Guangdong Provincial People’s Hospital: From Information Silos to an Integrated Smart Healthcare Platform
Recently, the 2016 China Hospital Information Network Conference was held at the Zhuhai International Convention and Exhibition Center. The significant expansion of the healthcare informatics market, driven by the overarching trend of tiered diagnosis and treatment, emerged as the most prominent topic. From tertiary hospitals to primary care institutions, there is an urgent need to optimize and upgrade healthcare informatics infrastructure. Digital health companies specializing in Clinical Data Repositories (CDR), Picture Archiving and Communication Systems (PACS), mobile health applications, regional health information platforms, telemedicine, medical cloud services, and healthcare big data are poised to enter a period of rapid expansion.
Available data indicate that the current overall scale of China’s healthcare informatization industry remains below RMB 40 billion. However, it is estimated that, driven by national policies and increased investment spurred by the momentum of smart healthcare development, the entire healthcare informatization sector will maintain an annual growth rate of 30%–40% over the next three to five years—nearly double the average growth rate of approximately 20% observed in recent years.
In terms of policy, for example, in March 2015, the State Council issued the “Notice on Printing and Distributing the Outline of the National Medical and Health Service System Plan (2015–2020),” which proposed to “launch the Healthy China Cloud Service Initiative, actively apply new technologies such as mobile internet, Internet of Things, cloud computing, and wearable devices, promote health information services and smart medical services that benefit the entire population, and advance the application of health big data. By 2020, achieve basic nationwide coverage of the three major databases—population information, electronic health records, and electronic medical records—with dynamic updates of information.”
In particular, regional health informatization and public health informatization have emerged as two new growth hotspots. In accordance with the “46312 Plan” issued by the National Health and Family Planning Commission (NHFPC) and a series of supportive policies, China will establish a four-tier health data platform at the national, provincial, municipal, and county levels, with an average budget of approximately RMB 30 million for municipal-level institutions.
A research report from BOC International points out that in 2014, only 8.77% of hospitals had implemented regional health information systems. It forecasts that the sales scale of the regional medical informatics market could reach RMB 500 billion, with annual operation and maintenance service fees amounting to RMB 40 billion, and derived businesses exceeding RMB 1 trillion.
So, in the process of moving towards a trillion-level market, what are the attitudes and considerations of the core of the healthcare ecosystem—the hospitals?
On May 27, at one of the sub-forums of the 2016 China Hospital Information Network Conference, titled “The Future Is Here: The Era of Integrated Smart Healthcare Informatics” and hosted by Peking University Medical Information Technology Co., Ltd. (hereinafter referred to as “PKU MedInfo”), Professor Liang Changhong, Director of the Information Management Department at Guangdong Provincial People’s Hospital, accepted an exclusive interview with VCBeat, sharing his reflections after nearly one year of overseeing the hospital’s informatization initiatives.
Liang Changhong
Notably, Liang Changhong also serves as the Director of the Department of Medical Imaging and Radiology at Guangdong Provincial People's Hospital, remaining an active clinician on the front lines. From his perspective, advancing healthcare informatization benefits from a dual viewpoint: that of both a hospital administrator and an end-user (medical staff).
Founded in 1946, Guangdong Provincial People's Hospital is the largest general hospital in Guangdong Province and ranks among the top nationwide in terms of scale and comprehensive strength. According to its official website, the hospital has a building area of nearly 230,000 square meters, with 5,226 employees, including 3,926 healthcare professionals and 685 senior-title staff. It currently has 2,729 inpatient beds, serves 101,100 discharged patients annually, and performs 105,000 surgeries per year. The hospital operates six outpatient departments, handling approximately 4.39 million outpatient visits in 2013.
Guangdong Provincial People’s Hospital began its informatization journey in 1993. According to Liang Changhong, the period from then until 2006 constituted the first phase, which was primarily characterized by PC-based informatization. During this time, various modules operated in isolation; while some data were generated to meet certain clinical and research needs, numerous information silos also emerged.
The hospital began constructing its integrated platform in 2011, a process that spanned four years until the platform went live in August 2015. Liang Changhong therefore designated 2015 as the “Year One of Integration” for Guangdong Provincial People’s Hospital. In fact, it was also in 2015 that Liang Changhong officially assumed responsibility for advancing the hospital’s medical informatization. He regarded the launch of the integrated platform as the beginning of the third phase of the hospital’s informatization journey. In his view, the development of the integrated platform represented a fundamental transformation compared with previous systems.
For example, in the absence of an integrated platform, a messaging mechanism could not be established, making it impossible to alert doctors and nurses to required actions; however, with the implementation of an integrated platform, previously isolated tasks can now be consolidated.
Liang Changhong also frankly stated that integration has both advantages and disadvantages. As the number of interactive relationships increases, the logical relationships become more complex, which may lead to a short-term decline in efficiency; however, this can be optimized. Meanwhile, there is a significant improvement in clinical application experience, quality control, and system scalability.
Nearly a year after the launch of the new integrated system, which is set to be rolled out to more sectors, Liang Changhong stated that the overall process has been relatively smooth. However, he also shared insights on how to address the challenges encountered.
For instance, regarding issues within the health IT management team, how should new requirements proposed by clinicians be evaluated?
In Liang Changhong’s view, effective management of emerging issues is crucial for the smooth deployment and operation of the system. He has defined four principles to address this: First, if existing knowledge clearly indicates that a task is mandatory, suppliers must be required to implement it. Second, if existing knowledge clearly shows that a task is unnecessary, one should not directly refuse physicians’ requests; instead, handle them tactfully to demonstrate respect for clinicians. Third, if uncertainty persists, seek input from multiple department directors before making a decision. Fourth, for particularly challenging issues, convene meetings to discuss and resolve them.
Furthermore, a real-time communication system should be established. Currently, WeChat is used to facilitate discussions among relevant stakeholders in a dedicated group. As young professionals are particularly adept at leveraging information, he has specifically mobilized the hospital’s Youth Committee to raise questions and articulate needs.
Transitioning from isolated information silos to an integrated, unified healthcare data platform requires a process of integration and implementation. How, then, should one choose among various models, such as exclusive vendor (e.g., Shengjing Hospital with Neusoft), in-house development (e.g., The First Affiliated Hospital of Wenzhou Medical University), multi-vendor, and lead vendor approaches?
Guangdong Provincial People’s Hospital has adopted PKU Healthcare IT as its primary vendor, a model reportedly representing a new approach that PKU Healthcare IT is piloting in response to the current state of hospital informatization.
Liang Changhong stated that, under the primary supplier model, more proactive communication and a service-oriented attitude were key reasons for the timely launch of the new system.
He explained that, given the hospital’s own circumstances, it lacked sufficient development capabilities. Furthermore, hospital data required unified management. The so-called “primary vendor” refers to a single supplier responsible for operating the integrated platform and the key information systems that run on it—such as the electronic medical record (EMR) system, pharmacy management system, clinical pathway system, and nursing system—which generate the most critical clinical data.
For instance, if a hospital’s electronic medical record (EMR) system is upgraded, the nursing information system must be simultaneously upgraded and adapted. If these two systems are provided by different vendors, achieving synchronized upgrades becomes extremely challenging. Thus, management considerations and data security are two key factors in opting for a primary vendor model.
Liang Changhong stated that, compared with other models, the primary supplier model is more replicable.
At present, various speculations about the potential of big data in healthcare are rampant. The gradual advancement of healthcare informatization has made it seem as though the steps leading to the big data era are becoming shorter. However, Liang Changhong’s reflections on the current state of medical big data development reflect the caution typical of a clinician.
“Current issues involve the hype around ‘Internet Plus’ and big data, which have failed to form a closed loop. This is not an IT failure but a management one, resulting in incomplete data and creating data gaps or ‘cliff-edge’ data. For instance, a patient may be hospitalized at the Provincial People’s Hospital but receive no follow-up care, making subsequent analysis impossible. To illustrate, consider a lung cancer patient who is hospitalized and then discharged. When did they die? It is unknown. Did they die in a car accident, by drowning, or from lung cancer? It is unclear. However, these distinctions significantly impact the analysis of treatment efficacy. Therefore, our current data quality is generally low,” stated Liang Changhong.
Overall, while Liang Changhong recognizes the value of intelligent knowledge systems, he believes it is still premature to engage in big data mining. First, there is a lack of true big data; although vast amounts of digital information exist, this does not equate to big data. Second, there is a shortage of experts who truly understand the data. Professional expertise is required—not only to access the data but also to know how to utilize it effectively. This is beyond the capability of IT professionals alone; it requires domain specialists, such as radiologists for radiology and pulmonologists or oncologists specializing in lung cancer. While there have been successful cases of big data mining in non-healthcare industries, the healthcare sector contains an excessive amount of unstructured data, driven by human heterogeneity. Medical practice is difficult to quantify using rigid rules, whereas IT experts tend to over-standardize complex phenomena into rule-based frameworks.
Liang Changhong stated that clinicians often have insights but lack the technical means, whereas data miners possess the tools but lack clinical insights; only by combining the two can we achieve both insight and capability. He also shared a previous project on imaging big data, which was based on CT image data mining. For instance, if a patient’s CT scan indicates colon cancer with surrounding lymphadenopathy, the enlargement could imply either metastatic spread of cancer cells or benign inflammation rather than a tumor. These two scenarios require completely different treatment approaches.
What did Professor Liang Changhong’s team accomplish?
By leveraging CT imaging data that extends beyond macroscopic size and morphology, tumor heterogeneity is quantified using computational methods for cluster analysis and the extraction of radiomic features. These features are integrated with clinical parameters, such as carcinoembryonic antigen (CEA) levels and the short-axis diameter of enlarged lymph nodes, to predict whether a lymph node is benign or malignant. This data-driven model improves prediction accuracy by more than 20%, while sparing patients from additional suffering or surgical interventions. Future integration of this model into Laboratory Information Systems (LIS) and Picture Archiving and Communication Systems (PACS) would significantly enhance diagnostic precision and patient experience.
Furthermore, according to Liang Changhong, he has also applied for a major big data project in medical imaging and was scheduled to travel to Beijing on the night of the interview to present his defense for this significant initiative under the Ministry of Science and Technology. Nevertheless, he emphasized that current big data mining is still limited to prediction and exploration, and it remains far too early for it to truly replace physicians.
From the hospital’s perspective, it is essential to build an internal information technology platform tailored to core clinical operations. At the same time, hospitals face the challenge of selecting among numerous third-party medical data companies that are eager to establish collaborative partnerships.
Liang Changhong first offered a cautionary note, emphasizing that the utilization of in-hospital data must be approached with extreme prudence to ensure compliance with national regulations. This involves a range of issues, including the security of the National Health and Family Planning Commission’s medical and health database. He also advised exercising caution when using free services, a point previously echoed by other hospital presidents.
At the end of the interview, Liang Changhong also shared his reflections from nearly one year in office.
First, ideals are often idealistic, while reality is stark. For instance, due to policy factors, many initiatives fail to materialize not because of technical issues.
Second, while technology can resolve process-related issues, IT personnel still need to deepen their understanding of business processes;
Third, when dealing with healthcare professionals—a group subjected to immense work intensity and pressure—it is essential to adopt appropriate communication strategies.
For instance, when Guangdong Provincial People’s Hospital implemented structured electronic medical records (EMRs) in its new system, frontline clinical staff had to change their usage habits, leading to significant resistance. Applying pressure solely on the grounds of facilitating research often proves counterproductive, as some physicians do not prioritize research. However, framing the initiative as a means to ensure the quality of medical care tends to yield better results, since no physician is indifferent to the quality of medical treatment.