should also focus on closed-loop management, process optimization, and intelligent navigation.

At this stage, hospitals require new information platforms that integrate the Internet of Things (IoT), “Internet Plus,” and, of course, artificial intelligence. Smart hospitals should be the product of integrating IoT, the Internet, and information platforms.

So, why do hospitals need to build an information platform?

Taking Wuxi People's Hospital as an example, its information system comprises 250 application modules. When the number of modules grows to such a scale, or when hospitals face multi-campus operations, how should they manage their systems?

By establishing a core clinical laboratory center, pathology center, and diagnostic imaging center, hospitals can achieve multi-campus sharing of clinical and operational resources. Specimens for laboratory testing can be transported by courier services, while imaging data and laboratory reports can be circulated digitally to enable virtual diagnostics. This approach significantly saves human and material resources, reduces costs, and ensures homogeneous care across multiple hospital campuses through unified management of personnel, finances, materials, and patients. Such integration requires the development of an information platform to provide technical support.

Nowadays, hospitals often build information platforms for two reasons: one is the requirement of the National Health Commission, and the other is the need for system upgrades.

During the process of upgrading and adjusting hospital information systems, hospitals have encountered numerous unscrupulous software vendors. To achieve interoperability, hospitals often need to open interfaces between systems; however, software companies frequently charge exorbitant fees for this access. As the number of hospital systems increases, so does the complexity and volume of interfaces. In this context, a unified information platform is required to standardize interfaces and services, decouple system dependencies, and enable interoperability based on the platform.

During the implementation of big data initiatives, hospitals have found that all collected data is disorganized and difficult to utilize. This issue stems from inconsistencies in code tables, dictionary tables, and terminology lists across different systems, as well as varying standards among various departments. Hospitals require an information platform capable of providing standardized code tables, dictionaries, and terminology lists, enabling data conversion through the platform to achieve standardized data utilization.

The construction goal of the new-generation information platform should center on electronic medical records to drive the development of hospital data centers. Meanwhile, it leverages knowledge bases to deliver multidimensional clinical decision support, enabling progressive artificial intelligence applications from basic to advanced levels. It aims to satisfy management decision-support needs focused on the operational aspects of healthcare—namely personnel, finances, and materials—and to ensure interoperability with regional health information platforms.

The information platform is a productized solution that should encompass various applications, such as master data management and interaction engines.

For information platforms, some companies split them into multiple products, while others offer a single product. Regardless of how it is presented, it must consist of a suite of tools with an ESB-based core oriented toward SOA, featuring business process management capabilities that support automated business process orchestration in hospitals and AI integration into workflows. Furthermore, it enables robust data management and utilization through event-driven architecture.

From the hospital’s perspective, we aim to leverage an information platform to achieve unified registration, indexing, portal access, interaction, communication, and data management and utilization. This includes implementing data integration and master data management, standardizing registration services along with associated code tables, dictionaries, and terminologies, while also advancing data resource management and the construction of a Clinical Data Repository (CDR). For instance, from the clinician’s viewpoint, the system should provide a comprehensive 360-degree patient view, displaying vital signs, examination reports, and medical records from previous visits.

Portal integration is achieved through unified registration, unified indexing, and related tools. This enables single sign-on (SSO) across all systems using a single entry of employee ID and password, while delivering personalized user interfaces.

In terms of application integration, we aim to implement an SOA-oriented architecture based on an Enterprise Service Bus (ESB), which supports mainstream data formats and various adapters, enables intelligent and orchestratable routing, and ensures real-time message traceability.

The visualization of closed-loop hospital management relies on the in-depth utilization of platform data and systems, including the construction of knowledge bases and the application of artificial intelligence and big data based on these knowledge bases. Examples include clinical early warning alerts, clinical auxiliary diagnostic navigation, and treatment plan navigation. This involves the extraction of information from electronic medical records (EMRs) or other heterogeneous clinical systems by the information platform. To achieve this, the information platform must establish backend databases and standardized medical terminology libraries, enabling data standardization through mapping techniques, ultimately providing clinical decision support for diagnosis and treatment.

We aim to achieve loose coupling through an information platform, facilitate interoperability between internal and external hospital systems via standardized interface processing, enable data accumulation and reuse, implement cross-system management and control within hospitals, and apply big data to scientific research and teaching.

Wuxi People’s Hospital has also undertaken a series of initiatives in this area, establishing a dedicated Artificial Intelligence Research Institute to conduct AI research in medical imaging, speech recognition, and intelligent navigation for electronic medical records. Its initial goal in big data applications is to establish a national lung transplantation big data center in China.

Information platform construction is a systematic project that requires comprehensive planning and phased implementation. For most hospitals, achieving Level 4-B interoperability is a suitable choice, which must be realized through a hybrid model of “platform + systems.” In contrast, attaining Level 4-A or Level 5-B requires integrating a large number of hospital applications into the platform, resulting in substantial transformation efforts. It takes at least three years for a platform to mature; therefore, hospitals should prioritize easier tasks before tackling more difficult ones, with standards established first. The process should begin with data standardization and the construction of a Clinical Data Repository (CDR), advancing step by step.

The true value of cloud computing lies not in migrating offline operations to the internet, but in providing a flexible and adaptable architecture, even when the cloud is deployed as a private cloud within a hospital.

Once sufficient data has been collected, the challenge shifts to leveraging it in a more valuable and rational manner, marking the stage of smart hospital development. Based on its assessment of hospitals’ developmental stages and its vision for intra-hospital informatization, Alibaba has proposed an overarching goal for future hospital construction: supporting the expansion and innovation of the entire hospital information system through a cloud-based architecture.

The greatest value of the cloud lies in its scalability and availability. Building in-hospital information platforms based on cloud architecture can enhance efficiency, support scientific research, facilitate clinical management and applications, and unlock the value of hospital data.

Second is to achieve intra-hospital collaboration, including interpersonal collaboration, IoT-enabled human-object collaboration, and collaboration between in-hospital and out-of-hospital settings.

The key distinction between cloud-based architecture and traditional architecture lies in placing the platform at the infrastructure layer, where all support required for changes in hospital operations is implemented at the bottom. However, at present, changes in hospital informatization services often occur at the application layer.

Traditional platforms extract data from various systems and then establish an upper-layer platform to aggregate and integrate the data, using messaging mechanisms for data transmission. While this is one approach, Alibaba believes that by building the platform at the infrastructure layer, many business processes can be clearly defined and mapped out, extracting foundational capabilities and achieving unification at the data level.

The physical architecture of the cloud delivers elasticity, scalability, availability, and security. Above it lies the middleware layer, followed by the IT infrastructure layer, which includes databases, application services, distributed services, and more.

This is a model of consulting services. Alibaba’s Tmall Supermarket and Hema Fresh, among others, are all supported by architectures based on middleware and data middle platforms.

For cloud-based Hospital Information System (HIS) solutions targeting large hospitals (tertiary hospitals), Alibaba adopts a private cloud model. This private cloud is not necessarily hosted on the public internet; it can also be deployed locally within the hospital premises. Furthermore, for numerous secondary hospitals and healthcare institutions with relatively limited IT budgets, Alibaba offers a Virtual Private Cloud (VPC) based on public cloud infrastructure, providing comprehensive architectural services to these facilities.

Electronic medical records (EMRs) serve as the central hub of a hospital’s information system, with patients as the primary beneficiaries. All patient-related information converges within the EMR, making this dataset the most valuable.

Previously, hospital electronic medical record (EMR) management relied on manual spot checks. Even where some automation was in place, it primarily focused on verifying completeness and documentation compliance. Currently, Alibaba Cloud’s EMR quality inspection system enables content-based quality control for electronic medical records, thereby ensuring data quality at the source.

Through intelligent automation, the burden of medical record quality control on hospitals can be alleviated, as well as that on relevant departments. With foundational data quality, AI can also assist physicians in certain data-driven tasks, such as imaging diagnosis. In this regard, Alibaba’s diagnostic solutions for medical imaging analysis—including tuberculosis, joint, and liver conditions—have become relatively mature.

Furthermore, regarding the standardization and normalization of clinical diagnosis and treatment, Alibaba addresses these issues through knowledge graphs to unlock the value of big data. Leveraging natural language processing and semantic extraction to ultimately construct knowledge graphs, Alibaba and Academician Ning Guang’s Research Laboratory at Ruijin Hospital have conducted research on intelligent diabetes diagnosis and treatment protocols. The system is now capable of significantly reducing physicians’ workload.

In supporting intelligent management decision-making, issues such as hospital costs, patient expenses for medical services, physician income, hospital expenditures and loss-making areas, as well as the decomposition of strategic objectives and allocation of diagnostic and treatment resources, can be further evaluated through a data cockpit.

Alibaba’s one-stop big data development platform meets the flexible and evolving needs of hospitals in data analysis. It includes metric-based data integration, real-time monitoring and operations maintenance, and real-time analytics, all supported by a comprehensive infrastructure that enables visual data presentation.

To tap into healthcare big data and drive fundamental changes in healthcare informatization, it is essential to thoroughly measure the depth of healthcare practices, enabling enterprises to truly deliver comprehensive, integrated healthcare solutions.

Currently, the application of artificial intelligence remains superficial, particularly in imaging services, which are still at a relatively shallow level. The core technology and foundation of artificial intelligence is big data; so, what exactly constitutes medical big data?

At present, 80%–90% of AI companies are engaged in medical image interpretation. The basic approach involves using machine learning to identify abnormal images and then comparing them with expert experience. However, from the perspective of AI’s core value, even the most skilled physicians can misinterpret imaging studies.

If the full-course big data of patients is correlated with imaging data for associative learning, the resulting artificial intelligence capabilities will undoubtedly far surpass those of human experts. The essence of medical big data lies in integrating seemingly unrelated information into a single platform, which is precisely what MedInfo Technology does.

In addition to data integration, Yihui Technology is also engaged in IoT integration, creating a unified open platform to aggregate data from all IoT devices.

To ensure compatibility with Wi-Fi, Bluetooth, 4G, and other networks, YiHui Technology has developed a highly integrated platform that converges various sensor networks to address data fusion challenges. Building on this foundation, YiHui Technology has constructed an Internet of Things (IoT) ecosystem centered on healthcare big data. This integration enables the provision of intelligent applications for hospitals, along with multi-dimensional visualizations and IoT-based big data services.

The purpose of the common open platform is to serve clinical practice. In hospitals, the entire solution can be integrated into every medical management process, whether it involves waste management, linen services, or the management of high- and low-value consumables.

In the past, hospitals frequently misattributed patient monitor data to the wrong patients. By leveraging IoT device positioning to pinpoint specific bed locations and integrating operational status and ambient temperature data through IoT methodologies, such errors can be eliminated. This is where the value of IoT data becomes evident.

Yihui Technology’s common open platform enables multi-dimensional management for healthcare professionals by leveraging various sensors, such as temperature and humidity sensors, body temperature sensors, and vital-sign-monitoring mattresses, thereby facilitating a fully closed-loop management system for doctors and nurses.

Technologies such as the Internet of Things (IoT), artificial intelligence (AI), big data, and blockchain form an interdependent, organic whole. Only through the integration of these technologies can hospitals truly realize their practical applications.

The integration of the Internet of Things (IoT) enables closed-loop management throughout the diagnostic and treatment process. Hospitals can interconnect all stages of care, including both in-hospital and out-of-hospital phases, via IoT infrastructure. This allows IoT-generated data to be leveraged through machine learning to provide advanced clinical decision support within diagnostic and therapeutic workflows. Ultimately, this facilitates intelligent quality control, where quality metrics are used to evaluate treatment efficacy and enhance the precision of physicians’ interventions.

Within such a framework, hospital information systems continuously evolve into an ecosystem. Clinical care is becoming increasingly precise and intelligent, enhancing physicians’ mastery over disease management and enabling the integration of research, treatment, and quality control for specific diseases.

To build such an ecosystem, YiHui Technology seeks to collaborate with enterprises possessing core technologies in the field, thereby playing a more significant role and delivering greater value within the ecological industry chain.