污染杂志

Consequences of Climate Variation on Malaria Incidence in Uganda

Abstract

FR Muwanika

Introduction: Temperature and rainfall are assumed to play an important role in the transmission of malaria. According to Zhou simultaneous analysis of the long-term time series of meteorological and parasitological data are critically needed to understand the effects of climate on malaria incidence. However, it would be more plausible to assess the effect of climate variation on the malaria incidence since increase or decrease in the number of malaria cases does not quantify the disease frequency.

Objective: This study therefore, seeks to examine the consequences of variation in climatic factors such as temperature and rainfall on the malaria incidence among the Ugandan population.

Methods: To account for variation and dynamics in monthly malaria incidence among the Ugandan population, data on environmental factors like minimum and maximum temperature, and rainfall were obtained from Uganda National Meteorological Authority (UNMA) while monthly malaria counts for the period (2006-2016) were obtained from Ministry of Health (MoH). Dynamically complete models were used to simultaneously investigate the impact of environmental factors on monthly malaria incidence in the population.

Findings: Findings revealed that the long-run impact of three months successive one percent increase in maximum temperature would result into an 8.1% reduction in monthly malaria incidence. Similarly, three months successive one percent increase in minimum temperature would increase monthly malaria incidence by about 16.7% in the long-run. The impact of rainfall was also significant in that successive one percent three months increase in rainfall would reduce malaria incidence by about 14% in the long-run.

Conclusion: Increasing maximum temperature is associated with a reduction in malaria incidence while increasing minimum temperature is associated with an increase in monthly malaria incidence. Lastly, increasing amount of rainfall is associated with a reduction in monthly malaria incidence.

Recommendations: In order to lower the incidence of malaria among the population, it’s imperative that malaria control and preventive interventions consider environmental modifications that lower minimum temperature and increase the amount of rainfall. These interventions include land use methods such as conversion of land from forests to settlements and agricultural activities, wetlands for settlements and other economic activities.

 Introduction

Uganda experiences stable endemic malaria in 95 % of the areas of altitude 1200 to 1600 m. Malaria transmission is perennial with two peaks, D�?er the rainy season, (April-May) and the period between October-November. Temperatures between 20 and 30 degrees centigrade and relative humidity of 60% provide optimal conditions for malaria transmission. Hence, the temperatures in Uganda which range between 16-36 degrees centigrade provide ideal conditions for malaria transmission. НLs implies that understanding the malaria dynamics in the country requires understanding its link with temperatures and rainfall in a given month. Нerefore, a model that links the monthly malaria cases with rainfall and temperature could be more accurate since these two factors have high correlation with increased disease burden, not only for malaria but also some other diseases. Previous  studies have attempted to link malaria transmission to temperature based on clinical trials mainly in eٹcDc\ studies. According to MacDonald, Bruce-Chwatt and Molineaux suggested that understanding the association between malaria incidence and environmental factors might be the most e�?ٴectLve way of predicting changes in malaria transmission dynamics and thus improve the impact of control e�?ٴorts [1-3]. Furthermore, many studies have attempted to analyze seasonality in malaria incidence and environmental factors using dL�?ٴerent approaches although there has not been a convincing approach that could be used across the continent including Uganda [4-8]. Нerefore, there is limited literature on the dynamics of malaria burden based on routine data obtained in normal health setting that could be used to improve the design of control interventions based on environmental factors in the communities.

In the Eastern part of Africa especially Uganda where malaria is more prevalent during seasons of peak agricultural activities, the disease not only excludes the sick ones from daily agricultural activities but also the healthy ones who have to take care of their sick family members and relatives. According to UBOS over 80% of the Ugandan population is still employed in agricultural sector. Agriculture in most parts of the country still depends on natural climatic conditions of rainfall and environmental temperature [9]. Increased burden of malaria does not only D�?ٴect their health but also their incomes as well as food security at household level. It is estimated that workers su�?ٴerLng from a malaria bout can be incapacitated for 5-20 days. Нe lack of adequate manpower during the peak of agricultural activities decreases productivity and hence, lowers income and aggravates food insecurity [10]. A poor malaria-stricken family may spend up to 25% of its income on malaria prevention and treatment. It is estimated that 40% of health expenditures in Sub-Saharan Africa are spent on malaria treatment [11]. Understanding the dynamics between rainfall and environmental temperature over time will not only help in their forecast but also accurately predict the burden of malaria. НLs study therefore, attempts to assess the variation in environmental factors and their consequences on observed malaria incidence in the population.

Temperature plays an important role in the survival of the parasites in the anopheles mosquito vector. A study by Crag established that these parasites have a short development cycle which lasts between 8 to 21 days. Нe parasites require an optimum temperature ranging between 27-31°C [12]. Furthermore, it was also established that temperatures below 19°C does not favor the Plasmodium falciparum species to complete their cycle and propagate malaria. НLs implies that in a country like Uganda with an average annual temperature ranging from 17°C to 36°C provides the best environment for the Plasmodium falciparum species to complete its cycle and propagate malaria in the population. Нerefore, any model to predict malaria incidence in the population need to incorporate the e�?ٴect of temperature over time [13].

Temperature and rainfall are assumed to play an important role in the malaria transmission. It has been established that rainfall Lnfluences the availability of mosquito larval habitats and thus mosquito demography. According to Zhou simultaneous analysis of the long-term time series of meteorological and parasitological data are critically needed to understand the effects of climate on malaria cases [14]. However, it would be more plausible to assess the effect of climate variation on the malaria incidence since increase or decrease in the number of malaria cases does not quantify the disease frequency. Patz reaffrims that climate variability is epidemiologically more relevant than the mean temperature increase [15]. НLs implies that the use of the epidemiological measure of disease frequency like malaria incidence in studying the effect of temperature and rainfall through an appropriate model will shade some insights on how climate variation is affecting malaria transmission in Uganda. НLs study therefore, seeks to examine the consequences of variation in climatic factors such as temperature and rainfall on the malaria transmission among the Ugandan population [16].

Methods and Materials

Data sources

To account for variation and dynamics in monthly malaria incidence among the Ugandan population, data on environmental factors like minimum temperature, maximum temperature, and rainfall were obtained from Uganda National Meteorological Authority (UNMA). UNMA collects daily maximum and minimum temperatures as well as rainfall from the diffٴerent weather stations across the districts in Uganda. Monthly malaria counts for the period (2006-2016) were obtained from Ministry of Health (MoH). MoH through its health facilities collects routine data using a standard Health Management Information System (HMIS) form. Нe HMIS 105 form is provided at all the health facilities across the country.

Data management and analysis

To explore the variation of malaria incidence and environmental factors, mean and standard deviations of each of the variables were computed. Furthermore, graphs of minimum temperature, maximum temperature, and rainfall and malaria incidence over time were plotted.

Test for stationarity

Each of the variables monthly malaria count, minimum temperature, maximum temperature, and rainfall were assessed to establish whether they were stationary before they were in the model. Augmented Dickey-Fuller test was used to test for stationary in the variables and establish the number of lagged variables in the model.

Multivariate analysis

To simultaneously estimate the e�?ٴect of environmental factors and monthly malaria incidence in the population, dynamically complete models were used. НLs framework was adopted to investigate the dynamics in both malaria incidences as well as in the environmental factors in predicting monthly malaria incidence among the population in Uganda because both malaria incidence and climate factors vary with time.

Model specification

where, A(L) is defined as the lagged values of monthly malaria incidence, B(L) is the lagged values of minimum temperature, C(L) is the lagged values of maximum temperature, D(L) is the lagged values of rainfall respectively.

Model Diagnostics

Нe dynamically complete model was assessed for normality of the errors, serial correlation and model inadequacies. For normality of the errors, a plot of the residuals against the fitted values of monthly malaria incidence was done. A random scattering of the points above and below the line with nearly all the data points being within the band defined by 2-standard deviations is expected if the assumptions are sDtLsfied. Нe model inadequacies are LdentLfied if the plot of the residuals against the fitted values gives asymmetric shape about zero.

Нe presence of a serial correlation in the residual was assessed using the run test. Нe test involved counting the number of runs or sequence of positive and negative residuals and comparing this result to expected number of runs under the null hypothesis of independence.

Results

Descriptive analysis

Table 1 shows the annual average malaria cases, rainfall, maximum temperature and minimum temperature observed across Uganda for the period 2006-2016. Нe average monthly malaria cases observed for the period 2006-2016 was about 1.1 million cases. Нe average monthly amount of rainfall observed was 117.6 mm. Нe average observed monthly minimum and maximum temperature were 17.3°C and 28.7°C respectively as seen in Table 1.