Systematic review on methodology: time series regression analysis for environmental factors and infectious diseases
10.2149/tmh.2014-21
- Author:
Chisato Imai
;
Masahiro Hashizume
- Publication Type:Journal Article
- Keywords:
time series;
seasonality;
infectious disease;
environmental factor;
weather;
review;
GLM;
GAM
- From:Tropical Medicine and Health
2014;():-
- CountryJapan
- Language:English
-
Abstract:
Background: Time series analysis is suitable forinvestigations of relatively direct and short-term effects of exposures on outcomes.In environmental epidemiology studies, this method has been one of the standardapproaches to assess impacts of environmental factors on acute non-infectious diseases(e.g. cardiovascular deaths), with conventionally generalized linear or additivemodels (GLM and GAM). However, the same manner of practices of this method is observedwith infectious diseases despite of the substantial differences fromnon-infectious diseases which may result in analytical challenges. Methods: Following Preferred ReportingItems for Systematic Reviews and Meta-Analyses guideline, systematic review wasconducted to elucidate important issues in assessing the associations betweenenvironmental factors and infectious diseases using time series analysis withGLM or GAM. Published studies in relation to associations between weatherfactors, and malaria, cholera, dengue, or influenza were targeted. Findings: Issues regarding theestimation of susceptible population and exposure lag times, adequacy ofseasonal adjustments, the presence of strong autocorrelations, and a lack of smallerobservation time unit of outcomes (i.e. daily data) were raised from our review.These concerns may be attributed to the features specific to infectious diseases,such as transmissions among individuals and complicated causal mechanisms. Conclusion: The consequence of not takingadequate measures to address these issues is distortion of the appropriate riskquantifications of exposures factors. The future studies are required careful attentionsto details, and recommended to examine alternative models or methods thatimprove studies with time series regression analysis for environmental determinantsof infectious diseases.
