value
| - Objectives: Health events emerge from a multifactorial milieu involving host, community, environment, and pathogen factors. Therefore, developing accurate forecasting models to improve epidemic prediction towards better prevention and capabilities management is a complex task. Here, we describe an exploratory analysis to identify non-health risk factors that could improve the forecast and events risk signals using a feasible and practical approach by combining surveillance report data with non-health data from open data sources. Methods: A line listing was developed using information from the World Health Organization Disease Outbreaks News from 2016-2018. A database was created merging the line listing data with non-health indicators from the World Bank. Poisson regression models employing forward imputations were used to establish relationships and predict values over the dependent variable (health event frequency); which are the health events reported by each country to WHO during 2016-2018. Findings: The resulting regression model provided evidence that changes in non-health factors important to community experiences impact the risk of the number of major health events that a country could experience. Three non-health indicators (extrinsic factors) were associated significantly to event frequency (population urban change, gross domestic product change per capita -a novel factor, and average forest area). An exploratory analysis of the current COVID-19 pandemic suggested similar associations, but confounding by global disease burden is likely. Conclusion: Continued development of forecasting approaches capturing available whole-of-society extrinsic factors (non-health factors); could improve the risk management process through earlier hazard identification, and as importantly inform strategic decision processes in multisectoral strategies to preventing, detecting, and responding to pandemic-threat events.
|