Impact of the 2013 Floods on the Incidence of Malaria in Almanagil Locality , Gezira State , Sudan

Authors Yasir Elfatih Abdelrahim Elsanousi Abbas Suleiman Elmahi

Besides the rise in global temperature, climate change also results in an increased frequency of extreme weather events such as floods and droughts .Floods are the most common natural disasters globally and have led to extensive morbidity and mortality throughout the world especially in low-resource countries .Impacts of floods on the health of populations depend not just on the magnitude of the flood, but also on geographic, socioeconomic and human factors including the vulnerability and adaptive capacity of the population affected .Climate change, resulting in rising temperatures, changing precipitation patterns and increase in extreme weather events, can affect infectious disease outbreaks by altering vector population, density and survival rates, and pathogen reproduction and development rate .This in turn will influence human exposure to bites from infected vectors .
Malaria has been identified as one of the diseases most sensitive to climatic factors .Conflicting results exist on the impact of extreme weather events on malaria transmission .Malaria epidemics attributed to prolonged precipitation and flooding have been reported in endemic areas world-wide .However, other studies showed a decrease in the malaria burden or did not provide evidence in support of causal relationship between malaria transmission and heavy rainfall or flooding .Although the two opinions seem to be contradictory, both viewpoints can be true.Heavy precipitations and floods may initially decrease vector populations by eliminating existing mosquito-breeding sites and, hence, lower malaria transmission .However, as the heavy rainfall stops and floods waters gradually recede, stagnant pools are created, providing ideal habitats for mosquitoes and resulting in an increase in the vector population and an upsurge of malaria transmission in the following weeks .Moreover, these differences in the impact of heavy rainfall and flooding may also be the result of the specific geographic setting of the affected area such as topography, distance to water bodies, altitude, climate and land type .The lag time, which may vary by geographic and climatic conditions, is usually around 4 to 8 weeks between the flooding and the onset of a malaria outbreak .
Malaria is considered a leading cause of mortality and morbidity around the world, with a global 2013 estimate of 198 million cases and about 584000 deaths.Africa accounts for an estimated 90 % of all deaths and children aged under five represented 78 % of all deaths .Malaria in Sudan is endemic and continues to be a major public health problem.The whole population is at varying degrees of risk.Climate models estimate that 75 % of the population of Sudan are at risk of endemic malaria and 25 % are at risk of epidemic malaria .Plasmodium

Funding Statement
Background 1, The objective of the study was to estimate the malaria incidence attributable to the 2013 flood in Almanagil Locality, Gezira State in central Sudan.

Study Area
Gezira State is located in the east-central region of Sudan, is crossed by the Blue Nile and is irrigated by two canals of the Gezira and Managil agriculture schemes.Almanagil Locality was in 2013 one of the 8 localities that constituted the Gezira State and was situated in the south western part of the state.Almanagil Town, the capital of the locality, is 62 km away from Wadmedani, the capital of Gezira State, and 156 km from Khartoum, the capital of Sudan (Figure 1).

Malaria case definition
The malaria case definition remained unchanged throughout the 3 years of study and was based on the national protocol for diagnosis and treatment of malaria: "A malaria case is confirmed by demonstration of asexual forms (trophozoite stage) of the parasite in a thick or thin peripheral blood film or by rapid diagnostic test (RDT) in the presence of fever" .

Data collection
The sources for data collection are twofold: the sentinel surveillance system and the routine health management information system operated by the Gezira State As heavy rain and subsequent flash flood occurred in the beginning of August and taking into account that the effect of these climatic factors on malaria is not immediate, a four-week lag effect was assumed .A new generation of infective vectors needs about 30 days to develop: 15 days for the preimaginal development of vector Anopheles, 4-7 days for the gonadotropic cycle for parous/nulliparous female mosquitoes and 12 days for the sporogonic cycle for the Plasmodium falciparum parasites in the vector mosquitoes .As a result, the study compared data of the months September, October and November (hospital data registered by Gezira State Ministry of Health) and the corresponding 12 weeks from 36th week to 47th week (data registered by 13 SMNSs) of the year of the flooding (2013) with data of the same periods in the two preceding years without flooding (2011 and 2012).
Estimates of the general population were obtained from the Preventive Medicine Department and under 5 years population from the Vaccination Department in Almanagil Locality.
To determine the impact of flooding on malaria burden in Almanagil Locality the following outcomes of interest were selected: (1) the malaria incidence rate expressed as the number of new malaria cases per person-time, (2) the SPR or test positivity rate (TPR) defined as the number of laboratory-confirmed cases (microscopy or RDT) per 100 suspected cases, (3) incidence proportion of uncomplicated malaria cases as the ratio of the number of malaria-related outpatient visits to the total of all-cause outpatient visits, (4) incidence proportion of severe malaria cases measured as the ratio of the number of malaria-related hospitalisations to the total hospital admissions.

Data analysis
The confirmed malaria cases and SPR data were extracted from the SMNSs' reports of the weeks 36 to 47 of each year representing the months September, October and November of each year under study.The registered data was stratified by age (under 5 years of age and above 5 years old) and gender.The confirmed malaria cases were extracted from the hospital reports of the months September, October and November of each year under study.The registered data was available in a stratified form by age groups (< 1 year, 1-4 years, 5-14 years, 15-24 years, 25-44 years and 45 years and older) and gender.As the time interval of the SMNS and hospital reports is not quite the same, i.e. 12 weeks or 84 days versus 3 months or 91 days, the malaria incidence rate is expressed as the number of new cases per 100,000 persons per day.
Data was exported to Microsoft Excel® 2007 spreadsheets and statistical analysis was performed using MedCalc® version 18. P ≤ 0.05 was considered significant for all tests.

Ethical considerations
Ethical clearance for the study was given by the Gezira State MOH.

Malaria incidence proportion in outpatient departments (OPD) of public hospitals
A noticeable increase (P < 0.0001) in the proportion of outpatient malaria cases in public hospitals of Almanagil Locality was observed for the months September, October, November 2013 in comparison to the same period of 2011 and 2012.Hospital OPD incidence proportion in the 2013 period was 19.7% (95% CI: 19.24-20.18%)compared to 12.85% (95% CI: 12.48-13.23%)in 2011, and 12.16% (95% CI: 11.82-12.51%) in 2012 (Table 5).Persons exposed to the flooding in 2013 and consulting the OPD of the public hospitals in Almanagil Locality had respectively 1.53 and 1.62 times the risk of contracting malaria compared to persons who were not exposed to flooding in 2011 and 2012.
A more detailed age -specific analysis was possible as hospital data were stratified by six age groups.Analysis of the age-specific proportion of the outpatient malaria cases revealed that the age groups responsible for the overall rise in the proportion of malaria cases in 2013 were the two under -5-year age groups (< 1 year and 1-4 years).Considering < 5 years of age as one group, the proportion of outpatient malaria cases to the total number of outpatients in 2013 study period was as high as 38.5% (95% CI: 37.2-39.8%)compared to 24.3 % (95% CI: 23.0-25.6 %; P < 0.0001) in the 2011 study period and 25.2% (95% CI: 23.9-26.6%;P < 0.0001) in the 2012 study period (Table 6).Analysis of the < 5 years of age group showed that the proportion of the hospital OPD malaria cases in the < 1 year old group more than doubled in the year of the flooding compared to the two non-flood years (Figure 6).
Children < 1 year of age exposed to the flooding in 2013 had respectively 2.5 (95% CI: 2.2-2.9) and 2.4 (95% CI: 2.1-2.7)times the risk of catching malaria compared to children of < 1 year who were not exposed to flooding in 2011 and 2012 (P < 0.001).
Analysis of age-specific proportion of the inpatient malaria cases (Table 8) showed that the increase in the proportion of severe malaria cases was due to a higher proportion of children < 5 years of age (P = 0.05) and especially to a higher proportion of children < 1 year of age (P = 0.01).According to the Malaria Programme Review 2001-2012 report of Sudan's NMCP, the national malaria incidence rate decreased from 38/100,000/day in 2000 to 9.9/100,000/day in 2011, based on both laboratory-confirmed and clinically diagnosed malaria cases reported by all outpatient departments .Our study in Almanagil Locality revealed a malaria incidence rate in the non-flood years 2011 and 2012 of respectively 6.1/100,000/day and 6.48/100,000/day based only on confirmed malaria cases reported by the SMNSs (outpatient department of 4 health centres and 9 hospitals).This lower malaria incidence rate may be caused by the fact that only laboratory confirmed cases are included in the study.
Sudan is frequently exposed to flash floods due to torrential rainfall or by overflow of the river Nile during the rainy season .Floods may play a major role in the emergence of malaria epidemics .Surveys showed that floods in endemic areas in Sudan were often associated with malaria outbreaks over and above the annual increase of malaria cases normally expected in the rainy season .Comprehending the interrelationship between malaria and flooding is essential for predicting epidemics, adequate preventive measures and appropriate response in order to minimize the morbidity of the affected populations .Malaria epidemics attributed to flooding depend on a number of factors such as the malaria endemicity and topography of the affected area, the severity of the flooding (damage to private houses and public infrastructure), the ecological change caused by the flood (propagation of malaria infected mosquitoes), displacement and vulnerability of the affected population and level and accessibility of healthcare services before and after the flood .The main sources of information on malaria incidence are disease surveillance systems and health information systems operated by ministries of health.In acute emergencies such as floods, surveillance is generally based on data collected by healthcare workers and reported at regular interval through health centres and hospitals and should be representative of the entire disaster area .It provides early warning of an epidemic and trends at geographical and temporal levels .The study used two sources of information to collect the malaria data: a sentinel-based (SMNSs) surveillance method advocated by the NMCP and the routine health information system of the Gezira State Ministry of Health.The SMNSs of the Almanagil Locality were selected to gather information representative of the entire locality based on a number of criteria, including the coverage of hospitals and health centres and geographical areas of earlier epidemics and natural disasters .The malaria incidence in this study was measured by passive case detection based on healthcare facility-based data.This assumes completeness of reporting, accurate laboratory confirmation of all malaria cases and all patients present to healthcare facilities .In order to minimize the impact of incomplete reporting, changes in healthcare utilization and errors in laboratory confirmation of malaria cases, it is recommended to focus on confirmed malaria cases, to assess trends in slide or test positivity rate (SPR/TPR) and to determine the proportion of cases that malaria make up out of all cause outpatient visits and hospital admissions .For surveillance purposes in a well-defined cohort (SMNSs), the SPR/TPR can provide a rapid and inexpensive method of estimating temporal changes in malaria incidence after a flood .They are less affected by change in reporting rates, diagnostic practices and healthcare facility utilization rate and consider only laboratory confirmed cases of malaria .Like SPR/TPR, the proportions of outpatient and inpatient malaria cases are less sensitive to changes in reporting rates and healthcare facility utilization rate and they can reflect the burden that malaria places on the healthcare system after the flood .Change in the proportions of outpatient and inpatient malaria cases as well as SPR/TPR can reflect relative change in the malaria incidence over time, they however cannot estimate the actual incidence of malaria in a target population at national or regional level .Identifying flood-related changes in malaria incidence needs the access to reliable baseline data.An effective surveillance system is essential for providing the baseline data in order to identify flood-related changes in malaria incidence .The malaria morbidity indicators in the study were extracted from surveillance and routine health information systems of the Gezira State Ministry of Health.As the analysis has been confined to SMNSs that recorded and reported consistently over time, the size of the catchment population and coverage of service utilization were stable over the evaluation period, as only laboratory-confirmed malaria cases were included in the study and no shortage in supply of microscopy supplies or RDTs in the SMNSs were observed , our health facility-based data may provide a reliable indication of the malaria incidence rates over time .
The 2013 flooding in Almanagil Locality was associated with a malaria epidemic resulting in a marked and concurrent increase in the malaria incidence rate, SPR and proportion of outpatient visits and admissions in public healthcare facilities in comparison with the same period in the two previous years without floods.Age group analysis demonstrated that young children (< 5 years) are at the greatest risk from malaria in the post-flooding period because of an immature immune system and possibly under-or malnutrition, consistent with previous published reports .In this study, the age-stratified hospital data showed that the post-flood increase in the proportion of uncomplicated and severe malaria cases in young children is mainly due to a large increase in the under-1-year age group.
Malaria epidemics following flooding are a well-known phenomenon in endemic areas worldwide .Most studies attribute the increase of malaria incidence to rainfall and floods resulting in the formation of new breeding sites for the Anopheles mosquitoes and in providing favourable conditions for the mosquito development and survival, particularly high humidity .This is what most probably occurred in Almanagil Locality after the 2013 flood, given the presence of several factors that favour development of standing water pools and subsequent rise in the mosquito population.
Besides changes in recording completeness, healthcare utilization and diagnostic accuracy, other potential contextual factors that can confound the true malaria incidence rate include IRS, ITNs, IPTp and case Given that the NMCP started a distribution of ITNs for target risk groups in Gezira State in 2012, it is unlikely that the use of ITNs would decrease further in 2013 .Since 2010 the IPTp was no longer in the national strategy and was limited to effective case management and prevention through ITNs .Malaria case management guidelines and coverage in the public health facilities were unchanged over the observation period .In summary, since no major changes in the possible confounding factors occurred, this study provides a strong evidence that flooding can lead to a significant increase in malaria incidence based on malaria data collected in public healthcare facilities.
The estimation of malaria morbidity based on healthcare facility-related data may be too low or too high due to the potential weakness of the quality of data used to measure the malaria incidence, and hence should be analysed and interpreted with caution .Possible causes include: reporting completeness, accuracy of confirmed malaria cases by microscopy or RDTs, asymptomatic and subpatent malaria infections, the extent to which patients seek treatment in public and private healthcare services or are treated at home .In order to minimize the influence of these sources of error and bias we used the strategy recommended by the World Health Organization: focusing on confirmed malaria cases, monitoring trends in SPR/TPR (microscopic or RDTs) and monitoring malaria outpatient visits and hospital admissions.Accuracy and completeness of data could not be fully verified as retrospective data were used.The surveillance sentinel sites of the public health centres and hospitals reported consistently over time.However, a spot check of reporting completeness of Almanagil Locality health centres not included in the SMNSs showed that data were not consistently reported to the Gezira State Ministry of Health.Although the private health sector is limited in Almanagil Locality, data from the private health sector were not available during the study period.This could lead to an underestimation of malaria incidence.
Overall, these results suggest that flooding following heavy precipitation has great potential to increase malaria burden in the affected population.It appears that the increase in malaria transmission occurs in the recovery phase of the flood disaster after a lag period of approximately 4 to 8 weeks.This initial delay between the flood and the post-flood malaria outbreak may provide the opportunity for vector control measures (IRS, ITNs/ Long-Lasting Insecticidal Nets [LLINs], larvicidal programmes) together with early case detection and management to mitigate the post-flood epidemic.

Table 4 :
Number of positive blood smears and slide positivity rate by age groups with 95% confidence intervals recorded in the SMNSs of Almanagil Locality in weeks 36-47 of years 2011-2013.

Fig. 5 :
Fig. 5: Slide positivity rate by age groups with 95 % confidence intervals recorded in the SMNSs of Almanagil Locality in weeks 36 to 47 of years 2011 to 2013.* P < 0.0001.

Fig. 6 :
Fig. 6: Comparison of the percentages of malaria cases out of all hospital outpatient visits with 95 % confidence intervals between <1 year old group and 1-4 years old age group during September to November in years 2011 to 2013.* P < 0.001.
All relevant data are within the manuscript and the public repository Figshare at https://figshare.com/s/d69b1011eb74a9361ee4.For more information, please contact the corresponding author: Yasir E A Elsanousi, yasir3@yahoo.comBSE, Blood Smears Examined CI, Confidence Interval FMOH, Sudanese Federal Ministry of Health IPTp, Intermittent Presumptive Treatment of Pregnant Women IR, Incidence Rate IRS, Indoor-household Residual Spraying ITNs, Insecticide Treated Nets LLINs, Long-Lasting Insecticidal Nets MIS, Malaria Indicators Survey MOH, Ministry of Health NMCP, National Malaria Control Program OPD, Outpatient Department RDT, Rapid Diagnostic Test SPR, Slide Positivity Rate SMNS, Sentinel Malaria Notification Site TPR, Test Positivity Rate UNDP, United Nations Development Program WHO, World Health OrganizationThe authors would like to sincerely thank Dr. Elfatih M. Malik, ex Minister of Health -Gezira State a scholar and an international malaria expert -and Prof. Mohamed Elsanousi, Vice chancellor of Gezira University for their invaluable comments and revision of an earlier draft.Thanks are extended also to the staff of the departments of Health Statistics, the Gezira State Malaria Control Program and the Preventive Medicine Department of Almanagil Locality in the MOH Gezira State, Sudan for their appreciated cooperation and efforts in extracting and providing the data of this study.The paper is the result of a thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Disaster Medicine (European Master in Disaster Medicine) 1. IPCC.Summary for Policymakers.In: Field CB, Barros V, Stocker TF, Qin D, Dokken DJ, Ebi KL, Mastrandrea MD, Mach KJ, Plattner G-K, Allen SK,Tignor M, Midgley PM, editors.Managing the risks of extreme events and disasters to advance climate change adaptation, editors.A Special Report of Working Groups I and II of the falciparum is the main parasite and Anopheles arabiensis is the primary vector .A National Malaria Control Program (NMCP) has been established under the Sudan Federal Ministry of Health (FMOH).The country endorsed the international Roll Back Malaria initiative in 1998 emphasizing efforts towards more attention on early detection, prompt treatment and various prevention measures .Collaboration of the NMCP with the United Nations Development Program (UNDP), the World Health Organization (WHO) and national Non-Governmental Organizations to fight malaria in Sudan, resulted in reducing the number of malaria cases from more than four millions in 2000 to less than one million in 2010, and 75% reduction in mortality due to malaria between 2001 and 2010 .According to the WHO criteria for risk of malaria transmission and the 2012 Malaria Indicators Survey (MIS), Gezira State is considered hypo endemic , entailing that the state is more liable to malaria outbreaks.Indoorhousehold residual spraying (IRS) and distribution of insecticide treated nets (ITNs) were implemented in Gezira State.The 2012 MIS showed that approximately 95% of the population was protected by IRS, but only 34% of the households owned at least one ITN and only 5% of the households members slept under an ITN .In 2013, the worst flash floods in 25 years hit Sudan .Continual heavy rains from early August 2013 and consequent flash floods caused extensive damage and loss of life in 15 states.Reports from the Humanitarian Aid Commission estimated 499,900 people countrywide were affected by the heavy rainfall and floods across Sudan since the onset of the events in early August 2013.Assessments have shown that the floods destroyed or damaged over 85,385 houses in the states resulting in the displacement of a large part of the affected population, disrupting the healthcare system, the provision of drinking water and access to sanitation .Gezira State was one of the most affected states: 52,975 affected people, 5,946 houses destroyed and 5198 houses damaged .There were concerns about epidemics of communicable diseases including vector-borne diseases such as malaria .

41 Methods Fig. 1: The location of Almanagil Locality in Gezira state in Sudan in 2013.
Ministry of Health (MOH).Data on confirmed malaria cases was collected from the sentinel malaria notification sites (SMNSs) of Almanagil Locality.The data was extracted from the weekly reports of 13 SMNSs (9 hospitals and 4 health centres) forwarded to the Department of Malaria Control Programme of the Gezira State MOH for the years 2011, 2012 and 2013.The SMNSs are facilities equipped with laboratories and trained clinical and laboratory staff capable of performing microscopy and RDTs for malaria.Hence the malaria incidence rate (IR) and the slide positivity rate (SPR) are calculated from this data.Hospital data of outpatient malaria cases was collected by the 9 hospitals of Almanagil Locality.Data of inpatient malaria cases was collected by 8 of the 9 hospitals (the ninth hospital was not yet fully operational in 2013 with only outpatient services).The data was extracted from the monthly reports forwarded to the Statistics Department of the Gezira State MOH for the years 2011, 2012 and 2013.The data consisted of the monthly numbers of all outand inpatient diagnoses, including the confirmed malaria cases.

Table 8 : Percentage of malaria cases out of all hospital admissions by age group with 95 % confidence interval of malaria in public hospitals of Almanagil Locality in the months September to November of the years 2011 to 2013.
IRS was implemented annually since 2000 in Gezira State.According to the NMCP, IRS in the irrigation schemes in Gezira State was consistent with the national standard method of implementation.The percentage of households covered in IRS campaigns was 97.7 % and the percentage of population protected was 98.9 % in Almanagil Locality in 2012 .According to the malaria indicators survey (MIS) 2009 and 2012, the percentage of households in Gezira State with at least one ITN decreased from 48 % in 2009 to 34.3 % in 2012 and the percentage of household members who slept under ITN decreased from 13.7 % in 2009 to 5 % in 2012 .