[Early warning technology for infectious disease outbreaks in primary and secondary school students based on integrating hospital visit data and monitoring data of school absence due to illness: an effectiveness analysis].

Objective

To analyze the association between hospital visit data and data of school absence due to illness in primary and secondary school students, and evaluate the performance of different monitoring data in early warning of infectious disease outbreaks.

Methods

Thirteen primary and secondary schools in Dapeng New District, Shenzhen, were selected to conduct the study. Data of infectious disease-related hospital visits in local students from January 1, 2015 to December 31, 2019 were collected from Chinese Disease Prevention and Control Information System, and the monitoring data of school absence due to illness during the period were collected from Shenzhen Student Health Monitoring System. Spearman rank correlation analysis was used to evaluate the time-series correlation between the two types of data. An early warning strategy was established by using the moving average method combined with a fixed-threshold method. The performance of different early warning strategies were evaluated by using indicators such as positive predictive value (PPV) and sensitivity, the school outbreak data reported in Shenzhen Disease Cluster Surveillance System were used as the gold standard.

Results

A total of 1 870 cases of school absence due to infectious diseases were found in local primary and secondary school students. The main diseases were influenza, varicella, and mumps, with fever, cough, and sore throat being the most common reasons for school absence. During the study period, 52 infectious disease outbreaks were reported in local primary and secondary schools, which were mainly caused by influenza and varicella. A lag correlation was observed between the hospital visit data and the school absence monitoring data (r=0.31, P<0.001), with the latter detecting outbreaks earlier. Hospital monitoring, school absence monitoring, serial monitoring and parallel monitoring triggered 385, 491, 298, and 876 early warning signals, respectively. The hospital monitoring alone achieved the highest positive prediction value (66.49%), the serial monitoring achieved the highest specificity (98.42%), and the parallel monitoring achieved the highest sensitivity (100.00%) and Youden index (80.52%).

Conclusions

The information about infectious disease obtained from hospital visit data lags behind monitoring data of school absence due to illness in primary and secondary school students. A parallel early warning strategy combining both monitoring methods enables accurate and rapid outbreak detection, providing data support for epidemiological investigation and prevention and control resource allocation.