Introduction

Malaria is a serious and fatal protozoan disease caused by plasmodium. People who get infected by malaria develop clinical manifestations such as fevers, shaking chills, and flu-like illnesses [1]. It is a key risk to human health, and it usually occurs in tropical and subtropical areas and has received much attention [2].Globally, 627 thousand deaths were caused by malaria [2], and the global burden of malaria has been increasing since the coronavirus disease (COVID-19) pandemic began [3]. According to the World Health Organization (WHO) 2022 report, around 249 million new cases and 608,000 malaria deaths were observed. The majority (94% and 95% of incidences and deaths, respectively) were observed in the WHO’s African region [4].

Approximately 80% of all deaths were due to malaria, and it’s mainly concentrated in African regions [5]. In Africa, 204 thousand under-five children’s deaths are caused by malaria [6]. In sub-Saharan Africa, malaria is a high-risk factor for the entire population, and 1.8 million under-five children developed malaria disease in 2016 [2, 7]. Shockingly, the five countries, including Pakistan, Ethiopia, Nigeria, Uganda, and Guinea contributed to the world’s 5 million new malaria cases observed in 2021 and 2022 [3]. Malaria deaths and infections are more likely to occur in the most vulnerable populations such as migrant people, children, and pregnant women [8].

In Ethiopia, 12% and 10% of outpatients are diagnosed with malaria from the total outpatients [9]. Six in ten (60%) of Ethiopians reside in malaria-prone areas, and 68% of the population is susceptible to malaria-infectious [10]. Malaria transmission is periodic, unpredictable, and depends on variations in rainfall and altitude [11]. High malaria transmission rates have been observed in Ethiopia in July, August, and September [10]. Moreover, malaria can occasionally spread during brief wet months in February and March [10, 12].

Although Ethiopia had made a significant decrease in malaria confirmed cases (47%) and death rate (58%), the malaria incidence in 2023 increased by a concerning 120% compared to 2022 [13]. Malaria transmission is significantly occur in high humidity, temperature, and heavy rainfall areas [14]. Despite these climatic factors, malaria transmission is also affected by sociodemographic characteristics, knowledge and awareness of the population, and utilization of malaria prevention tools [15]. Malaria infection and transmission are relatively higher in rural than urban areas due to poor housing conditions, higher vector density, and poor drainage systems [15, 16]. This skyrocketing change in malaria is also caused by conflicts, displacement, and the COVID-19 pandemic [17]. Additionally, malaria cases are rapidly increasing and becoming a fatal disease [18].

Although numerous primary studies have investigated malaria prevalence and its risk factors, its nationally condensed evidence has not been identified. Hence, malaria cases are rapidly increasing in Ethiopia since the COVID-19 pandemic began [18], and synthesizing up-to-date information about the pooled prevalence of malaria and its determinants are fundamental interventions for proper resource allocation and malaria prevention. Moreover, the COVID-19 pandemic interrupted the malaria response and busted the burden of malaria globally, the recent increase of global and local climate and weather changes, and conflict and population displacements were major problems in Ethiopia. Thus, gathering the latest pooled malaria prevalence and its determinant information is valuable for malaria prevention and control mechanisms. Studies about systematic review and meta-analysis of the prevalence of malaria and its risk factors concerning vulnerable populations such as pregnant women, children, and migrants are limited. The findings of this study could benefit stakeholders and policymakers in the malaria eradication and elimination program, and prioritization in budget allocation for malaria prevention and control.

Methods

Protocol and registration

The systematic review protocol is registered in the PROSPERO database, and its registration is accessible from https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42024517273.

Study design

A systematic literature review and meta-analysis were used for Primary studies conducted about the prevalence of malaria and its risk factors in Ethiopia. To map the identified studies, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were used [19].

Inclusion criteria

All primary studies conducted among the most vulnerable populations such as children, pregnant women, and migrants were included in this study. Children, pregnant women, and migrant people were included regardless of their age, health status, and sex. Primary studies that have been published in peer-reviewed journals since 2020, and written in the English language were included.

Exclusion criteria

Duplicated studies, studies posted at preprint, conferences, reviewed papers, books, diaries, commentaries, and letters about malaria were not included. Additionally, Primary studies that have been done among participants other than children, pregnant women, and migrants were excluded. Moreover, studies written in non-English languages, and published before the 2020 Gregorian calendar were excluded. This is essential to synthesize evidence and report updated information about malaria and its related factors for policymakers and stakeholders.

Information source and search strategy

Electronic databases such as PubMed, Google Scholar, Web of Science, Semantic Scholar, and Scopus were used for searching from November 01, 2023. Gray literature and hand searching using the references of the included studies were done. All peer-reviewed Ethiopian journals, health institutions, and Universities’ websites were considered for article searching. Searching was performed by following systematic review searching procedures. Snowballing was used to look through the references of recognized publications for studies that might be relevant. For studies that might be pertinent, snowballing was employed to search through the references of reputable publications. Possible search words (synonyms) and MeSH terms for each keyword were defined. Each keyword or MeSH term was searched using Boolean operators (OR, AND) in combinations: ((((((((((Magnitude) OR (Prevalence)) AND (Malaria)) OR (Plasmodium falciparum)) OR (Plasmodium vivax)) AND (risk factors)) OR (associated factors)) OR (related factors)) AND (Children)) OR (Under five children) AND (Expectant mothers) OR (Pregnant women) AND (Displaced people) OR (Migrants)) AND (Ethiopia). The MeSH terms of the PubMed database are presented below.

figure a

The domain of the study

Common symptoms of malaria are fever, headache, chills, extreme tiredness and fatigue, impaired consciousness, multiple convulsions, difficulty breathing, dark or bloody urine, jaundice (yellowing of the eyes and skin), and abnormal bleeding [20]. All the suspected cases of malaria need to be confirmed using parasite-based diagnostic testing such as microscopy (blood film) or a rapid diagnostic test [21]. Therefore, if the person is suspected of a malaria parasite, then the person has malaria in his/her blood. Otherwise, the person did not have malaria infection.

Review and selection process

A Preferred Reporting Item for Systematic Review and Meta-Analysis Protocol (PRISMA-P was used for article screening [19]. Duplicates were removed using Endnote X9 version software. Titles and abstracts of the studies were used as selection criteria and were independently examined for the probability of eligibility by authors (AWD, AAS, WWJ, and ADW). During the phase of article screening, studies were removed if the titles and abstracts of the studies were inadequate to meet eligibility criteria. Eligible studies had undergone full-text screening. At the phase of article eligibility, full-text screening was done independently by authors (AWD, GWK, GB, and TB). At each stage of study selection, any disagreement between authors was resolved through discussion. Finally, a study that met the inclusion criteria was included for review.

Data extraction

Three authors (AWD, WG, and ZW) were independently involved in data extraction using predefined criteria. The studies’ characteristics, such as authors’ names, publication year, study design, data source, sample size, and sampling technique were extracted independently to describe the included studies. Moreover, the malaria prevalence among vulnerable populations in each study with their respective adjusted odds ratios and 95% confidence interval (CI) were extracted to estimate the pooled prevalence and pooled determinant factors.

Data processing and analysis

The extracted data were first exported from Microsoft Excel and imported into STATA version 15 software. Estimates of each study’s pooled effect size and effect with their respective 95% CI were made using a random-effect model meta-analysis. To visualize the data, forest plots were utilized to assess the pooled impact size and weight with 95% CI of each recruited study [22, 23]. The degree of heterogeneity between the included studies was assessed using the indicator of heterogeneity (I2 statistics). A random-effect model was employed to accommodate for significant heterogeneity. To check for publication bias in the meta-analysis, Egger’s test, and funnel plot analysis were used.

Quality assessment and critical appraisal for the included studies

Standard quality assessment criteria were used to assess the quality of the included studies to explain variations in the results of included studies. A modified version of the Newcastle–Ottawa Scale (NOS) was used to assess the risk of bias, and the methodological and other aspects of each study were assessed [24,25,26]. Issues that led to possible bias at all stages of the review process were reduced and addressed by authors (AWD, WWJ, AAC, and ZW). The quality of the studies was assessed by the authors independently. The authors met periodically to discuss biased concepts, and a consensus was reached through discussion. According to the NOS quality assessment scale, a study is relevant if its quality assessment score is ≥ 7 [27].

Results

Number of articles searched in the included information database

From all electronic databases, a total of 722 studies were identified. 302 studies were eligible for full-text assessment after 420 duplicated studies were removed. A total of 176 studies were excluded after reviewing the titles and abstracts due to being unable to meet the inclusion criteria. 126 full-text articles were assessed for eligibility based on the predefined criteria, and 121 articles were excluded with reason. A total of 12e studies met the eligibility criteria and were included in a systematic review and meta-analysis (Fig. 1).

Fig. 1
figure 1

PRISMA flow chart showing screening of the identified studies

Full size image

Characteristics of the included studies

A total of 7597 study participants, from 12 articles, were estimated. All studies were done cross-sectional including community-based, and facility-based cross-sectional study designs. From the total included studies, seven studies were done in the Amhara region [28,29,30,31,32,33,34] with a total of 5579 study participants. From the total of included studies, four studies are done among children [28,29,30, 35], six studies are done among pregnant women [32, 33, 36,37,38,39], and another two studies are done among migrant people [31, 34]. Six studies were published in 2021 [29,30,31,32,33, 37]. The prevalence of malaria ranges from a minimum of 3.6% [37] to a maximum of 24.10% [36] (Table 1). According to NOS quality assessment criteria, all included studies had met the specified quality with a quality score ≥ 8 (Supplementary file 1).

Table 1 Characteristics of the included studies among vulnerable populations in Ethiopia
Full size table

The pooled prevalence of malaria in Ethiopia

As shown in Fig. 2, the pooled prevalence of malaria was 11.10% (95% CI: 6.10–16.11). A random-effects model shows the absence of heterogeneity among the included studies with 95% CI (I2 = 0.0%% and P-value = 0.781). Even if there is no significant heterogeneity between the included studies, subgroup analysis was carried out to know the prevalence of malaria among study participants. Thus, the pooled prevalence of malaria among children, migrant people, and pregnant women was 15.36% (95% CI: 5.32, 25.39), 16.78% (95% CI: 1.83, 31.72), and 8.45% (95% CI: 2.19, 14.72) respectively (Fig. 3).

Fig. 2
figure 2

Random effect analysis for the pooled prevalence of malaria in Ethiopia

Full size image
Fig. 3
figure 3

Random effect analysis for subgroup analysis of malaria among vulnerable populations

Full size image

Publication bias

A funnel plot and Egger's test were performed for publication bias assessment with a 5% significance level. The funnel plot indicates the absence of statistically significant bias among the included studies (Fig. 4). Egger’s and Begg’s test reports show that the standardized effect is precisely concentrated on the slope and horizontal lines, which indicates the absence of statistically significant publication bias among the included studies (Figs. 5 and 6) with a p-value of 0.347 (Table 2).

Fig. 4
figure 4

Funnel plot for spotting publication bias among the included studies

Full size image
Fig. 5
figure 5

Egger’s funnel plot for spotting publication bias among the included studies

Full size image
Fig. 6
figure 6

Begg’s funnel plot for spotting publication bias among the included studies

Full size image
Table 2 Publication bias assessment using the Egger test
Full size table

Factors associated with malaria prevalence among vulnerable population in Ethiopia

This study examined the pooled significant factors associated with malaria prevalence among vulnerable populations in Ethiopia. From the 12 included studies; 11 studies, that have at least two common significant variables, were used to assess the pooled risk factors of malaria infection among vulnerable populations. Therefore, stagnant water (AOR: 3.84, 95% CI: 2.45, 6.02), staying outdoors at night (AOR: 3.39, 95% CI: 1.33, 8.67), vulnerable populations who did not use insecticide-treated net (AOR: 4.57, 95% CI: 2.32, 9.02), and low age group (AOR: 0.57, 95% CI 0.36, 0.89) were significantly associated with malaria infection and its prevalence in Ethiopia (Fig. 7).

Fig. 7
figure 7

Graphical illustration of pooled significant factors associated with malaria, where ITN stands for insecticide-treated net

Full size image

Discussion

In this study, the most vulnerable groups such as children, pregnant women, and migrants’ people were included. From the twelve studies included, the pooled prevalence of malaria among the most vulnerable populations in Ethiopia was 11.10%. This evidence is consistent with studies done in Kenya [40], and Mozambique [41]. However, the current evidence was lower than studies done in Malawi [15], Ethiopia [42], and Ghana [43]. In general, the pooled prevalence of malaria and its risk factors among vulnerable populations in Ethiopia has increased. This might be due to a lack of health education on malaria transmission methods, malaria prevention and control measures, lack of knowledge and behavioral interventions for malaria [42]. Additionally, societal norms, cultural beliefs, and behaviors may influence the use of malaria prevention strategies and promote malaria infection, where malaria is not seen as a major risk factor for health [44, 45]. Moreover, poor personal cleanliness, conflict, internal community displacement, low food access, and starvation might be contributors to malaria infection and its high incidence [46]. The Ethiopian malaria eradication and elimination program is planned to achieve zero malaria in districts by 2025 [47], the current pooled prevalence of malaria among the most at-risk populations is higher than the targeted plan. Therefore, performing the seed action plans that lead to zero malaria is an essential intervention.

This study examined the pooled significant factors associated with malaria infection and its prevalence among vulnerable populations in Ethiopia. Therefore, vulnerable populations who live near stagnant water were 3.8 times more likely infected with malaria a key risk factor for malaria prevalence. This might be due to accidental irrigation arrangements, and water collection reservoirs may facilitate the spread of disease [48, 49]. Additionally, the presence of stagnant water and the short distance of settlement from a water body such as a river is critical for high malaria vector density and transmission [48]. Draining stagnant water, and arranging the settlements far away from the water bodies might reduce the infection of malaria.

Vulnerable populations who did not utilize insecticide-treated nets were 4.6 times more likely to increase malaria infection and its prevalence. This might be because of misconceptions about insecticide-treated bed nets, improper utilization, and inadequate frequency of home visits by health extension workers [50]. In addition, an insufficient supply of bed nets and, the wrong perception that “malaria is not present in the area” might be the possible reasons [51]. Moreover, population growth and low economic status of the population, discomfort, overheating, and perceived low mosquito density might limit the access and frequency of using insecticide-treated bed nets [52]. However, vulnerable populations who used insecticide-treated bed nets were 1.6 times more likely not infected with malaria. So, swift scaling up of insecticide-treated bed net coverage can be realized through mass net delivery campaigns, and governments and health administrative bodies need to be keen to increase the use of nets, especially for children. Plus, local production is encouraged, and manufactured nets are better sold on the open market for better access and utilization of insecticide-treated bed nets.

Vulnerable populations who stayed outdoors at night were 5.2 times more likely infected with malaria and associated with the prevalence of malaria. This might be due to vulnerable populations who stay outdoors at night might be more likely exposed to cold and cloudy weather that is related to mosquito breeding spots, and a situation in which malaria transmission could be accelerated [53, 54]. Moreover, vulnerable populations such as pregnant women and migrants might have lots of outdoor activities, and hence more likely to be bitten by female Anopheles mosquitoes [55]. Hence, eradicating mosquito breeding sites, and minimizing the night outdoor activities might minimize the probability of being infected by malaria.

Younger age-vulnerable populations were 57% more likely infected with malaria in Ethiopia. This might be because placenta malaria most frequently occurs in young pregnant women [56]. Additionally, high parasite densities and low capability of immunity resistance might be associated with high malaria infection in young populations including children [57, 58]. Moreover, migrant people, pregnant women, and children's nutritional status might not be at its adequate and appropriate level, which further makes them infected with infectious diseases [59].

Conclusions and recommendations

The pooled prevalence of malaria is high among vulnerable populations in Ethiopia. The presence of stagnant water, no use of the insecticide-treated bed net, and being young are risk factors for malaria infection. Therefore, arranging settlements far away from rivers and draining stagnant water, distributing adequate bed nets, empowering young vulnerable populations with education, and encouraging local net manufacturers could be possible interventional measures to reduce the prevalence of malaria.

Strengths and limitations of the study

This study provides synthesized and compiled evidence about malaria prevalence. This research can help stakeholders make decisions about malaria-related problems. Plus, this systematic review was limited by the studies' publication dates. Additionally, as long as the included articles are studies from three regions (Amhara, Oromia, and Benishangul Gumuz) based on our search abilities, and publication date, the findings may not represent the prevalence of malaria in Ethiopia.