Table 2.
Characteristics of studies on the association between climatic variables and malaria transmission
|
Study & Language |
Study area & period |
Data Collection |
Statistical Methods |
Main findings |
Comments |
|
|---|---|---|---|---|---|---|
| Risk factors | Disease/vector | |||||
| Huang et al. (2011) English [19] |
Anhui, Henan, Hubei Provinces 1990-2009 |
Normalized annual temperature, relative humidity and rainfall |
Cases counts |
-Bayesian Poisson models |
-Rainfall played a more important role in malaria transmission than other meteorological factors. |
-Spatial-temporal models were developed |
| - GIS |
-Socioeconomic factors were not taken into account. |
|||||
| Huang et al. (2011) English [20] |
Motuo County, Tibet 1986-2009 |
Monthly average temperature, maximum temperature, minimum temperature, relative humidity and total amount of rainfall |
Monthly incidence of malaria |
-Spearman correlation analysis |
-Relative humidity was more sensitive to monthly malaria incidence. |
-Several statistical methods were applied |
| -Cross-correlation analysis |
-The relationship between malaria incidence and rainfall was not directly and linearly. |
-Only one county was considered |
||||
| -SARIMA model | ||||||
| -Inter-annual analysis | ||||||
| Zhou et al. (2010) English [21] |
Huaiyuan County of Anhui and Tongbai County of Henan Province 1990-2006 |
Monthly and annual average temperature, maximum temperature, minimum temperature, relative humidity and rainfall |
Monthly and annual incidence of malaria Vectorial capacity |
-Spearman correlation |
-Temperature and rainfall were major determinants for malaria transmission. However, no relationship between malaria incidence and relative humidity was observed. |
-Entomological investigate was conducted to determine the vectorial effect of malaria re-emergency. |
| -Stepwise regression analysis | ||||||
| -Curve fitting | ||||||
| -Trend analysis |
-Only two counties were examined |
|||||
| - Entomological investigation | ||||||
| Zhang et al. (2010) English [22] |
Jinan city, Shangdong Province 1959-1979 |
Monthly average maximum temperature, minimum temperature, relative humidity and rainfall |
Cases counts |
-Spearman correlation |
-Temperature was greatest relative to the transmission of malaria, but rainfall and relative humidity were not. |
-Only one city was included |
| -Cross-correlation |
-Socioeconomic factors ware ignored. |
|||||
| -SARIMA model | ||||||
| Yang et al. (2010) English [23] |
The P.R. China 1981-1995 |
Yearly growing degree days (YGDD), annual rainfall and relative humidity |
Malaria-endemic strata |
-A Delphi approach |
-Relative humidity was found to be the most important environmental factor, followed by temperature and rainfall. However, temperature was the major contributor of malaria intensity in regions with relative humidity >60%, |
-National-level analysis |
| -Multiple logistical regression |
-Risk maps of malaria based on different climatic factors were developed |
|||||
| -GIS | ||||||
| -Annual indicators were used | ||||||
| Xiao et al. (2010) English [24] |
Main island of Hainan province 1995-2008 |
Monthly average temperature, maximum temperature, minimum temperature, relative humidity and accumulative rainfall |
Monthly incidence of malaria |
-Cross correlation and autocorrelation analysis |
- Temperature during the previous one and two months were observed as major predictors of malaria epidemics. |
-Spatial-temporal analysis |
| -Poission regression | ||||||
| -GIS |
-Countermeasure and socioeconomic circumstances ware not taken into account. |
|||||
| -It was not necessary to consider rainfall and relative humidity to make malaria epidemic predictions in the tropical province. | ||||||
| Hui et al. (2009) English [25] |
Yunnan Province 1995-2005 |
Monthly average temperature, maximum temperature, minimum temperature, relative humidity and rainfall |
Monthly incidence of P. vivax malaria Monthly incidence of P. falciparum malaria |
-Spearman correlation analysis |
-Obvious associations between both P. vivax and P. falciparum malaria and climatic factors with a clear 1-month lagged effect, especially in cluster areas. |
-Analysis of both P. vivax malaria and P. falciparum malaria |
| -Temporal distribute analysis | ||||||
| -Spatio-temporal analysis | ||||||
| -Spatial autocorrelation | ||||||
| -Minimum temperature was most closely correlated to malaria incidence | ||||||
| -Spatial cluster analysis | ||||||
| - GIS | ||||||
| Clements et al. (2009) English [26] |
Yunnan Province 1991-2006 |
Monthly average rainfall, maximum temperature and minimum temperature |
Monthly incidence of P. vivax malaria Monthly incidence of P. falciparum malaria |
-Corss-correlation |
-Significant positive relationships between malaria incidence and rainfall and maximum temperature for both P. vivax and P.falciparum malaria |
-Analysis of both P. vivax malaria and P. falciparum malaria |
| -Bayesian Poisson regression | ||||||
| -Spatial-temporal analysis | ||||||
| -GIS | ||||||
| -Socioeconomic factors were ignored. | ||||||
| -High-incidence clusters located adjacent the international borders were not explained by climate, but partly due to population migration. | ||||||
| Tian et al. (2008) English [27] |
Mengla County, Yunnan Province 1971-1999 |
Monthly rainfall, minimum temperature, maximum temperature, relative humidity, and fog day frequency |
Monthly incidence of malaria |
-ARIMA models |
-Temperature and fog day frequency were key predictors of monthly malaria incidence. However, relative humidity and rainfall were not. |
-Fog day frequency used -P. vivax malaria and P. falciparum malaria were pooled together when malaria incidence was calculated. |
| -The annual fog frequency was the only weather predictor of the annual incidence of malaria | ||||||
| Bi et al. (2005) English [28] |
Anhui province 1966-1987 |
Monthly EI-Nino Southern Oscillation Index (ENSO) |
Monthly malaria cases |
-Spearman correlation |
-A positive correlation between ENSO and the incidence of malaria with no lag effect was found. |
-The impact of ENSO on malaria was analysed -Other meteorological variables were not considered. |
| -Only used correlation method | ||||||
| Liu et al. (2006) English [29] |
Twenty-one townships of 10 counties in Yunnan province 1984-1993 |
Monthly minimum temperature, maximum temperature, rainfall, sunshine duration, NDVI. |
Monthly incidence of malaria and vector density. |
-Principle component analysis |
-Remote sensing NDVI and climatic variables had a good correlation with Anopheles density and malaria incidence rate. |
-Both environmental and vector factors were analysed. |
| -Factor analysis | ||||||
| -Grey correlation analysis | ||||||
| Bi et al. (2003) English [30] |
Sunchen County in Ahui Province 1980-1991 |
Monthly maximum temperature, minimum temperature, relative humidity and rainfall |
Monthly incidence of malaria |
-Spearman correlation |
-Monthly average minimum temperature and total monthly rainfall, at one-month lag were major determinants in the transmission of malaria. |
-Non-climatic factors were neglected |
| -Cross-correlation | ||||||
| -Only one county considered | ||||||
| -ARIMA models | ||||||
| Hu et al. (1998) English [31] |
Yunnan Province 1991-1997 |
Annual rainfall, annual mean temperature |
Annual incidence of malaria |
- Multiple regression |
-Malaria incidence rates are higher in areas with temperature above 18°C, rainfall of more than 1000 mm |
-Socioeconomic factors such as income of farmers were taken into account. |
| -GIS | ||||||
| -Every one degree increase in temperature corresponds to 1.2/10,000 higher malaria incidence and when rainfall increase by 100 mm, malaria will increase to 100.0/10,000 |
-Annual data were used |
|||||
| Liu et al. (2011) Chinese [32] |
Pizhou City, Jiangsu province 2001-2006 |
Monthly mean temperature, maximum temperature, minimum temperature, rainfall days, relative humidity, evaporation, total cloud cover, sunlight time and low cloud. |
Monthly incidence of malaria |
-Correlation analysis |
-The incidence of malaria was passive relative to temperature, rainfall, relative humidity, evaporation and total cloud cover, but no relation with low cloud and sunlight. |
-Various meteorological variables were considered |
| -Multiple regression | ||||||
| -Only one city was analysed based on a relative short study period | ||||||
| -The monthly minimum temperature and relative humidity were two major factors influencing malaria transmission. | ||||||
| Wu et al. (2011) Chinese [33] |
Dianjiang county, Chongqing 1957-2010 |
Monthly mean temperature, maximum temperature, minimum temperature, rainfall days, relative humidity, absolute humidity, duration of sunshine, air pressure and wind speed. |
Case counts |
-Principal Component Analysis |
-Significant associations between malaria incidence and monthly mean temperature, rainfall and duration of sunshine were observed. |
-Various meteorological variables were considered |
| -Multiple regression | ||||||
| -Long-term data from a fifty-four-years period-Only one county considered | ||||||
| -Temperature was greatest relative to malaria transmission | ||||||
| | ||||||
| Huang et al. (2009) Chinese [34] |
Tongbai and Dabie mountain areas, Huibei Province 1990-2007 |
Monthly mean temperature, maximum temperature, minimum temperature, rainfall. |
Case counts |
Descriptive study |
-Temperature and rainfall were major determinants for malaria transmission and the yearly peak of cases occurred one month after the rainy season. |
-Not enough statistical methods |
| Wang et al. (2009) Chinese [35] |
Anhui Province 2004-2006 |
Annually mean temperature and rainfall NDVI and elevation. |
Cases counts |
-Principal Component Analysis |
-Malaria transmission intensity was positively associated with the NDVI, but negatively associated with minimum temperature, rainfall and elevation. |
-Annual indicators were used |
| -Logistic regression |
-A two-years short period of study. |
|||||
| -GIS | ||||||
| Wen et al. (2008) Chinese [36] |
Hainan Province May-Oct in 2002 |
Monthly mean temperature, maximum temperature, minimum temperature, rainfall, relative humidity, land use, land surface temperature (LST) and elevation. |
Monthly incidence of malaria |
-Spearman correlation |
-No associations between meteorological factors and malaria incidence were observed. However, land use, elevation and LST appeared to be good contributors of malaria transmission. |
-Various environmental variables were collected |
| -Negative binomial regression analysis | ||||||
| -A six-month short period of study. | ||||||
| Su et al. (2006) Chinese [37] |
Hainan Province 1995 |
Monthly mean temperature, maximum temperature, minimum temperature, rainfall, relative humidity and NDVI. |
Monthly incidence of malaria |
-Factor Analysis |
-Rainfall and the NDVI may be used to explain the malaria transmission and distribution. |
-A one-year short period of study. |
| -Principal Component Analysis | ||||||
| -Multiple liner regression analysis | ||||||
| Fan et al. (2005) Chinese [38] |
Ailao mountain of Yuxi city in Yunnan Province 1993-2002 |
Annual man temperature and rainfall |
Anopheles minimus density |
-Correlation analysis |
-Significant relationship between malaria incidence and abundance of Anopheles minimus. However, no significant correlations between abundance of Anopheles minimus and climatic variables. |
-No disease data |
| -Annual data used | ||||||
| Wen et al. (2005) Chinese [39] |
Hainan Province Feb 1995- Jan 1996 |
NDVI |
Monthly incidence of malaria |
-Spearman correlation |
-Malaria prevalence was highly associated with NDVI value which could be used for malaria surveillance in Hainan province. |
-A short study period |
| -GIS | ||||||
| -No other climatic indicators used | ||||||
| Huang et al. (2004) Chinese [40] |
Luodian county 1951–2000 Libo county 1958–2000 Sandu county 1960–2000 Pintang county 1961–2000 Dushan county1951-2000 Guizhou Province |
Monthly mean temperature, rainfall, relative humidity |
Monthly incidence of malaria |
-Correlation analysis |
-Significant relationship between malaria incidence and climatic factors, but the influences of different climatic variables were not consistent among the eight study counties. |
-Relative long study periods |
| -Path analysis |
-Direct and indirect effects of climate were analysed by Path analysis |
|||||
| -The influence of climate on malaria was greater in Libo, Sandu, Dushan counties than in Luodian and Pintang counties | ||||||
| Gao et al. (2003) Chinese [41] |
Yunnan Province 1994-1999 |
Monthly mean temperature, maximum temperature, minimum temperature, rainfall, relative humidity, rain day, evaporation and sunshine hours |
Monthly incidence of malaria |
-Back Propagation Network Model |
-The efficiency of malaria forecasting was 84. 85% based on meteorological variables. |
-Descriptions of associations between malaria and climate was inadequate |
| -A five-years short study period | ||||||
| Wen et al. (2003) Chinese [42] |
Hainan Province 1995-2000 |
Monthly average temperature, maximum temperature, minimum temperature, rainfall, relative humidity |
Monthly incidence of malaria |
-Correlation analysis |
-Temperature and rainfall were relative to malaria transmission with various lag times, but relative humidity was not. |
-Analysis of high epidemic area and the whole province -Social-economic factors were neglected |
| -Stepwise regression analysis | ||||||
| -The influence of climatic variables on malaria was more obvious in high epidemic area than that in the whole province | ||||||
| | ||||||
| Huang et al. (2002) Chinese [43] |
Jiangsu Province 1973-1983 |
Monthly rainfall, rain days, relative humidity, evaporation and NDVI |
Monthly incidence of malaria |
-Correlation analysis |
-The NDVI positively correlated with precipitation and relative humidity. |
-No temperature data included |
| -GIS |
-Only correlation method used |
|||||
| -The NDVI may be a good indicator to predict the distribution and transmission of malaria. | ||||||
| Huang et al. (2001) Chinese [44] |
Gaoan city, Jiangxi Province 1962-1999 |
Annually average rainfall during April to June, annually average temperature during July to August, annual average rainfall and temperature |
Case counts |
-Circular distribution method |
-Malaria cases increased with increase of average temperature from July to August and rainfall from April to June. |
-Annual index were used |
| -Descriptive study | ||||||
| Kan et al. (1999) Chinese [45] |
Anhui Province 1969-1999 |
Annual temperature and rainfall |
Annual incidence of malaria |
-Descriptive study |
-Annual incidences of malaria in 1975, 1977, 1980 in Madian, Lixin County increased with increase of rainfall, while decreased in 1976, 1978, 1981 with decreased rainfall |
-Not enough explanation on effects of climate factors on malaria. |
| -No statistical methods used | ||||||
| Yu et al. (1995) Chinese [46] |
Libo County, Guizhou Province 1958-1993 |
Monthly average temperature, rainfall, relative humidity |
Monthly incidence of malaria |
-Correlation analysis |
-Positive associations between malaria incidence and climatic factors were observed. |
-Relative long study periods |
| -Path analysis | ||||||
| -Direct and indirect effects of climate were analysed |
||||||
| -Direct effect of relative humidity was greatest on malaria incidence compared with temperature and rainfall. | ||||||