Elsevier

Acta Tropica

Volume 166, February 2017, Pages 81-91
Acta Tropica

Modelling the influence of temperature and rainfall on malaria incidence in four endemic provinces of Zambia using semiparametric Poisson regression

https://doi.org/10.1016/j.actatropica.2016.11.007Get rights and content

Highlights

  • In Zambia, the role of climatic factors on malaria, has not been determined in combination of space and time in modelling.

  • The reversal in malaria reduction after the year 2010 and its variation by transmission zones make it critical.

  • Semiparametric Poisson regression modelled a strong positive association between malaria incidence and precipitation.

  • Malaria risk was 95% and 68% lower in Lusaka and Western (APR = 0.05 and 0.31, 95% CI = 0.04 – 0.06 and 0.25 – 0.41 respectively) compared to Luapula.

  • The unique behaviour and effects of minimum and maximum temperatures are an indication of geographical region effects.

Abstract

Although malaria morbidity and mortality are greatly reduced globally owing to great control efforts, the disease remains the main contributor. In Zambia, all provinces are malaria endemic. However, the transmission intensities vary mainly depending on environmental factors as they interact with the vectors. Generally in Africa, possibly due to the varying perspectives and methods used, there is variation on the relative importance of malaria risk determinants. In Zambia, the role climatic factors play on malaria case rates has not been determined in combination of space and time using robust methods in modelling. This is critical considering the reversal in malaria reduction after the year 2010 and the variation by transmission zones. Using a geoadditive or structured additive semiparametric Poisson regression model, we determined the influence of climatic factors on malaria incidence in four endemic provinces of Zambia. We demonstrate a strong positive association between malaria incidence and precipitation as well as minimum temperature. The risk of malaria was 95% lower in Lusaka (ARR = 0.05, 95% CI = 0.04–0.06) and 68% lower in the Western Province (ARR = 0.31, 95% CI = 0.25–0.41) compared to Luapula Province. North-western Province did not vary from Luapula Province. The effects of geographical region are clearly demonstrated by the unique behaviour and effects of minimum and maximum temperatures in the four provinces. Environmental factors such as landscape in urbanised places may also be playing a role.

Introduction

The burden of malaria has greatly reduced globally (Moss et al., 2012) owing to great control efforts. However, the disease remains the main contributor to morbidity and mortality. Malaria was responsible for an estimated 214 million morbidity cases and 438,000 deaths worldwide in 2015, with 88% of the morbidity cases and 90% of the deaths occurring in Africa (World Health Organisation, 2015). Considering the estimated 262 million morbidity cases and 839,000 deaths recorded in 2000, the 2015 statistics represent a decline of 18% in the estimated malaria cases and of 48% in the number of deaths. Nevertheless, the decline in malaria burden has not been uniform among countries in the world as some countries experience resurgence (Moss et al., 2012).

Consistent with the global picture, Zambia reported a decline in malaria cases over the recent years (Masaninga et al., 2012). However, the burden of malaria continues to be high in the country (Chanda et al., 2012). An estimated four million suspected cases were reported in 2007 (Roll Back Malaria, 2011) increasing to an estimated 4,300,000 cases in 2010 (Government of the Republic of Zambia, 2013a). Within the country, all provinces are malaria endemic (Roll Back Malaria, 2011) although the transmission intensities vary mainly depending on environmental factors (Ghebreyesus et al., 1997, Lieshout et al., 2004) as they interact with the vectors. Following the break in interventions due to socio-economic problems in Zambia in 2009, the impact of malaria control efforts was reversed (Masaninga et al., 2012). Nonetheless, the reversal was not uniform across all provinces. The reversal in malaria reduction in Zambia observed during the 2009–2010 period was mainly attributed to decreased donor funding towards malaria control activities (Masaninga et al., 2012).

Malaria is a vector-borne disease and eco-environmental factors such as climate, landscape, housing structure (World Health Organisation, 2015) and proximity of households to vector breeding sites contribute to the burden in two major ways; either by affecting the vector and parasite development (Lieshout et al., 2004) or by facilitating exposure of community members to vectors (Lenntech, 2015). The larval development stage of mosquitoes occurs in various types of water bodies depending on their water-ecological requirements (World Health Organisation, 2001). In Africa, the Anopheles gambiae vector breeds in numerous small pools of water (Centre for Diseases Control and Prevention, 2016). The main sources of the water pools are rainfall (Centre for Diseases Control and Prevention, 2016), artificial means or even natural disasters (Lenntech, 2015). The occurrence of water pools has been shown to expose people to epidemics of flood-linked water borne diseases such as malaria (Lenntech, 2015).

With regards to climate, a number of studies have shown variations in climatic factors influencing malaria. Among them, some suggest minimum temperature only (Nkurunziza et al., 2010), others rainfall only (Huang et al., 2011) while others a combination of minimum, maximum temperature and rainfall (Mabaso et al., 2006). The variation in climatic factors influencing malaria in these studies could largely be due to the different regions experiencing varied intensity of prevailing eco-environmental factors. Eco-environmental factors relate to both the ecology of vector breeding sites as well as the environment including climate and housing structures where humans get exposed to vectors or favouring conditions.

The malaria control program in Zambia is one of the leading programs in Africa (Roll Back Malaria, 2011, Ashraf et al., 2010) although research in eco-environmental factors still has room for expansion. A number of studies have been conducted in Zambia to describe the epidemiology of malaria in the country (Masaninga et al., 2012, Chimumbwa, 2003). These studies have suggested that malaria is endemic throughout the country (Chimumbwa, 2003), varying in transmission intensity across three distinct transmission zones (Masaninga et al., 2012). Studies in Zambia have also indicated a marked change in sibling species composition over time, mainly due to change in ecology. The change in sibling species composition corresponds to malaria transmission rates in some places. Current findings suggest that the malaria vectors in Zambia consist of An. gambiae ss., An. arabiensis, and An. funestus (Chanda et al., 2012). Other studies conducted in Zambia have been focussed on land ecology as it relates to vector development (Ricotta et al., 2014, Clennon et al., 2010). One of such studies used an image processing software called ImageJ, to analyse Google Earth satellite imagery. This procedure was a way to evaluate and enumerate information on vegetation cover such as number of plants, total amount of vegetation and its percentage of the total area and in relation to malaria cases (Ricotta et al., 2014). The study obtained and modeled data to evaluate local vegetation as a risk factor and they obtained proper usable images in southern Zambia using the software (Ricotta et al., 2014). The landscape model was another tool developed for malaria control in Zambia. This tool was successful at ruling out potential locations of vector breeding sites although it was limited in predicting the type of vector species that inhabited the sites (Clennon et al., 2010). An early warning system which conformed to seasonal incidence patterns was also developed in Zambia although it required longer periods of surveillance data to perform better (Davis et al., 2011). Additionally, some climatic variables have also been demonstrated to be significantly related to malaria transmission based on geographic patterns of risk in Zambia, namely: low altitude, high normalised difference vegetation index (NDVI), and high day and night land surface temperatures (Riedel et al., 2010).

In a study by Martens et al. (1995), the authors conclude that future climate scenarios from models can be used to calculate the potential impact of climate change on malaria transmission. This conclusion was in consideration of the significantly differing estimates of future populations at risk of malaria between regions and between climate scenarios (Martens et al., 1995).

Generally in Africa, many perspectives and methods are used to study climate-malaria links (Bouma, 2003), and literature on the matter suggests variation in the relative importance of specific factors

with region (Teklehaimanot et al., 2004). For Zambia no current studies have demonstrated the relative importance of factors over space and time using robust methods in modelling. It is important for the country to undertake these studies considering the reversal in malaria reduction after the year 2010 (Masaninga et al., 2012) and the malaria burden variation by transmission zones.

Patz and Olson (2006) recommend cross-cutting research activities in fields of health, environment, sociology and economics if risk assessment and risk management of the synergistic processes of climate and land-use change are to be improved. (Patz and Olson, 2006). Therefore, this study sought to determine the influence of climatic factors on malaria incidence in four endemic provinces of Zambia.

Section snippets

Study area

Zambia is a landlocked country surrounded by eight countries (Africa-EU Energy Partnership, 2013). The country is located in southern Africa between latitudes −8° and −18° South and longitudes 22° and 34° East (Africa-EU Energy Partnership, 2013). It has a tropical climate with hot episodes at times. The climate in Zambia is generally made up of three seasons: a cool dry season with temperatures ranging from 25° to 30 °C in April to 29° to 33 °C in July, a hot dry season with temperatures ranging

Results

In this study, we explored the dynamics of the evolution of malaria incidence over time in relation to temperature, and precipitation. Fig. 2 presents the space-time evolution of malaria incidence over time in the study region. The results suggest that malaria risk increased significantly over time in all the provinces in the study region.

Table 1 presents summary statistics for the annual malaria risk, maximum temperature, minimum temperature and precipitation. Malaria risk was highest in

Discussion

This study modelled the influence of selected climatic variables on the incidence of malaria in the four provinces in Zambia. We employed a geoadditive or structured additive Semiparametric Poisson regression model and utilised standardised data to avoid the effect of scale. It has been suggested that although scientific evidence with regards to the relationship between vectors and parasites with raw temperature data exists, it is not recommended to use raw temperature data in spatial

Conclusions

Our study proposes a geoadditive or structured additive Semiparametric Poisson regression model to assess climatic factors associated with malaria incidence based on 2009–2012 data in four endemic provinces of Zambia. We demonstrate a strong positive association between malaria incidence and precipitation as well as minimum temperature. The effects of region are clearly demonstrated by the unique behaviour and effects of minimum and maximum temperatures in the four provinces. Environmental

Ethical approval

The study protocol was approved by University of Zambia (UNZA) Biomedical Research Ethics Committee (IRB00001131 of IORG0000774).

Competing interests

The authors as well as the funders declare that they have no competing interests in the manuscript.

Authors’ contributions

All authors conceived the study and contributed to the study design; NMS-M designed the study, collected and analysed the data. SM, ET-M and MG guided in the considerations for data collection and analysis; TNOA led the statistical analysis component; NMS-M drafted the manuscript; All authors contributed to the interpretation and presentation of data and read, edited and approved the final manuscript.

Authors’ information

NMS-M (Nzooma M. Shimaponda-Mataa) is a Medical Parasitologist and Lecturer at the University of Zambia and holds a PhD in Parasitology; SM (Samson Mukaratirwa) is a Parasitologist and Professor, Dean and Head of School and Lecturer at the University of KwaZulu-Natal in Durban, South Africa and holds a PhD in Parasitology; MG (Michael Gebreslasie) is a Geographer (geospatial methods) and Lecturer at the University of KwaZulu-Natal in Durban, South Africa and holds PhD in Geography; ET-M (Enala

Acknowledgements

This work was part of the requirements for the fulfilment of a PhD degree in Parasitology at the University of KwaZulu-Natal, South Africa and was funded partly by University of KwaZulu-Natal.

We further acknowledge the University of Zambia, Zambia National Malaria Control Centre, Ministry of Health Headquarters and Provincial Health offices for approval of the study, logistical and data support; Ministry of Finance, Planning and Economic Development for logistical and data support; Central

References (46)

  • P. Dambach et al.

    Utilisation of combined remote sensing techniques to detect environmental variables influencing malaria vector densities in rural West Africa

    Int. J. Health Geogr.

    (2012)
  • R.G. Davis et al.

    Early detection of malaria foci for targeted interventions in endemic southern Zambia

    Malar. J.

    (2011)
  • K.L. Ebi et al.

    Climate suitability for stable malaria transmission in Zimbabwe under different climate change scenarios

    Clim. Change

    (2005)
  • Food and Agriculture Organisation

    Country Pasture/Forage Resource Profiles, ZAMBIA

    (2009)
  • P.W. Gething et al.

    Modelling the global constraints of temperature on transmission of Plasmodium falciparum and P. vivax

    Parasites Vectors

    (2011)
  • T.A. Ghebreyesus et al.

    Pilot studies on the possible effects on malaria of small-scale irrigation dams in Tigray regional state, Ethiopia

    J. Public Health Med.

    (1997)
  • Government of Zambia

    Central Statistical Office 2012

    (2010)
  • Government of the Republic of Zambia

    MDG Zambia Profile

    (2013)
  • Government of the Republic of Zambia

    MDG Luapula Profile

    (2013)
  • Government of the Republic of Zambia

    MDG Western Profile

    (2013)
  • Government of the Republic of Zambia

    MDG North-western Profile

    (2013)
  • Government of the Republic of Zambia

    MDG Lusaka Profile

    (2013)
  • F. Huang et al.

    Meteorological factors-based spatio-temporal mapping and predicting malaria in Central China

    Am. J. Trop. Med. Hyg.

    (2011)
  • Cited by (0)

    View full text