EXPLORING 23 YEARS DATA OF RAINFALL AND MALARIA POSITIVE FRACTION IN AREAS UNDER KILIFI HEALTH AND DEMOGRAPHIC SURVEILLANCE SYSTEM
Abstract
Malaria has been identified as the disease most likely to be affected by climatic changes.
Identifying climatic factor that influence the incidence of malaria by modifying the
abundance of mosquito populations or the length of the extrinsic parasite cycle in the
mosquito, and the emergence of epidemics in areas of low endemicity is an important step in
controlling episodes of malaria during periods of intense transmission. This study therefore,
focused on one climatic factor, rainfall.
The study was conducted in Kilifi County. It aimed at exploring the patterns of monthly
rainfall and the monthly proportion of children who were positive for malaria parasites in
areas under Kilifi Health and Demographic Surveillance System (KHDSS) for the past 23
years. Analysis of the data was done in R version 3.4.1. Spearman Rank correlation was used
to test for association while Auto Regressive Integrated Moving Average model (ARIMA
model) was used for forecasting.
This study presented a similarity in rainfall patterns in both the South and North regions of
Kilifi County. Spearman rank correlation analysis between rainfall and Malaria Positive
Fraction (MPF) at one, two, three and four months lagged period was conducted and it was
observed that all the lags gave a positive correlation. Even though, there was a weak
relationship between contemporaneous rainfall and Malaria Positive Fraction (MPF), at two
months lag i.e. comparing rainfall 2 months before the MPF, there was a stronger correlation
of rainfall preceding malaria in both the Southern and Northern region of the county with the
Southern region having p-value= 0.1x10-5 and r2= 0.08 while the Northern region had a pvalue=
0.0x10-5 and r2= 0.10. ARIMA model proved to be effective in predicting future
values.
Rainfall is still a threat which is likely to affect spatial and temporal distribution of malaria
even if declining figures of malaria is being recorded at the Coastal region. A 2 months’ time
lag between rainfall and malaria is enough for Anopheline mosquitoes to complete their life
xii
cycle. This information could form an insight basic foundation for policy formulation to lay
down in advance appropriate measures for the control of malaria during periods of intense
transmission.
Keywords: Malaria Positive Fraction (MPF), Rainfall, Kilifi Health and Demographic
Surveillance System (KHDSS), Correlation, Forecast.