心血管疾病危害严重，风险预测对疾病防治意义重大，预测模型在国际上发展迅速，但针对中国人群的风险评估工具较少、预测模型较老，公共卫生筛查策略有待完善。因此，本研究拟利用宁波市鄞州区区域信息平台的健康大数据，通过该地区的全人群自然队列，结合北京市房山区的前瞻性队列研究，建立中国人群的疾病风险预测模型。首先选取QRISK2、Framingham风险评分等经典预测工具，对原始评分和校准后评分在现阶段中国人群中进行直接对比评估；然后在本课题组前期英国研究的基础上，通过随机森林等多种方案在数据库中筛选合适的因子，开发更为精准的基于竞争风险的预测模型，再利用北京市房山区的前瞻性队列研究对新开发的模型进行验证；最后，提出适合我国现阶段国情的公共卫生序贯筛查策略，应用预测模型结合贝叶斯统计方法对管理与决策导向的健康大数据公共卫生筛查策略进行评估，为我国人群心血管疾病一级预防及患者个体化疾病风险管理提供依据。

Cardiovascular disease (CVD) remains the grip reaper’s primary calling card. Disease prevention must be seen as a life-long effort. Major strategies in CVD prevention require measurement of CVD risk. Several models for CVD risk prediction have been developed and updated in the Western population while the algorithms for Chinese population were based on the cohort ended before 2002. The CVD risk prediction model for the Chinese population under current economic levels and the public health modelling of the screening strategies are limited. We aim to develop the population-based CVD risk prediction model for the Chinese population using the big data from the healthcare information system in Ningbo as the discovery data and validated in the independent cohort study established in Beijing. Several well-known models derived from the Western population will be firstly re-calibrated in the population-based discovery cohort using the clinical practice records of residents in the town in Ningbo. The original and the re-calibrated scores will be directly compared and evaluated. We will then develop and evaluate novel scores by combining features not used in concert by previous studies, and by exploiting extensive candidate and discovery data using multiple methods including traditional parametric regression and machine learning algorithms such as random forest. To help avoid “overfitting”, we will use cross-validation in the discovery cohort and external validation using the epidemiological cohort study established in Beijing. Finally, we will assess various screening options including the proposed sequential screening strategy, comparing their consequent estimated effects in CVD prevention in a wide range of circumstances. The project will provide the latest evidence for population-level CVD primary prevention and tools for individualized CVD risk management in China.