Summary of Key Points
China leads the world in both new cancer cases (4.82 million in 2022) and deaths (2.57 million in 2022). However, the current tumor registration system, which relies on manual operations, has three major issues: data lag (more than five years), limited coverage (reaching at most 37% of the population), and incomplete monitoring dimensions (lack of prevalence and survival rates). Experts suggest integrating multiple large datasets from sources such as healthcare insurance and cause-of-death surveillance, combined with AI technology, to establish a dynamic monitoring system. Pilots have shown that this approach can reduce reporting delays to half a year and complement the shortcomings of traditional data. Nevertheless, implementing the new system requires overcoming challenges in data sharing, technical integration, and nationwide validation. The ultimate goal is to provide timely and accurate decision-making support for cancer prevention and control efforts, contributing to the construction of a healthier China.
1. Why is Traditional Cancer Monitoring Slow and Incomplete?
The traditional tumor registration system has been in use for over 60 years and has provided valuable information for prevention and control efforts. However, it can no longer meet current needs:
- Data Lag: Cases must be collected, reviewed, and reported manually, resulting in a national report that takes several years to release. For example, data from 2022 may not be fully available until 2027, leaving policymakers with outdated information.
- Limited Coverage: Registration points that meet international standards cover only 13% of the population; even if all qualified domestic sites are included, the coverage still amounts to 37%. There are significant differences in data quality due to regional variations in personnel and technology.
- Incomplete Monitoring Dimensions: The system only counts new cases and deaths but does not provide information on the current number of patients alive (prevalence), their survival times, or quality of life. These critical pieces of data are often obtained through occasional research projects rather than regular monitoring.
2. Are Healthcare Insurance Data a Reliable New Tool for Cancer Monitoring?
Healthcare insurance data has become a promising solution for several reasons:
- Wide Coverage and Real-time Data: Healthcare insurance covers nearly the entire population, and patient information such as diagnoses, hospital records, treatments, and medications is automatically recorded in databases.
- International Validation: Countries like Canada and Germany have used healthcare insurance data for cancer monitoring with consistent results compared to traditional systems, and it has helped identify cases missed by traditional methods.
- Positive Pilot Results in China:
- A pilot involving 7 million people in different regions reduced reporting delays to half a year and was much cheaper than manual registration.
- In Beijing, the use of healthcare insurance data supplemented cases missing from traditional records (e.g., patients who sought treatment at multiple hospitals).
- Combining with cause-of-death data, it achieved for the first time comprehensive monitoring of "onset-death-survival," solving issues with prevalence and survival rate measurement.
3. What Challenges Must Be Overcome to Implement the New System?
Moving from pilots to nationwide adoption faces several practical obstacles:
- Data Sharing: Cancer data is scattered across health, healthcare insurance, disease control, and public security departments, each with different standards and no established sharing mechanisms. For example, healthcare insurance data is managed by the healthcare insurance bureau, while death data is handled by disease control agencies, creating barriers to data integration.
- Technical Integration: Data formats and disease codes vary across systems (e.g., the same cancer may have different codes in healthcare insurance and hospitals), requiring extensive data cleaning, deduplication, and identification of duplicate cases.
- Privacy and Institutional Issues: Patient data must be protected, and clear authorization rules are needed. While AI can improve efficiency, it cannot replace well-established institutional frameworks; without a coordinated approach, even advanced technology is ineffective.
- National Validation: Currently, these are only local pilots, and there is no nationwide validation of the system's effectiveness. For instance, differences in healthcare insurance policies across regions may affect data accuracy, which needs further testing.
4. How to Progress?
To establish the new system, a step-by-step approach is recommended:
- Establish a Coordination Mechanism: The National Health Commission should lead the effort, working with healthcare insurance, disease control, and other departments to develop unified data standards and sharing rules.
- Leverage Technology: Use AI (e.g., natural language processing and large models) to automatically clean and identify cases, improving efficiency and accuracy, but only after establishing a solid institutional foundation.
- Pilot Projects First: Select several regions for trials to integrate tumor registration, healthcare insurance, and cause-of-death data, evaluating feasibility, accuracy, and cost-effectiveness before rolling out nationwide.
- Ultimate Goal: Create a nationwide, dynamically updating monitoring system that enables policymakers to promptly understand cancer trends (e.g., which types of cancer are growing fastest), assess the effectiveness of screening/treatment policies (e.g., whether certain drugs truly improve survival rates), and allocate resources efficiently (e.g., identifying areas in need of more oncologists).
If implemented successfully, this new system will not only address data lag but also make cancer prevention and control more precise, becoming a valuable tool in reducing the burden of cancer.