- Permanent Link:
- Impacts of Socioeconomic Factors on Malaria Transmission: Urbanization, Armed Conflicts and Economic Conditions
- Qi, Qiuyin
- University of Florida
- Publication Date:
- Doctorate ( Ph.D.)
- Degree Grantor:
- University of Florida
- Degree Disciplines:
- Committee Chair:
- Committee Co-Chair:
- FIK,TIMOTHY J
- Committee Members:
- Graduation Date:
- Subjects / Keywords:
- Area surveys ( jstor )
Cities ( jstor )
Countries ( jstor )
Datasets ( jstor )
Diseases ( jstor )
Economic conditions ( jstor )
Falciparum malaria ( jstor )
Malaria ( jstor )
Rank tests ( jstor )
Urbanization ( jstor )
- General Note:
- Over the past decade, although malaria transmission has been substantially reduced by the scaling up of malaria interventions, it remains a major global health problem. As a vector-borne disease, the spatial distribution and intensity of malaria transmission are determined by a range of socioeconomic and environmental factors. Yet few have quantitatively addressed the impacts of these factors, particularly the socioeconomic ones, on the transmission of malaria over large spatial scales, mostly due to the lack of globally consistent disease and socioeconomic variable data. This dissertation aims to fill these gaps by focusing on the effects of three socioeconomic factors on malaria transmission, utilizing recently assembled global Plasmodium falciparum parasite prevalence (PfPR) and Plasmodium vivax parasite prevalence (PvPR) data. Specifically, a range of global datasets were assembled to 1) examine the effects of urbanization on global P. vivax malaria transmission at different spatial scales and by dominant vector species; 2) explore the association between armed conflicts and P. falciparum malaria transmission in Africa; 3) investigate the relationship between economic conditions and malaria prevalence at a variety of spatial scales.
- Source Institution:
- Rights Management:
- Copyright Qi, Qiuyin. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
- Embargo Date:
This item has the following downloads:
1 IMPACT S OF SOCIOECON O MIC FACTORS ON MALARIA TRANSMISSION: URBANIZATION, ARMED CONFLICTS AND ECONOMIC CONDITIONS By QIUYIN QI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 201 4
2 2014 Qiuyin Qi
3 To m y family
4 ACKNOWLEDGMENTS I would like to acknowledge and thank the following people for their continuous sup port, encouragement and advice d uring my time as a doctoral student at University of Florida. First, I want to expr ess my sincerely gratitude to Dr. Andrew Tatem, my doctoral committee chair and advisor. W ithout his ongoing guidance, great patience in reading my dissertation drafts and insightful responding to my questions, this dissertation would not be possible. I t has been a privilege to work with such a c onsiderate and brilliant person, who has been always ready and available to help me with various academic problems. I offer special thanks to Dr. Timothy Fik, my committee co chair, who has been so generously guiding, supporting and helping me throughout m y doctoral study. I also want to thank Dr. Liang Mao and Dr. Xiaohui Xu, my committee members. T heir valuable suggestions and feedbacks to my dissertation make this a quality product. M any thanks to my friends, fellow graduate students and colleagues for all the support, encouragement, inspiring ideas and amazing time spent together. M y graduate life at University of Florida would be different without them. L astly, I would like to thank the members of my family for their unconditional love, respect and encouragement. They are so important to me and never ceased to support and love me. Very special thanks to my husband for all the support, dedication and motivation to my dissertation.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 9 LIST OF ABBREVIATIONS ........................................................................................... 11 ABSTRACT ................................................................................................................... 12 CHAPTER 1 INTRODUCTION .................................................................................................... 13 2 THE EFFECTS OF URBANIZATION ON GLOBAL PLASMODIUM VIVAX MALARIA TRANSMISSION .................................................................................... 18 Chapter Summary ................................................................................................... 18 Background ............................................................................................................. 19 Data ........................................................................................................................ 21 The MAP Pv PR D atabase ................................................................................ 21 Global U rb an M ap ............................................................................................ 22 Dominant Anopheles V ector M aps ................................................................... 23 Methods .................................................................................................................. 23 Urbanization and P. viva x M alaria T ransmission .............................................. 23 Dominant Anopheles V ectors ........................................................................... 25 Results .................................................................................................................... 25 Urbanization and P. viva x M alaria T ransmission .............................................. 25 Dominant Anopheles V ectors ........................................................................... 27 Discus sion .............................................................................................................. 28 3 ARMED CONFLICTS AND PLASMODIUM FALCIPARUM MALARIA TRANSMISSION IN AFRICA .................................................................................. 40 Chapter Summary ................................................................................................... 40 Background ............................................................................................................. 41 Data ........................................................................................................................ 44 The M AP Pf PR D atabase ................................................................................. 44 Pf Rc M ap .......................................................................................................... 45 Armed C onflict L ocation and E vent D ata .......................................................... 45 Methods .................................................................................................................. 46 Armed C onflicts and P. f alciparum M alaria T ransmission ................................ 47 Space time C lusters of A rmed Co nflict ............................................................. 48
6 Results .................................................................................................................... 49 Armed C onflicts and P. f alciparum M alaria T ransmission ................................ 49 Space time C lusters of A rmed C onflict ............................................................. 50 Discussion .............................................................................................................. 52 4 A GLOBAL ASSESSMENT OF ECONOMIC CONDITIONS AND MALARIA PREVALENCE ........................................................................................................ 65 Chapter Summary ................................................................................................... 65 Background ............................................................................................................. 66 Data and Methods .................................................................................................. 67 P. falciparum M alaria E ndemicity M ap ............................................................. 67 Geographically B ased Economic D ata ............................................................. 68 Satellite Night Time Lights ( NTL ) B ased E conomic D ata ................................. 69 Analyses ........................................................................................................... 70 Model S pecification .......................................................................................... 71 Results .................................................................................................................... 74 Discussion .............................................................................................................. 78 5 CONCLUSION ........................................................................................................ 86 APPENDIX A SUPPLEMENTARY TABLE S FOR CHAPTER 2 .................................................... 90 B SUPPLEMENTARY TABLE S FOR CHAPTER 3 .................................................... 93 C SUPPLEMENTARY FIGURE S FOR CHAPTER 4 .................................................. 95 LIST OF REFERENCES ............................................................................................. 104 BIOGRAPHICAL SKETCH .......................................................................................... 121
7 LIST OF TABLES Table page 2 1 Summary of the Pv PR surveys by region ........................................................... 34 2 2 Results of Wilcoxon Signed Rank tests on Pv PR values between GRUMP UE defined urban and rural survey pairs for countries, regions and the World ... 35 2 3 Robustness analysis of the Wilcoxon Signed Rank tests on urbanrural Pv PR value pairs derived from various spatial and temporal limits .............................. 36 3 1 Results of Wilcoxon Signed Rank tests on Pf PR values between before and after conflict survey pairs across Africa and by country ...................................... 5 9 3 2 Results of Wilcoxon Signed Rank tests on Pf PR values between before and after conflict survey pairs by Pf Rc class .............................................................. 59 3 3 Results of Wilcoxon Signed Rank tests on Pf PR values between before and after conflict survey pairs by country with different time lim its ............................ 60 3 4 Space time clusters of armed conflicts ............................................................... 61 4 1 Estimation results of malaria models .................................................................. 82 4 2 Estimation results of economic models .............................................................. 82 A 1 Results of Wilcoxon Signed Rank tests on Pv PR values between GRUMP UE defined urban and rural survey pairs for the dominant Anopheles vectors in Asia Pacific region .......................................................................................... 90 A 2 Results of Wilcoxon Signed Rank tests on Pv PR values between GRUMP U E defined urban and rural survey pairs for the dominant Anopheles vectors in Africa, Europe and the Middle East ................................................................ 91 A 3 Results of Wilcoxon Signed Rank tests on Pv PR values bet ween GRUMP UE defined urban and rural survey pairs for the dominant Anopheles vectors in the Americas ................................................................................................... 91 A 4 Results of Wilcoxon Signed Rank tests on Pv PR values between MODI S defined urban and rural survey pairs for continents, countries and the World .... 92 B 1 Results of Wilcoxon Signed Rank tests on Pf PR values with conflict pair s that occurred during ETa Anomalies excluded .......................................................... 93 B 2 Results of Wilcoxon Signed Rank tests on Pf PR values between before and after conflict survey pairs by conflict type ........................................................... 93
8 B 3 Results of Wilcoxon Signed Rank tests on Pf PR values between before a nd after conflict survey pairs by conflict level ........................................................... 94
9 LIST OF FIGURES Figure page 1 1 The epidemiology and ecology of malaria transmission ..................................... 17 2 1 The global spatial limits of P. vivax malaria transmission in 2009. ..................... 37 2 2 Boxplots showing the differences in Pv PR values between GRUMP UE defined urban and rural surveys for cities. .......................................................... 38 2 3 Bar charts showing the test results for the dominant Anopheles vectors of human malaria . .................................................................................................. 39 3 1 P lasmodium falciparum malaria endemicity in 2010 ........................................... 62 3 2 Space time clusters of armed conflicts. .............................................................. 63 3 3 Boxplots showing the differences in Pf P R values among surveys taken within five years before, during and within five years after the conflict clusters ............ 64 4 1 Geographically based and night time light based economic data ....................... 83 4 2 Boxplots showing the di fferences in lnP and lnG among a reas with no, unstable and stable risk a n d low intermediate and high risk o f m a l a r i a globally a n d b y r e g i o n ........................................................................................ 84 4 3 Plots showing the lnP differences among areas with no, unstable and stable, and low, intermediate and high risk of Plasmodium falciparum malaria by country ................................................................................................................ 85 C 1 Countries with highest and lowest GDP per capita and Plasmodium falciparum malaria endemicity ............................................................................ 96 C 2 Boxplots showing the differences in lnP among areas with no, unstable and stable risk of Plasmodium falciparum malaria by country ................................... 97 C 3 Boxplots showing the differences in lnP among areas with low intermediate and high risk of Plasmodium falciparum malaria by country ............................... 98 C 4 Boxplots showing the differences in lnG among areas with no, unstable and stable risk of Plasmodium falciparum m alaria by country ................................... 99 C 5 Boxplots showing the differences in lnG among areas wit h low intermediate and high risk of Plasmodium falciparum malaria by country ............................. 100 C 6 Boxplots showing the differences in NLDI among areas with no, unstable and stable risk of Plasmodium falciparum malaria globally and by region. .............. 101
10 C 7 Boxplots showing the differences in NLDI among areas with no, unstable and stable risk of Plasmodium falciparum malaria by country ................................. 102 C 8 Boxplots showing the differences in NLDI among areas with low inte rmediate and high risk of Plasmodium falciparum malaria globally, by region and by country ....................................................................................... 103
11 LIST OF ABBREVIATIONS ACLED Armed Conflict Location and Event Dataset Africa+ Africa, Saudi Arabia and Yemen CSE Asia Central and South East Asia DRC Democratic Republic of the Congo DVS dominant vector species GCP Gross Cell product GDP Gross Domestic Product G Econ Geographically based Economic data GRUMP Global Rural Urban Mapping Project GRUMP UE Global Rural Urban Mapping Project urban extent ITNs insecticide treated nets MAP Malaria Atlas Project NLDI Night Light Development Index NTL night time light Pf PR Plasmodium falciparum parasite rate Pf PR 2 10 P. falciparum parasite rate in 2 10 years old Pf API P. falciparum annual parasite incidence Pf MECs P. falciparum malaria endemic countries Pf R 0 P. falciparum basic reproductive number Pf R c P. falciparum basic reproductive number under control Pv PR Plasmodium vivax parasite rate Pv API P. vivax annual parasite incidence Pv MECs P. vivax malaria endemic countries UPDF Uganda People's Defense Force
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IMPACT S OF SOCIOECON O MIC FACTORS ON MALARIA TRANSMISSION: URBANIZATIO N, ARMED CONFLICTS AND ECONOMIC CONDITIONS By Q i u y i n Q i May 2014 Chair: Andrew J Tatem C ochair: Timothy J Fik Major: Geography Over the past decade, although malaria transmission has been substantially reduced by the scaling up of malaria interventions, it remains a major global health problem. As a vector borne disease, the spatial distribution and intensity of malaria transmission are determined by a range of socioeconomic and environment al factors. Yet few have quantitatively addressed the impacts of these factors, particularly the socioeconomic ones, on the transmission of malaria over large spatial scale s, mostly due to the lack of global ly consistent disease and socioeconomic variable dat a This dissertation aims to fill the se gaps by focusing on the effects of three socioeconomic factors on malaria transmission, utilizing recently assembled global P lasmodium falciparum parasite prevalence ( Pf PR) and P lasmodium vivax parasite prevalence ( Pv PR) data. Specifically, a range of global datasets were assembled to 1) examine the effects of urbanization on global P. vivax malaria transmission at different spatial scales and by dominant vector species; 2) explore the association between armed conflicts and P. falciparum malaria transmission in Africa ; 3) investigate the relationship between economic conditions and malaria prevalence at a variety of spatial scales
13 CHAPTER 1 INTRODUCTION Malaria is a vector borne disease caused by parasites of the genus Plasmodium and transmitted to humans by the bites of female Anopheles mosquitoes. These insects inject a life stage of the parasite called sporozoites into the human body, where they replic ate in the liver, and then infect the red blood cells, causing symptoms including fever, headache and vomiting. Five species of Plasmodium can affect humans and P. falciparum is the most common and clinically serious parasite species which is responsible for the majority of malaria deaths. P. vivax is less dangerous but more widespread, and the other three species ( P. ovale, P. malariae and P. knowlesi ) are found much less frequently . As a major global public health problem, nearly half of the worlds population (3.3 billion) is at risk of malaria infection. An estimated 300 to 500 million clinical malaria cases occur every year, and ap proximately 650,000 deaths, mostly children under the age of five in subSaharan Africa . The epidemiology and ecology of vector borne disease s can be described as a triangle of host (human or animal), pathogen and vector populations interacting within a permissive environment [3 4] (Figure 1 1 ). As a vector borne disease, the spatial distribution and intensity of malaria transmission are determined by the complex interactions between human host, arthropod vectors and pathogens that are affected by a range of socioeconomic and environmental factors. Many recent reviews have aimed to identify primary drivers for the emerging, reemerging and transmission of vector borne and other infectious diseases, includi ng, amongst others, climate change, land cover change, urbanization, agricultural development, poverty, human movement, armed conflicts [5 8] However, few have quantitatively addressed the impact of these
14 factors on the transmission of malaria at large scales, mostly due to the lack of global ly consistent disease and socioeconomic variable dat a Therefore, this research aims to fill some of the se gaps by focusing on the effects of three socioeconomic factors (urbanization, armed conflicts and economic conditions ) o n malaria transmission, m aking use of recently assembled global P. falciparum parasi te rate ( Pf PR) and P. vivax parasite rate ( Pv PR) database s [9 11] Ch apter 2 focuses on the effects of urbanization on global P. vivax malaria transmission. The need for accurate and contemporary descripti ons of populations at risk has lead to several attempts to quantify the impact of urbanization on P. falciparum malaria transmission and a general trend of reduced transmission in urban areas was found [12 13] However, knowledge is lacking regarding the relationship between urbanization and P. vivax malaria, which is the most widely distributed malaria species  G eo referenced Pv PR surveys and urban extent maps were integrated to examine the relationships between Pv PR in urban and rural settings at various spatial scales (global, regional, national and at city level). Furthermore, the regions of highest P. vivax transmission in Asia are composed of a considerably greater range of vector species and species complexes than seen in Africa [15 17] where P. falciparum malaria is principally concentrated, and urbanization may impact each of these vector species differently. Therefore, distribution maps of dominant vector species were used to explore the effects of urbanization on P. vivax transmission by dominant vector species to discern whether differential impacts we re evident. Chapter 3 focuses on the association between armed conflicts and P. falciparum malaria transmission in Africa. Malaria is often a significant health problem during and
15 after armed conflicts as large numbers of vulnerable people are displaced to environment s that favor vector breeding, malaria intervention programs are disrupted, health systems collapse and access to medical supplies is hindered  Several studies have investigated the impacts of conflicts on malaria transmission  and strategies for malaria interventions in conflict affected regions  However, there is a lack of research quantitatively investigating the associations between armed conflicts and malaria transmission over large areas. Here, geo referenced Pf PR surveys and armed conflict events datasets were integrated to examine the associations between armed conflicts and P. falciparum mal aria transmission in Africa which is the continent with the most serious P. falciparum malaria problems and most affected by armed conflicts over the past two decades. Furthermore, areas within significant armed conflict spatial/temporal clusters were identified to explore P. falciparum malaria prevalence before, during and after specific major conflict events. Chapter 4 focuses on the relationship between economic conditions and malaria prevalence. Malaria is consistently described as a disease of poverty with the global distribution of malaria transmission showing a striking coincidence with poverty  The relationship between malaria and poverty, however, is complex and can be explained in several ways. On the one hand, poverty may increase people s susceptibility to malaria infection, consequently sustai ning the transmission  On the other hand, malaria causes economic burden and social costs on individuals and government s  Furthermore, the causal relationship may work in both directions. However, the findings from existing studies are not consistent regarding the causal relationship  and only a few recent studies have explored this relationship over large
16 scales and considered possible endogeneity quantitatively. Here, gridded economic datasets and malaria endemicity maps were used to investigate the relationship between economic conditions and malaria endemic ity classes at various spatial scales. Furthermore, econometric mod els (OLS, 2SLS and 3SLS) were developed to quantify the magnitude and direction of t his causal relationship, by including a set of primary covariates and taking endogeneity into account.
17 Figure 1 1. The epidemiology and ecology of malaria transmission
18 CHAPTER 2 THE EFFECTS OF URBANIZATION ON GLOBAL PLASMODIUM VIVAX MALARIA TRANSMISSION Chapter Summary Many recent studies have examined the impact of urbanization on Plasmodium falciparum malaria endemicity and found a general trend of reduced transmission in urban areas. However, none has examined the effect of urbani zation on Plasmodium vivax malaria, which is the most widely distributed malaria species and can also cause severe clinical sy ndromes in humans. In this chapter a set of 10,003 community based P. vivax parasite rate ( Pv PR ) surveys are used to explore the relationships between Pv PR in urban and rural settings The Pv PR surveys were overlaid onto a map of global urban extents to derive an urban/rural assignment. The differences in Pv PR values between urban and rural areas were then examined. Groups of Pv PR surveys inside individual city extents (urban) and surrounding areas (rural) were identified to examine the local variations in Pv PR values. Finally, the relationships of Pv PR between urban and rural areas within the ranges of 41 dominant Anopheles vectors were examined. Significantly higher Pv PR values in rural areas were found globally. The relationship was consistent at continental scales when focusing on Africa and Asia only, but in the Americas, significantly lower values of Pv PR in rural areas were found, though the numbers of surveys were small. Moreover, except for the countries in Americas, the same trends were found at national scales in African and Asian countries, with significantly lower values of Pv PR in urban areas. However, the patterns at city scales among 20 specific cities where sufficient data were available were less clear, with seven cities having significantly lower Pv PR values in urban areas and two cities showing significantly lower Pv PR in rural areas The urbanrural Pv PR differences
19 within the ranges of the dominant Anopheles vectors were generally, in agreement with the regional patterns found. Except for the Americas, the patterns of significantly lower P. vivax transmission in urban areas have been found globally, regionally, nationally and by dominant vector species here, following trends observed previously for P. falciparum To further understand these patterns, more epidemiological, entomological and parasitological analyses of the disease at s maller spatial scales are needed. Background The world population has undergone unprecedented growth along with rapid urbanization. Slightly more than 50% of the population (3.4 billion) is now living in urban areas compared to only 30% (0.7 billion) in 1 950  By 2050, it is projected that urban dwellers will account for approximately 67% (6.3 billion) of the world total population, while most of the estimated growth will be concentrated in less developed regions, particularly in Asia and Africa  These substantial transitions have significant public health implications associated with changes in the social and physical environment and access to public health services [26 30] Although large heterogeneity exists, it is commonly accepte d that the process of urbanization reduces malaria transmission, primarily because urban environments (e.g. the lack of suitable breeding sites, the pollution of existing larval habitats, etc.) are generally unsuitable for malaria vectors [13, 3132] Other explanations include better access to health care services and an increased ratio of humans to mosquitoes [31, 33 34] However, there is concern regarding urban malaria in less developed regions, typically those undergoing rapid and unprecedented urbanization [35 36] Between the two dominant parasite species of human malaria, Plasmodium falciparum has attracted the focus of most research because of its hi gh mortality and
20 intensive transmission in Africa  Plasmodium vivax malaria, in contrast, is commonly considered as a benign infection and largely overlooked by researchers, government, and funding agencies. Increasing evidence has shown that P. vivax is neither rare nor benign, however. It is estimated that 2.85 billion people were at risk of P. vivax infection in 2009, with 91% (2.59 billion) of them living in Central and South East Asia  and that P. vivax is the most widely distributed (geographically) malaria species of humans. Furthermore, although the infection with P. vivax malaria is rarely directly fatal, it can cause severe clinical syndromes [38 39] Recent studies have examined the impact of urbanization on P. falciparum malaria endemicity and disease burden estimation [12 13, 31 32, 36] Various urban extent maps have been used to compare the differences in P. falciparum malaria endemicity between urban settlements and rural areas  ; exclude the urban extents of cities identified as malaria free in the mapping of malaria tra nsmission limits [40 41] ; downgrade endemic classes in estimates of malaria burden [13, 42] ; and predict P. falciparum malaria endemicity based on geostatistical models [43 44] However, according to our best knowledge, no known research has examined the effect of urbanization on P. vivax malaria over similarly large sca les. In addition, the regions of highest P. vivax transmission in Asia are composed of a considerably greater range of vector species and species complexes than seen in Africa, where P. falciparum transmission is principally concentrated [15 17, 45] and urbanization may impact each of these vector species differently, dependent on their preferences and bionomics. For example, Anopheles culicifacies was reported to be the vector responsible for 6065% malaria cases in urban environments of India  and shows significant environment
21 tolerance and adaptability [47 48] while larvae of Anopheles stephensi were found in various domestic containers and collections of water related to construction and industrial sites in cities [49 50] Therefore, there is a need to examine the effects of urbanization on P. vivax transmission by dominant vector species to discern whether differential impacts are evident. Here georeferenced P. vivax parasite rate ( Pv PR) surveys and urban extent maps are integrated to examine the impact of urbanization on P. vivax malaria transmission at various spatial scales (global, regional, national and at the city level). Furthermore, distribution maps of dominant Anopheles vectors are used to explore the relationships between urbanization, Anopheles vectors and P. vivax malaria transmission. Data The MAP Pv PR D atabase As with P. falciparum malaria, parasite rate (PR) is the most commonly reported and consistent metric of P. vivax malaria endemicity  A total of 10,003 community based Pv PR surve ys taken between 1985 and 2010 were obtained by the Malaria Atlas Project (MAP  ). The logistically intensive process of searching for, identifying and geo locating the Pv PR surveys has been documented elsewhere  All these P v PR surveys were georeferenced to precise locations and not duplicated within three months at the same site. A summary of some of the key features of the Pv PR survey data is presented in Table 2 1. Of the surveys, 410 (4.1%) were in America, 1651 (16.5%) in Africa, Saudi Arabia and Yemen (Africa+), and 7942 (79.4%) in Central and South East Asia ( CSE Asia ). Approximately half (51%) of the Pv PR values are zero and the majority of the surveys were undertaken after 2000. The sample sizes of these
22 surveys varies, with most of them (76%) are being larger than 50. Among the 95 P. vivax malaria endemic countries ( Pv MECs)  Pv PR data were available for 53 (12 in Americ a, 19 in Africa+ and 22 in CSE Asia ). There ar e 8588 discrete Pv PR survey locations and the distribution of them are shown in Figure 2 1, overlaid on the international limits of P. vivax malaria transmission  with most of the survey points located in Southeast Asia and the Horn of Africa. Global U rban M ap Although urbanization has been one of the most important transformations of our world for decades, there is still little consensus on the definitions of what consists an urban area and urbanization among nation al and international bodies  Such ambiguity has lead to the construction of several global urban maps (e.g., Digital Chart of the World (DCW)  Global Rural Urban Mapping Project (GRUMP)  Advance d Very High Resolution Radiometer (AVHRR) Global Land Cover Classification urban land cover class  Defense Meteorological Satellite Program Operational Linescan System (DMSP OLS)  and Moderate Resolution Imaging Spectroradiometer (MODIS) Land Cover Product Binary Data [56 57] ) derived from satellite imagery  As discussed in Tatem et al  all of these global urban maps demonstrated a different range of inaccuracies and limitations, however, the GRUMP urban extent map was considered to be most accurate in matching original urban assignments o f P. falciparum malaria surveys  Therefore, the GRUMP urban extent (GRUM P UE) map was used here to distinguish Pv PR surveys taken in urban areas from those in rural areas. This urban extent map was developed by the Centre for International Earth Science Information Network (CIESIN) at 1km 1km spatial resolution in 2004, util izing
23 information from satellite night time lights (NTL), Landsat satellite sensor imagery and other geographical data  Dom inant Anopheles V ector M aps The distributions and bionomics of dominant Anopheles vectors play an important role in malaria transmission and are the targets of vector control  Vector species normally display a range of ecological and behavioral characteristics. For example, unlike other malaria vectors, an urban environment is favored by the urban vector Anopheles stephensi  To assess the impact of urbanization on P. vivax malaria transmission by dominant Anopheles vectors, expert opinion distribution maps of global dominant vector species (DVS) of malaria were obtained from the M AP [10, 1517] These maps were constructed through exhaustive searches of literature and refinement through opinion and experience by Anopheles experts  A total of 41 maps of DVS were available, of which, 19 were in the AsiaPacific region  13 in Africa, Europe and the Middle East  and nine in the Americas  Methods Urbanization and P. vivax M alaria T ransmission To quantify the patterns of P. vivax malaria transmission between urban and rural areas at global, regional and national scales, sets of spatially and temporally associated urbanrural pairs of Pv PR values were obtained and tested. Firstly, all the georeferenced Pv PR surveys were overlaid onto the GRUMP UE map to derive an urban/rural assignment. Following previous approaches  for each Pv PR survey assigned as urban, all the rural Pv PR surveys taken within 100km and five years were identified. Then, the identified rural Pv PR values were averaged and assigned to that urban Pv PR survey to make spatially and temporally associated urbanrural Pv PR value
24 pairs  Given the highly skewed distribution of Pv PR values in the MAP database  the Wilcoxon signed rank  a nonparametric test for paired variables, was used to d etermine if significant differences between Pv PR values in urban and rural areas existed. These tests were undertaken globally, by region (Africa+, Americas, CSE Asia ) and by country (those for which at least ten urbanrural Pv PR survey pairs existed) to e xamine if the patterns of P. vivax malaria transmission between urban and rural areas were significant1As the choice of spatial and temporal limits (100km and five years) is arbitrary in obtaining urbanrural pairs of Pv PR values, a robustness analysis w as conducted. Sets of urban rural Pv PR pairs were obtained through applying various spatial and temporal limits (100km and two years; 50km and five years; 50km and two years), and tested under the Wilcoxon Signed Rank test, respectively. In addition, the m ean number of rural surveys paired to each urban survey and the overlap rate ( surveys paired to each urban survey / total number of rural surveys) for each spatial and temporal limit were calculated to assess the effects of overlapping rural surveys in the sample pairs. To examine local variations (city scale) in Pv PR, groups of Pv PR surveys inside individual city extents (urban) and surrounding areas (rural) were identified and tested. Cities where more than eight Pv PR surveys (to provide a reasonable number of cities for testing) fell inside their urban extents were first identified. For each city, rural Pv PR surveys that fell within 100km of the centroid of the urban extent were found and 1 Multiple testing adjustment s are not necessary here as our test s do not incorporate any simultaneous hypotheses for an overall result and there is no need to partition the rejection regions. Although the Pv PR su rveys may be used multipl e times in the analysis, t hey were averag ed to create urbanrural pairs of Pv PR values
25 assigned to that city. Following this, for each c ity, Pv PR values within its urban extent and surrounding rural area were compared and tested using the Wilcoxon rank sum test. Dominant Anopheles V ectors The impact of urbanization on malaria endemicity may vary by dominant Anopheles vectors of human malaria. To test this, Pv PR values between urban and rural areas within the extents of 41 dominant Anopheles vector were examined. Sets of spatially and temporally associated urbanrural pairs of Pv PR values within the extents of each dom inant Anopheles vector were extracted and tested separately. The georeferenced Pv PR surveys were firstly overlaid onto the GRUMP UE map to derive an urban/rural assignment. For each dominant Anopheles vector, all the geo referenced Pv PR surveys that fell within its extent were extracted. For each urban Pv PR survey, all of the rural Pv PR surveys taken within 100km and five years were again identified, averaged and assigned to that urban Pv PR survey to make a set of spatially and temporally associated urbanrural Pv PR value pairs  This set of urbanrural Pv PR value pairs were then subject to the Wilcoxon signed rank test to determine if significant differences in Pv PR between urban and rural areas existed. Results Urbanization and P. vivax M alaria T ransmission Among the Pv PR surveys, 1260 were classified as urban and 8973 were cl assified as rural based on the GRUMP UE map (Table 2 1). The mean sample size was 278 for the urban surveys and 230 for the rural surveys, which are comparable. Table 2 2 shows the results of the Wilcoxon Signed Rank tests between urban and rural pairs of Pv PR values defined by GRUMP UE. Significantly higher Pv PR values in rural areas were found globally and in the Africa+ and CSE Asia regions, while in the
26 Americas, significantly lower values of Pv PR in rural areas were found. The Z values indicate, however, that the differences observed in the Americas are weaker than in other regions. Moreover, the numbers of surveys available were much smaller in the Americas. Those countries with at least ten urbanrural Pv PR value pairs and the other countries combined for each reg ion (Africa+, Americas and CSE Asia ) were tested further and the results are presented in Table 2 2. The trends found in most of the countries in Africa+ ( Ethiopia, Yemen ) and CSE Asia ( Bangladesh Indonesia, India Cambodia, Nepal Thailand, Vietnam ) were consistent with the global and regional findings, with significantly lower values of Pv PR in urban areas. The relationships found between urban and rural Pv PR values for the other countries in Africa+ ( Sudan and ot her African countries) and CSE Asia ( Afghanistan, China Pakistan and other Asian countries ) were not significant. There are two countries (Ghana and Zambia) in Africa that have sufficient Pv PR surveys but are of entirely zero values, so were not listed. The results for the Americas are certainly not as conclusive as the relationships found in the other regions, with one country (Brazil) showing significant higher urban Pv PR values, another country (Mexico) showing the reverse and the other American countries showing insignificant differences, though each were only based on a small number of Pv PR pairs. The robustness analysis (Table 2 3) suggests that the overlap rate of rural surveys decreases as the spatial and temporal limits contract, while the patterns of Pv PR between urban and rural areas at global and regional scales are generally consistent. Thus, the method used to determine the relationship of P. vivax malaria
27 transmission between urban and rural areas is robust and the effects of overlapping rural surveys on the results ar e minimal. Figure 2 2 shows the boxplots for urban and rural Pv PR surveys for individual cities whose extents were defined by the GRUMP UE. The results indicate that the patterns among the 20 cities examined were less consistent with the global, regional and national patterns found. Seven cities (Alamata, Ethiopia; Jakarta, Bat am, Kupang, Jambi and Ambon, Indonesia; Rourkela, India) were found to have significantly lower Pv PR values in their urban extents than the surrounding rural areas; two cities (Qandahar, Afghanistan; Ariquemes, Brazil) were found to have significantly lower Pv PR values in their surrounding rural areas (though again, the numbers of surveys were small). The remainder were either insignificant or of zero Pv PR values. Dominant Anopheles V ectors Figure 2 3A presents the results of Wilcoxon Signed Rank tests on P vPR values between GRUMP UE defined urban and rural survey pairs stratified by the dominant Anopheles vectors of human malaria in the AsiaPacific region. In this region, the patterns of lower P. vivax malaria transmission in urban areas are noticeable and consistent, with significantly higher rural Pv PR values found for most of the dominant Anopheles vector distributions (17 out of 19). Furthermore, insignificant differences between urban and rural areas ( Anopheles balabacensis and Anopheles lesteri ) were found in regions with small numbers of survey pairs. Figure 2 3B shows the results of Wilcoxon Signed Rank tests on Pv PR values between urban and rural survey pairs stratified by the dominant Anopheles vector distributions in Africa, Europe and the Middle East. Pv PR s urveys were only available for nine (of the 13) dominant Anopheles vectors. The consistent patterns of lower Pv PR
28 values in urban areas are not as evident as in AsiaPacific region. The differences of Pv PR between urban and rural areas are found to be stat istically significant for only four (out of 9) dominant Anopheles vectors ( Anopheles arabiensis Anopheles funestus Anopheles nili# and Anopheles sergentii ). The others were insignificant, while two of th em ( Anopheles melas and Anopheles sacharovi ) have insufficient number of Pv PR surveys. Figure 2 3C presents the results of Wilcoxon Signed Rank tests for analyses stratified by dominant Anopheles vectors in the Americas. For two (out of nine) of the dominant Anopheles vectors no Pv PR surveys fell within their extents. Unlike the patterns exhibited in the other regions, consistently higher Pv PR values in urban areas were observed in this region, with most of the dominant Anopheles vectors ( Anopheles albitars is Anopheles darlingi Anopheles marajoara and Anopheles nuneztovari ) showing significantly higher urban Pv PR values. However, the numbers of survey pairs in this region are generally small. More detailed statistical results for the three regions are prov ided in Appendix A Discussion The rapid urban transformation of the developing world  has and will continue to have a profound influence on the malaria landscape. The need for accurate and contemporary descriptions of populations at risk (PAR) has lead to several attempts to quantify the impact of urbanizati on on P. falciparum malaria transmission [12 13, 36] Knowledge is lacking however regarding the relationship between urbanization and P. vivax malaria transmis sion. In this study, the most contemporary and comprehensive database of Pv PR surveys was used to explore the differences in P. vivax transmission between urban and rural areas.
29 Lower P. vivax malaria transmission in urban areas than surrounding rural areas was found global ly, and in the Africa+ and CSE Asia regions (Table 2 2), which corroborates previous findings that the urban environment is typically not suitable for malaria mosquito vectors [13, 3132] The consistent patterns of significantly lower urban Pv PR values found at the national scale in most of the countries in Africa+ and CSE Asia further supports these findings (Table 2 2). Howev er, the urbanrural survey pairs for each region are dominated by a few countries (e.g., Indonesia accounts for 65% of the Asia pairs and Sudan accounts for 45% of the Africa pairs), which make the patterns found at regional scale less informative. Distinc t and inconsistent results were found in the Americas, with higher Pv PR values in urban areas at the continental scale and for one particular country (Brazil) at the national scale. This result is probably due to the lack of Pv PR surveys in this region, as surveys from the region only account for 4.1% of the Pv PR global database. Getting extreme results is more likely when the numbers of surveys are small and only the rural Pv PR surveys were averaged. There is also evidence suggesting that higher malaria transmission in some areas of Brazil was actually a result of rapid urbanization, during which settlements were built close to forest boundaries or along riversides and thus resulting in greater exposure to the malaria parasite for residents  Figure 2 2 indicates that considerable heterogeneity exists when examining individual cities, with two cities (out of twenty) showing significantly lower Pv PR in their surrounding rural areas, and seven cities showing significantly lower prevalence in urban areas. Thus, only nine of the twenty cities examined showed significant differences in transmission between urban and rural areas, and three showed zero
30 prevalence both within and around the urban areas. Compared to P. falciparum  therefore, the patterns of Pv PR between urban and rural areas exhibit a higher level of heterogeneity. Several possible reasons include: 1) the wider transmission limits of P. vivax  but lower transmission intensity with many zero Pv PR values in the database; 2) the wide distribution i n Asia and high prevalence of Duffy negativity in Africa [65 66] ; 3) relatively fewer Pv PR surveys available in the MAP database compared with a total of 22,212 P. falciparum parasite rate ( Pf PR ) surveys in 2010  The Pv PR differences between urban and rural settings within the ranges of the dominant Anopheles vectors generally follows the patterns found in each region. This is partly because vector species that had sufficient urbanrural Pv PR pairs within their extents usually cover a large portion of the region. An issue raised here is that the distributions of most of the vector species overlap substantially with each other. Thus, drawing conclusions about the patterns of individual vec tor species is difficult without considering such overlap. However, according to expert opinion distribution maps of global DVS [15 17] the spatial relationships among those vector species are extremely complex and the interaction effects of them is beyond the scope of this analysis. The GRUMP UE was used to define urban areas here, though several alternative global urban maps exist  Every global map suffers from different errors and uncertainties  and the GRUMP UE map exhibits overestimation of large urban area extents, due to the blooming effect of NTL imagery [58, 67] This suggests that the Pv PR urban values that were significantly higher than nearby rural ones found in the Americas and several other individual cities could actually be located in surrounding lower population density areas, as significantly higher mal aria prevalence and
31 entomologic inoculation rates in peri urban areas compared to urban centers have been found in a number of studies [12 13, 36] To assess br iefly this potential bias in the GRUMP UE map, urban extents mapped using Moderate Resolution Imaging Spectroradiometer (MODIS) satellite sensor imagery [56 57] were utilized to derive an alternative, more conservative, urban assignment for the Pv PR surveys. Again, sets of spatially and temporally associated urbanrural pairs of Pv PR values were extracted and tested. The results show that, due to the more conservative nature of the classification, and the fact that only intensely urban areas were mapped [56 57] far fewer Pv PR surveys were identified as urban and the differences in Pv PR between urban and rural areas were generally not significant ( Appendix A ). Such results highlight the differing outcomes that can occur through using differing definitions of urban, and that the effects of urbanization on P. vivax transmission may extend beyond the borders of intensely urban areas for most of the regions as a general trend of decreased Pv PR was found in urban areas. Another issue is that the GRUMP UE map was produced in 2004 and some Pv PR surveys may be misclassified as the urban extent changes through time. However, global urban maps that are updated regularly or that quantify urban extent change do not currently exist. Furthermore, the majority of the Pv PR surveys were conducted between 2000 and 2010 (Table 2 1). Thus, it is reasonable to use the single timepoint GRUMP UE map in this analysis. A range of humaninduced environmental changes (e.g., deforestation, urbanization, water control projects and climate change) have been identified as drivers of 'emerging' and 'reemerging' diseases and the transmission of vector borne and other infectious diseases [5 7, 68] Urbanization is usually recognized as one of the primary
32 factors affecting vector borne diseases  as it can not only provide residents with better access to healthcare and interventions [28 29] and an environment generally less favorable for many disease vectors [31 32] but can also modify land uses to expose humans to new pathogens and vectors  While global and regional scale results here show a general trend of decreased P. vivax transmission in urban areas, the heterogeneous impacts of urbanization on P. vivax malaria tr ansmission at the city scale found in these analyses support increasing concerns of urban malaria problems in developing countries. Urbanization in these regions is often associated with poverty, poor water supplies and sanitation in peri urban areas, prov iding breading sites for certain vectors  Although malaria vectors are generally not favoured by urban environments, there is evidence highlighting the potential of malari a vectors in adapting to urban environments [71 73] For example, Anopheles gambiae s.s. was found breeding in polluted water bodies in Lagos, Nigeria  Furthermore, many studies suggested that urban agriculture is another important source for providing favourable breeding sites for malaria vectors in cities [74 77] Increased malaria prevalence is often found in communities within a distance of 1km from irrigated urban agriculture in Accra, Ghana  for example. Thus, malaria transmission in urban areas exhibits considerable spatial heterogeneity both between and within cities, depending on factors such as proximity to possible vector breeding habitats, urbanization level and socioeconomic status [31, 78] Future work should aim to elucidate these drivers through examination of the disparity of P. vivax malaria transmission between and within cities using detailed household prevalence surveys and higher resolution urban maps
33 In general, the results here highlight a consistent relationship at large scales between urban areas and lower P. vivax transmission, mirroring results found for P. falciparum and pointing towards global declines in P. vivax transmission as urbanization permanently alters the receptivity of many areas. The findings suggest that these trends will likely continue to catalyze malaria declines on the path to a malaria free future.
34 Table 21. Summary of the Pv PR surveys by region Africa+ Americas CSE Asia Total Pv PR values Number of zero records 1,299 193 3,631 5,123 Mean Pv PR (%) 0.60 3.25 3.55 3.05 Median Pv PR (%) 0.00 0.61 0.51 0.00 Time period of surveys 1985 1999 225 223 1328 1776 2000 2010 1426 187 6614 8227 Sample size 1 50 911 151 1316 2378 >50 740 259 6626 7625 Median (IQR) 48 (34 109) 87 (37 210) 120 (67 281) 107 (53 236) Records of surveys GRUMP UE defined urban 444 61 755 1260 GRUMP UE defined rural 1203 349 7241 8793 Discrete geographic locations 1424 291 6873 8588 Total 1,651 410 7,942 10,003 Africa+=Afric a, Saudi Arabia and Yemen; CSE Asia =Central and South East Asia
35 Table 22 Results of Wilcoxon Signed Rank tests on Pv PR values between GRUMP UE defined urban (U) and rural (R) survey pairs for countries, regions and the World Region No. pairs U>R U
36 Table 23 Robustness analysis of the Wilcoxon Signed Rank tests on urbanrural Pv PR value pairs derived from various spatial and temporal limits Region 100km 5years 100km 2years 50km 5years 50km 2years Z P value Z P value Z P value Z P value Africa+ 5.670 < 0.001*** 5.623 <0.001*** 5.644 <0.001*** 5.397 <0.001*** Americas 2.307 0.021** 1.680 0.094* 0.486 0.631 0.730 0.471 CSE Asia 11.194 <0.001*** 11.065 <0.001*** 9.080 <0.001*** 9.005 <0.001*** World 11.732 <0.001*** 11.555 <0.001*** 10.052 < 0.001*** 9.757 <0.001*** No. pairs 1189 1106 1156 1106 Mean No. R 49.653 42.287 31.813 27.061 Overlap rate 6.752 5.349 4.206 3.423 Africa+=Afric a, Saudi Arabia and Yemen; CSE Asia =Central and South East Asia; Mean No. R= Mean number of rural surveys for each urban rural pair; Overlap rate= number of rural surveys paired to each urban survey / total number of rural surveys (***=P<0.01, **=P<0.05, *=P<0.1)
37 Figure 2 1. The global spatial limits of P. vivax malaria transmission in 2009  A ) T he spatial limits of P. vivax malaria risk defined by P. vivax annual parasite incidence ( Pv API ) data. Areas were defines as stable (dark grey, where Pv API 0.1 per 1,000 pa), unstable (medium grey, where Pv API < 0.1 per 1,000 pa) and no risk (light grey, where Pv API = 0 per 1,000 pa). The 10,003 community based Pv PR surveys are plotted and colored based on their values (red, where Pv PR >7%; yellow, 3% < Pv PR <7%; light blue, Pv PR<3%) with zerovalued surveys shown in white. B ) C lose ups for regions with plenty of Pv PR surveys around Jakarta, Indonesia. C ) C lose ups for areas around Sorong, Indonesi a.
38 Figure 2 2 Boxplots showing the differences in Pv PR values between GRUMP UE defined urban and rural surveys for cities. (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
39 Figure 2 3 Bar charts showing the test results for the dominant Anopheles vectors of human malaria A ) R esults of Wilcoxon Signed Rank tests on Pv PR values bet ween urban and rural survey pairs for the dominant Anopheles vectors in Asia Pacific region. B ) R esults for the dominant Anopheles vectors in Africa, Europe and the Mi ddle East. C ) R esults for the dominant Anopheles vectors in the Americas. (#) denotes that a vector species is now recognized as a species complex. (*) denotes the significant level of t he test (***=P<0.01, **=P<0.05, *=P<0.1).
40 CHAPTER 3 ARMED CONFLICTS AND PLASMODIUM FALCIPARUM MALARIA TRANSMISSION IN AFRICA Chapter Summary Malaria is often a significant health problem during and after armed conflicts. Several studies have examined the impacts of conflicts on malaria transmission and strategies for malaria intervention in conflict affected areas. There is a lack of research investigating the associations between armed conflicts and malaria transmission quantitatively over large areas, h owever. In this study, georeferenced databases of P. falciparum parasite rate ( Pf PR) surveys and conflict events were integrated to explore the relationships between armed conflicts and P. falciparum malaria transmission in Africa from 1997 to 2010. The differences in Pf PR before and after conflict events were examined across Africa and by country to determine if consistent and significant increases in malaria transmission after conflicts occurred. The datasets were then stratified by basic reproducti ve number under control ( Pf Rc) class to investigate if any relationships were affected by transmission intensity. To explore the possible impacts of major conflict events, spatial and temporal clusters of armed conflicts were identified and the differenc es in Pf PR values before, during and after the time period of each cluster were also examined. Finally, literature and news reports were obtained to provide some background context to the patterns seen in malaria transmission for each cluster. Significant ly lower P. falciparum malaria transmission after armed conflicts was found across Africa and in most African countries, though five countries showed the opposite trend. Pf Rc class was found to affect the relationship between armed conflict and malaria transmission and areas with high transmission tended to be more sensitive
41 to armed conflicts. Among the nine identified clusters of specific major conflicts, increased P. falciparum malaria transmission during conflicts followed by decreased transmission afte r conflicts was observed in six clusters, each representing sudden and disruptive conflict events. While the results found here of significantly lower P. falciparum malaria transmission after conflicts are contrary to many observations from conflict affe cted regions, the analyses are set against a general background of continued declines in transmission for many countries, indicating progress against malaria can still be made even during conflicts. When specific major conflict events were examined, the pattern of increased prevalence during significantly disruptive events and rapid recovery afterward appears clear, highlighting the need for efforts to maintain intervention and healthcare coverage in conflict situations. Background Despite our strong perceptions, humanity has been becoming less violent on both long and short time scales  The number of ongoing armed conflicts in the world has declined steadily through the 1990s and early 2000s, declining over 40% from the peak years shortly after the end of the Cold War [80 82] However, this trend ended in the mid 2000s and the annual frequency of major armed conflicts has stabilized at around 35 in recent years, with most concentrated in Asia and Africa [80 82] Over the past two deca des, at least 20 African countries have been involved in armed conflicts of various types and levels of intensity (e.g., civil wars, interstate wars and violence against civilians)  These conflicts continue to exert assorted detrimental effects, including massive deaths, substantial economic losses and large numbers of forcibly displaced people [84 85] It is estimated that armed conflicts cost Africa approximately
42 $18bn per year and have shrunken each conflict edafflicted African nations economy by 15% on average since 1990  By the end of 2011, Africa hosted more than a quarter (2.7 milli on) of the worlds 10.4 million refugees  and one third (9.7 million) of the worlds 26.4 million internally displace persons  while only constituting 15% of the world total population  The challenges that armed conflicts pose on public health are widely acknowledged not only the direct injuries and deaths among military personnel and civilians, but also the indirect effects on the physical and socioeconomic environments that exacerbate morbidity and mortality [19, 9091] In fectious disease, including malaria, is often a significant health problem during and after conflicts [92 94] as multiple risk factors flourish that enhance disease emergence and transmission, including displacement of large nonimmune populations to endemic areas [95 96] resettlement of refugees to deteriorated environments that fav or vector breeding (e.g., inadequate sanitation, marginal land)  disruption of disease control programs, breakdown of health systems [97 98] and impeded access to population for timely delivery of medical supplies [18, 99100] For example, the civil war in Tajikistan during 19921993 lead to an increase in the annually reported malaria cases from 200 in 1992 to almost 30,000 in 1997  Plasmodium falciparum is responsible for the majority of malaria deaths, mostly children under the age of five in subSaharan Africa  Africa is the continent with most serious P. falciparum malaria problems characterized by intense transmission and high mortality [13, 43, 51] In 2010, an estimated 1.44 billion people worldwide were at risk of stable P. falciparum transmission ( P. falciparum annual parasite incidence, Pf API > 0.1
43 per 1,000 pa) and 343 million people lived in regions with high P. falciparum endemicity ( P. falciparum parasite rate in 210 years old, Pf PR2 10>5%), of which 52% and 95% were distributed in Africa, respectively  Additionally, there were an estimated 174 million (110 242 million) cases and 0.596 million (0.4510.813 million) deaths in Africa in 2010, which accounted for 79% and 90% of the global totals  respectively. Several studies have examined the effects of armed conflicts on malaria transmission [19, 103104] and explored barriers and strategies for malaria interventions and control in conflict situations [20, 99] The majority of theories and findings suggest that armed conflicts are associated with increased malaria risk [18 19, 99, 105106] Furthermore, the absence of internal and external conflicts has been considered as a crucial factor affecting t he operational feasibility of malaria elimination  However, there are also studies indicating a negative association between level of conflicts and malaria risk  ; describing successful malaria control in conflict affected regions such as Sri Lanka, which have almost eliminated malaria despite nearly 30 year of civil war [20, 108] Generally, most studies examining the relationship between armed conflicts and malaria transmission are descriptive or limited to individual countries. There is a lack of research investigating the association between conflicts and malaria quantitatively over large areas. In this study, georeferenced P. falciparum parasite rate ( Pf PR) surveys, P. falciparum basic reproductive number under control ( Pf Rc) map and armed conflict events datasets were integrated to examine the associations between armed confli cts and P. falciparum malaria transmission across Africa, by country and by Pf Rc class. Furthermore, areas with significant armed conflict clusters in space and time were
44 identified to explore P. falciparum malaria prevalence before, during and after speci fic major conflict events, and background literature was examined to provide context. Data The M AP Pf PR D atabase Among the various metrics of malaria transmission, parasite rate (PR) is the most commonly reported and reliable metric for P. falciparum malaria endemicity  and sensitive across a broad range of the transmission spectrum  At the time of analysis, a total of 22,212 Pf PR surveys undertaken between 1985 and 2010 were obtained from the Malaria Atlas Project (MAP [10 11] ), of whic h 15,606 (70.3%) were collected in Africa and used in this analysis. The logistically intensive process of searching for, identifying and georeferencing the Pf PR surveys has been documented elsewhere [9, 11] with all of them geolocated to precise locations and not duplicated within three months at the same site. As Pf PR follows a pattern related to age and is generally reported across different age ranges [110 111] an algorithm described by Smith et al.  was applied to standardize the values of Pf PR to a single and epidemiologically important age group (210 years). Of the surveys in Africa, in particular, the majority were conducted after 2000 (79%) and diagnosed through microscopy (71%) The sample size of these surveys varies and more than half of them (52%) are larger than 50. These Pf PR surveys came from 47 P. falciparum malaria endemic countries ( Pf MECs) with 9,433 surveys representing unique survey locations in Africa. The spatial distribution of these data overlaid on the spatial limits of P. falciparum malaria transmission [44, 51] is sh own in Figure 3 1A.
45 Pf Rc M ap The basic reproductive number of P. falciparum ( Pf R0) is a crucial metric of malaria transmission intensity, which describes the potential of disease spread within a naive population and the effort required to eliminate it  However, its generally impossible to measure P f R0 directly from field and mathematical models have been developed to estimate Pf R0 from Pf PR  As contemporary maps of malaria control coverage are generally not available  the basic reproductive number under some existing level of control, Pf Rc, is often estimated and used. Here, the Pf Rc map in 2010 (Figure 3 1B) was obtained from the MAP [10 11] The process of generating this Pf Rc map has been described in Gething et al  Briefly, transmission models that link Pf PR to Pf Rc were first developed and validated with field data. Then, those models were used to create predictions of Pf Rc from the 2010 Pf PR map  The majority of the pixels have a predicted Pf Rc value of less than two and around 10% exceed 10  In this analysis, areas with stable malaria transmission in Africa were thus categorized into three Pf Rc classes: lo w, Pf Rc <2; intermediate, 2 Pf Rc 10; high, Pf Rc >10. Armed C onflict L ocation and E vent D ata The armed conflict location and event data were obtained from the Armed Conflict Location and Event Dataset (ACLED)  which assembles and codes reported political violent events in unstable and warring states. This dataset, which cover all of the countries in Africa from 1997 to present, provides detailed information on the dates, locations, event type, groups involved, information sources and fatalities for armed conflicts  Specifically, it focuses on tracking rebel, militia and government activities, identifying territorial transfers and collecting information on rioting, protesting and nonviolent events  ACLED is the most comprehensive georeferenced and
46 m icro level conflict event dataset and has advantages in several aspects  : 1) the disaggregated event data at local level enables the studies of conflict patterns and dynamics spatially and/or temporally  ; 2) by reconciling various forms of conflict into one coherent entity, it avoids potential biases introduced by other datasets using prior definitions  ; 3) it can be further aggregated to any desired level and fits the scale of the community based Pf PR surveys well for thi s analysis. By 2012, more than 60,000 events in Africa had been recorded by ACLED with 42% being battles between government, rebels and militias, 38% violence against civilians, 14% riots and protests, and 6% nonviolent events  As each entry in this disaggregated dataset is atomic, in the sense th at events which took place over multiple days are recorded as consecutive events on a specific day and in an exact location  multiple conflict events were aggregated into a single event based on the following criteria for the purpose of this analysis: 1) in the same location, 2) belonging to the same event type and 3) occurred in consecutive days or within one month. Therefore, the initial 64,900 conflicts events in the database at the time of analysis were aggregated into 36,445 events, and this was t he final sample used in this analysis. Methods Two analysis approaches were undertaken: (i) continent wide statistical examinations of relationships between armed conflict events and Pf PR values before and after them, and (ii) more detailed analyses of specific major conflicts, with examination of the literature on each conflict to provide context to relationships with Pf PR values found.
47 Armed C onflicts and P. falciparum M alaria T ransmission To investigate the patterns of P. falciparum malaria transmission before and after armed conflicts across Africa and by country, sets of spatially and temporally associated beforeafter (before and after conflict event) pairs of Pf PR values were obtained and tested. First, following a similar approach used previously [12, 118] we identif ied all of the Pf PR surveys that were conducted within 50 km distance and two years for each armed conflict. Then, the averages of the Pf PR values that were measured before and after the conflict event, respectively, were calculated to create beforeafter Pf PR value pairs for each conflict event. After that, the differences between Pf PR values before and after the conflicts were compared and tested across Africa and by country2The relationship between armed conflicts and P. falciparum malaria transmission may vary by Pf Rc class. To test this hypothesis, the spatially and temporally associated beforeafter Pf PR pairs were stratified by Pf Rc class (low, intermediate and high) and subject to the Wilcoxon signed rank test to determine if differences in Pf PR values between before and after conflicts were significant. The patterns of P. falciparum malaria transmission before and after armed conflicts among various Pf Rc class were also compared to investigate if Pf Rc affected the relationship between conflict and malaria transmission. Given the highly skewed distribution of Pf PR values in the MAP database  Wilcoxon signed rank tests  were used to determine if significant increases in Pf PR values were observed after the outbreak of conflicts. 2 Multiple testing adjustment s are needed here as our test s do not incorporate any simultaneous hypotheses for an overall result and there is no necessary to partition the rejection regions. Although the Pf PR su rveys may be used multiple times in the analysis, t hey were averag ed to create beforeafter conflict pairs of Pf PR values
48 Only a single time limit of two years was used in obtaining the before after pairs of Pf PR values and the temporal limits may play a role in the associations of armed conflicts with malaria prevalence. To explore these temporal effects, sets of beforeafte r pairs of Pf PR were obtained by changing the time limits (one year, three years and five years) and keeping the spatial limits constant (50km). In brief, for each armed conflict, all the Pf PR surveys that were within 50km distance were identified and the average of Pf PR values that were measured within one year, three years and five years before and after the conflict event were calculated to create three sets of beforeafter Pf PR pairs and test these. Spacetime C lusters of A rme d Co nflict To explore the changes in P. falciparum malaria transmission before, during and after major conflict events, spatial and temporal clusters of armed conflict were first identified using space time scan statistics implemented in SaTScan software (version 9.1.1) [119 120] The space time permutation model was applied as it requires only case data (armed conflict event) with information about its location and time  The maximum spatial cluster size was set to 500 km, because the default setting (50% of all the cases) would lead to clusters with radius larger than 100 0 km, distracting our focus on the local spacetime clusters of armed conflicts. The output from this method was a set of spacetime clusters with information indicating their spatial extents (circles with locations and radius) and time periods (starting and ending time). In the next step, for each armed conflict cluster, the Pf PR surveys that fell within its spatial extent were identified. Then, Pf PR surveys taken within five years before, during and five years after the time period of the individual arme d conflict cluster were compared and tested. The Kruskal Wallis rank sum test  a nonparametric test for
49 more than two variables, was used to determine if the differences in Pf PR values among before, during and after conflict were significant. Finally, literature and news stories on each major conflict event, disease interventions and health system impacts were obtained to provide context in explaining the patterns seen in P. falciparum malaria transmission for each cluster of armed conflicts. Results Armed C onflicts and P. falciparum M a laria T ransmission The results of the Wilcoxon Signed Rank tests between before and after conflict pairs of Pf PR values across Africa and by country are shown in Table 3 1. Significantly higher Pf PR values before armed conflicts were found across Africa. Countries with more than ten beforeafter Pf PR value pairs were further tested and these results are also presented in Table 3 1. Among those countries, patterns of lower P. falciparum malaria transmission after conflict events were found to be significant in most of the countries (ten out of 18), including Burkina Faso, Burundi, Ethiopia, Gambia, Ghana, Mozambique, Somalia, Sudan, Tanzania and Zambia. However, five countries, including Cte d'Ivoire, Malawi, Nigeria, Uganda and Zimbabwe, sho wed the opposite results, and significantly higher P. falciparum malaria transmission was found after conflict events. The other three countries (Cameroon, Democratic Republic of the Congo (DRC) and Kenya) showed insignificant differences in Pf PR values be fore and after conflicts. Table 3 2 presents the results of Wilcoxon Signed Rank tests on beforeafter conflict pairs of Pf PR values stratified by Pf Rc class. Within the stable transmission areas, significantly higher P. falciparum malaria transmission was found before conflicts
50 for areas with low Pf Rc, while areas with intermediate and high Pf Rc were found to have significant higher Pf PR values after conflict events. The results of Wilcoxon Signed Rank tests on beforeafter pairs of Pf PR with different time limits were unclear and indicative of some nonrobust relationships ( Table 3 3 ). Among the 12 countries with sufficient Pf PR pairs (more than ten pairs), when the time limits increased from one year to five years, four countries ( DRC, Kenya, Uganda and Nigeria) showed changes in the direction of the conflict effects, one country (Cte d'Ivoire) showed a decline in significance (Z value) first followed by increased significance and the rest generally showed increasing significance i n one direction. Spacetime C lusters of A rmed C onflict Among the 65 spacetime clusters of armed conflicts indicated by the spacetime scan statistic using SaTScan software, 32 of them had more than 100 cases and a spatial radius greater than zero. After linking with the Pf PR surveys, nine clusters were found to have sufficient numbers of Pf PR surveys falling within their spatial and temporal extents (more than two Pf PR surveys found in at least two time intervals: before, during and after conflict). The s patial extents and locations of them overlaid onto the distributions of armed conflict events and Pf PR surveys are shown in Figure 3 2A and Figure 3 2B, respectively. The space time clusters of armed conflicts were matched to specific major conflict events that happened in associated areas and time periods, and the results are presented in Table 34 Cluster1 relates to the conflicts between the Lords Resistance Army (LRA) and Uganda military, the Uganda People's Defense Force (UPDF), which lasted fr om 1987 to 2006 in the northwest region of Uganda [123 124] There was a peak of violence in 2002 as the UPDF initiated a military offensive against the LRA 
51 Cluster2 represents the 20072008 Kenya crisis that erupted after the presidential election, which led to a series of violent events in the southwest and other parts of t he country [126 127] Cluster3 relates to the Burundian civil war that started in 1993 and continued through the early 2000s [128 129] Cluster4 refers to the Somali civil war that started in 2006 with Ethiopian troops involved. It was concentrated in s outhern Somalia and the boundary area of Ethiopia  Cluster5 reflects the Fast Track Land Reform during 20002003 in Zimbabwe, which led to severe deterioration of the economic and health system in the country and displaced over one million civilians  Cluster6 represents the Eritrean Ethiopian War that occurred during 19982000 in Eritrea and the Ethiopia border region  Cluster7 covers the region of Liberia and Southwest Cte d'Ivoire and represents the Second Liberian Civil War (19992003) and First Ivorian Civil War (2002 2004). Cluster8 encompasses the region of Southern Congo and border area of Gabo n and DRC and covers the Republic of the Congo Civil War (1997 to 1999) in the Brazzaville area  Cluster9 relates to two conflicts: the Casamance conflict since 1982 between the Senegal Government and the Movement of Democratic Forces of Casamance (MFDC) in southwest Senegal  and the GuineaBissau Civil war during 19981999  Figure 3 3 shows the boxplots of Pf PR values within the spatial extent and five years before, during and five years after the time period of individual armed conflict cluster. The most noticeable pattern that is seen for six of the nine clusters is that of P. falciparum malaria transmission increasing during the time periods of conflict, then decreasing within five years after the conflicts had ended. Four of these clusters were found to have significantly higher Pf PR values during the conflict compared to those
52 within five years before the conflict and significantly lower Pf PR v alues within five years after the conflict compared to those during the conflict (C2: Southwest Kenya; C5: Zimbabwe; C7: Liberia and Southwest Cte d'Ivoire; C9: Southwest Senegal, Gambia and GuineaBissau). Two clusters exhibited significantly lower Pf PR values during the conflict compared to those before the conflict, with one showing increased Pf PR values after the conflict (C1: Northwest Uganda) and the other one having no data after the conflict (C4: Southern Somalia and border area of Ethiopia). The r emaining three clusters (C3: Burundi; C6: Eritrea and border area of Ethiopia; C8: Southern Congo and border area of Gabon and DRC) showed insignificant differences in malaria transmission across different time periods of conflicts. Discussion Although increased malaria prevention and control measures are thought to have brought down malaria mortality substantially over the past decade, more efforts are required to achieve global malaria control and eradication targets  Armed conflicts pose a significant challenge to malaria control and elimination as they displace large numbers of vulnerable populations to deteriorated envir onments, disrupt malaria intervention programs and hinder access to medicine supplies [18, 9598] However, only a handful of studies have examined quantitatively the associations between armed conflicts and malaria transmission [19, 103] and there is a substantial lack of research examining such patterns at continental scales. In this study, the most comprehensive geo referenced databases of Pf PR surveys and armed conflict events were integrated and used to investigate if significant inc reases in malaria transmission during or after conflicts existed in Africa.
53 Contrary to the trends observed in many conflict affected areas (e.g., increased malaria morbidity and mortality in Afghanistan  malaria outbreaks in Burundi  and high malaria mortality rate in Af rican refugees  ), significantly lower P. falciparum malaria transmission after armed conflicts was found across Africa and in most African countries (Table 3 1). This unexpected result can be explained by several possible reasons. First, the burden of malaria in many African countries has declined substantially in the past decade, coinciding with (i) the scaling up of malaria interventions supported by increased international funding for malaria control [137 139] and (ii) increasing urbanization and development  In 2000, only 1.8% African children slept under insecticidetreated nets (ITNs) in stable endemic areas, and this rose to 18.5% by 2007  and continues to rise today  The substantial decreases in malaria incidence in southern Africa (South Africa and Mozambique) and the Horn of Africa (Ethiopia and Eritrea) have been linked to the introduction of specific interventions  Therefore, the changes in malaria transmission brought by expanding coverage of malaria intervention likely outweigh the negative impacts of armed conflicts over the timescales examined here. Secondly, conflicts often result in improved coordination and effort among key actors in health and bring more attention from humanitarian organizations [142 143] For example, improved coordination among government, international organizations and local representatives were found in Nepal and Timor Leste during conflict situations [104, 142] ; a consider able decline in casefatality among hospitalized children reported in GuineaBissau during a war was partly attributed to improved access to drugs funded by humanitarian organizations  Thirdly, armed conflicts in many African countries were characterized by long duration
54 or high frequency, which further makes it difficult to define a baseline for comparison. The approach used to create beforeafter conflict Pf PR pairs in this study would be less reliable if they are overlapped over time (e.g., the same Pf PR survey could be counted as an after conflict survey of one pair while as a beforeconflict survey of another pair). Nonetheless, this issue is minimal due to the facts that malaria incidence in many African countries has been reduced by more than 75% since 2000  Finally, the beforeafter conflict Pf PR pairs identified for each country depend on the availability of both Pf PR surveys and armed conflict event data. The results would be biased if only a few Pf PR surveys were taken during or after conflict events. However, for those identified beforeafter conflict pairs of Pf PR, the average numbers of Pf PR surveys taken within two years before and after conflict are 19 and 23, respectively. The results are thus n ot likely to be affected by the availability of Pf PR surveys after conflict. Nevertheless, five countries (Cte d'Ivoire, Malawi, Nigeria, Uganda and Zimbabwe) showed significantly higher P. falciparum malaria transmission after conflict events (Table 3 1 ), which is consistent with previous findings that armed conflicts create environments that increase malaria risk, morbidity and mortality [18 19, 99] For example, deteriorated sanitation, reduced availability of protective measure against mosquitoes and limited access to healthcare infrastructure were found in western Cte d'Ivoire shortly after the 2002/2003 Ivorian civil war  ; this conflict also resulted in serious health system failures in the northern, western and central regions of the co untry, with more than 60% of the highly trained health personnel fleeing  Another possible explanation is that malaria funding and intervention cov erage are still unequal across Africa and deficient in several countries [42, 141] For example, among
55 the 89.6 million children living in conditions of stable malaria transmission and unprotected by ITNs by 2007, Nigeria alone accounted for 25%; Cte d'Ivoire, Nigeria and Uganda were among the seven countries that had national ITN coverage less than 15% in 2007  The unstable political and insecure situations may thus have contributed to the inadequate intervention coverage and increased malaria risk after conflict s in those conflict affected countries. Furthermore, climate could be an important confounder as precipitation and temperature extremes are considered to be associated with changes in conflict risk  which also affect malaria endemicity For example, excessive rainfall is likely to raise the risk of conflict [14 7] and also the risk of malaria transmission. To assess this potential confounder, monthly Evapotranspiration (ETa) Anomaly Products (2001present) were used to exclude conflict pairs of Pf PR that occurred during ETa Anomalies  Briefly, the before after conflict Pf PR pairs were removed from the dataset if any ETa Anomaly value in the same location is larger than 150 or less than 50 wit hin four months. Then, the remaining dataset was subject to the Wilcoxon signed rank test across Africa and by country. Th e results ( Appendix B ) are similar to that of the original dataset and only a few countries ( Ethiopia Cte d'Ivoire and Uganda) turned insignificant in the differences of Pf PR before and after conflict, indicating the effects of climate is negligible. Pf Rc was found to affect the relationships between armed conflict and Pf PR values, with significantly increased malaria transmission after conflicts found in intermediate and high Pf Rc regions, while lower malaria transmission was found after conflicts in low Pf Rc regions (Table 3 2). This result matches with the analysis by country, where most of the countries (four out of five, including Cte d'Ivoire, Malawi,
56 Nigeria and Uganda) with higher Pf PR values after conflicts show intermediate to high Pf Rc values  Therefore, areas with high malaria transmission are probably more sensitive to the changes brought by armed conflicts as Pf Rc reflects the potential of disease spread and high Pf Rc regions generally respond to malaria control or other influences more quickly and to a great er extent than those with lower Pf Rc [44, 114] Additionally, the effects of armed conflicts on P. falciparum malaria transmission likely vary by conflict type and level. To test this, the beforeafter Pf PR pairs were stratified by conflict types (battles, nonviolen t activity, rioting and violence against civilians) and levels (fatalities and length) and subject to the Wilcoxon Signed Rank test. The results showed that there were no differences among various conflict types and levels with most of the groups showing s ignificantly lower Pf PR values after conflicts, while the remainder was insignificant ( Appendix B ). There are uncertainties in the analyses presented around choices of thresholds. An arbitrarily defined single temporal limit of two years was used to obtain spatially and temporally associated beforeafter pairs of Pf PR. The temporal effects analysis indicated that this approach exhibited limited robustness ( Table 33 ). Another issue raised here is that armed conflicts often cause large numbers of refugees and internally displaced people [87 88] With a spatial constraint of 50 km, Pf PR surveys undertaken after the conflict may survey a very different population compared to the population before the conflict. However, there is no available data to quantify the biases that may result from such population movements. The pattern of increased P. falciparum malaria transmission during conflicts followed by decreased transmissio n after conflicts was observed in six of the clusters of
57 the nine major conflicts, with four of these being significant (Figure 3 3) This result was further supported by previous findings and evidence of interrupted healthcare and higher mortality rates d ue to major conflict events in several of the countries within the spatial extents of those clusters (e.g., Kenya, Zimbabwe and Cte d'Ivoire) [19, 145, 150151] For example, the Zimbabwean health services sector has been worse off since the Fast Track Land Reform in 2000, with increasing shortages of qualified personnel because of the undesirable political environment  ; elevated child mortality rates were reported in Western Kenya after the post election violence in 2008 with malaria accounting for the greatest increases which may have resulted from a shortage of malaria drugs and disruption of health services during the conflict  ; lower clinic attendance and medication adherence for HIV infected children were also observed during the post election crisis in Western Kenya due to concerns over personal safety, shortages of resour ces and hopelessness  The significant decline in malaria prevalence after those conflicts may attribute to the charitable medical assistance provided by international organizations during or after specific major conflict. However, such information is not readily available. The decline in P. falciparum prevalence during conflicts in Northwest Uganda is likely partially due to the lack of Pf PR surveys in this region, with only four surveys before the conflict and three surveys during the conflict. The decreased malaria transmission in Southern Somalia and border area of Ethiopia during the Somali civil war may be attributed to the decreasing trend of malaria incidence in the Horn of Africa  Furthermore, clusters characterized by sudden, short and se rious conflicts are more likely to be disruptive and thus increase malaria prevalence, whereas clusters with long lasting or low level conflicts tend to
58 show little effects on malaria transmission. For example, the 20072008 Kenyan crisis lasted only a few months but caused more than 1,000 deaths and 300,000 displaced six weeks after the eruption  ; Cte d'Ivoire was a model of political stability and economic prosperity in Africa before the 2002 civil war that suddenly split the nation into rebel held north and government controlled south  ; the Second Liberian Civil War was ruinous and left approximately 250,000 people dead, many thousands displaced and the infrastructure in ruins  In contrast to these conflicts in Kenya, Cte d'Ivoire and Liberia, those in northwest Uganda, Burundi and Somal ia were generally protracted civil wars or occurring at low levels or in isolated areas (e.g., the conflicts between LRA and Uganda government lasted 19 years [1231 24] ; the Burundian civil war lasted 12 years  ; the civil war in Somalia started in 1991 and is still ongoing  ). Therefore, the effects of these types of conflicts on malaria transmission remain more difficult to quantify. In conclusion, consistently lower P. falciparum malaria transmission after armed conflicts were found across Africa, in most African countries and in low Pf Rc regions, as all of these conflicts were occurring against a background of generally declining malaria transmission in Africa brought down by scaleup interventions and other factors in the past decade. However, significantly increased malaria transmission during/after conflicts was found in areas affect ed by sudden major conflicts or countries with intermediate to high transmission of malaria. This result highlights that the maintenance of intervention coverage and provision of healthcare in conflict situations to protect vulnerable populations can maint ain gains in even the most difficult of circumstances.
59 Table 3 1 Results of Wilcoxon Signed Rank tests on Pf PR values between before (B) and after (A) conflict survey pairs across Africa and by country Region No. pairs B>A BA B
60 Table 3 3 Results of Wilcoxon Signed Rank tests on Pf PR values between before (B) and after (A) conflict survey pairs by country with different time limits (one year, three years and five years) Region 50km 1year 50km 3year 50km 5year No. pairs Z P value No. pairs Z P value No. pairs Z P value Africa+ 1496 9.297 <0.001 *** 3093 18.966 <0.001 *** 3841 25.864 <0.001 *** Burundi 56 4.062 <0.001 *** 73 5.830 <0.001 *** 97 8.353 <0.001 *** DRC 51 0.641 0.525 53 0.461 0.649 86 3.951 <0.001 *** Kenya 336 3.253 0.001 *** 647 1.937 0.053 779 9.698 <0.001 *** Malawi 8 2.521 0.014 ** 16 2.637 0.009 *** 25 3.484 <0.001 *** Somalia 699 18.019 <0.001 *** 1422 18.352 <0.001 *** 1455 20.080 <0.001 *** Sudan 105 2.460 0.014 ** 162 10.091 <0.001 *** 181 10.922 <0.001 *** Tanzania 58 4.587 <0.001 *** 101 6.733 <0.001 *** 121 8.322 <0.001 *** Uganda 45 1.259 0.210 104 2.050 0.041 ** 207 4.273 <0.001 *** Zimbabwe 20 3.920 <0.001 *** 136 4.571 <0.001 *** 168 4.534 <0.001 *** Ethiopia 18 1.633 0.107 33 2.746 0.006 *** 70 2.054 0.040 ** Cte d'Ivoire 18 2.896 0.003 *** 60 0.096 0.927 122 4.250 <0.001 *** Nigeria 37 2.135 0.033 ** 69 0.677 0.499 158 6.476 <0.001 *** (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
61 Table 3 4 Spacetime clusters of armed conflicts ID Radius Time period No. of c ases Study area Related e vent Cluster1 222.99 2002/42005/ 4 1819 Northwest Uganda The Lords Resistance Army (LRA)(1987 2006) Cluster2 220.12 2007/10 200 8/1 398 Southwest Kenya 2007 2008 Kenyan crisis Cluster3 161.25 2000/1 200 3/1 1742 Burundi Burundian civil war(1993 2005) Cluster4 434.83 2007/4201 2/12 5319 Southern Somalia and border area of Ethiopia Somali civil war(2006present) Cluster5 388.35 2001/10 2002/ 4 767 Zimbabwe Fast Track Land Reform(2000 2003) Cluster6 443.49 1998/42000/6 493 Eritrea and border area of Ethiopia Eritrean Ethiopian War(19982000) Cluster7 406.96 2002/42003/ 7 448 Liberia and Southwest Cte d'Ivoire Second Liberian Civil War(1999 2003); First Ivorian Civil War(2002 2004) Cluster8 498.83 1997/11999/ 7 384 Southern Congo and border area of Gabon and DRC The republic of the Congo civil war(1997 1999) Cluster9 268.28 1997/72002/ 7 414 Southwest Senegal, Gambia and Guinea Bissau Casamance conflict(1982); Guinea Bissau Civil War(1998 1999)
62 Figure 3 1. P lasmodium falciparum malaria endemicity in 2010  A ) The spatial limits of P. falciparum malaria risk defined by Pf API data in Africa. Areas were defines as stable (dark grey, where Pf API 0.1 per 1,000 pa), unstable (medium grey, where Pf API < 0.1 per 1,000 pa) and no risk (light grey, where Pf API = 0 per 1,000 pa). The georeferenced Pf PR surveys are plotted and colored based on their values (red, where Pf PR2 10 >40%; yellow, 5% < Pf PR2 10 <40%; light blue, 0< Pf PR2 10<5%; white, Pf PR2 10=0). B ) T he Pf Rc map of Africa in 2010. The color scale is logarithmic to allow better display of the highly positively skewed values.
63 Figure 3 2 Space time clusters of armed conflicts. A ) S patial extents and locations of the conflict clusters. The georeferenced armed conflict events were plotted and colored based on their time (red, 19972000; yellow, 20012005; light blue, 200 6 2010; blue, 20112013). B ) C onflict clusters overlaid onto the distri butions of the Pf PR surveys.
64 Figure 3 3 Boxplots showing the differences in Pf P R values among surveys taken within five years before, during and within five years after the conflict clusters. NW=northwest; SW=southwest; BA=border area; GNB=GuineaBissau. (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
65 CHAPTER 4 A GLOBAL ASSESSMENT OF ECONOMIC CONDITIONS AND MALARIA PREVALENCE Chapter Summary Malaria is often described as a disease of poverty and the relationship between the two is complex. Poverty may increase peoples vulnerability to malaria, consequently sustaining transmission, while the socioeconomic costs related to the disease may lead to further impoverishment. However, findings from exist ing studies are not consistent regarding this relationship and few studies assessing the link between economic condition and malaria exist over large areas. A global map of Plasmodium falciparum endemicity was combined with v arious gridded economic datase t s to explore the relationship between malaria and economic conditions across spatial scales. Endemicity was classified into two groups: i ) no, unstable and stable risk; ii) low, intermediate and high risk within the extents of stable risk. The differences in economic conditions among endemicity classes of each group were compared and tested globally, by region and by country. Moreover, econometric models were developed to quantify the magnitude and direction of this relationship. Significantly higher economic conditions, measured by multiple datasets, in malaria free and low risk areas were found globally, continentally and in several countr ies, while just a few countries showed the opposite pattern. The overall negative association between economic conditions and malaria transmission suggests that economic development is potentially an effective intervention against malaria. Equally, the control and elimination of malaria may facilitate economic development.
66 Background Nearly half of worlds population (3.3 billion) is at risk of malaria, with Plasmodium falciparum being responsible for the majority of malaria deaths  Malaria is consistently described as a disease of poverty, with the global distribution of transmission showing a striking coincidence with levels of poverty, principally concentrated in tropical and subtropical regions [21, 23] In terms of national Gross Domestic Product (GDP) per capita in 2005, the wealthiest twenty countries are all free of malaria transmission, whereas the poorest twenty countries are all hav e endemic malaria (Appendix C) [44, 160] Evidence from a cross country analysis suggests that malaria reductions in Africa increase economic productivity and every $1 invested per capita in suppressing malaria led to an increase of $6.75 in per capita GDP bet ween 2007 and 2011  The relationship between malaria and economic conditions is complex and can be explained in several ways  On the one hand, poverty may limit the use of interventions [24, 163] obstruct access to health care  cause malnutrition that increase malaria susceptibility  and expose people to environments pron e to vector proliferation (e.g., poor quality housing  and certain work conditions such as forestry and agriculture [22, 169] ), consequently sustaining the transmission of malaria  For example, higher rates of use of more effective preventions such as insecticide sprays has been found among those of higher socioeconomic status [171 172] ; richer people are more likely to use orthodox health services, rather than self treatment [173 174] ; regular activity in the forests increased the odds of malaria infection by 10fold in a village of central Vietnam  On the other hand, malaria has negative effect on economic development as it results in burdens and social costs 
67 on individuals and government, and impedes trade and investment  Furthermore, the causal relationship may work in both directions. However, the findings from existing studies are not consistent regarding this relationship  and little evidence of the link between economic conditions and malaria exists over large areas. In addition, only a few recent studies have explored the bi directional relationship and considered possible endogeneity quantitativ ely [21, 177] Thus, there is a need to elucidate this relationship over large areas using consistent data and systematic approaches. The recent construction of a global P. falciparum endemicity map, a global map of per grid cell economic output  and satellite night time light (NTL) based economic indices  enables the exploration of the relationship between economic conditions and malaria transmission at global and subnational scales for the first time. Here, these datasets were integrated to explore the relationship between P. falciparum malaria endemicity classes and economic conditions at different spatial scales (global, regional and national) and further examine the differences among various economic indices. Furthermore, econometric models were devel oped to quantify the magnitude and direction of this causal relationship, by including a set of primary covariates and taking endogeneity into account. Data and Methods P. falciparum M alaria E ndemicity M ap The most likely endemicity class map for P. falcip arum malaria in 2010 (Appendix C) was obtained from the Malaria Atlas Project (MAP)  The process of generating this endemicity class map has been described in Gething et al  Briefly, the process can be divided into two steps. The first step is to delineate the spatial limits of endemic transmission based on the initial ident ification of Pf MECs and sub national reported
68 Pf API data. Areas were thus defined as stable transmission ( Pf API 0.1 per 1,000 pa), unstable transmission ( Pf API < 0.1 per 1,000 pa) and no risk ( Pf API = 0 per 1,000 pa). Then, these limits were further refin ed by additional medical information and environmental constraints (temperature and aridity). The second step involves geostatistical modeling of malaria endemicity within the boundaries of stable transmission, reliant on georeferenced P. falciparum parasite rate ( Pf PR) surveys and a large suite of environmental covariates. Based on the predicted probabilities of class membership, the grid cells in the stable transmission areas were further categorized into their most likely membership of three endemicity classes, namely, low ( Pf PR2 10), intermediate ( 5% < Pf PR2 10< 40% ) and high ( Pf PR2 10 ) risk. Geographically B ased Economic D ata Although poverty is a multidimensional concept involving a range of socioeconomic factors, GDP per capita is commonly used as an indicator for macroeconomic analyses at country level and household income, consumption or expenditure for micro analyses [181 182] The per grid cell economic output map obtained from the Yale Geographically based Economic data (G Econ) project provides estimates of gross output (conceptually similar to GDP) at a 1 degree longitude by 1 degree latitude for global terrestrial regions [183 184 ] This resolution is approximately 100km100km, which is about the size of most thirdlevel political entities, such as counties in the United States  The concept and method of rescaling political economic output data to gridded economic out put has been documented elsewhere  The economic output measures from different countries were converted into a common currency using both market exchange rates (MER) and purchase power parity (PPP) adjustments. The newest dataset (GEcon 4.0) includes a suite of economic variables for each grid cell
69 and the Gross cell product for 2005 converted to 2005 US$ at purchasing power parity exchange rates (referred to as PPP2005) was used in this analysis (Figure 41A)  For each grid cell, this database also provides a series of geographic and environmental variables, including area of grid cell, country name, distance to coast, elevation, grid cell population, precipitation, and temperature, amongst others. Sa tellite Night Time Lights ( NTL ) B ased E conomic D ata Recognizing the shortcomings and challenges in economic data collection, several economic datasets have been developed from NTL imagery data collected by US Air Force Defense Meteorological Satellite Prog ram Operational Linescan System (DMSP OLS) to provide more readily available and spatially and temporally comparable measure of socioeconomic conditions [164, 185187] Two sets of NTL based economic data were obtained from NOAAs National Geographical Data Center (NGDC)  The first dataset was a map of estimated total economic activity derived from NTL satel lite imagery and the LandScan population grid  It provides globally gridded GDP data at 1km1km resolution for 2006, and referred to as NGDP hereafter (Figure 4 1B). The methods and models used to generate this spatially disaggregated map have been described elsewhere  The second dataset is a Night Light Development Index (NLDI) that was derived from NTL satellite imagery and a population density grid [190 191] and based on Lorenz curves  It is inversely correlated with the Human development index (HDI) and provides a spatial depiction of development differences within countries  The index is available at 0.25 degree by 0.25 degree resolution globally for 2006 (Figure 41C), and the process of constructing it has been documented elsewhere 
70 Analyses The economic datasets and the malaria endemicity map described above were obtained at different spatial resolutions. To overcome this incompatibility where it existed, each dataset was resampled to 11 degree resolution, which is the coarsest resolution among all the datasets. Specifically, the P. falciparum malaria endemicity class map was re sampled to 11 degree resoluti on using a majority algorithm; the NTL based GDP map (NGDP) was resampled to the same resolution using a summation algorithm; and the NLDI data was resampled using a mean algorithm. To explore the relationship between economic conditions and P. falcipar um malaria transmission and also the differing representations of economic conditions to see if common patterns arise, the per grid cell economic output map (PPP2005), the NTL based GDP map (NGDP) and the NLDI data were used and the patterns of them within the malaria endemicity classes were examined. First, the economic datasets were overlaid onto the P. falciparum malaria endemicity map to derive an endemicity class for each grid cell. Then, a set of exclusion criteria were applied to remove population de nsity biases. Cells with (i) population density lower than 0.1 per sqkm and (ii) areas smaller than 100 sqkm were excluded and the remaining cells were used in following analyses. Additionally, to better present the economic condition of each cell and alleviate the impacts of outliers, the original variable PPP2005 was transformed from total cell economic output to the natural log of per capita economic output as follows: = ( ) (4 1) The NGDP was converted to the natural log of per capita NGDP as:
71 = ( ) (4 2) Finally, the patterns of economic conditions among P. falciparum malaria endemicity classes were investigated at various scales. The endemicity classes were categorized into two groups: i) no risk, unstable risk and stable risk; ii) low risk, intermediate risk and high risk w ithin the extents of stable risk. The trends of lnP lnG and NLDI among endemicity classes of each group were compared and tested separately. The Kruskal Wallis rank sum test  a nonparametric test for more than two variables, was used to determine if the differences in lnP lnG and NLDI among malaria endemicity classes were significant. The tests were conducted globally by region (Africa+ (Africa, Saudi Arabia and Yemen); the Americas; Central and South East (CSE) Asia) and by country. The Wilcoxon rank sum test was used if samples were only available in two classes. M odel S pecification The primary challenge in understanding the relationship between economic conditions and malaria transmission is the endogeneity bias caused by the possible bi direct ional causation between them. Simple ordinary least square (OLS) regression models assume that the model error term is unrelated to the explanatory variables and the presence of endogeneity would thus violate this assumption and produce biased estimates of explanatory variables. The simultaneous equation model (with twostage least square estimation(2SLS); and threestage least square estimation (3SLS)), which composes multiple equations with simultaneous feedbacks and considers endogeneity using instrument al variables, is introduced to address this problem and examine the bi directional relationship between economic conditions and malaria transmission.
72 To build up the simultaneous equation model, we begin with simple OLS regression models for both malaria a nd economic condition as: = + + + + + + + (4 3) = + + + + + + + (4 4) Equation (43) is the malaria OLS regression model and equation (4 4) is the economic OLS regression model. represents Pf PR2 10 (for more details about this variable see Data in Chapter 3) at location i ; represents the natural log of per capita e conomic output ( lnP ), which was matched to based on their locations. Briefly, for each Pf PR2 10 survey, the nearest was assigned to it. denotes minimum monthly precipitation; denotes average temperature; denotes minimum monthly temperature; denotes distance to coast in kilometers from location i These four variables are also from the G Econ database and they are corresponding to denotes normalized difference vegetation index (NDVI), which was obtained from the Global Inventory Modeling and Mapping Studies (GIMMS) dataset [193 195] denotes the status of urban or rural for location i This variable was abstracted from the GRUMP UE dataset (for more details about this variable see Data in Chapter 2); denotes the completion rate of primary education; denotes foreign direct investment (% of GDP); denotes net capital account (% of GDP), including government debt forgiveness, inv estment grants in cash or in kind by a government entity, and taxes on capital transfers; denotes total natural resource rents (% of GDP), representing the difference between the value of crude natural resource (oil, natural gas, coal, mineral and f orest) production at world prices and total costs of production; those four variables come from the World Development Indicator (WDI) database  All va riables are for
73 2005. Initially, the suite of environmental and socioeconomic covariates for both economic conditions (e.g., education, trade, investment and economic policy) and malaria transmission (e.g., precipitation, temperature, NDVI and health serv ices) were collected based on literature and matched to individual Pf PR2 10 surveys. Then, a stepwise approach was applied to remove insignificant explanatory variables and colinearity. The remaining covariates are presented in equations (43) and (44). The simultaneous equation model estimates a system of structural equations, where some equations contain endogenous variables in the explanatory variables. In this study, there are two equations (the malaria model and the economic model) and the endogenous explanatory variables are dependent variables from the other equations in the system. All other variables are treated as exogenous to the system. Specifically, in the malaria model, is endogenous and it will be estimated by instrumental variables ( , and ), which are exogenous explanatory variables in the economic model. The first stage regression for this model is: = + + + + + + (4 5) The second stage regression uses the fitted values of economic conditions from the first stage regression as independent variables to estimate malaria: = + + + + + + + (4 6) In the economic model, is endogenous and it was estimated by instrumental variables ( , and ), which are exogenous explanatory variables in the economic model. The first and second stage regression models are: = + + + + + + (4 7) = + + + + + + + (4 8)
74 The 2SLS estimation considers the endogenous variables individually for each equation and no covariance is estimated between the parameters of the equations, while for the 3SLS estimation, it considers the endogeneity simultaneously for both equations and uses generalized least squares (GLS) to estimate correlation across the equations [197 198] The 3SLS approach is thus more efficient in obtaining the estimates. R esults Figure 42a shows the boxplots of lnP for areas with no, unstable and stable P. falciparum malaria transmission and the results of the Kruskal Wallis tests across the world and by region. Significantly higher values of lnP in areas with no malaria risk were found globally and by regions (Africa+, Americas and CSE Asia), while the differences o f lnP values between unstable and stable transmission areas were relatively small (Figure 4 2a). Countries with more than nine samples in at least two classes of P. falciparum malaria endemicity (no, unstable and stable risk) were further tested and the re sults are shown in Figure 43a. Among the 30 countries examined, 20 countries were found to have significant differences in lnP among no, unstable and stable risk of malaria areas. However, the pattern of the relationships between lnP and malaria endemcity showed by those 20 countries were complicated (Figure 43a). Four countries (China, Iran, Mauritania and Namibia) were found with significantly higher values of lnP in areas with no risk and similar lnP values between unstable and stable areas; two countr ies (Chad and Ethiopia) were found to have significantly lower values of lnP in stable risk areas and similar lnP between no risk and unstable risk regions; three countries (Brazil, Niger and Sudan) were found to have progressively and significantly decreasing values of lnP from no risk to stable risk areas; two countries
75 (Algeria and Saudi Arabia) were found with significantly lower lnP values in unstable risk areas compared to that of no risk areas and no data available for stable risk areas. Furthermore, four countries (Bangladesh, Peru, Philippines and Thailand) showed the opposite trend with significantly higher values of lnP in stable transmission areas compared to that of unstable and stable areas. The other five countries showed either higher (Mali and Papua New Guinea) or lower (Colombia, Indonesia and South Africa) values of lnP in unstable risk areas than that of no risk and stable risk areas. Boxplots showing the differences in lnP among areas with no, unstable and stable risk of P. falciparum mal aria by country is presented in Appendix C. Figure 42c presents the boxplots of lnP among areas with low, intermediate and high P. falciparum malaria transmission and results of the Kruskal Wallis tests globally. Significantly higher lnP values were found in low risk areas globally with no significant differences between intermediate and high risk areas. The results for the regions were not listed here due the lack of samples for intermediate and high transmission in CSE Asia and the Amer icas. Countries with more than seven samples in at least two classes within the stable transmission areas (low, intermediate and high risk) were further tested, of which 10 countries (out of 20) showed significant differences of lnP values among malaria endemicity classes (Figure 4 3b). The trends of the relationship between lnP and malaria endemicity classes demonstrated by those countries were unclear. Five countries (Central African Republic, Madagascar, Mali, Mozambique and Nigeria) showed significantly lower values lnP in high transmission areas compared with that of intermediate transmission areas, while the other five countries exhibited the opposite pattern, with Cameroon and Tanzania showing significantly lower lnP values in
76 intermediate risk areas compared to that for high risk areas, and Chad, Malawi and Indonesia showed lower values in low risk areas compared to that for high risk areas. Boxplots showing the differences in lnP among areas with low, intermediate and high risk of P. falciparum malar ia by country is presented in Appendix C. Two NTL based economic indicators ( lnG and NLDI) were also utilized to investigate the relationship between economic condition and malaria endemicity. The results based on lnG are shown in Figure 42b, 4 2d and App endix C. Similar to the results based on lnP significantly higher values of lnG in zero risk areas were found globally and by region (Africa+, Americas and CSE Asia) (Figure 42b). However, the patterns shown by the country samples showed differences (Appendix C). 13 countries were found to have significant differences in lnG among no, unstable and stable risk areas, of which five countries (Brazil, Colombia, Indonesia, Peru and Venezuela) were found to have lower values in stable risk areas, while four co untries (Chad, Mali, Mauritania and South Africa) were found to have higher values in stable risk areas. Within the stable transmission areas, the trend at the global scale was also similar to that of the lnP (Figure 4 2d). At country level, only five of t he 20 examined countries showed significant differences in lnG values among low, intermediate and high risk areas (Appendix C). Significantly higher values of lnG in low transmission areas were found in four countries (Central African Republic, Democratic Republic of Congo, Mozambique and Uganda), while significantly higher values of lnG in high transmission areas were found in one country (Senegal). The results based on NLDI are shown in Appendix C. Due to the large number of missing values in this dataset many fewer samples were involved in this set of analyses.
77 Significantly higher values of NLDI in stable risk areas compared to that of no and unstable risk areas were found globally and regionally. Among the 12 examined the countries, significantly lower values of NLDI in no risk areas were found in Brazil, Indonesia and Pakistan, and significantly lower values of NLDI in unstable risk areas were found in Bangladesh, Colombia and India. Within the stable transmission areas, significantly lower values of N LDI in low risk areas were found globally, while no significant differences were found by region or in the three identified countries. Table 41 presents the estimation results of the malaria models (OLS, 2SLS and 3SLS), suggesting coherent relationships between malaria transmission and economic conditions ( R2 2SLS and 3SLS regression models. Economic conditions were found to have a significantly negative impact on malaria prevalence, with negative coefficient estimates in all models. The coefficient of 0.260 estimated by the 3SLS suggests that a 10% increase in per capital economic output would reduce malaria prevalence by approximately 0.02 unit Mo 4.727E 0 4 ), average 0.057) and malaria transmission. Such results are consistent with findings in the liter ature [44, 51, 199] The estimation results of the economic models are shown in Table 4 2. Negative impacts of malaria prevalence on economic conditions were also found to be statistically significant. The coefficient value of 0.728 estimated from the 3SLS regression suggests if ma laria prevalence increased by 0.1 unit we would expect econ omic output to decline
78 by approximately 2.787E 04 be significantly associated with economic conditions and all the estimated coefficients are comparable among different models. These results corroborate previous findings [21, 200 201] D iscussion Where malaria prospers most, human societies have prospered least  Remaining one of the Worlds most serious public health problems, malaria poses not only a direct health burden, but also negative social and economic effects on endemic countries  The correlation between malaria and poverty has long been acknowledged. However, findings from existing studies are not consistent regarding this relationship [24, 173, 177, 203] and there is a lack of research examining the linkage between economic conditions and malaria over large areas. In this study, a recently constructed global map of P. falciparum endemicity per grid cell economic output data and NTL based economic indices were used to explore the relationship between economic conditions and P. falciparum malaria transmission at global and subnational scales. Significantly higher values of lnP in malaria fr ee and low risk areas were found globally, continentally and in several countries (Figure 42a, 42c and 43). This corroborates previous findings that malaria is mainly in poorer areas and those areas experience slower economic growth rates [21, 23] Furthermore, similar patterns of lnG values were found at global and regional scales (Figure 42b and 42d), suggesting a robust relationship of low economic conditions with high malaria transmission globally and regionally, no matter which way the subnational economic conditions are
79 measured. The conspicuous concurrence of poverty and malaria in tropical and subtropical regions have led to several attempts to examine t he effects of malaria on economic development [200 201] and most literature suggests significant negative impacts of malaria on economic developm ent, and that malaria can impede economic growth in multiple ways [21, 23, 204] For example, farmers tend to adopt less labor intensive crops instead of more profitable ones due to the perceived risk of malaria  ; high infant mortality and fertility rates caused by malaria in poor families may lower educational investment in children  Furthermore, poverty is frequently considered as a factor sustaining malaria transmission, though evidence is often not consistent, with some studies suggesting significant negative associations [163, 199] while others find no significant relationship or mixed results [173, 203] Although the causal relationship between poverty and malaria transmission remains tangled, the a ssociation between them is evident. Nevertheless, the patterns of lnP among various P. falciparum malaria endemicity classes exhibit substantial heterogeneity at national levels with several countries showing the opposite trends (e.g., higher lnP in stabl e risk areas; lower lnP in intermediate transmission areas than high transmission areas). This result can likely be explained by the fact that the dynamics of both malaria and economic conditions are determined by multiple factors and it is difficult to at tribute changes to a single cause. As a vector borne disease, unlike other directly transmitted diseases, malaria parasites spend much of their life cycle in Anopheles vectors [3 4] The transmission of malaria is therefore mainly dependent on external environmental factors such as temperature, precipitation, vector al capacity and other ecological conditions [21, 204] Intensive
80 malaria transmission in poor areas is not necessarily a consequence of poverty and the escape from poverty is not principally explained by malaria elimination, but through ot her factors such as economies opening up, improved labor productivity and technical revolutions  Another obvious reason is that some of those countries are largely covered by inhabitable desert (e.g., Chad) or mountainous areas (e.g., Peru) and populations are constrained to the parts of the countries that are most habitable and fertile, which coincide with where malaria is. The G Econ database provides economic output data at a resolution of 1 degree longitude by 1 degree latitude, which is much finer than most of the economic activity indices measured at national levels (e.g., GDP) and enable the integration with other gridded environmental datasets [178, 183, 206] However, the accuracy of the dataset varies among countries as economic data are usually sparse in low income countries  To assess potential bias and uncertainties in the G Econ dataset, two NTL based economic datasets were also used to explore the relationship between economic conditions and malaria transmission. Similar results were found at global and regional scales between the G Econ dataset and the NTL based GDP dataset, while the patterns at country level showed differences, suggesting considerable uncertainties inherent in those datasets. The NLDI dataset was generally not comparable due to the large number of grid cells with no data. Therefore, future research may need to look at other economic measures for more detailed analyses and explore the changes over time. Moreover, the econometric models explored only a limited se t of covariates for both malaria and economic conditions in a single period of time. Other explanatory factors (e.g., health care access, intervention coverage and economic policy) and their time
81 series data, if available in future, should be collected to refine the model and decipher the dynamics of this relationship in a more comprehensive manner. In the past decade, malaria incidence and mortality have been significantly reduced by the tremendous expansion in the financing and coverage of malaria interventions  However, this trend has slowed down in recent years  There is a risk of malaria resurgence due t o increasing resistance to antimalarials and insecticides, and fatigue of financial aid for malaria interventions  The negative association between economic conditions and malaria transmission found globally, regionally and in some of the countries suggests that economic development is potentially an effective intervention against malaria as wealth is associated with better access to health care, improved nutrition status and housing quality [24, 163, 209] For example, malaria elimination in Europe and North America was achieved as a by product of improved socioeconomic conditions  Nevertheless, economic development should not be a standalone strategy, but rather an essential component of malaria intervention strategies  Mutually, the control and elimination o f malaria would facilitate economic development as malaria poses serious social and economic burdens on endemic countries.
82 Table 41 Estimation results of malaria models Dependent variable: Pf PR2 10. Independent variables Parameter estimates (standard error) OLS 2SLS 3SLS lnP 0.151 (0.012)*** 0.323 (0.015)*** 0.260 (0.014)*** Minimum precipitation 4.4 15 E 04 ( 1.2 29 E 04 )*** 8.2 73 E 04 ( 2.0 36 E 04 )*** 4.7 27 E 04 ( 1. 7 8 0 E 04 )*** Average temperature 0.005 (0.001)*** 0.005 (0.001)*** 0.003 (0.001)*** Minimum temperature 0.026 (0.002)*** 0.053 (0.003)*** 0.039 (0.002)*** Urban 0.023 (0.014) 0.006 (0.017) 0.057 (0.014)*** NDVI 0.250 (0.035)*** 0.071 (0.042)* 0.127 (0.034)*** Constant 0.474 (0.074)*** 1.143 (0.096)*** 1.062 (0.092)*** N o. observation 1323 1156 1156 R 2 0.322 0.323 0.339 Adjusted R 2 0.319 --***=P<0.01, **=P<0.05, *=P<0.1 Table 42 Estimation results of economic models Dependent variable: lnP. Independent variables Parameter estimates (standard error) OLS 2SLS 3SLS Pf PR 2 10 0.246 (0.060)*** 0.707 (0.128)*** 0.728 (0.120)*** Distance to coast 3. 390 E 04 ( 5. 50 E 05 )*** 3.8 16 E 04 (4.5 10 E 05 )*** 2. 787 E 04 ( 3.6 30 E 05 )*** Foreign direct invest 0.081 (0.006)*** 0.094 (0.006)*** 0.077 (0.005)*** Net capital account 0.026 (0.004)*** 0.012 (0.005)** 0.023 (0.004)*** Natural resources 0.032 (0.001)*** 0.029 (0.001)*** 0.030 (0.001)*** Primary education 0.019 (0.001)*** 0.019 (0.001)*** 0.016 (0.001)*** Constant 4.859 (0.119)*** 5.232 (0.129)*** 5.195 (0.114)*** N o. observation 1165 1156 1156 R 2 0.735 0.718 0.702 Adjusted R 2 0.734 --***=P<0.01, **=P<0.05, *=P<0.1
83 Figure 4 1. Geographically based and night time light based economic data. Panel A shows the distribution of PPP2005 from the G Econ dataset at 1 1 degree resolution with close up around east Brazil. Panel B shows the distribution of NGDP from the NTL based GDP map at 1 1 km resolution with closeup around east Brazil. Panel C shows the distribution of NLDI data at 0.25 0.25 degree resolution with close up around east Brazil.
84 Figure 42. Boxplots showing the differences in a) lnP and b) lnG among areas with no (N), unstable (U) and stable (S) risk of Plasmodium falciparum malaria globally and by region; c) lnP and d) lnG among areas with low (L), intermediate (I) and high (H) risk globally. (*) denotes the significant level of the test resul ts (**=P<0.05, *=P<0.1). Africa+=Africa, Saudi Arabia and Yemen; CSE Asia=Central and South East Asia.
85 Figure 4 3 Plots showing the lnP differences among areas with a) no, unstable and stable, and b) low, intermediate and high risk of Plasmodium falciparum malaria by country (*) denotes the significant level of the test results (**=P<0.05, *=P<0.1). In each case, scatter plot of lnP difference between no and unstable risk versus difference between unstable risk is shown on the left panel, and sca tter plot of lnP for no/unstable risk areas versus unstable/stable risk areas is shown on the right with one to one lines overlaid (for countries with only two endemic groups). The ISO country abbreviation for country name is used on the scatter plots ( http://www.worldatlas.com/ aatlas/ctycodes.htm).
86 CHAPTER 5 CONCLUSION Although increas ing levels of malaria prevention and control measures are thought to have brought down malaria mortality substantially over the past decade, more efforts are required to achieve global malaria control and eradication targets. The recent construction of global Pf PR and Pv PR databases enables the exploration of the relationships between malaria transmission and various determinants at global and subnational scales for the first time. T h is study presents initial efforts to integrate various global datasets to quantitatively examine the effects of urbanization, armed conflicts and economic conditions on malaria transmission at various spatial scales. The documented results facilitate an improved understanding of the global epidemiology of malaria transmission providing inf ormation to support more accurate malaria risk estimation and for designing malaria intervention and control strategies. R apid urb anization in the developing world has and will continue to have a profound influence on the malaria landscape T his research demonstrates consiste nt relationships between urban areas and lower P. vivax transmission at large scales ( globally, regionally, nationally and by dominant vector species ) following trends observed previously for P. falciparum malaria S uch results indica te that urban environment is generally not favorable for malaria vectors and an overall coverage of mala ria intervention is not necessary in urban areas However, a substantial level of spatial heterogeneity of malaria transmission wa s found to exist between and within cities U rban malaria has received increasing concerns in developing countries suggesting that targeted malaria intervention is more appropriate in urban areas.
87 Contrary to many observations from conflict affected regions, consistently lower rates of P. falciparum malaria transmission after armed conflicts were found across Africa, in most African countries and in low Pf Rc regions as the analyses are set against a general background of substantial declines in transmission for many count ries in the past decade, indicating progress can still be made even during the most difficult of circumstances. However, significantly increased malaria transmission during/after conflicts was found in areas affected by sudden major conflicts or in countries with intermediate to high levels of transmiss ion of malaria, highlighting the need for efforts to maintain intervention and healthcare coverage in conflict situations Also, the priority for malaria intervention and funding should be allocated to areas with intensive malaria transmis sion and severe armed conflicts. Significantly higher economic output in malaria free and low endemic areas w as found globally, continentally and in several countries, while a few countries showed the opposite patt erns. The evident negative association between economic conditions and malaria transmission suggests that economic development is potentially an effective intervention against malaria. Equally, the control and elimination of malaria may facilitate economic development. T he inconsistent and intricate relationships between malaria transmission and economic conditions at the country level are probably caused by the fact that a single variable of economic condition is unable to fully capture the socioeconomic status of a grid cell and this variable masks the plausible within cell heterogeneity in terms of economic development There are still some limitations in this study and several potential fields for future research. First, only an urban map at a singletime point was used to distinguish urban
88 surveys from rural ones I t is likely that some of the surveys were misclassified as urban extent changes over time F uture research may look at longitudinal population density maps that cover multiple time periods to delineate the dynamics of urban extent S econd, the beforeafter conflict pairs of Pf PR values were generated by averaging th e surveys that were taken before and after the conflicts respectively To make these values more accurate, population and distance weights can be added to take underling population density and distance effects into account. Further more, a spatial constraint of 50km may be problematic as refugees tends to be unstable and Pf PR surveys undertaken after t he conflict may survey a very different population compared to that before the conflict T herefore, bias resulting from such population movement, if data available in the future, should be rectified T hird, the econometric models used to quantify the relationship between malaria and economic c onditions only included a limited set of covariates. O ther explanatory factors should be collected to refine the models. M oreover, spatially lagged variables and spatial error autocorrelation can be incorporat ed to account for the spatial dependence inherently embedded in the georeferenced Pf PR dataset Finally it should be recognized that signif icant heterogeneity in the associations between socioeconomic factors and malaria transmission exist at city levels or among countries and the causal relationship between them remains unsolved T his study presents a first step to wards understanding the general picture of such relationships and improved allocation of malaria intervention resources worldwide To further understand these patterns and untangle the causation, more epidemiological,
89 parasitological and behavioral analyses of the disease at smaller spatial scales are needed.
90 APPENDIX A SUPPLEMENTARY TABLE S FOR CHAPTER 2 This document shows the supplementary tab le s used in chapter 2. Table A 1. Results of Wilcoxon Signed Rank tests on Pv PR values between GRUMP UE defined urban (U) and rural(R) survey pairs for the dominant Anopheles vectors in Asia Pacific region Dominant Anopheles vector s pecies No. pairs U>R U
91 Table A 2 Results of Wilcoxon Signed Rank tests on Pv PR values between GRUMP UE defined urban (U) and rural(R) survey pairs for the dominant Anopheles vectors in Africa, Europe and the Middle East Dominant Anopheles ve ctor species No. pairs U>R UR U
92 Table A 4 Results of Wilcoxon Signed Rank tests on Pv PR values between MODIS defined urban (U) and rural(R) survey pairs for continents, countries and the World Region No. pairs U>R U
93 APPENDIX B SUPPLEMENTARY TABLE S FOR CHAPTER 3 This document shows the supplementary tables used in chapter 3 Table B 1 Results of Wilcoxon Signed Rank tests on Pf PR values between before (B) and after (A) conflict survey pairs across Africa and by country (with conflict pairs that occurred during ETa Anomalies excluded) Region No. pairs B>A BA B
9 4 Table B 3 Results of Wilcoxon Signed Rank tests on Pf PR values between before (B) and after (A) conflict survey pairs by conflict level ( fatalities &length) Event level No. pairs B>A B10 104 58 2739 43 2412 0.554 0.581 L ength (day) 1 1787 1173 99626 2 558 50278 5 11.863 <0.001*** 2 342 234 3947 4 97 1547 3 6.888 <0.001*** 3 10 287 175 23195 97 13933 3.566 <0.001*** >10 43 29 598 11 222 2.527 0.012** (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
95 APPENDIX C SUPPLEMENTARY FIGURE S FOR CHAPTER 4 This document shows the supplementary figure s used in chapter 4
96 Figure C 1. Countries with highest and lowest GDP per capita and Plasmodium falciparum malaria endemicity. Panel A shows the wealthiest and poorest twenty countries in terms of GDP per capita in 2005  Panel B shows the P. falciparum malaria endemicity class in 2010. Areas were defined as stable transmission ( Pf API 0.1 per 1,000 pa), unstable transmission (medium grey, Pf API < 0.1 per 1,000 pa) and no risk (light grey, Pf API = 0 per 1,000 pa). Within the stable transmission, areas were further categorized into low (light red, Pf PR2 10), intermediate (medium red, 5% < Pf PR2 10< 40% ) and high (dark red, Pf PR2 10 ) risk 
97 Figure C 2 Boxplots showing the differences in lnP among areas with no (N), unstable (U) and stable (S) risk of Plasmodium falciparum malaria by country (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
98 Figure C 3 Boxplots showing the differences in lnP among areas with low (L), intermediate (I) and high (H) risk of Plasmodium falciparum malaria by country (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
99 Figure C 4 Boxplots showing the differences in lnG among ar eas with no (N), unstable (U) and stable (S) risk of Plasmodium falciparum malaria by country (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
100 Figure C 5 Boxplots showing the differences in lnG among areas with low (L), intermediate (I) and high (H) risk of Plasmodium falciparum malaria by country (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
101 Figure C 6 Boxplots showing the differences in NLDI among areas with no (N), unstable (U) and stable (S) risk of Plasmodium falciparum malaria globally and by region. (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1). Africa+=Africa, Saudi Arabia and Yemen; CSE Asia=Central and South East Asia.
102 Figure C 7 Boxplots showing the differences in NLDI among areas with no (N), unstable (U) and stable (S) risk of Plasmodium falciparum malaria by country (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1). Africa+=Africa, Saudi Arabia and Yemen; CSE Asia=Central and South East Asia.
103 Fig ure C 8 Boxplots showing the differences in NLDI among areas with low (L), intermediate (I) and high (H) risk of Plasmodium falciparum malaria globally, by region and by country (*) denotes the significant level of the test results (***=P<0.01, **=P<0.05, *=P<0.1).
104 LIST OF REFERENCES 1. World Health Organization: World Malaria Report 2011. Geneva: World Health Organization, 2011. 2. Hay SI, Okiro EA, Gething PW, Patil AP, Tatem AJ, Guerra CA, Snow RW: Estimating the global clinical burden of Plasmodium falciparum malaria in 2007. PLoS Med 2010, 7: e1000290. 3. Sutherst RW: Global change and human vulnerability to vector borne diseases. Clin Microbiol Rev 2004, 17: 136173. 4. Reisen WK: Landscape epidemiology of vector borne diseases. Annu Rev Entomol 2010, 55: 461 483. 5. McMichael AJ: Environmental and social influences on emerging infec tious diseases: past, present and future. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 2004, 359: 10491058. 6. Patz JA, Daszak P, Tabor GM, Aguirre AA, Pearl M, Epstein J, Wolfe ND, Kilpatrick AM, Foufopoulos J, M olyneux D, Bradley DJ: Unhealthy landscapes: policy recommendations on land use change and infectious disease emergence. Environ Health Perspect 2004, 112: 10921098. 7. Weiss RA, McMichael AJ: Social and environmental risk factors in the emergence of infectious diseases. Nat Med 2004, 10: 70 76. 8. Harrus S, Baneth G: Drivers for the emergence and reemergence of vector borne protozoal and bacterial diseases. Int J Parasitol 2005, 35: 13091318. 9. Guerra CA, Hay S, Lucioparedes L, Gikandi P, Tatem A, Noor A, Snow R: Assembling a global database of malaria parasite prevalence for the Malaria Atlas Project. Malar J 2007, 6: 17. 10. Malaria Atlas Project [ www.map.ox .ac.uk ] 11. Moyes C, Temperley W, Henry A, Burgert C, Hay S: Providing open access data online to advance malaria research and control. Malar J 2013, 12: 161. 12. Tatem A, Guerra C, Kabaria C, Noor A, Hay S: Human population, urban settlement patterns and t heir impact on Plasmodium falciparum malaria endemicity. Malar J 2008, 7: 218. 13. Hay SI, Guerra CA, Tatem AJ, Atkinson PM, Snow RW: Urbanization, malaria transmission and disease burden in Africa. Nat Rev Micro 2005, 3: 81 90.
105 14. Guerra CA, Howes RE, Patil AP, Gething PW, Van Boeckel TP, Temperley WH, Kabaria CW, Tatem AJ, Manh BH, Elyazar IRF, Baird JK, Snow RW, Hay SI: The international limits and population at risk of Plasmodium vivax transmission in 2009. PLoS Negl Trop Dis 201 0, 4: e774. 15. Sinka M, Bangs M, Manguin S, Chareonviriyaphap T, Patil A, Temperley W, Gething P, Elyazar I, Kabaria C, Harbach R, Hay S: The dominant Anopheles vectors of human malaria in the AsiaPacific region: occurrence data, distribution maps and bionomic precis. Parasit Vectors 2011, 4: 89. 16. Sinka M, Bangs M, Manguin S, Coetzee M, Mbogo C, Hemingway J, Patil A, Temperley W, Gething P, Kabaria C, Okara R, Van Boeckel T, Godfray HC, Harbach R, Hay S: The dominant Anophele s vectors of human malaria in Africa, Europe and the Middle East: occurrence data, distribution maps and bionomic precis. Parasit Vectors 2010, 3: 117. 17. Sinka M, Rubio Palis Y, Manguin S, Patil A, Temperley W, Gething P, Van Boeckel T, Kabaria C, Harbach R, Hay S: The dominant Anopheles vectors of human malaria in the Americas: occurrence data, distribution maps and bionomic precis. Parasit Vectors 2010, 3: 72. 18. Gayer M, Legros D, Formenty P, Connolly MA: Conflict and emerging infectious diseases. Emerg Infect Dis 2007, 13: 1625163 1. 19. Frst T, Raso G, Acka CA, Tschannen AB, N'Goran EK, Utzinger J: Dynamics of socioeconomic risk factors for neglected tropical diseases and malaria in an armed conflict. PLoS Negl Trop Dis 2009, 3: e513. 20. Reilley B, Abeyasinghe R, Pakianathar MV: Barriers to prompt and effective treatment of malaria in northern Sri Lanka. Trop Med Int Health 2002, 7: 744 749. 21. Gallup JL, Sachs JD: The economic burden of malaria. Am J Trop Med Hyg 2001, 64: 8596. 22. Bates I, Fenton C, Gruber J, Lalloo D, Lara AM, Squire SB, Theobald S, Thomson R, Tolhurst R: Vulnerability to malaria, tuberculosis, and HIV/AIDS infection and disease. Part 1: determinants operating at individual and household level. The Lancet Infectious Diseases 2004, 4: 267 277. 23. Sachs J, Malaney P: The economic and social burden of malaria. Nature 2002, 415: 680 685. 24. Worrall E, Basu S, Hanson K: Is malaria a disease of poverty? A review of the literature. Trop Med Int Health 2005, 10: 1047 1059.
106 25. United Nations: World Urbaniz ation Prospect, the 2011 Revision. New York: United Nations, 2011. 26. Vlahov D, Galea S: Urbanization, urbanicity, and health. J Urban Health 2002, 79: S1 S12. 27. Vlahov D, Freudenberg N, Proietti F, Ompad D, Quinn A, Nandi V, Galea S: Urban as a determin ant of health. J Urban Health 2007, 84: 1626. 28. Dye C: Health and urban Living. Science 2008, 319: 766 769. 29. Utzinger J, Keiser J: Urbanization and tropical health: then and now. Ann Trop Med Parasitol 2006, 100: 517 533. 30. Moore M, Gould P, Keary BS: Global urbanization and impact on health. Int J Hyg Envir Heal 2003, 206: 269 278. 31. Robert V, Macintyre K, Keating J, Trape J F, Duchemin J B, Warren M, Beier JC: Malaria transmission in urban subSaharan Africa. Am J Trop Med Hyg 2003, 68: 169 176. 32. Donnelly M, McCall PJ, Lengeler C, Bates I, D'Alessandro U, Barnish G, Konradsen F, Klinkenberg E, Townson H, Trape J F, Hastings I, Mutero C: Malaria and urbanization in subSaharan Africa. Malar J 2005, 4: 12. 33. Trape JF: Mala ria and urbanization in Central Africa: the example of Brazzaville: Part IV. Parasitological and serological surveys in urban and surrounding rural areas. Trans R Soc Trop Med Hyg 1987, 81: 2633. 34. Trape JF, Zoulani A: Malaria and urbanization in Central Africa: the example of Brazzaville: Part III: Relationships between urbanization and the intensity of malaria transmission. Trans R Soc Trop Med Hyg 1987, 81: 1925. 35. Knudsen AB, Sloff R: Vector borne disease problems in rapid urbanization: new approaches to vector control. Bull World Health Organ 1992, 70: 1 6. 36. Keiser J, Utzinger J, De Castro MC, Smith TA, Tanner M, Singer BH: Urbanization in subSaharan Africa and implication for malaria control. Am J Trop Med Hyg 2004, 71: 118 127. 37. Mendis K, Sina B, Marchesini P, Carter R: The neglected burden of Plasmodium vivax malaria. Am J Trop Med Hyg 2001, 64: 97 106. 38. Price RN, Tjitra E, Guerra CA, Yeung S, White NJ, Anstey NM: Vivax malaria: neglected and not benign. Am J Trop Med Hyg 2007, 77: 79 87.
107 39. Baird JK: Neglect of Plasmodium vivax malaria. Trends Parasitol 2007, 23: 533 539. 40. Guerra CA, Snow RW, Hay SI: Defining the global spatial limits of malaria transmission in 2005. In Adv Parasitol. Volume 62. Edited by Simon I. Hay AG, David JR: Acad emic Press; 2006: 157179 41. Guerra CA, Snow RW, Hay SI: Mapping the global extent of malaria in 2005. Trends Parasitol 2006, 22: 353 358. 42. Snow RW, Guerra CA, Noor AM, Myint HY, Hay SI: The global distribution of clinical episodes of Plasmodium falcipa rum malaria. Nature 2005, 434: 214 217. 43. Hay SI, Guerra CA, Gething PW, Patil AP, Tatem AJ, Noor AM, Kabaria CW, Manh BH, Elyazar IRF, Brooker S, Smith DL, Moyeed RA, Snow RW: A world malaria map: Plasmodium falciparum endemicity in 2007. PLoS Med 2009, 6: e1000048. 44. Gething P, Patil A, Smith D, Guerra C, Elyazar I, Johnston G, Tatem A, Hay S: A new world malaria map: Plasmodium falciparum endemicity in 2010. Malar J 2011, 10: 378. 45. Foley DH, Rueda LM, Wilkerson RC: Insight into global mosquito biogeography from country species records. J Med Entomol 2007, 44: 554 567. 46. Sharma VP: Fighting malaria in India. Current science (Bangalore) 1998, 75: 1127. 47. Surendran SN, Ramasamyb R: Some characteristics of the larval breeding sites of Anopheles culicifacies species B and E in Sri Lanka. J Vector Dis 2005, 42: 3944. 48. Roberts D: Mosquitoes (Diptera: Culicidae) breeding in brackish water: female ovipositional preferences or larval survival? J Med Entomol 1996, 33: 525 530. 49. Batra CP, Mittal PK, A dak T, Subbarao SK: Efficacy of Agnique MMF monomolecular surface film against Anopheles stephensi breeding in urban habitats in India. J Am Mosquito Contr 2006, 22: 426432. 50. Biswas D, Dutta R, Ghosh S, Chatterjee K, Hati A: Breeding habits of Anopheles stephensi Liston in an area of Calcutta. Indian J Malariol 1992, 29: 195 198.
108 51. Guerra CA, Gikandi PW, Tatem AJ, Noor AM, Smith DL, Hay SI, Snow RW: The limits and intensity of Plasmodium falciparum transmission: implications for malaria control and elimination worldwide. PLoS Med 2008, 5: e38. 52. Danko DM: The digital chart of the world project. Photogramm Eng Rem S 1992, 58: 11251128. 53. Balk DL, Deichmann U, Yetman G, Pozzi F, Hay SI, Nelson A: D etermining global population distribution: methods, applications and data. Adv Parasitol 2006, 62: 119156. 54. Hansen M, DeFries R, Townshend JRG, Sohlberg R: 1 km land cover classification derived from AVHRR 55. Elvidge CD, Baugh KE, Kihn EA, Kroehl HW, Davis ER: Mapping city lights with nighttime data from the DMSP Operational Linescan System. Photogramm Eng Rem S 1997, 63: 727734. 56. Schneider A, Friedl MA, Potere D: A new map of global urban extent from MODIS data. Environ Res Lett 2009, 4: 4. 57. Schn eider A, Friedl MA, Potere D: Monitoring urban areas globally using MODIS 500m data: New methods and datasets based on urban ecoregions. Remote Sens Environ 2010, 114: 17331746. 58. Tatem A, Noor AM, Hay SI: Assessing the accuracy of satellite derived global and national urban maps in Kenya. Remote Sens Environ 2005, 96: 8797. 59. Center for Internation Earth Science Information Network(CIESIN), Columbia University, International Food Policy Research Institute (IFPRI), The World Band, Centro Internation al de Agricultura Tropical (CIAT): Global Rural Urban Mapping Project (GRUMP): Urban Extents 60. Hay SI, Sinka ME, Okara RM, Kabaria CW, Mbithi PM, Tago CC, Benz D, Gething PW, Howes RE, Patil AP, Temperley WH, Bangs MJ, Chareonviriyaphap T, Elyazar IRF, Harbach RE, Hemingway J, Manguin S, Mbogo CM, RubioPalis Y, Godfray HCJ: Developing global maps of the dominant Anopheles vectors of human malaria. PLoS Med 2010, 7: e1000209. 61. Gupta D, Bhatt R, Sharma R, Gautam A, Rajnikant: Intradomestic mosquito bree ding sources and their management. Indian J Malariol 1992, 29: 4146. 62. Wilcoxon F: Individual comparisons by ranking methods. Biometrics 1945, 1: 761 764.
109 63. Montgomery MR: The urban transformation of the developing world. Science 2008, 319: 761 764. 64. Goncalves MJF, Alecrim WD: Non planned urbanization as a contributing factor for malaria incidence in ManausAmazonas, Brazil. Rev salud pblica 2004, 6: 156 166. 65. Miller LH, Mason SJ, Clyde DF, McGinniss MH: The resistance factor to Plasmodium vivax in Blacks. New Engl J Med 1976, 295: 302 304. 66. Howes RE, Patil AP, Piel FB, Nyangiri OA, Kabaria CW, Gething PW, Zimmerman PA, Barnadas C, Beall CM, Gebremedhin A, Menard D, Williams TN, Weatherall DJ, Hay SI: The global distribution of the Duffy blood group. Nat Commun 2011, 2: 266. 67. Sutton PC: A scaleadjusted measure of Urban sprawl using nighttime satellite imagery. Remote Sens Environ 2003, 86: 353 369. 68. Vora N: Impact of anthropogenic enviromental alterations on vector borne diseases. Medscape J Med 2008, 10: 238. 69. Molyneux DH: Patterns of change in vector borne diseases. Ann Trop Med Parasitol 1997, 91: 827840. 70. Norris DE: Mosquitoborne diseases as a consequence of land use change. EcoHealth 2004, 1: 19 24. 71. Chinery WA: Effects of eco logical changes on the malaria vectors Anopheles funestus and the Anopheles gambiae complex of mosquitoes in Accra, Ghana. J Trop Med Hyg 1984, 87: 191206. 72. Awolola TS, Oduola AO, Obansa JB, Chukwurar NJ, Unyimadu JP: Anopheles gambiae s.s. breading in polluted water bodies in urban Lagos, southwestern Nigeria. J Vector Dis 2007, 44: 241 244. 73. AntonioNkondjio C, Simard F, AwonoAmbene P, Ngassam P, Toto J C, Tchuinkam T, Fontenille D: Malaria vectors and urbanization in the equatorial forest region of south Cameroon. Trans R Soc Trop Med Hyg 2005, 99: 347 354. 74. Afrane YA, Klinkenberg E, Drechsel P, Owusu Daaku K, Garms R, Kruppa T: Does irrigated urban agriculture influence the transmission of malaria in the city of Kumasi, Ghana? Acta Trop 2004, 89: 125 134. 75. Matthys B, Vounatsou P, Raso G, Tschannen AB, Becket EG, Gosoniu L, Cisse G, Tanner M, Ngoran EK, Utzinger J: Urban farming and malaria risk factors
110 in a medium sized town in Cte d'Ivoire. Am J Trop Med Hyg 2006, 75: 12231231. 76. Klinkenber g E, McCall P, Wilson M, Amerasinghe F, Donnelly M: Impact of urban agriculture on malaria vectors in Accra, Ghana. Malar J 2008, 7: 151. 77. Stoler J, Weeks JR, Getis A, Hill AG: Distance threshold for the effect of urban agriculture on elevated self reported malaria prevalence in Accra, Ghana. Am J Trop Med Hyg 2009, 80: 547 554. 78. Mendez F, Carrasquilla G, Muoz A: Risk factors associated with malaria infection in an urban setting. Trans R Soc Trop Med Hyg 2000, 94: 367 371. 79. Pinker S: The Better Angels of our Nature. New York: Viking; 2011. 80. Themnr L, Wallensteen P: Armed conflict, 1946 2010. Journal of Peace Research 2011, 48: 525 536. 81. Themnr L, Wallensteen P: Armed Conflicts, 1946 2011. Journal of Peace Research 2012, 49: 565 575. 82. Uppsala University, Peace Research Institute Oslo (PRIO): UCDP/PRIO Armed Conflict Dataset v.42013, 19462012 [ http://www.pcr.uu.se/research/ucdp/datasets/ucdp _prio_armed_conflict_dataset/ ] 83. Armed Conflict Location and Events Dataset (ACLED) [ http://www.acleddata.com/ ] 84. Li Q, Wen M: The immediate and lingering effects of armed conflict on adult mortality: a time seri es crossnational analysis. Journal of Peace Research 2005, 42: 471 492. 85. Pedersen D: Political violence, ethnic conflict, and contemporary wars: broad implications for health and social well being. Soc Sci Med 2002, 55: 175 190. 86. Hillier D: Africa's m issing billion: International arms flows and the cost of conflict. The International Action Network on Small Arms and Oxfam International, 2007. 87. United Nations High Commissioner for Refugees (UNHCR): Statistical Yearbook 2011. 2012. 88. Internal Displa cement Monitoring Centre (IDMC), Norwegian Refugee Council: Global overview 2011: People internally displaced by conflict and violence. Geneva, 2012.
111 89. United Nations: World Population Prospects, the 2012 Revision. New York: United Nations, 2013. 90. Lev y BS, Sidel VW: War and public health. Washington, DC: Ameican Public Health Association; 2002. 91. Utzinger J, Weiss MG: Armed conflict, war and public health. Trop Med Int Health 2007, 12: 903 906. 92. Connolly MA, Gayer M, Ryan MJ, Salama P, Spiegel P, Heymann DL: Communicable diseases in complex emergencies: impact and challenges. Lancet 2004, 364: 1974 1983. 93. Salama P, Spiegel P, Talley L, Waldman R: Lessons learned from complex emergencies over past decade. Lancet 2004, 364: 18011813. 94. Christophers SR: Malaria in war. Trans R Soc Trop Med Hyg 1939, 33: 277 292. 95. Prothero R: Forced movements of population and health hazards in tropical Africa. Int J Epidemiol 1994, 23: 657 664. 96. Toole M, Waldm an R: The public health aspects of complex emergencies and refugee situations. Annu Rev Public Health 1997, 18: 283 312. 97. Garfield RM, Frieden T, Vermund SH: Health related outcomes of war in Nicaragua. Am J Public Health 1987, 77: 615 618. 98. Beyrer C, Villar JC, Suwanvanichkij V, Singh S, Baral SD, Mills EJ: Neglected diseases, civil conflicts, and the right to health. Lancet 2007, 370: 619 627. 99. Rowland M, Nosten F: Malaria epidemiology and control in refugee camps and complex emergencies. Ann Trop M ed Parasitol 2001, 95: 741754. 100. Berrang Ford L: Civil conflict and sleeping sickness in Africa in general and Uganda in particular. Confl Health 2007, 1: 6. 101. Aliev SP: Malaria in the Republic of Tajikistan [in Russian]. Med Parazitol 2000, 2: 27 29. 102. World Health Organization: World Malaria Report 2012. Geneva: World Health Organization, 2012. 103. Messina J, Taylor S, Meshnick S, Linke A, Tshefu A, Atua B, Mwandagalirwa K, Emch M: Population, behavioural and environmental drivers of malaria preva lence in the Democratic Republic of Congo. Malar J 2011, 10: 161.
112 104. Martins J, Zwi A, Martins N, Kelly P: Malaria control in Timor Leste during a period of political instability: what lessons can be learned? Confl Health 2009, 3: 11. 105. World Health Organization: Malaria control in complex emergencies: An inter agency field handbook. World Health Organization, 2005. 106. Kolaczinski J: Roll Back Malaria in the aftermath of complex emergencies: the example of Afghanistan. Trop Med Int Heal th 2005, 10: 888 893. 107. Tatem A, Smith DL, Gething PW, Kabaria CW, Snow RW, Hay SI: Ranking of elimination feasibility between malaria endemic countries. Lancet 2010, 376: 15791591. 108. Abeyasinghe RR, Galappaththy GNL, Smith Gueye C, Kahn JG, Feachem R GA: Malaria control and elimination in Sri Lanka: documenting progress and success factors in a conflict setting. PLoS ONE 2012, 7: e43162. 109. Hay SI, Smith DL, Snow RW: Measuring malaria endemicity from intense to interrupted transmission. Lancet Infect Dis 2008, 8: 369378. 110. Gupta S, Snow R, Donnelly C, Newbold C: Acquired immunity and postnatal clinical protection in childhood cerebral malaria. Proc R Soc Lond Ser B Biol Sci 1999, 266: 33 38. 111. Baird J, Purnomo, Basri H, Bangs M, An dersen E, Jones T, Masbar S, Harjosuwarno S, Subianto B, Arbani P: Age specific prevalence of Plasmodium falciparum among six populations with limited histories of exposure to endemic malaria. The American Journal of Tropical Medicine and Hygiene 1993, 49: 707 719. 112. Smith D, Guerra C, Snow R, Hay S: Standardizing estimates of the Plasmodium falciparum parasite rate. Malar J 2007, 6: 131. 113. Macdonald G: Epidemiological basis of malaria control. Bull World Health Organ 1956, 15: 613 626. 114. Dietz K: T he estimation of the basic reproduction number for infectious diseases. Stat Methods Med Res 1993, 2: 2341. 115. Raleigh C, Linke A, Hegre H, Karlsen J: Introducing ACLED: an Armed Conflict Location and Event Dataset. Journal of Peace Research 2010, 47: 651 660.
113 116. Restrepo JA, Spagat M, Vargas JF: The severity of the Colombian conflict: crosscountry datasets versus new micro data. Journal of Peace Research 2006, 43: 99115. 117. Gleditsch NP, Wallendsteen P, Eriksson M, Sollenberg M, Strand H: Armed conflict 19462001: A new dataset. Journal of Peace Research 2002, 39: 615 637. 118. Qi Q, Guerra C, Moyes C, Elyazar IR, Gething P, Hay S, Tatem A: The effects of urbanization on global Plasmodium vivax malaria transmission. Malar J 2012, 11: 403. 119. Kulldorff M, Athas WF, Feurer EJ, Miller BA, Key CR: Evaluating cluster alarms: a spacetime scan statistic and brain cancer in Los Alamos, New Mexico. Am J Public Health 1998, 88: 13771380. 120. SaTScan [ http://www.satsca n.org/ ] 121. Kulldorff M, Heffernan R, Hartman J, Assuno R, Mostashari F: A space time permutation scan statistic for disease outbreak detection. PLoS Med 2005, 2: e59. 122. Hollander M, Wolfe DA: Nonparametric Statistical Methods. New York: John Wiley & Sons; 1973. 123. Akhavan P: The Lord's Resistance Army case: Uganda's submission of the first state referral to the International Criminal Court. The American Journal of International Law 2005, 99: 403 421. 124. Nannyonjo J: Conflicts, poverty and human de velopment in Northern Uganda. The Round Table 2005, 94: 473 488. 125. Van Acker F: Uganda and the Lords Resistance Army: the new order no one ordered. African Affairs 2004, 103: 335 357. 126. United Nations: Data Exchange Platform for the Horn of Africa (DE PHA): Post election violence, Kenya [ http://www.depha.org/Post_election_Violence.asp ] 127. Gettleman J: Displaced Kenyans live in limbo as aid lags after election strife. The New York Times 6 A pril 2008. 128. United Nations: Report of the Secretary General on the Situation in Burundi. Security Council, S/1996/660. New York: United Nations, 1996.
114 129. Verwimp P: Undernutrition, subsequent risk of mortality and civil war in Burundi. Econ Hum Biol 2012, 10: 221 231. 130. Somalia conflict kills more than 2,100 this year. International Herald Tribune, 29 March 2009. 131. Green S: Zimbabwe land conflict [ http://www1.american.edu/ted/ice/zimbabwe.htm] Updated 5 May 2004. 132. BBC News: Ethiopia says 'war is over' [ http://news.bbc.co.uk/2/hi/africa/771972.stm] Updated 31 May 2000. 133. GlobalS ecurity.org: Republic of Congo Civil War [ http://www.globalsecurity.org/military/world/war/congob.htm] Updated 11 July 2011. 134. Fall A: Understanding The Casamance Conflict: A Background. Accra: Kofi Annan International Peacekeeping and Training Centre (KAIPTC); 2010. 135. Uppsala Conflict Data Program Conflict Encyclopedia: Guinea Bissau [ http://www.ucdp.uu.se/gpdatabase/gpcountry.php?id=68®ionSelect=2Southern_Africa#] Cited 12 July 2013. 136. Protopopoff N, Van Herp M, Maes P, Reid T, Baza D, D'Alessandro U, Van Bortel W, Coosemans M: Vector control in a malaria epidemic occurring within a complex emergency situation in Burundi: a case study. Malar J 2007, 6: 93. 137. Pigott D, Atun R, Moyes C, Hay S, Gething P: Funding for malaria control 20062010: A comprehensive global assessment. Malar J 2012, 11: 246. 138. O'M eara WP, Mangeni JN, Steketee R, Greenwood B: Changes in the burden of malaria in sub Saharan Africa. Lancet Infect Dis 2010, 10: 545 555. 139. Murray CJL, Rosenfeld LC, Lim SS, Andrews KG, Foreman KJ, Haring D, Fullman N, Naghavi M, Lozano R, Lopez AD: Glo bal malaria mortality between 1980 and 2010: a systematic analysis. Lancet 2012, 379: 413 431. 140. Tatem A, Gething P, Smith D, Hay S: Urbanization and the global malaria recession. Malar J 2013, 12: 133. 141. Noor AM, Mutheu JJ, Tatem AJ, Hay SI, Snow RW: Insecticide treated net coverage in Africa: mapping progress in 2000 07. Lancet 2009, 373: 58 67. 142. Devkota B, van Teijlingen E: Understanding effects of armed conflict on health outcomes: the case of Nepal. Confl Health 2010, 4: 20.
115 143. Human Security R eport Project: Human Security Report 2009/2010: The Causes of Peace and the Shrinking Costs of War. New York: Oxford University Press, 2011. 144. Sodemann M, Veirum J, Biai S, Nielsen J, Bale C, Jakobsen MS, Gustafson P, Aaby P: Reduced case fatality among hospitalized children during a war in Guinea Bissau: a lesson in equity. Acta Paediatr 2004, 93: 959 964. 145. Betsi NA, Koudou BG, Ciss G, Tschannen AB, Pignol AM, Ouattara Y, Madougou Z, Tanner M, Utzinger J: Effect of an armed conflict on human resourc es and health systems in Cte d'Ivoire: Prevention of and care for people with HIV/AIDS. AIDS Care 2006, 18: 356365. 146. OLoughlin J, Witmer FDW, Linke AM, Laing A, Gettelman A, Dudhia J: Climate variability and conflict risk in East Africa, 1990 2009. Proceedings of the National Academy of Sciences 2012, 109: 18344 18349. 147. Hendrix CS, Salehyan I: Climate change, rainfall, and social conflict in Africa. Journal of Peace Research 2012, 49: 35 50. 148. Famine Early Warning Systems Network (FEWS NET): [ http://earlywarning.usgs.gov/fews/datadownload.php?product=eta&periodicity=m onthly®ion=aa ] 149. Allen R, Tasumi M, Trezza R: Satellite based energy balance for mapping evapotranspiration with internalized calibration (METRIC) Model. Journal of Irrigation and Drainage Engineering 2007, 133: 380 394. 150. Shikanga O T, Mutonga D, Abade M, Amwayi S, Ope M, Limo H, Mintz ED, Quick RE, Breiman RF, Feikin DR: High mortality in a cholera outbreak in Western Kenya after Post election violence in 2008. Am J Trop Med Hyg 2009, 81: 10851090. 151. Feikin DR, Adazu K, Obor D, Ogwang S, Vulule J, Hamel MJ, Laserson K: Mortality and health among internally displaced p ersons in western Kenya following post election violence, 2008: novel use of demographic surveillance. Bull World Health Organ 2010, 88: 601 608. 152. Nyazema NZ: The Zimbabwe crisis and the provision of social services: health and education. Journal of Developing Societies 2010, 26: 233 261. 153. Hamel MJ, Adazu K, Obor D, Sewe M, Vulule J, Williamson JM, Slutsker L, Feikin DR, Laserson KF: A reversal in reductions of child mortality in western Kenya, 20032009. The American Journal of Tropical Medicine and Hygiene 2011, 85: 597 605.
116 154. Vreeman R, Nyandiko W, Sang E, Musick B, Braitstein P, Wiehe S: Impact of the Kenya post election crisis on clinic attendance and medication adherence for HIV infected children in western Kenya. Confl Health 2009, 3: 5. 155. Internatinal Crisis Group: Kenya in crisis. Africa Report N137 21 February 2008. 156. Africa: Ivory Coast profile. BBC News 18 June 2013. 157. Africa: Liberia profiles. BBC News 19 June 2013. 158. Africa: Burundi profile. BBC News 5 June 2013. 159. GlobalSecurity.org: Somalia civil war [ http://www.globalsecurity.org/military/world/war/somalia.htm] Updated 10 May 2013. 160. The world Bank: World Development Indicators (WDI) [ http://data.worldbank.org/products/wdi ] 161. Jobin WR: Suppression of malaria transmission and increases in economic productivity in African countries from 2007 to 2011. MWJ 2014, 5: 4. 162. Teklehaimanot A, Mejia P: Malaria and poverty. Ann N Y Acad Sci 2008, 1136: 3237. 163. Tusting LS, Willey B, Lucas H, Thompson J, Kafy HT, Smith R, Lindsay SW: Socioeconomic development as an intervention against malaria: a systematic review and metaanalysis. Lancet 2013, 382: 963 972. 164. Noor A, Alegana V, Gething P, Tatem A, Snow R: Using remotely sensed night time light as a proxy for poverty in Africa. Population Health Metrics 2008, 6: 5. 165. Stillwaggon E: The ecology of poverty: Nutrition, par asites, and vulnerability to HIV/AIDS. AIDS, Poverty and Hunger: Challenges and Responses International Food Policy Research Institute, Washington 2006 : 167180. 166. Hess F, Nukuro E, Judson L, Rodgers J, Nothdurft H, Rieckmann K: Anti malarial drug resistance, malnutrition and socio economic status. Trop Med Int Health 1997, 2: 721728. 167. Plsson K, Jaenson TGT, Dias F, Laugen AT, Bjrkman A: Endophilic Anopheles Mosquitoes in Guinea Bissau, West Africa, in Relation to Human Housing Conditions. J M ed Entomol 2004, 41: 746 752.
117 168. Ye Y, Hoshen M, Louis V, Seraphin S, Traore I, Sauerborn R: Housing conditions and Plasmodium falciparum infection: protective effect of iron sheet roofed houses. Malar J 2006, 5: 8. 169. Butraporn P: Social epidemiology of malaria. In Social and economic aspects of malaria control. Edited by Sornmani S, Fingladda W. Bangkok: MRCTROPMED; 1991: 14 35 170. United Nations: The Millennium Development Goals Report 2010. New York: United Nations, 2010. 171. Ettling M, McFarland DA Schultz LJ, Chitsulo L: Economic impact of malaria in Malawian households. Tropical medicine and parasitology : official organ of Deutsche Tropenmedizinische Gesellschaft and of Deutsche Gesellschaft fur Technische Zusammenarbeit (GTZ) 1994, 45: 7479. 172. Macintyre K, Keating J, Sosler S, Kibe L, Mbogo C, Githeko A, Beier J: Examining the determinants of mosquitoavoidance practices in two Kenyan cities. Malar J 2002, 1: 14. 173. Filmer D: Fever and its treatment among the more and less poor in subSaharan Africa. Health Policy Plan 2005, 20: 337 346. 174. Biritwum RB, Welbeck J, Barnish G: Incidence and management of malaria in two communities of different socioeconomic level, in Accra, Ghana. Ann Trop Med Parasitol 2000, 94: 771 778. 175. Erhart A, Thang ND, Hung NQ, Toi LV, Hung LX, Tuy TQ, Cong LD, Speybroeck N, Coosemans M, D'Alessandro U: Forest malaria in Vietnam: a challenge for control. The American journal of tropical medicine and hygiene 2004, 70: 110118. 176. Hoek W: How can bette r farming methods reduce malaria. Acta Tropica 2004, 89: 9597. 177. Somi MF, Butler JR, Vahid F, Njau JD, Kachur SP, Abdulla S: Use of proxy measures in estimating socioeconomic inequalities in malaria prevalence. Trop Med Int Health 2008, 13: 354 364. 178. Nordhaus WD: Geography and macroeconomics: New data and new findings. Proc Natl Acad Sci U S A 2006, 103: 3510 3517. 179. An Estimate of Gross Domestic Product (GDP) Derived From Satellite Data [ http://ngdc.noaa.gov/eog/dmsp/download_gdp.html]
118 180. Night Light Development Index (NLDI) [ http://ngdc.noaa.gov/eog/dmsp/download_nldi.html] 181. Deaton A, Zaidi S: Guidelines for constructing consumption aggregates for welfare analysis. Living Standards Measurement Study Working Paper, No. 135. Washington D.C: World Bank; 2002. 182. Falkingham J, Namazie C: Measuring health and poverty: a review of approaches to identifying the poor. London: DFID Health Systems Resource Centre; 2002. 183. Nordhaus W, Azam Q, Corderi D, Hood K, Victor NM, Mohammed M, Miltner A, Weiss J: The G Econ database on gridded output: methods and data. Yale University 2006. 184. Geographically based Economic d ata (G Econ) [ http://gecon.yale.edu/ ], 185. Elvidge CD, Sutton PC, Ghosh T, Tuttle BT, Baugh KE, Bhaduri B, Bright E: A global poverty map derived from satellite data. Computers & Geosciences 2009, 35: 16521660. 186. Doll CNH, Muller J P, Morley JG: Mapping regional economic activity from night time light satellite imagery. Ecological Economics 2006, 57: 7592. 187. Doll CNH, Muller J P, Elvidge CD: Night time imagery as a tool for global mapping of socioeconomic parame ters and greenhouse gas emissions. AMBIO: A Journal of the Human Environment 2000, 29: 157 162. 188. National Geographical Data Center (NGDC) [ http://ngdc.noaa.gov/ngdc.html ] 189. Ghosh T, Powell RL, Elvidge CD, Baugh KE, Sutton PC, Anderson S: Shedding light on the global distribution of economic activity. The Open Geography Journal 2010, 3: 148161. 190. Dobson JE, Bright EA, Coleman PR, Durfee RC, Worley BA: LandScan: a global population database for estimating populations at risk. Photogrammetric engineering and remote sensing 2000, 66: 849 857. 191. Bhaduri B, Bright E, Coleman P, Dobson J: LandScan. Geoinformatics 2002, 5: 34 37. 192. Elvidge C, Baugh K, Anderson S, Sutton P, Ghos h T: The Night Light Development Index (NLDI): a spatially explicit measure of human development from satellite data. Social Geography 2012, 7: 23 35.
119 193. Tucker CJ, Pinzon JE, Brown ME: Global Inventory Modeling and Mapping Studies, NA94apr15b.n11VIg, 2.0. Global Land Cover Facility, University of Maryland, College Park, Maryland, 04/15/1994. 194. Pinzon J, Brown ME, Tucker CJ: Satellite time series correction of orbital drift artifacts using empirical mode decomposition. In Hilbert Huang transform: intr oduction and applications. Edited by Huang N: Hackensack, NJ: World Scientific; 2005 195. Tucker CJ, Pinzon JE, Brown ME, Slayback DA, Pak EW, Mahoney R, Vermote EF, El Saleous N: An extended AVHRR 8 km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. International Journal of Remote Sensing 2005, 26: 44854498. 196. The World Bank: World Development Indicators [ http://databank.worldbank.org/data/views/variableSelection/selectvariables.aspx ?source=worlddevelopment indicators ] 197. Davidson R, MacKinnon JG: Estimation and Inference in Econometrics. New York: Oxford University Press; 1993. 198. Greene WH: Econometr ic Analysis. 6th ed. Upper Saddle River, NJ: Prentice Hall; 2008. 199. Koram K, Bennett S, Adiamah J, Greenwood B: Socio economic risk factors for malaria in a peri urban area of Gambia. Trans R Soc Trop Med Hyg 1995, 89: 146 150. 200. Bowden S, Michailidou DM, Pereira A: Chasing mosquitoes: An exploration of the relationship between economic growth, poverty and the elimination of malaria in Southern Europe in the 20th century. Journal of International Development 2008, 20: 10801106. 201. McCarthy FD, Wolf H C, Wu Y: Malaria and growth. World Bank Policy Research Working Paper 2000. 202. Tusting LS, Willey B, Lucas H, Thompson J, Kafy HT, Smith R, Lindsay SW: Socioeconomic development as an intervention against malaria: a systematic review and metaanalysis. T he Lancet 2013. 203. Schellenberg JA, Victora CG, Mushi A, de Savigny D, Schellenberg D, Mshinda H, Bryce J: Inequities among the very poor: health care for children in rural southern Tanzania. The Lancet 2003, 361: 561 566. 204. Bonds MH, Dobson AP, Keenan DC: Disease ecology, biodiversity, and the latitudinal gradient in income. PLoS Biol 2012, 10: e1001456.
120 205. Yamada T: Causal relationships between infant mortality and fertility in developed and less developed countries. Southern Economic Journal 1985, 52: 364 370. 206. Doll CN, Muller J P, Morley JG: Mapping regional economic activity from night time light satellite imagery. Ecological Economics 2006, 57: 7592. 207. World Health Organization: World malaria report 2013. Geneva: World Health Organization, 2013. 208. Cohen JM, Smith DL, Cotter C, Ward A, Yamey G, Sabot OJ, Moonen B: Malaria resurgence: a systematic review and assessment of its causes. Malar J 2012, 11: 122. 209. Barat LM, Palmer N, Basu S, Worrall E, Hanson K, Mills A: Do malaria control inte rventions reach the poor? Am J Trop Med Hyg 2004, 71 210. Garca Martn G: Status of malaria eradication in the Americas. The American journal of tropical medicine and hygiene 1972, 21: 617. 211. Bruce Chwatt LJ, De Zulueta J: The rise and fall of malaria in Europe: a historico epidemiological study. London: Oxford University Press; 1980.
121 BIOGRAPHICAL SKETCH Qiuyin Qi was born in Yangzhou, China, where she also grew up. She earned her Bachelor of Management in t ourism r esource m anagement in 2008 and Master of Science in g eography in 2010 from Nanjing University. In 2010, she entered the Ph.D program in Department of Geography at University of Florida Her primary research interest is in spatial epidemiology and spatial modeling of disease dynamics, with particular focuses on global health, infectious disease, disease mapping and impacts of environmental and socioecon omic factors.
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EP9SVUN0Q_ZK8L6I INGEST_TIME 2014-10-03T22:27:11Z PACKAGE UFE0046597_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC