Spatio-temporal Analysis of Dengue Fever Cases: A Retrospective Study

Abstract

Introduction: In India, during the last ten years, the dengue fever increased in incidence and in territorial extent. Several studies already described the dengue fever incidence in terms of time, place and person But in spatio-temporal analysis, space and time will be analysed together to identify the high priority areas and during which time period there is chance for occurrence of outbreaks in these high priority districts. Aim: To identify the districts with high log likelihood ratio for high dengue fever incidence. Materials and Methods: The current study design was a retrospective observational study conducted using the secondary data provided by Director of Public Health of Kerala state on a public domain. All the suspected dengue fever cases reported for a period of six months from 1st May 2017 to 4th November 2017 was used for the study. Weekly incidence rates of dengue fever cases and moving averages of dengue fever cases were calculated using MS Excel software. Retrospective spatio-temporal analysis was done with SaT Scan software using discrete poisson model. Results: Total number of dengue fever cases reported was 52371. The incidence of dengue fever rose from May 1-7 to June 26-July 02 and later showed declining trend. Moving averages graph of actual data shows peaks and troughs with an approximate 7 day pattern and the troughs coincided with public holidays. Spatio-temporal analysis revealed that, ten districts out of fourteen districts were identified having significantly (p<0.005) high relative risk and high likelihood ratio for dengue fever cases. Conclusion: Epidemic of dengue fever cases will start in the month of May, reaches peak in June and July, declines thereafter and reaches to normal by September. Time series analysis revealed the lacunae in surveillance system. Spatio-temporal analysis identified ten high priority districts out of which top four are in south Kerala.

Keywords

• disease mapping
• discrete poisson model
• relative risk