Real world data collection is difficult. In China, there are additional challenges to overcome. The main obstacles to real world data collection in China revolve around data privacy and data integration and quality. This article highlights below these obstacles, as well as possible solutions that could address this.
Data Privacy
Data privacy comes into play from multiple angles in China. For one, it is unclear who owns the data and to what extent hospitals can use, or even distribute, the data to others. In 2017, the Chinese government issued the ‘The Cybersecurity Law of the People’s Republic of China’ and ‘Regulations for the Application of Electronic Medical Records. According to these regulations, data sharing can only be performed when the safety of patients’ electronic data is ensured. Essentially, the language means the hospital is responsible when an issue arises, with the definition of what is an “issue” open to interpretation. This creates hesitancy and hospitals are extra cautious to share data not only with other parties, including third-party vendors, and even other hospitals. The data/electronic record sharing among hospitals seemingly would be important for patients when transferring hospital systems, but data privacy combined with lack of needed infrastructure (see below) makes it difficult.
Even from a genomic standpoint, there are also specific regulations. For genomics/biospecimens, the Human Genetic Resource Administration (HGRAC) of China is responsible for enforcing what can and cannot be collected. Human genetic resources are defined as “any materials of and from human beings that contain human genome, genes or gene products, or parts thereof”. These include when the data will be used abroad, typically the case for global pharma and diagnostic companies. In addition, the data itself cannot leave China. As a result, companies need to either have the resources, internal or vendors, to analyze the data in China, or use systems that restrict analyses to cohorts of patients rather than at the patient level. Cohort analyses can be used and interpreted outside of China.
Data Integration & Quality
While 90% of Chinese Hospitals now have EMR systems, most of these EMR systems usually evolve independently. In China, there is no centralized national health record systems or major vendor(s) as seen in many other countries. These EMR systems have their own information standards, leading to the development of information silos. Additionally, healthcare classification and coding are not standardized. Hospitals create their own disease codes, billing codes, drugs, herbs and supplements databases. The system simply isn’t set-up for data aggregation and analysis, with little public or private guidance for quality control and data collection.
Further, the EMR structures can create problems in regards to data quality. For instance, EMR/ structured data cannot be directly downloaded into research files. That requires manual transcription (into Excel for instance) which can often lead to errors. Although not specific to China, many laboratory/imaging test data is stored in PDF format or image files and must be manually transcribed. If done properly (with experts), however, manual transcription can be a powerful tool, as it may be the only way to capture longitudinal data from disparate EMR systems.
Possible Solutions
The environment for Real World Data use in China is changing for the positive. The NMPA (The Chinese FDA/EMA) has accepted RWE for regulatory approval for drugs and devices and launched new guidelines. There has also been traction on the local level, with the Hainan Province (as part of the Hainan Pilot Zone) launching the Real-World Data Research Institute.
Still, it will take some time for hospital systems and data privacy laws to catch up. What are the possible current workarounds? Data quality standards can be provided by sponsor companies especially those looking to collect data prospectively. From a data privacy standpoint, Pharma/Biotech companies can work directly with hospitals and clinicians in the form of a research study. Since data cannot leave China however, a third-party vendor would be needed to help analyze the data and provide insights.
The longitudinal data issue is a bit trickier, as there isn’t an automated system or specialized software to address this. Intriguingly, it seems the patients themselves can address the issue of longitudinal data collection. According to one study, patients in China are diligent in collecting and keeping their medical records at home, obtaining them from multiple hospitals. Given that data privacy is a concern, perhaps patients hold the key in helping with real world evidence collection?