Optimization of absenteeism indicators for a face recognition based syndromic surveillance system: a longitudinal study, China, September 2021 to June 2022

Abstract

Abstract Background Although smart attendance can only collect all-cause absenteeism currently, whichis conductive to the modernization of school-oriented syndromic surveillance system (SSS).It is of great significance to optimize indicators of all-cause absenteeism based on smart attendance. Aim To choose an appropriate time standard for absenteeism, and explore more effective indicators for the face-recognition-based SSS (FRSSS). Methods Two primary schools in Hangzhou, China participated in the pilot study. Grade 1-2 (DARL), 3-6 (DARH), and school-wide (DARX) all-cause-absenteeism reported by FRSSS, and all-cause (DARY) and sickness absenteeism (DARZ) reported by school physicians, were daily collected from September 1, 2021, to June 24, 2022, and these five indicators' effectiveness of epidemic detection were compared by correlations, time series, and control charts. Results The time standard of absenteeism was "≥ 24 hours" for DARY and DARZ, while "≥ one hour" for DARX, DARL and DARH. DARY and DARZ only were 32.6% and 25.2% of DARX. The correlation coefficient between DARY and DARZ was 0.843 (P<0.001) in school A and 0.933 (P<0.001) in school B. In school A, Yoden indexes of DARL, DARH, DARX, DARY and DARZ were 83.0%, 85.0%, 80.6%, 78.2% and 80.4%, respectively. In school B, Yoden indexes of these five indicators were 89.3%, 91.0%, 83.9%, 76.8% and 81.0%, respectively. Conclusions The effectiveness of outbreak detection for the smart attendance based indicators could be raised to a considerable level by setting reasonable time standard and adopting multi-level indicators. It is feasible and effective to popularize smart attendance in school-oriented SSSs.