Investigation of health care-associated infections surveillance and early-warning variables and assessment of their efficiencies in the era of big data and artificial intelligence

OBJECTIVE To investigate the current status of quality of health care-associated infections (HAIs) surveillance data in secondary or above medical institutions， analyze the early-warning variables in application of big data and artificial intelligence， and observe the efficiencies so as to provide supporting data for construction of intelligentized and precise HAIs surveillance and early-warning system and guidance for improvement. METHODS By mean of questionnaire survey and onsite verification， the current status of surveillance of HAIs cases in the medical institutions of 2023 was studied. The informatization level， quality of data， early-warning efficiency and human resource investment were evaluated. RESULTS The research showed that 89.19% of the hospitals have achieved the information-based surveillance， the percentage of the tertiary hospitals with the information-based surveillance was 96.55%， higher than that of the secondary hospitals(P=0.026). Regarding to the efficiency of surveillance， the median number of full-time surveillance personnel per 150 beds per day was 0.65 (IQR: 0.41~1.23)， and the number of the personnel was remarkably larger in the secondary hospitals than in the tertiary hospitals (P=0.045). The time consumed for surveillance was approximately same in the tertiary hospitals and the secondary hospitals， that was 0.3 hour per 150 beds per person per day， and the median number of early-warning cases was 3.70 cases per 150 beds per day. With the respect to the quality of data， 94.12% of the hospitals achieved various degree of evaluation of data quality， 62.16% of the hospitals controlled the missing report rate within 5%; 35.14% of the hospitals had the accuracy of self-assessment ranging between 90% and 95%， with higher than 95% in 32.43% of the hospitals; the timeliness of data of the tertiary hospitals was superior to that of the secondary hospitals (P=0.031). CONCLUSIONS The intelligent surveillance and early-warning system can substantially raise the early warning efficiency in the era of big data and artificial intelligence， but it imposes higher demands for the quality of data. Although the popularizing rate of informatization of HAIs surveillance is high， it is necessary to attach more importance to the quality of data and raise the early-warning efficiency so as to push forward the intelligentized and precise informatization of HAIs surveillance and prevent the false early warning due to the poor quality of data.