Malaria positivity rate trend analysis at water resources development project of Wonji Sugar Estate Oromia, Ethiopia

Evidence on the trends of the proportion of malaria infections detected by routine passive case detection at health facilities is important for public health decision making especially in areas moving towards elimination. The objective was to assess nine years of trends on clinical malaria infections detected at health facility and its associated climate factors, in the water resource development set up of Wonji sugar estate, Oromia, Ethiopia. Retrospective data were collected from malaria-suspected patient recording logbook at Wonji sugar factory’s primary hospital. Monthly average meteorological data were obtained from the estate meteorological station. Data were collected from April through June 2018 and January 2022. The data were analyzed using Stata version 16.0 software for Chi-square and regression analysis. Over the last nine years, 34,388 cases were legible for analysis with complete data. Of these, 11.75% (4039/34388) were positive for clinical malaria. Plasmodium vivax test positivity was the highest proportion (8.2%, n = 2820) followed by Plasmodium falciparum (3.48%, n = 1197) and mixed infections (P. falciparum and P. vivax, 0.06%, n = 21). The odds of being positive for malaria was highest in males (AOR = 1.46; 95%CI = 1.36–1.52; P < 0.001) compared to females and in older individuals of above 15 years old (AOR = 4.55, 95%CI = 4.01–5.17, P < 0.001) followed by school-aged children (5–15 years old) (AOR = 2.16; 95%CI = 1.88–2.49, P < 0.001). There was no significant variation in the proportion of malaria-positive cases in the dry and wet seasons (P = 0.059). Malaria test positivity rates were associated with average monthly rainfall (AdjIRR = 1.00; 95%CI = 1.00–1.001, P < 0.001) while negatively associated with average monthly minim temperature (adjIRR = 0.94; 95%CI = 0.94–0.95; P < 0.001) and average monthly relative humidity (adjIRR = 0.99, 95%CI = 0.99–1.00, P = 0.023). There was year-round malaria transmission, adults especially males and school children frequently tested malaria positive. Hence, alternative vector management tools like larval source management have to be deployed besides ITNs and IRS in such water development areas to achieve the malaria elimination goal.


Introduction
Malaria is one of the most serious public health problems in the developing world. Globally, there was an increasing number of malaria-related death increased by 12% from 558,000 in 2019 to an estimated 627,000 in 2020 with the African region contributing 95% of the global burden (WHO 2017). In Ethiopia, an estimated 60% of the population lives in areas at risk of malaria transmission with Plasmodium falciparum and P. vivax, the two major parasites contributing 65-70% and 30-35% of the malaria burden, respectively (FMOH 2020;PMI 2020). Malaria infection has marked seasonality in transmission and geographic variation in intensity which is mainly bimodal following heavy from (June to August) and small rainy (February to March) seasons (FMOH 2020).
From 2005 to 2015, the malaria burden has reduced substantially in Ethiopia due to the concerted effort of roll back malaria (RBM) initiative and scale-up of control and prevention interventions such as prompt diagnosis and treatment of cases with artemisinin lumefantrine (AL), epidemic prevention and control, and use of long-lasting insecticide nets (LLINS) and indoor residual spray (IRS) (Gari and Lindtjørn 2018;Taffese et al. 2018). Considering this achievement, the country has launched a national malaria elimination roadmap to be achieved for 2030 by stratifying epidemiologic settings based on annual parasite index (API)/1000 people as high (API ≥ 50), moderate (API ≤ 10-< 50), low (API > 5-< 10), very low (AP > 0-≤ 5), and free (API = 0) (FMOH 2020).
Malaria is heterogeneous in Ethiopia, ranging from seasonal to perennial transmission largely depending on altitude, climate, and rainfall patterns requiring ecoepidemiologic-specific tailored approaches. Following the expansion of water resource development projects in Ethiopia, year-round malaria transmission is a challenge in areas around irrigation schemes and dams (Kibret et al. 2019;Yewhalaw et al. 2014;Hawaria et al. 2019). Moreover, malaria transmission is influenced by rainfall, temperature, and humidity as these climate factors affect parasite development in the vector (Shapiro et al. 2017;Villena, et al. 2020) as well as vector development, abundance, and longevity in a given area (Kulkarni et al. 2016;Farajzadeh et al. 2015;Yamana and Eltahir 2013). Therefore, climate and weather data integrated with routine malaria case detection at health facilities could be important predictors of future disease burden and transmission for appropriate public health decisions (Xiang et al. 2018;Abeku et al. 2004;Murdock et al. 2016;Kajeguka and Tarmo 2017).
As part of successful malaria control and elimination efforts, understanding trends of a proportion of malaria infections detected by routine passive case detection at health facilities is important for public health decision making especially in countries moving towards elimination. This study aimed to assess the trend of proportions of malaria infections detected at a primary hospital of Wonji sugar estate and its associated environmental factors using nine years of retrospective data, in the water resource development set up of Wonji sugar estate, Oromia, Ethiopia.

Study area
The study was conducted in Wonji sugar Estate, Oromia, Ethiopia. Data were collected in April through June 2018 and January 2022. Wonji sugar Estate is located between latitudes 39° 20′ N and longitudes 8° 30′ to 8° 35′ E and at an altitude between 1510 and 1560 m above sea level (Fig. 1). The area is characterized by a bimodal rainfall pattern with an average annual rainfall of 850 mm per year, annual precipitation of 830 mm, and mean maximum and minimum temperature of 27 °C and 15 °C, respectively (Wonji Sugar estate metrological office data). According to the estate's annual report, there is an estimated 48,000 population, of which 18,000 were permanent residents of factory workers and the rest were seasonal laborers who were residents in the estate for about eight months. The study populations are of different ethnic compositions with varied economic and educational statuses. In the study area, an irrigation canal has leakages which created many small swamps and water puddles in the Sugar cane fields suited for mosquito breeding (Supplementary Figure 1). The estate has four health institutions one primary hospital, two health posts, and one polyclinic. Malaria is a major public health problem in the area which occurs year-round with peaks after the two annual rainy seasons September to November and May to July (Wonji sugar estate Primary Hospital unpublished data). The major malaria vector in the area is Anopheles arabiensis with P. falciparum and P. vivax co-endemic.

Study design
Nine years of retrospective malaria data (from January 2013 to December 2021) was collected from the malaria-suspected patient recording logbooks of Wonji Sugar Estate's primary Hospital.

Data collection
The data were collected independently by two laboratory professionals who were trained before the study. Data was double entered into the excel sheet, checked and supervised daily by the principal investigator for completeness and consistency. Similarly, all data from the meteorological station were checked for completeness and cleaned. Suspected cases with incomplete records of important variables and examination results were excluded from the study. At the primary hospital, microscopy was used to detect Plasmodium species from Giemsa-stained peripheral blood films following WHO protocol (WHO 2020) and malaria patients were treated based on the Ethiopian national malaria guideline (PMI 2020). Monthly averages of minimum and maximum temperatures (in ℃), relative humidity (in %), and rainfall (in millimeters) data were obtained retrospectively from the estate meteorological stations. The malaria patient data were collected from the primary hospital due to its larger catchment population and presence of quality, complete, and consistent data based on microscopy diagnosis besides malaria patients were referred to the primary hospital from the aforementioned health facilities.

Statistical analysis
Data analysis was performed using Stata version 16.0 (Stata Corp LLC, College Station, Texas, USA). Data were double entered into excel and checked for consistency. Descriptive statistics were used to summarize the overall malaria prevalence across the age and gender of participants. Proportions were compared using Pearson's chi-squared test when appropriate. Logistic regression was used to test the association between malaria test positivity and characteristics of study participants and seasonality. Univariate and multivariate negative binomial regression was used to explore the association between malaria test positivity rates and meteorological variables. P-values at 0.05 were considered statistically significant in all cases.

Association of malaria case positivity rate and meteorological variables over the study years (2013-2021)
Over the past nine years (January 2013 to December 2021), the median maximum temperature of the area was 28.9 ℃ (Interquartile range (IQR) = 27.3-30.20 ℃) while the median rainfall was 28.7 mm (IQR = 0-108.3 mm). The median relative humidity was 40.8% (IQR = 34.03-50.93%). The proportion of monthly malaria-positive cases was higher in the years from 2013 to 2017 with significant epidemic-like reports in February 2016; it then declined onwards in the consecutive years of 2018-2021 (Fig. 2).
In this study, during the last nine years, clinical malaria case occurrence was observed with moderate fluctuation (Fig. 3), where only 4 months were registered with malaria case free out of 108 studied months.

Discussion
This study aimed to investigate the trend of passive malaria case positivity rate in the past nine years, in the water resource development site of Wonji sugar estate primary hospital. The overall malaria case positivity was 11.74% in the past nine years in the area which was higher than the national malaria prevalence of 0.5% in 2015 (PMI 2022).This is possibly attributed to the sugar cane plantation providing ideal breeding site for the vector(s) to sustain transmission throughout the year. The majority of the cases were P. vivax infections (8.3%) while P. falciparum was detected in 3.48% of the suspected cases. Individuals above 15 years old and school children shared the majority of the infection burden compared to preschool children (< 5 years old). Rainfall was positively associated with the malaria test positivity rate in the area. The overall proportion of malaria test positivity was very low when compared with the study conducted under a similar scenario in the Arjo sugar factory in Oromia, Ethiopia (Hawaria et al. 2019), as well as other studies in the western, northern, and southern parts of the country (Aregawi et al. 2014;Girum et al. 2019;Tesfa et al. 2018) that could be emanated due to eco-epidemiological variations, where our study site is arid compared with these study sites. This difference could also be due to the heterogeneous and  unstable community influx from around the countryside to this newer sugar factory. Plasmodium vivax and P. falciparum are the two parasite species recorded in this study; however, the proportion of P. vivax infection (69.78%) was significantly higher than P. falciparum (29.61%) unlike the national attributed infection burden of 60% and 40% for P. falciparum and P. vivax, respectively (FMOH 2020), and other studies that reported a higher proportion of P. falciparum in the country (Aregawi et al. 2014;Girum et al. 2019;Tesfa et al. 2018;Shamebo and Petros 2019;Karunamoorthi and Bekele 2012). Our report was similar to studies conducted elsewhere in Ethiopia (Hawaria et al. 2019;Shamebo and Petros 2019;Sani Dedgeba 2017) and previous studies from Adama city and Dugda district adjacent to our study site that highlighted the changing trend of P. vivax infection dominance over P. falciparum in the districts (Getachew and Teshome 2022;File et al. 2019). Similarly, studies in different malaria-endemic countries depict that there was an epidemiological diversion from P. falciparum to P. vivax infection (Ogwang et al. 2018;Wang et al. 2016). This might partly be explained due to the aggressive measure/intervention undertaken to eliminate falciparum with ACT (FMOH 2020; Gari and Lindtjørn 2018;Girum et al. 2019;Taffese et al. 2018). Importantly, P. vivax has unique hypnozoite stages that cause periodic relapses within weeks to several months that contribute to over 80% of infections in endemic settings (Robinson et al. 2015;Price, et al. 2020). This feature makes P. vivax more refractory to conventional vector control strategies in coendemic settings that are efficient for P. falciparum reduction (Price, et al. 2020).
Similar to other findings (Hawaria et al. 2019;Tesfa et al. 2018;Sena et al. 2014;Kenea et al. 2011;Gemechu et al. 2015), male individuals were more infected than females during the study periods. Moreover, adults and school children were disproportionately infected compared to preschool children which might be explained related to agricultural practices that mainly involved adults spent in sugar cane production activities till midnight and school-aged children also stay outdoors, which coincides with the active biting time of malaria vectors during dusk and dawn (Taye, et al. 2016;Kenea et al. 2016;Kibret et al. 2010). On the other hand, the lower risk of malaria positivity rate among females and preschool children might be due to the going prioritized successful indoor-based ITNs and IRS malaria vector control tools massively deployed at the national level to protect those high-risk population segments (FMOH 2020).
A peak in malaria case increase occurred in 2016 similar to studies reported from other parts of Ethiopia that revealed the highest number of malaria cases in 2016 was recorded (Hawaria et al. 2019). During the peak malaria case detection months, the average rainfall pattern was very low which could favor the occurrence of more mosquito breeding sites in the area. Malaria positivity rate was more strongly correlated with rainfall than other weather variables as has been reported elsewhere (Hawaria et al. 2019;Gemechu et al. 2015). Moreover, as a consequence of the water project in the study community, we have observed year-round malaria transmission, unlike bimodal malaria transmission of the country in general (FMOH 2020;File et al. 2019), and associated with climate variabilities similar to other studies (Kenea et al. 2011;Getachew et al. 2020). This year-round malaria transmission in the setup might also be due to permanent potential breeding habitats created for the mosquitoes as stated elsewhere (Hawaria et al. 2019;Haileselassie et al. 2021).
In conclusion, in Ethiopia, there was a considerable decline in malaria infection due to the country's aggressive intervention efforts conducted throughout the regions in the past two decades (FMOH 2020; Girum et al. 2019) with an ambitious goal of a malaria elimination strategic plan by the end of 2030. However, in the study area, the majority of the infection was detected in adults especially males and school-aged children. In such water development sites, larval source management interventions integrated with ITNs and IRS intervention are still key vector control tools (Tesfa et al. 2018;Sena et al. 2014). Moreover, to reduce the vulnerability of workers to mosquito bites and other arboviral infections additional protective measures such as mosquito repellant cream, and full goggle aprons might be important for the workers according to their job characteristics. Furthermore, quantification of community asymptomatic malaria reservoirs and screening for mass treatment might be feasible in such water irrigation sites as the site can act as a potential hot spot to maintain transmission for the neighboring seasonal districts without irrigation schemes.