Spatiotemporal dynamics and forecasting of dengue incidence in Northeastern Thai border provinces, 2014-2023
Kulchaya Loyha , Worasorn Netthip , Surat Haruay , Denduangdee Srisura , Kornwika Harasarn , Chanapong Kuasiri , Panita Khampoosa
Asian Pacific Journal of Tropical Medicine ›› 2026, Vol. 19 ›› Issue (3) : 111 -119.
Objective: To identify the underexplored nature of localized, longterm spatiotemporal patterns (2014-2023) and predictive modeling feasibility in these regions. The analysis specifically explored the potential for robust forecasting of dengue incidence trends.
Methods: Dengue incidence trends were analyzed across five northeastern Thai provinces: Sisaket (SSK), Ubon Ratchathani (UBN), Yasothon (YST), Amnat Charoen (ACR), and Mukdahan (MDH) over a ten-year period (2014-2023). Analysis of Kruskal-Wallis H test followed by post hoc Dunn’s tests with Bonferroni adjustment were used for inter-provincial comparisons. Seasonal- Trend Decomposition using Loess (STL) and Seasonal Index calculations were applied to identify seasonal patterns, and Simple Linear Regression to assess the influence of monthly rainfall on dengue seasonality. Additionally, a negative binomial regression analysis was conducted, and spatial autocorrelation was investigated using Moran's I analysis.
Results: The analysis revealed significant spatiotemporal heterogeneity and distinct cyclical patterns in dengue incidence. Pronounced annual peaks consistently occurred during the rainy season (June-September), with major epidemics observed in 2015, 2019, and 2023. Inter-provincial comparisons revealed statistically significant differences in monthly dengue incidence across the study area (H=10.08, P=0.039). Post-hoc Dunn’s tests indicated that UBN and SSK had higher transmission burden compared to YST and MDH. Seasonal Index calculations confirmed July as the predominant overall peak month (indices 1.4-2.5), with minimal dengue activity from January to March. While rainfall significantly influenced seasonality across all provinces. A negative binomial regression analysis indicated that all of the variables included in the model had no statistically significant relationship with dengue incidence. Furthermore, a Moran's I analysis revealed a significant clustering effect only in YST (1=0.314, P=0.037).
Conclusions: The finding underscores the critical requirement for adopting a highly focused approach, necessitating a shift toward region-specific interventions to effectively combat dengue and sustain progress toward good health and well-being for all.
Dengue / Dengue forecasting / Epidemic preparedness / Spatiotemporal analysis / Seasonality / Northeastern Thailand / Predictive modeling
| [1] |
|
| [2] |
|
| [3] |
|
| [4] |
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
|
| [19] |
|
| [20] |
|
| [21] |
|
| [22] |
|
| [23] |
|
| [24] |
|
| [25] |
|
| [26] |
|
| [27] |
|
| [28] |
|
| [29] |
|
| [30] |
|
| [31] |
|
| [32] |
|
| [33] |
|
| [34] |
|
| [35] |
|
| [36] |
|
| [37] |
|
| [38] |
|
| [39] |
|
| [40] |
|
| [41] |
|
| [42] |
|
| [43] |
|
| [44] |
|
| [45] |
|
| [46] |
|
/
| 〈 |
|
〉 |