Objective: The objective of this study was to establish the combination of time and space parameters for spatiotemporal rearrangement scanning. The spatial and temporal rearrangement of dengue cluster events was performed by using dengue syndrome to evaluate the effects of different spatiotemporal parameter combinations on the early warning effect. This study aimed to select the best spatiotemporal parameter combination for early warning of dengue syndrome and establish a perfect early warning model of dengue syndrome based on spatiotemporal rearrangement scan statistics.
Methods: The data of dengue syndrome in eighty medical institutions (including township hospitals and village clinics) in the symptom monitoring system of medical institutions in Mengla County in 2017 and the data of reported cases of dengue in the direct reporting system of infectious diseases were collected using Excel to describe the time, region and population distribution of dengue fever. The Java server page was scanned to establish different spatiotemporal parameter combinations (X, Y). X was the time parameter, and Y was the space parameter. The maximum number of early warning days was set to four, and when including the date of the gathering event, a total of five days. The dengue fever clusters in Mengla County in 2017 were scanned day by day in time and space, and the sensitivity and the number of days of advance warning were used as the warning indicators to comprehensively evaluate the warning effects of nine warning schemes.
Results: 1. A total of two hundred forty-seven cases of dengue fever were reported in Mengla County in 2017, with an incidence rate of 85.79/100,000. The peak incidence was mainly from August to November; cases of dengue fever reported by the infectious disease direct reporting system were distributed over eight medical institutions. Mengla County People's Hospital had the largest number of reported cases; the ratio of males to females was roughly balanced (approximately 1:1). The largest population distribution was 100 cases in farmers (40.49%), followed by 50 cases of business service personnel (20.24%). 2. Nine different spatiotemporal parameter combinations were used to scan a total of 225 times, which were performed for five dengue fever clusters in Mengla County in 2017. The aggregation region, LLR(Log Likelihood Ratio), RR and P values of the warning under each scan combination were obtained. 3. Sensitivity analysis of nine parameter combinations for five dengue aggregation events included the following: the (3, 10), (4, 10), (5, 10), (4, 15), (5, 15) combination alerted to five dengue aggregation events, and the sensitivity was 100%; the (3, 15) combination alerted to four dengue aggregation events, and the sensitivity was 80%; the (3, 5), (4, 5), (5, 5) combination alerted to three dengue aggregation events, and the sensitivity was 60%. 4. Analysis of the early warning times of five dengue aggregation events by nine parameter combinations resulted in the following: the (3, 10) combination early warning times of five dengue aggregation events were three days, four days, four days, one day and one day, with an average early warning time of two point six days; the (4, 10), (5, 10) combination early warning times of five dengue aggregation events were two days, three days, four days, one days and zero day, respectively, with an average early warning time of two days; the (3, 15) combination early warning times of five dengue aggregation events were zero day, four days, four days, one day and one day, with an average early warning time of two days; the (5, 15) combination early warning times of five dengue aggregation events were one day, four days, four days, one day and zero day, with an average early warning time of 2 days; the (4, 5), (5, 5), (4,15) combination early warning times of five dengue aggregation events were zero day, four days, three days, one day and zero day, with an average early warning time of one point six days; the (3, 5) combination early warning times of five dengue aggregation events were one day, four days, three days, zero day and zero day, with an average early warning time of one point six days. 5. Analysis of the duration warning time of five dengue aggregation events by nine parameter combinations included the following: for the (3, 10) combination among the five dengue aggregation events, the average duration warning days was two point two days; for the (4, 10), (5, 10), (5, 15) combination among the five dengue aggregation events; the average duration warning days was two days, and the average duration warning days of other parameter combinations were less than two days.
Conclusions: 1. Dengue fever has an obvious seasonal trend in Mengla County, and the cases are mainly concentrated in summer and autumn. In this time period, we should strengthen the corresponding prevention and control of dengue fever, especially the main gathering areas of farmers and commercial service groups. 2. Based on the data of dengue syndrome, time-space rearrangement scanning of dengue clustering events in Mengla County was carried out, and the optimal time-space clustering combination was selected. In the future, the combination of the (3, 10) parameters can be preferentially used for the early warning of dengue outbreaks in this area by using dengue syndrome. In addition, the combination of the (4, 10), (5, 10), and (5, 15) parameters can be used as an alternative parameter combination. 3. In the future, when the optimal parameter combination is used for the early warning of dengue fever, a graded response can be adopted according to the actual situation, which maximally ensures that real aggregation events are detected while avoiding excessive waste of resources.