Study on sensitivity of climatic factors on influenza A (H1N1) based on classification and regression tree and wavelet analysis.
- Author:
Hong XIAO
1
;
Xiao-ling LIN
;
Xiang-yu DAI
;
Li-dong GAO
;
Bi-yun CHEN
;
Xi-xing ZHANG
;
Pei-juan ZHU
;
Huai-yu TIAN
Author Information
- Publication Type:Journal Article
- MeSH: China; epidemiology; Climate; Humans; Influenza A Virus, H1N1 Subtype; Influenza, Human; epidemiology; virology; Regression Analysis; Risk Factors; Seasons; Temperature
- From: Chinese Journal of Preventive Medicine 2012;46(5):430-435
- CountryChina
- Language:Chinese
-
Abstract:
OBJECTIVETo analyze the periodicity of pandemic influenza A (H1N1) in Changsha in year 2009 and its correlation with sensitive climatic factors.
METHODSThe information of 5439 cases of influenza A (H1N1) and synchronous meteorological data during the period between May 22th and December 31st in year 2009 (223 days in total) in Changsha city were collected. The classification and regression tree (CART) was employed to screen the sensitive climatic factors on influenza A (H1N1); meanwhile, cross wavelet transform and wavelet coherence analysis were applied to assess and compare the periodicity of the pandemic disease and its association with the time-lag phase features of the sensitive climatic factors.
RESULTSThe results of CART indicated that the daily minimum temperature and daily absolute humidity were the sensitive climatic factors for the popularity of influenza A (H1N1) in Changsha. The peak of the incidence of influenza A (H1N1) was in the period between October and December (Median (M) = 44.00 cases per day), simultaneously the daily minimum temperature (M = 13°C) and daily absolute humidity (M = 6.69 g/m(3)) were relatively low. The results of wavelet analysis demonstrated that a period of 16 days was found in the epidemic threshold in Changsha, while the daily minimum temperature and daily absolute humidity were the relatively sensitive climatic factors. The number of daily reported patients was statistically relevant to the daily minimum temperature and daily absolute humidity. The frequency domain was mostly in the period of (16 ± 2) days. In the initial stage of the disease (from August 9th and September 8th), a 6-day lag was found between the incidence and the daily minimum temperature. In the peak period of the disease, the daily minimum temperature and daily absolute humidity were negatively relevant to the incidence of the disease.
CONCLUSIONIn the pandemic period, the incidence of influenza A (H1N1) showed periodic features; and the sensitive climatic factors did have a "driving effect" on the incidence of influenza A (H1N1).