Application of medical information system to identify dengue outbreak factors: Insights from a hyperendemic city in Malaysia

•Data from the Malaysian e-Dengue system is analysed to distinguish factors that drive dengue outbreaks.•A unique “wet-dry steps” rainfall pattern precedes outbreak events.•Data visualisation enables the identification of dengue hotspots and “reservoir” zones for targeted vector control measures.•High human settlement density, guerrilla gardening practises, and proximity to forest or oil palm plantation edges are plausible factors that cause outbreaks. Dengue is widespread globally, but it is more severe in hyperendemic regions where the virus, its vectors, and its human hosts naturally occur. The problem is particularly acute in cities, where outbreaks affect a large human population living in a wide array of socio-environmental conditions. Controlling outbreaks will rely largely on systematic data collection and analysis approaches to uncover nuances on a city-by-city basis due to the diversity of factors. The main objective of this study is to consolidate and analyse the dengue case dataset amassed by the e-Dengue web-based information system, developed by the Ministry of Health Malaysia, to improve our epidemiological understanding. We retrieved data from the e-Dengue system and integrated a total of 18,812 cases from 2012 to 2019 (8 years) with meteorological data, geoinformatics techniques, and socio-environmental observations to identify plausible factors that could have caused dengue outbreaks in Ipoh, a hyperendemic city in Malaysia. The rainfall trend characterised by a linearity of R2 > 0.99, termed the “wet-dry steps”, may be the unifying factor for triggering dengue outbreaks, though it is still a hypothesis that needs further validation. Successful mapping of the dengue “reservoir” contact zones and spill-over diffusion revealed socio-environmental factors that may be controlled through preventive measures. Age is another factor to consider, as the platelet and white blood cell counts in the “below 5” age group are much greater than in other age groups. Our work demonstrates the novelty of the e-Dengue system, which can identify outbreak factors at high resolution when integrated with non-medical fields. Besides dengue, the techniques and insights laid out in this paper are valuable, at large, for advancing control strategies for other mosquito-borne diseases such as malaria, chikungunya, and zika in other hyperendemic cities elsewhere globally.