Tags

Type your tag names separated by a space and hit enter

Developing a dengue prediction model based on climate in Tawau, Malaysia.
Acta Trop. 2019 Sep; 197:105055.AT

Abstract

Dengue is fast becoming the most urgent health issue in Malaysia, recording close to a 10-fold increase in cases over the last decade. With much uncertainty hovering over the recently introduced tetravalent vaccine and no effective antiviral drugs, vector control remains the most important strategy in combating dengue. This study analyses the relationship between weather predictors including its lagged terms, and dengue incidence in the District of Tawau over a period of 12 years, from 2006 to 2017. A forecasting model purposed to predict future outbreaks in Tawau was then developed using this data. Monthly dengue incidence data, mean temperature, maximum temperature, minimum temperature, mean relative humidity and mean rainfall over a period of 12 years from 2006 to 2017 in Tawau were retrieved from Tawau District Health Office and the Malaysian Meteorological Department. Cross-correlation analysis between weather predictors, lagged terms of weather predictors and dengue incidences established statistically significant cross-correlation between lagged periods of weather predictors-namely maximum temperature, mean relative humidity and mean rainfall with dengue incidence at time lags of 4-6 months. These variables were then employed into 3 different methods: a multivariate Poisson regression model, a Seasonal Autoregressive Integrated Moving Average (SARIMA) model and a SARIMA with external regressors for selection. Three models were selected but the SARIMA with external regressors model utilising maximum temperature at a lag of 6 months (p-value:0.001), minimum temperature at a lag of 4 months (p-value:0.01), mean relative humidity at a lag of 2 months (p-value:0.001), and mean rainfall at a lag of 6 months (p-value:0.001) produced an AIC of 841.94, and a log-likelihood score of -413.97 establishing it as the best fitting model of the methodologies utilised. In validating the models, they were utilised to develop forecasts with the model selected with the highest accuracy of predictions being the SARIMA model predicting 1 month in advance (MAE: 7.032, MSE: 83.977). This study establishes the effect of weather on the intensity and magnitude of dengue incidence as has been previously studied. A prediction model remains a novel method of evidence-based forecasting in Tawau, Sabah. The model developed in this study, demonstrated an ability to forecast potential dengue outbreaks 1 to 4 months in advance. These findings are not dissimilar to what has been previously studied in many different countries- with temperature and humidity consistently being established as powerful predictors of dengue incidence magnitude. When used in prognostication, it can enhance- decision making and allow judicious use of resources in public health setting. Nevertheless, the model remains a work in progress- requiring larger and more diverse data.

Authors+Show Affiliations

Center for Disease Control, Tawau District Health Office, Sabah State Health Department, Ministry of Health, Malaysia. Electronic address: vivekjason1987@gmail.com.Department of Community and Family Medicine, Faculty of Medicine and Health Sciences, University Malaysia Sabah, Malaysia. Electronic address: drrichardavoi@gmail.com.Center for Disease Control, Tawau District Health Office, Sabah State Health Department, Ministry of Health, Malaysia. Electronic address: kaps6962@gmail.com.Ministry of Health, Malaysia. Electronic address: dr.dhesi@gmail.com.Center for Disease Control, Tawau District Health Office, Sabah State Health Department, Ministry of Health, Malaysia. Electronic address: hagku74@gmail.com.

Pub Type(s)

Journal Article

Language

eng

PubMed ID

31185224

Citation

Jayaraj, Vivek Jason, et al. "Developing a Dengue Prediction Model Based On Climate in Tawau, Malaysia." Acta Tropica, vol. 197, 2019, p. 105055.
Jayaraj VJ, Avoi R, Gopalakrishnan N, et al. Developing a dengue prediction model based on climate in Tawau, Malaysia. Acta Trop. 2019;197:105055.
Jayaraj, V. J., Avoi, R., Gopalakrishnan, N., Raja, D. B., & Umasa, Y. (2019). Developing a dengue prediction model based on climate in Tawau, Malaysia. Acta Tropica, 197, 105055. https://doi.org/10.1016/j.actatropica.2019.105055
Jayaraj VJ, et al. Developing a Dengue Prediction Model Based On Climate in Tawau, Malaysia. Acta Trop. 2019;197:105055. PubMed PMID: 31185224.
* Article titles in AMA citation format should be in sentence-case
TY - JOUR T1 - Developing a dengue prediction model based on climate in Tawau, Malaysia. AU - Jayaraj,Vivek Jason, AU - Avoi,Richard, AU - Gopalakrishnan,Navindran, AU - Raja,Dhesi Baha, AU - Umasa,Yusri, Y1 - 2019/06/08/ PY - 2019/02/21/received PY - 2019/06/03/revised PY - 2019/06/04/accepted PY - 2019/6/12/pubmed PY - 2019/11/23/medline PY - 2019/6/12/entrez KW - Dengue fever KW - Early warning KW - Epidemic KW - Forecasting model KW - Rainfall KW - Temperature SP - 105055 EP - 105055 JF - Acta tropica JO - Acta Trop. VL - 197 N2 - Dengue is fast becoming the most urgent health issue in Malaysia, recording close to a 10-fold increase in cases over the last decade. With much uncertainty hovering over the recently introduced tetravalent vaccine and no effective antiviral drugs, vector control remains the most important strategy in combating dengue. This study analyses the relationship between weather predictors including its lagged terms, and dengue incidence in the District of Tawau over a period of 12 years, from 2006 to 2017. A forecasting model purposed to predict future outbreaks in Tawau was then developed using this data. Monthly dengue incidence data, mean temperature, maximum temperature, minimum temperature, mean relative humidity and mean rainfall over a period of 12 years from 2006 to 2017 in Tawau were retrieved from Tawau District Health Office and the Malaysian Meteorological Department. Cross-correlation analysis between weather predictors, lagged terms of weather predictors and dengue incidences established statistically significant cross-correlation between lagged periods of weather predictors-namely maximum temperature, mean relative humidity and mean rainfall with dengue incidence at time lags of 4-6 months. These variables were then employed into 3 different methods: a multivariate Poisson regression model, a Seasonal Autoregressive Integrated Moving Average (SARIMA) model and a SARIMA with external regressors for selection. Three models were selected but the SARIMA with external regressors model utilising maximum temperature at a lag of 6 months (p-value:0.001), minimum temperature at a lag of 4 months (p-value:0.01), mean relative humidity at a lag of 2 months (p-value:0.001), and mean rainfall at a lag of 6 months (p-value:0.001) produced an AIC of 841.94, and a log-likelihood score of -413.97 establishing it as the best fitting model of the methodologies utilised. In validating the models, they were utilised to develop forecasts with the model selected with the highest accuracy of predictions being the SARIMA model predicting 1 month in advance (MAE: 7.032, MSE: 83.977). This study establishes the effect of weather on the intensity and magnitude of dengue incidence as has been previously studied. A prediction model remains a novel method of evidence-based forecasting in Tawau, Sabah. The model developed in this study, demonstrated an ability to forecast potential dengue outbreaks 1 to 4 months in advance. These findings are not dissimilar to what has been previously studied in many different countries- with temperature and humidity consistently being established as powerful predictors of dengue incidence magnitude. When used in prognostication, it can enhance- decision making and allow judicious use of resources in public health setting. Nevertheless, the model remains a work in progress- requiring larger and more diverse data. SN - 1873-6254 UR - https://www.unboundmedicine.com/medline/citation/31185224/Developing_a_dengue_prediction_model_based_on_climate_in_Tawau_Malaysia_ L2 - https://linkinghub.elsevier.com/retrieve/pii/S0001-706X(19)30186-X DB - PRIME DP - Unbound Medicine ER -