Household malaria risk perception in Ethiopia: multilevel and spatial analysis using the 2022-2023 National Health Equity Survey.

Household malaria risk perception in Ethiopia: multilevel and spatial analysis using the 2022–2023 National Health Equity Survey - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice BMC Public Health . 2026 Apr 17;26:1731. doi: 10.1186/s12889-026-27402-4 Household malaria risk perception in Ethiopia: multilevel and spatial analysis using the 2022–2023 National Health Equity Survey Yitayh Leul 1, ✉ , Desalegn Ararso 1 , Aderajew Mekonnen Girmay 1 , Arega Zeru 1 , Girum Taye 1 , Weldemariam Bahre 1 , Kelem Berhanu 1 , Akberet Lemlem 1 , Tsegaye Getachew 1 , Hiwot Achamyeleh 1 , Yirgalem Shibiru 2 , Afewerk Alemu 1 , Ashenif Tadele 1 , Abebaw Wasie 1, 6 , Gebeyaw Molla 1 , Fikreselassie Getachew 1 , Wogayehu Tadele 1 , Senait Alemayehu Beshah 1 , Hanim Tesfaye 1 , Desalew Zelalem 1 , Seboka Abebe Sori 1 , Sinafikish Ayele 1 , Tesfaye Dagne 1 , Teshome Kabeta 3 , Yihalem Abebe Belay 4 , Bililign Terefe 1 , Lelisa Fekadu 5 , Gemu Tiru 5 , Getachew Tollera 1 Author information Article notes Copyright and License information 1 Ethiopian Public Health Institute, Addis Ababa, Ethiopia 2 Success HealthEcon Consultancy plc, Addis Ababa, Ethiopia 3 Jimma University, Jimma, Ethiopia 4 Department of Public Health, College of Medicine and Health Sciences, Debre Markos University, Debre Markos, Ethiopia 5 Ministry of Health, Addis Ababa, Ethiopia 6 Wolkite University, Wolkite, Ethiopia ✉ Corresponding author. Received 2025 Oct 21; Accepted 2026 Apr 10; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13220557  PMID: 41998604 Abstract Background Malaria remains a major public health challenge in Ethiopia, affecting more than two‑thirds of the population living in endemic regions. Although local studies have explored community perceptions, nationally representative evidence on household‑level malaria risk perception is limited. Objectives To examine socioeconomic and geographic correlates of household malaria risk perception in Ethiopia. Methods We analyzed data from 9,157 households in the 2022–2023 Ethiopian National Health Equity Survey (NHES) using multilevel mixed‑effects logistic regression to account for clustering at region, zone, and woreda‑levels. Spatial analysis was conducted using Global Moran's I, and Getis Ord Gi* was used to identify geographic patterns. Results Overall, 76.4% of households reported feeling at risk of malaria (95% CI: 74.9–77.9). Rural households had 54% higher odds of perceiving risk compared to urban households (AOR = 1.54, 95% CI: 1.22–1.95). Households with higher education were nearly twice as likely to perceive risk (AOR = 1.89, 95% CI: 1.36–2.63), while older household heads were about half as likely (AOR = 0.49, 95% CI: 0.35–0.67). Households without a mobile phone (AOR = 0.59, 95% CI: 0.51–0.69) or bank/microfinance account (AOR = 0.59, 95% CI: 0.50–0.70) were less likely to perceive risk. Approximately 38% of the variability in malaria risk perception occurred between woredas, indicating strong local geographic clustering. Approximately 38% of the variability in perception occurred between woredas, with a Median Odds Ratio of 2.62, indicating strong local clustering. Spatial autocorrelation was significant (Global Moran’s I = 0.42, p < 0.001), with 11% of areas identified as High–High clusters. Conclusions Household malaria risk perception in Ethiopia is shaped by socioeconomic disparities and geographic clustering. Integrating perception indicators into malaria surveillance and control strategies will strengthen precision public health and support equitable progress toward malaria elimination. Supplementary Information The online version contains supplementary material available at 10.1186/s12889-026-27402-4. Keywords: Malaria risk, Preventive practices, Multilevel analysis, Health equity, Spatial autocorrelation, Ethiopia Introduction Malaria remains a major public health challenge in low‑ and middle‑income countries, with Sub‑Saharan Africa carrying the greatest burden [ 1 , 2 ]. In 2023, an estimated 263 million malaria cases and 597,000 deaths occurred worldwide, over 90% of which were in the WHO African Region [ 3 , 4 ]. Children under five and pregnant women remain especially vulnerable [ 5 ]. Although mortality declined between 2000 and 2015 due to widespread use of insecticide‑treated nets, indoor residual spraying, rapid diagnostics, and artemisinin‑based therapies [ 6 , 7 ], progress has stalled in recent years because of drug and insecticide resistance, climate change, and health system disruptions [ 8 , 9 ]. The Health Belief Model (HBM) provides a framework for understanding malaria risk perception. Perceived susceptibility and severity drive awareness, while perceived benefits, barriers, cues to action (such as health campaigns or peer norms), and self‑efficacy situate risk perception within broader health behaviors [ 10 ]. Household perceptions are critical because they reflect awareness rather than biological incidence [ 11 , 12 ]. At the individual level, perceptions are shaped by sociodemographic characteristics, household resources, and prior experiences [ 13 , 14 ]. At the community level, shared malaria ecology, health messaging, and information environments influence collective awareness [ 15 , 16 ]. Ethiopia remains a high‑burden country, with about 70% of the population living in areas conducive to transmission [ 17 – 20 ]. The spread of Anopheles stephensi and persistent socioeconomic barriers sustain transmission despite national interventions [ 21 – 23 ]. Because perceptions cluster within communities due to shared environments and collective information exposure, multilevel and spatial analysis is essential to capture both household‑level associations and community‑level effects. Identifying geographic hotspots and coldspots of perception provides critical evidence for precision public health [ 24 – 27 ]. Despite extensive research on malaria epidemiology in Ethiopia, nationally representative evidence on household‑level risk perception and its geographic clustering remains limited. This study addresses this gap by applying multilevel modeling and spatial autocorrelation to examine both household‑level correlates and community‑level clustering of malaria risk perception. Specifically, the study aims to (1) quantify household‑level correlates of perceived malaria risk and (2) assess geographic clustering across administrative levels. Methods and materials Study area description Ethiopia lies in northeastern Africa between 3.3°–15.0°N and 33.0°–48.0°E, covering about 1.1 million km². Its diverse topography, ranging from 125 m below sea level (Kobar Sink) to 4,620 m at Mount Ras Dashen, creates ecological variation that strongly influences malaria transmission [ 24 – 29 ]. Approximately 70% of the population lives in areas conducive to transmission, with 75% of the landmass classified as endemic [ 17 ]. Transmission intensity varies by altitude and rainfall, with peaks after the major rains (September–December and March–June) [ 30 ]. Vector ecology is heterogeneous. Anopheles arabiensis remains dominant, but the spread of Anopheles stephensi in urban and peri‑urban areas has altered transmission dynamics [ 31 , 32 ]. Malaria is perennial in lowland and riverine zones, while highland regions generally experience lower transmission, though localized outbreaks occur under favorable microclimates [ 33 ]. Study design and sampling We conducted a cross‑sectional study using nationally representative household data from the 2022–2023 Ethiopian National Health Equity Survey (NHES), following STROBE guidelines for transparency and rigor. The NHES used a two‑stage stratified cluster design based on the 2019 Population and Housing Census frame. In stage one, enumeration areas (EAs) were selected using probability proportional to size. In stage two, households within each EA were systematically sampled. Data collection in Tigray was delayed until security was restored, requiring adjustments to survey timing. The final sample included 9,157 households across all regions and city administrations. Sampling weights, clustering, and stratification ensured representativeness at national, regional, and urban–rural levels. Post stratification weights were scaled to account for the hierarchical survey design, ensuring estimates reflect the Ethiopian population distribution. Full details of survey procedures are available in the NHES main report [ 34 ]. Data sources The National Health Equity Survey (NHES), managed by the Ethiopian Public Health Institute (EPHI), provided geo‑referenced household‑level data on sociodemographic characteristics, health service access, and equity indicators. Data were accessed through the official recode file. Formal permission to use the data was granted by EPHI, and ethical clearance was obtained before analysis. Missing data were minimal (< 0.5%). We applied complete case analysis, excluding households with missing values on key covariates. Given the low level of missingness, imputation was not required. All analyses incorporated sampling weights, clustering, and stratification to ensure representativeness at national, regional, and urban–rural levels. Variables The primary outcome variable was perceived malaria risk, measured using the NHES household question: “Do you feel your household is at risk of malaria?” Responses were originally ordinal (Yes/No/Don’t know). For analysis, responses were dichotomized into Yes = high perceived risk and No = low perceived risk, with “Don’t know” responses excluded ( n < 0.5%). Unlike “No,” which indicates low perceived risk, “Don’t know” reflects uncertainty or lack of awareness or lack of awareness. This exclusion was necessary to maintain interpretability of the binary outcome and aligns with national malaria strategy indicators. Independent variables were grouped into socio‑demographic, structural/behavioral, and contextual domains, guided by theoretical relevance and prior literature. Community‑level variables were constructed by aggregating household responses within EAs and then grand mean‑centered to express values relative to the national average. This approach improves interpretability and reduces collinearity between household‑ and community‑level predictors. For statistical modeling, variables were recoded into categorical or dummy formats as appropriate. Table 1 summarizes all variables, their categories, operational definitions, and supporting references. Table 1. Variables, categories, definitions, and supporting literature Variable Name Category Definition / Operationalization Supporting Literature Household head age Socio-demographic Categorized as 18–24, 25–44, 45–59, and ≥ 60. [ 11 , 35 ] Household head sex Socio-demographic Male vs. Female [ 14 , 36 ] Education level Socio-demographic No education, primary (Grade 1–8), secondary (Grade 9–12), technical/vocational, Higher education. [ 36 ] Occupation Socio-demographic Classified into farmer, employed, merchant, unemployed [ 12 ] Household size Socio-demographic Categorized into 1–2, 3–4, and ≥ 5 members [ 14 , 15 ] Residence Socio-demographic Urban vs. Rural [ 17 , 18 ] Wealth index Socio-demographic Constructed using PCA of household assets and housing characteristics [ 19 , 20 ] Communication assets Structural/Behavioral Ownership of a radio, television, landline phone, and mobile phone [ 21 , 22 ] Financial inclusion Structural/Behavioral Household access to a bank/microfinance account; Received cash/food support [ 23 ] PSNP participation Structural/Behavioral Household participation in the Productive Safety Net Programme [ 35 ] Altitude Contextual/Environmental Categorized into < 1,000 m, 1,000–1,500 m, 1,500–2,300 m, and > 2,300 m [ 24 , 25 ] Administrative region Contextual/Environmental Ethiopia’s 11 regions and 2 city administrations (Tigray, Afar, Amhara, Oromia, Somali, Benishangul‑Gumuz, SNNP, Sidama, Southwest Ethiopia, Gambela, Harari, Addis Ababa, and Dire Dawa) [ 26 , 34 ] Perceived malaria risk Outcome variable Household response to NHES question: “Do you feel your household is at risk of malaria?” (Yes/No) [ 34 ] Open in a new tab Statistical analysis Household‑level predictors of perceived malaria risk were examined using mixed‑effects logistic regression. Variables were included in the multivariable model if they were statistically significant at p < 0.05 in bivariate analyses or had strong theoretical justification. A four‑level mixed‑effects logistic regression framework was specified to account for hierarchical clustering, with households (level 1) nested within woredas (level 2), zones (level 3), and regions (level 4): The model was specified as: Where: probability that household in woreda , zone , region perceives malaria risk. fixed intercept. -th household or community-level covariate. fixed-effect coefficients. random intercepts for woreda, zone, and region‑level. Random effects assumptions: Where: random intercept at the woreda‑level. random intercept at the zone‑level. random intercept at the region‑level. normal distribution with mean 0 and variance . variance components for woreda, zone, and region‑levels. The intraclass correlation coefficient (ICC) quantified the proportion of total variance attributable to clustering at higher administrative levels. To improve interpretability, Median Odds Ratios (MORs) were also computed: where: variance of the random effect for a given level. 0.6745 = the 75th percentile of the standard normal distribution. By definition, MOR ≥ 1, with larger values indicating stronger between‑cluster heterogeneity. Survey sampling weights were incorporated into the multilevel logistic regression using a weight scaling method. Household weights were normalized within clusters (woredas, zones, regions), ensuring unbiased fixed‑effect estimates and correct variance partitioning. Weighted models were retained for inference, while sensitivity analyses using unweighted models yielded consistent estimates. Model diagnostics included assessment of multicollinearity (variance inflation factors < 2), evaluation of influential clusters via residual diagnostics (none detected), and comparison of model fit using Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and likelihood ratio tests. Spatial clustering of perceived malaria risk was evaluated using Global Moran’s I and Local Moran’s I (LISA) at the woreda level, with each woreda treated as a spatial unit. An eight‑nearest‑neighbor (k = 8) Euclidean distance spatial weights matrix with row standardization was applied. Global Moran’s I was computed with 9,999 Monte Carlo permutations, while LISA identified statistically significant local clusters (High–High, Low–Low, High–Low, Low–High). Local p‑values were adjusted using the Benjamini–Hochberg false discovery rate procedure. GPS displacement rules applied in the NHES dataset were explicitly considered. Sensitivity analyses confirmed robustness despite random displacement of household coordinates. All statistical analyses were conducted in Stata 17.0 , with spatial clustering and maps generated in ArcGIS 10.8 . Result Household socio-demographic profile A total of 9,157 households were analyzed across Ethiopia. The largest shares were from Oromia (40.5%), Amhara (27.4%), and SNNPR (12.6%). This reflects the population distribution across major regions. Nearly three‑quarters of households were rural (73.1%), and most were male‑headed (88.7%). Educational attainment was generally low: 42.8% of household heads had no formal education, while only 5.7% had higher education. Household size was typically large, with more than half reporting five or more members. Farming was the dominant occupation (51.6%), followed by unemployment (31.8%). Access to communication tools and financial services varied. Mobile phones were the most widely owned (69.3%), while radios (25.1%) and televisions (24.3%) were less common. Financial inclusion was moderate, with 51.8% of households reporting access to a bank or microfinance account. Weighted percentages account for survey design and population distribution, explaining discrepancies between raw frequencies and weighted percentages across regions. Detailed distributions are presented in Table 2 . Table 2. Socio-demographic characteristics of surveyed households Variable Category Frequency ( n ) Weighted (%) Region Tigray 1,080 0.97* Afar 633 0.98 Amhara 914 27.35 Oromia 955 40.51 Somali 424 3.19 Benishangul-Gumuz 708 0.91 SNNP 811 12.55 Sidama 724 3.38 Southwest Ethiopia 701 4.22 Gambela 538 0.39 Harari 399 0.19 Addis Ababa 736 4.95 Dire Dawa 534 0.41 Residence Rural 5,557 73.11 Urban 3,600 26.89 Sex Male 7,785 88.7 Female 1,369 11.3 Age category of household heads 18–24 404 4.35 25–44 3,456 37.44 45–59 3,495 38.59 ≥ 60 1,800 19.62 Marital Status Married 8,580 94.98 Not Married 576 5.02 Education No Education 3,365 42.8 Primary (Grade 1–8) 3,123 35.4 Secondary (Grade 9–12) 1,474 13.1 Technical/Vocational 403 3.0 Higher education 791 5.7 Occupation Farmer 3,727 51.61 Employed 1,652 11.25 Merchant 672 5.38 Unemployed 3,105 31.77 Family Size 1–2 106 1.42 3–4 3,972 42.49 ≥ 5 5,076 56.1 Variable Category Response Weighted (%) Ownership of assets Radio Yes 25.11 No 74.89 Television Yes 24.28 No 75.72 Landline phone Yes 3.32 No 96.68 Mobile phone Yes 69.29 No 30.71 Bank /Microfinance Account Yes 51.82 No 48.18 Financial Access Received cash/food support Yes 6.11 No 93.89 Open in a new tab *Weighted percentages reflect sampling weights, clustering, and stratification. Differences from raw frequencies arise from survey design. Data for Tigray were collected separately and integrated later; therefore, estimates for this region should be interpreted with caution Bivariate associations between household characteristics and malaria risk perception Malaria risk perception varied significantly across socio‑demographic and environmental groups. Male‑headed households were more likely to perceive risk than female‑headed households. Rural households were more likely to perceive risk than urban households (53.9% vs. 22.6%). Larger households (≥ 5 members) reported higher perceived risk (41.9%) compared to smaller households (≤ 2 members, 0.6%). Educational attainment showed a strong gradient: households with no education reported the highest proportion of low perceived risk (13.9%), whereas households with higher education reported much lower levels (3.8%), reflecting stronger awareness among educated households. Perceived risk was higher among farmers and unemployed households, while employed and merchant households reported lower levels. Ownership of communication assets and financial inclusion were consistently linked to higher perceived risk. The environmental context also influenced perception. Households in mid‑highland zones (1,500–2,300 m) reported the highest perceived risk (39.6%), consistent with ecological conditions favorable to malaria transmission. Lowland households (< 1,000 m) reported the lowest perceived risk (2.5%), despite being biologically more vulnerable, possibly reflecting differences in awareness or health messaging. Highland households (> 2,300 m) reported intermediate levels of perceived risk (26.5%). These bivariate associations are observational and demonstrate socio‑economic and environmental patterns of awareness that inform targeted communication and community‑level interventions. Associations were assessed using survey‑adjusted Wald tests, reported as design‑based F statistics, which account for sampling weights, clustering, and stratification (see Table 3 ). Table 3. Bivariate associations between household characteristics and malaria risk perception (Survey‑adjusted Wald F tests) Variable Category High perceived risk (%) Low perceived risk (%) F statistic p -value Sex of household head Male 68.9 19.8 20.02 0.001 Female 7.5 3.8 Residence Urban 22.6 4.3 46.09 0.001 Rural 53.9 19.3 Family size 1–2 0.6 0.9 20.65 0.001 3–4 34.0 8.5 ≥ 5 41.9 14.2 Education No Education 28.9 13.9 40.54 0.001 Primary (Grade 1–8) 28.2 7.2 Secondary (Grade 9–12) 11.4 1.8 Technical/Vocational 2.8 2.7 Higher Education 5.3 3.8 Age category 18–24 3.2 1.2 2.37 0.001 25–44 28.4 9.0 45–59 30.4 8.2 ≥ 60 14.5 5.2 Marital status Married 73.4 21.5 24.36 0.001 Not Married 3.0 2.0 Occupation Farmer 36.5 15.1 Employed 10.1 1.2 32.92 0.001 Merchant 4.6 0.7 Unemployed 25.2 6.6 Radio ownership Yes 21.2 3.9 36.75 0.001 No 55.3 19.6 TV ownership Yes 21.7 2.6 114.22 0.001 No 54.7 21.0 Mobile ownership Yes 58.2 11.1 217.62 0.001 No 18.2 12.5 Bank/microfinance Yes 44.5 7.3 163.63 0.001 No 31.9 16.3 PSNP support Yes 4.34 1.77 3.74 0.053 No 72.10 21.79 Altitude < 1000 m (Lowlands) 2.46 0.66 2.75 0.043 1000–1500 m (high lowlands) 7.97 2.44 1500–2300 m (mid-highlands) 39.56 13.50 > 2300 m (highlands) 26.45 6.96 Perceived malaria status No 1966 23.56 22.06–25.13 Yes 7191 76.44 74.87–77.94 Open in a new tab Reported statistics are design‑based F values from survey‑adjusted Wald tests, accounting for sampling weights, clustering, and stratification p < 0.05 indicates statistical significance Altitude and spatial patterns of malaria risk perception Households were stratified into four altitude categories to reflect Ethiopia’s diverse topography. The majority were located in highland areas (> 2,300 m; 53.1%) and mid‑highland areas (1,500–2,300 m; 33.4%), with smaller proportions in lowland zones (< 1,000 m; 3.1%) and lower‑mid altitudes (1,000–1,500 m; 10.4%) (Table 4 ; Fig. 1 ). Table 4. Malaria risk perception by altitude category Elevation (m a.s.l.) Frequency ( n ) Weighted % Low perceived risk (%) High perceived risk (%) < 1000 m 2,134 10.4 18.6 81.4 1000–1500 m 756 3.1 16.8 83.2 1500–2300 m 1,916 33.4 19.8 80.2 > 2300 m 4,351 53.1 24.4 75.6 Total 9,157 100.0 23.6 76.4 Open in a new tab m meter Fig. 1. Open in a new tab Elevation map of Ethiopia with surveyed household locations Perceived malaria risk varied significantly across altitude zones (Pearson χ²(3) = 23.15; design‑based F (2.92, 26,765.34) = 2.75; p = 0.043). Households in the 1,000–1,500 m range reported the highest proportion of high perceived risk (83.2%), followed by those in lowland areas (< 1,000 m; 81.4%). Highland households (> 2,300 m) reported 75.6% high perceived risk, while mid‑highland households (1,500–2,300 m) reported comparatively lower levels (80.2%). These results reflect perceptions of malaria risk rather than actual transmission intensity. Elevated perception in highland zones may reflect increased awareness, concern about ecological changes, or recognition of emerging vectors such as Anopheles stephensi. These findings show the need for geographically tailored communication strategies that strengthen awareness while avoiding conflation of perceived vulnerability with biological incidence. Global Moran’s I = 0.42 ( p < 0.001) indicated significant positive spatial autocorrelation, meaning households with similar malaria risk perceptions were spatially clustered. Moran’s I values range from − 1 to + 1, where positive values indicate clustering, negative values indicate dispersion, and values near zero indicate randomness. Local Moran’s I identified 997 High–High clusters (10.9%), 97 Low–Low clusters (1.1%), 211 High–Low clusters (2.3%), and 791 Low–High clusters (8.7%), while 7,038 areas (77.0%) were not significant after FDR correction (Table 5 ; Figs. 2 and 3 ). High–High clusters were concentrated in western and northwestern Ethiopia (Oromia, Benishangul‑Gumuz, Amhara), consistent with known malaria‑prone ecological zones. Low–Low clusters were found in high‑altitude regions of Amhara, Tigray, and central Oromia, where cooler climates reduce vector survival. Table 5. Local Moran’s I cluster classification for perceived malaria risk Cluster Type Count ( n ) Percent (%) Avg LMi Index Avg LMi Z Score Avg LMi p Value High–High (HH) 997 10.9 2.10 4.15 0.001 Low–Low (LL) 97 1.1 0.35 2.05 0.025 High–Low (HL) 211 2.3 0.95 1.75 0.001 Low–High (LH) 791 8.7 0.55 2.60 0.003 Not Significant 7,038 77.0 0.08 0.30 0.335 Open in a new tab HH and LL clusters are associated with positive indices and high Z scores, confirming strong clustering. HL and LH clusters, despite positive average indices, represent spatial outliers because they indicate contrasting local patterns (high surrounded by low, or low surrounded by high). Not Significant areas have near‑zero indices and high p‑values, confirming randomness H High–High cluster, LL Low–Low cluster, HL High–Low cluster, LH Low–High cluster Fig. 2. Open in a new tab Getis‑Ord Gi hot‑spot analysis of household malaria risk perception in Ethiopia. Colors indicate statistically significant hot and cold spots at 90%, 95%, and 99% confidence levels Fig. 3. Open in a new tab Local Moran’s I cluster analysis of household malaria risk perception in Ethiopia. Colors indicate cluster types (High–High, Low–Low, High–Low, Low–High, and Not Significant) after FDR correction Outliers (HL and LH clusters) occurred in transitional interfaces and areas with strong local health programs, reflecting localized disparities in perception. Although HL and LH clusters indicated positive average LMi values, they function as spatial outliers because they represent households whose perceptions diverge from those of surrounding communities. Geographic clustering: null multilevel model The null multilevel logistic regression model revealed substantial geographic variation in malaria risk perception. The largest share of variance was observed at the woreda‑level, with an intraclass correlation coefficient (ICC) of 0.38 (95% CI: 0.32–0.45), indicating that 38% of variation was explained by differences between woredas. Zones accounted for 19% of the variance (ICC = 0.19; 95% CI: 0.12–0.28), while regions contributed a smaller proportion (ICC = 0.04; 95% CI: 0.01–0.17). The fixed intercept represents the baseline log‑odds of perceiving malaria risk when all predictors are at their reference categories. Model fit statistics confirmed the appropriateness of the multilevel specification, with the likelihood ratio test strongly supporting the inclusion of random effects (χ²(3) = 1,552.62, p < 0.001). Median Odds Ratios (MORs) provided an intuitive measure of community‑level heterogeneity. The MOR for woreda‑level effects was 2.62, meaning that moving from a low‑perception to a high‑perception woreda would typically increase the odds of perceiving malaria risk by about 162%. The MOR for zones was 2.28, while the MOR for regions was 1.55, confirming that woreda‑level differences contributed most strongly to household perceptions. Table 6 presents variance estimates, ICC values, and MORs. Table 6. Null model, ICC, and MOR estimates for malaria risk perception. Model Component Parameter / Level Estimate Std. Error 95% CI Fixed Effect Intercept (Constant) AOR = 5.78 1.11 3.97–8.42 Random Effects (Variance σ²) Region 0.21 0.17 0.04–1.01 Zone (within region) 0.80 0.24 0.45–1.43 Woreda (within zone) 1.02 0.15 0.77–1.35 Intraclass Correlation (ICC) Region 0.04 0.03 0.01–0.17 Zone (within region) 0.19 0.04 0.12–0.28 Woreda (within zone) 0.38 0.03 0.32–0.45 Model Fit Likelihood Ratio Test χ²(3) = 1552.62 p < 0.001 Open in a new tab Abbreviations: AOR Adjusted Odds Ratio CI Confidence Interval, ICC Intraclass Correlation Coefficient, MOR Median Odds Ratio Household-level predictors of malaria risk perception The multilevel mixed‑effects logistic regression model confirmed several household‑level predictors. Rural households were about 50% more likely to perceive malaria risk compared to urban households (AOR = 1.54, 95% CI: 1.22–1.95). Education was positively associated with perception: Households with higher education were nearly twice as likely to perceive risk (AOR = 1.89, 95% CI: 1.36–2.63). Household heads over 50 years were less than half as likely to perceive risk compared to younger adults (AOR = 0.49, 95% CI: 0.35–0.67). Middle‑income households reported lower odds compared to the poorest group (AOR = 0.73, 95% CI: 0.60–0.89). Access to communication and financial services was strongly correlated with perception. Households without a mobile phone (AOR = 0.59, 95% CI: 0.51–0.69) or bank account (AOR = 0.59, 95% CI: 0.50–0.70) were significantly less likely to perceive malaria risk. Random effects remained significant at all administrative levels. Woreda‑level variance (0.96) accounted for the largest share of unexplained variation, with an ICC of 0.38. The corresponding MOR was 2.55, indicating that households in high‑perception woredas had more than double the odds of perceiving malaria risk compared to those in low‑perception woredas, even after adjusting for household‑level predictors. Zone‑level MOR was 2.21, and region‑level MOR was 1.55, again confirming that woreda‑level heterogeneity was most influential. Detailed, refined multilevel results are presented in Table 7 . Table 7. Refined multilevel model: Significant predictors of malaria risk perception Predictor OR Std. Err z p -value 95% CI Rural residence 1.54 0.18 3.60 0.001 1.22–1.95 Secondary education 1.32 0.15 2.48 0.013 1.06–1.64 Technical/Vocational 1.55 0.30 2.27 0.023 1.06–2.27 Higher education 1.89 0.32 3.77 0.001 1.36–2.63 Adult age (45–59 yrs) 0.68 0.11 -2.47 0.013 0.50–0.92 Old age (≥ 60 yrs) 0.49 0.08 -4.46 0.001 0.35–0.67 Middle wealth 0.73 0.07 -3.12 0.002 0.60–0.89 No radio 0.77 0.07 -2.96 0.003 0.65–0.92 No mobile 0.59 0.05 -6.82 0.001 0.51–0.69 No bank account 0.59 0.05 -6.36 0.001 0.50–0.70 Random effects Region variance 0.21 0.16 0.05–0.94 Zone variance 0.69 0.21 0.38–1.26 Woreda variance 0.96 0.14 0.73–1.28 Model fit χ²(15) = 296.66, p < 0.001; LR χ²(3) = 1233.8, p < 0.001 Open in a new tab AOR Adjusted Odds Ratio, CI Confidence Interval HH Household ICC Intraclass Correlation Coefficient MOR Median Odds Ratio Discussion This study provides nationally representative evidence on household‑level associations with malaria risk perception in Ethiopia. Given its cross‑sectional design, the findings are associative rather than causal and should be interpreted as patterns of awareness shaped by socioeconomic and contextual correlates. Rural households consistently reported higher perceived risk, reflecting both limited access to preventive interventions and heightened awareness in endemic settings. This is consistent with findings from other African contexts [ 15 , 37 ]. Education showed a positive gradient, with households attaining secondary, vocational, or higher education reporting greater awareness, underscoring the role of health literacy and information exposure. Middle‑income households reported lower perceived risk than the poorest group, which may reflect possible complacency associated with improved living conditions. This interpretation aligns with prior evidence linking socioeconomic status to perceived disease risk rather than limited access to communication [ 1 , 13 ]. Older household heads similarly reported lower perceived risk, potentially due to habituation from lifetime experience with malaria or reduced engagement with newer health messaging [ 38 , 39 ]. These findings highlight the need to sustain engagement among older adults and address complacency among better‑off households. Ownership of communication assets such as radios, televisions, and mobile phones increased exposure to health messaging and served as cues to action, thereby strengthening awareness. Financial inclusion, through access to banking or microfinance services, enhanced empowerment, facilitated information flow, and supported engagement with health services. Conversely, households with repeated prior malaria experience may normalize risk, reducing perception despite ongoing vulnerability. These mechanisms emphasize the importance of integrating communication infrastructure and financial inclusion into malaria awareness strategies. Findings align with studies in Mozambique [ 15 ], Madagascar [ 37 ], and Nigeria [ 40 ], which emphasize the importance of digital and financial inclusion in empowering malaria-related knowledge. Spatial analysis revealed pronounced geographic inequalities in malaria risk perception. Nearly 38% of variation was explained at the woreda‑level, with additional clustering at the zone (19%) and regional (4%) levels. Hotspots in western and northwestern Ethiopia corresponded to ecological and socio‑environmental conditions favorable to transmission, while coldspots in highland regions suggested natural protection but potential complacency in awareness. Spatial outliers provided further insights: High–Low clusters (areas of high perceived risk surrounded by low‑perception neighbors) may indicate localized outbreaks or service gaps, whereas Low–High clusters (areas of low perceived risk surrounded by high‑perception neighbors) highlight successful local interventions embedded within high‑risk landscapes. Similar clustering has been observed in Kenya and Nigeria, reinforcing the regional relevance of geographically differentiated communication strategies [ 27 , 32 , 39 , 41 – 43 ]. These HL and LH clusters underscore areas where household perception diverges from surrounding communities, emphasizing the need for targeted communication and localized interventions. The significant variance observed at woreda and zone levels suggests that perceptions are shaped by shared malaria ecology (vector distribution, altitude, rainfall), shared health messaging (campaign intensity, NGO presence), and community norms (peer influence, collective memory of outbreaks). The high Median Odds Ratios (MORs) for woreda‑level effects confirm that households in different localities may perceive malaria risk very differently, even when their individual characteristics are similar. This reinforces the need for geographically tailored communication strategies and community‑level interventions that align perception with epidemiological realities. These findings highlight priority groups for targeted health education, particularly older and middle‑income households, and emphasize the importance of expanding communication infrastructure and financial inclusion. Geographic clustering further supports precision public health approaches, ensuring interventions are tailored to local contexts. Practical policy actions include strengthening the role of health extension workers in community‑level education, leveraging SMS messaging for rapid dissemination of malaria awareness, and implementing woreda‑level targeting to align perception with transmission risk. By integrating household perception into malaria surveillance and policy, Ethiopia can advance toward more equitable and effective malaria elimination strategies, aligned with the national roadmap and the WHO Global Technical Strategy. This study benefits from nationally representative NHES data, ensuring broad generalizability across Ethiopia. The integration of multilevel logistic regression with spatial autocorrelation methods provides robust insights into both socio‑economic and geographic correlates of malaria risk perception. Its novel focus on household‑level perception adds an important dimension to malaria research, where perception is often overlooked compared to biological incidence. Several limitations should be acknowledged. First, malaria risk perception was self‑reported, which introduces the possibility of reporting bias due to social desirability, recall error, or misunderstanding of survey questions. Second, the cross‑sectional design restricts temporal inference, preventing conclusions about changes in perception over time or causal pathways. Third, despite adjustment for multiple socio‑demographic and contextual variables, residual confounding may remain, as unmeasured factors such as prior malaria episodes, local program intensity, or community health campaigns could influence perception. Fourth, there is potential spatial misalignment between perceived risk and actual malaria transmission, given that perception reflects awareness rather than biological incidence. Another limitation is the perception-behavior gap, as households may perceive risk but not necessarily adopt preventive practices. This gap limits the extent to which perception can be assumed to influence actual malaria outcomes. This misalignment has policy implications, as areas with low perception but high transmission may be underserved, while areas with high perception but lower transmission may receive disproportionate attention. GPS displacement rules applied in the NHES dataset may further influence spatial precision despite sensitivity checks. Finally, the absence of entomological or clinical data prevents direct comparison between perception and transmission intensity, and the generalizability of findings may be restricted to Ethiopia during the 2022–2023 survey period. In summary, this study demonstrates that malaria risk perception in Ethiopia is shaped by socioeconomic disparities and geographic clustering, underscoring the importance of tailored communication strategies. By prioritizing older and middle‑income households, expanding communication infrastructure, and leveraging financial inclusion, interventions can better align awareness with epidemiological realities. Integrating household perception into surveillance and policy offers a pathway toward more equitable and effective malaria elimination, consistent with Ethiopia’s national roadmap and the WHO Global Technical Strategy. Conclusion This study provides nationally representative evidence on household‑level malaria risk perception in Ethiopia, showing that perceptions reflect awareness rather than incidence and are shaped by both individual sociodemographic factors and community‑level contexts. Nearly 38% of the variability in perception was attributable to woreda‑level differences, underscoring the importance of local context in shaping awareness. Areas with low perceived risk but high transmission potential are particularly vulnerable to delayed preventive action, whereas high‑perception areas may reflect successful communication or heightened concern. Targeted interventions should prioritize populations with lower awareness, especially older and middle‑income households, and focus on woredas where perception does not align with transmission risk. Overall, these findings demonstrate the value of integrating behavioral perception indicators into malaria surveillance and control strategies. Doing so will strengthen precision public health approaches, ensure interventions are tailored to local realities, and support Ethiopia’s progress toward equitable and effective malaria elimination in line with the national roadmap and the WHO Global Technical Strategy. Supplementary Information 12889_2026_27402_MOESM1_ESM.pdf (759.4KB, pdf) Supplementary Material 1. EPHI Institutional Review Board approval certificate protocol EPHI-IRB-403-2021. Acknowledgements The authors thank the Ethiopian Public Health Institute (EPHI) and the Ministry of Health with true gratitude for providing financial and technical assistance in the entire National Health Equity Survey process. The Regional State Health Bureaus and Regional Public Health Institutes are credited with providing a critical physical presence in the assignment of data collectors, regional coordinators, and with sustained provision of support and supervision in the field implementation. The members of the Steering Committee and Technical Working Group are of great interest to us as well, and they guided us and helped in carrying out the survey. We are grateful to health officials at the zonal and other administrative levels, household respondents, and key informants in Ethiopia whose input and views were priceless. We also recognize the Ethiopian Public Health Institute (EPHI) Management Council and staff as leaders and supporters of its operations. The data collectors are greatly valued in terms of their dedication and professionalism. Lastly, we would like to thank everyone who participated in the survey process in one form or another but were not mentioned in this manuscript. Abbreviations AOR Adjusted Odds Ratio CI Confidence Interval EA Enumeration Area EPHI Ethiopian Public Health Institute FDR False Discovery Rate GIS Geographic Information System HH Household Head ICC Intraclass Correlation Coefficient LISA Local Indicators of Spatial Association NHES National Health Equity Survey OR Odds Ratio PSNP Productive Safety Net Programme RCT Randomized Controlled Trial WHO World Health Organization Authors' contributions YL led the conceptualization, study design, and manuscript drafting. AZ served as the principal investigator and lead coordinator of the National Health Equity Survey, overseeing protocol development, field implementation, and data integrity. DA, AMG, GT, WB, KB, AL, TG, HA, AA, AT, AW, GM, FG, WT, SAB, HT, DZ, SA, SAy, TD, BT, and GT contributed to data acquisition, quality assurance, and preliminary analysis. YS provided methodological input and supported the interpretation of health economic implications. TK and YA contributed to contextual analysis and manuscript review. LF and GTi provided policy alignment guidance and supported institutional coordination. All authors reviewed and approved the final manuscript. Funding No funding was received for this study. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate This study received ethical approval from the Ethiopian Public Health Institute Institutional Review Board (EPHI-IRB), under protocol number EPHI-IRB-403-2021. The research was conducted in accordance with the principles outlined in the Declaration of Helsinki. Verbal informed consent was obtained from all participants before data collection. Consent for publication Not applicable. No personally identifiable information was collected. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Aychiluhm SB, Gelaye KA, Angaw DA, Dagne GA, Tadesse AW, Abera A, et al. Determinants of malaria among under-five children in Ethiopia: Bayesian multilevel analysis. BMC Public Health. 2020;20(1):1468. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 2. Debash H, Bisetegn H, Ebrahim H, Feleke DG, Gedefie A, Tilahun M, et al. Prevalence and associated risk factors of malaria among febrile under-five children visiting health facilities in Ziquala district, Northeast Ethiopia: A multicenter cross-sectional study. PLoS ONE. 2022;17(10):e0276899. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. WHO. WHO Malaria Policy Advisory. Group (MPAG) meeting report, 4, 5 and 7 March 2024. World Health Organization; 2024. https://www.who.int/publications/i/item/9789240095298 . 4. Fikadu M, Ashenafi E. Malaria: an overview. Infect Drug Resist. 2023:3339–47. [ DOI ] [ PMC free article ] [ PubMed ] 5. Hunter PJ, Awoyemi T, Ayede AI, Chico RM, David AL, Dewey KG, et al. Biological and pathological mechanisms leading to the birth of a small vulnerable newborn. Lancet. 2023;401(10389):1720–32. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Gelband H, Carshon-Marsh R, Ansumana R, Swaray IB, Pandey A, Aimone A, et al. Could vaccinating adults against malaria materially reduce adult mortality in high-transmission areas? Malar J. 2023;22(1):278. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Touré M, Keita M, Kané F, Sanogo D, Kanté S, Konaté D, et al. Trends in malaria epidemiological factors following the implementation of current control strategies in Dangassa, Mali. Malar J. 2022;21(1):65. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Domingos ES. Rethinking malaria elimination: a perspective on challenges and solutions in Angola. Malar J. 2025;24:214. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Okore W, Ouma C, Okoth RO, Yeda R, Ingasia LO, Mwakio EW, et al. Increased sensitivity of malaria parasites to common antimalaria drugs after the introduction of artemether-lumefantrine: Implication of policy change and implementation of more effective drugs in fight against malaria. PLoS ONE. 2024;19(6):e0298585. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Ochepo PA. Perceptions and attitudes to health-seeking delays for malaria treatment in Makurdi, Nigeria. 2022. 11. Getahun GK, Shewamare A, Andabob WA, Duressa EM, Birhanu MY. Healthcare professionals perceptions towards the determinants of effective emergency health care services in public health centres of Addis Ababa, Ethiopia. Afr J Emerg Med. 2024;14(4):240–5. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Feleke DG, Gebretsadik D, Gebreweld A. Analysis of the trend of malaria prevalence in Ataye, North Shoa, Ethiopia between 2013 and 2017. Malar J. 2018;17(1):323. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Assebe LF, Belete WN, Alemayehu S, Asfaw E, Godana KT, Alemayehu YK, et al. Economic evaluation of health extension program packages in Ethiopia. PLoS ONE. 2021;16(2):e0246207. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Ewing VL, Tolhurst R, Kapinda A, Richards E, Terlouw DJ, Lalloo DG. Increasing understanding of the relationship between geographic access and gendered decision-making power for treatment-seeking for febrile children in the Chikwawa district of Malawi. Malar J. 2016;15(1):521. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 15. Hutchinson P, Zulliger R, Butts JK, Candrinho B, Saifodine A, Eisele TP, et al. Interpersonal communication, cultural norms, and community perceptions associated with care-seeking for fever among children under age five in Magoé district, Mozambique. Malar J. 2023;22(1):279. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. Ngutu M, Omia DO, Ngage TO, Oduor CA, Ouko NO, Oingo B, et al. Gender-related factors affecting community malaria-related perceptions and practices in Migori County, Kenya. Malar J. 2025;24(1):196. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Adugna A. Malaria in Ethiopia. Ethio Demogr Heal org. 2011. 18. Endriyas M, Kassa M, Chisha Y, Mekonnen E, Misganaw T, Loha E, et al. Low long-lasting insecticidal net use in malaria elimination areas in Southern Ethiopia: results from community based cross-sectional study. Malar J. 2024;23(1):94. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Lindblade KA, Li Xiao H, Tiffany A, Galappaththy G, Alonso P. Supporting countries to achieve their malaria elimination goals: the WHO E-2020 initiative. Malar J. 2021;20(1):481. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Quaye IK, Aleksenko L, Paganotti GM, Peloewetse E, Haiyambo DH, Ntebela D, et al. Malaria elimination in Africa: rethinking strategies for Plasmodium vivax and lessons from Botswana. Trop Med Infect Dis. 2023;8(8):392. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. WHO. Malaria – Ethiopia. Disease Outbreak News. 2024. Available from: https://www.who.int/emergencies/disease-outbreak-news/item/2024-DON542 . 22. Ashine T, Eyasu A, Asmamaw Y, Simma E, Zemene E, Epstein A, et al. Spatiotemporal distribution and bionomics of Anopheles stephensi in different eco-epidemiological settings in Ethiopia. Parasit Vectors. 2024;17(1):166. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Hassen J, Dinka H. Urban malaria prevalence and its associated risk factors: the case of Batu town, East Shoa. Ethiopia: Oromia Regional State; 2020. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Legendre E, Lehot L, Dieng S, Rebaudet S, Thu AM, Rae JD, et al. Malaria temporal dynamic clustering for surveillance and intervention planning. Epidemics. 2023;43:100682. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Zegeye AF, Mekonen EG, Gebrehana DA, Tekeba B, Tamir TT. Spatial variation and multilevel determinants of malaria infection among pregnant women in Sub-Saharan Africa: using malaria indicator surveys. BMC Infect Dis. 2025;25(1):654. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. Alonso D, Bouma MJ, Pascual M. Epidemic malaria and warmer temperatures in recent decades in an East African highland. Proc R Soc B Biol Sci. 2011;278(1712):1661–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Birhanu Z, Abebe L, Sudhakar M, Dissanayake G, Yihdego YY, ebiyo, Alemayehu G, et al. Malaria related perceptions, care seeking after onset of fever and anti-malarial drug use in malaria endemic settings of southwest Ethiopia. PLoS ONE. 2016;11(8):e0160234. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Dabaro D, Birhanu Z, Negash A, Hawaria D, Yewhalaw D. Effects of rainfall, temperature and topography on malaria incidence in elimination targeted district of Ethiopia. Malar J. 2021;20(1):104. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Teferi M. Geography of Ethiopia and the Horn. 2019. 30. Jalilian A, Ayana GM, Ashine T, Hailemeskel E, Ebstie YA, Molla E, et al. Waning success: a 2013–2022 spatial and temporal trend analysis of malaria in Ethiopia. Infect Dis Poverty. 2024;13(1):93. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Woyessa D, Yewhalaw D. Anopheles mosquito fauna, blood meal sources and transmission intensity from high and moderate malaria endemic areas of Ethiopia. Sci Rep. 2025;15(1):10636. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Teka H, Golassa L, Medhin G, Balkew M, Sisay C, Gadisa E, et al. Trend analysis of malaria in urban settings in Ethiopia from 2014 to 2019. Malar J. 2023;22(1):235. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Haileselassie W, Zemene E, Lee MC, Zhong D, Zhou G, Taye B, et al. The effect of irrigation on malaria vector bionomics and transmission intensity in western Ethiopia. Parasit Vectors. 2021;14(1):516. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 34. EPHI and MoH. Ethiopia National Health Equity Survey 2022/23. Addis Ababa; 2024. 35. Awuah RB, Asante PY, Sakyi L, Biney AAE, Kushitor MK, Agyei F, et al. Factors associated with treatment-seeking for malaria in urban poor communities in Accra, Ghana. Malar J. 2018;17(1):168. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Griffin T, Pabon-Rodriguez F, Ayodo G, Zhuang Y. Strengthening health information systems and inherent statistical outputs for improved malaria control and interventions in western Kenya. Front Epidemiol. 2025;5:1591261. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Favero R, Dentinger CM, Rakotovao JP, Kapesa L, Andriamiharisoa H, Steinhardt LC, et al. Experiences and perceptions of care-seeking for febrile illness among caregivers, pregnant women, and health providers in eight districts of Madagascar. Malar J. 2022;21(1):212. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 38. Haileselassie W, Parker DM, Taye B, David RE, Zemene E, Lee MC, et al. Burden of malaria, impact of interventions and climate variability in Western Ethiopia: an area with large irrigation based farming. BMC Public Health. 2022;22(1):196. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Chiziba C, Mercer LD, Diallo O, Bertozzi-Villa A, Weiss DJ, Gerardin J, et al. Socioeconomic, demographic, and environmental factors may inform malaria intervention prioritization in urban Nigeria. Int J Environ Res Public Health. 2024;21(1):78. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 40. Mann DM, Swahn MH, McCool S. Undernutrition and malaria among under-five children: findings from the 2018 Nigeria demographic and health survey. Pathog Glob Health. 2021;115(6):423–33. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Madani AF, Triani E, Hasbi N. Environmental risk factors that influence malaria incidence: Literature review. Media Publ Promosi Kesehat Indones. 2024;7(6):1405–13. [ Google Scholar ] 42. Arisco NJ, Sewe MO, Bärnighausen T, Sié A, Zabre P, Bunker A. The effect of extreme temperature and precipitation on cause-specific deaths in rural Burkina Faso: a longitudinal study. Lancet Planet Heal. 2023;7(6):e478–89. [ DOI ] [ PubMed ] [ Google Scholar ] 43. Richmond AK, Malcomb D, Ringler K. Household vulnerability mapping in Africa’s Rift Valley. Appl Geogr. 2015;63:380–95. [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials 12889_2026_27402_MOESM1_ESM.pdf (759.4KB, pdf) Supplementary Material 1. EPHI Institutional Review Board approval certificate protocol EPHI-IRB-403-2021. Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Articles from BMC Public Health are provided here courtesy of BMC ACTIONS View on publisher site PDF (3.7 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Anopheles stephensi bionomics and epidemiology in Ethiopia: A systematic review and meta-analysis with implications for urban malaria control#223917
• Socioeconomic determinants of malaria in Ugandan children: An interpretable machine learning approach for public health policy.#653505
• Socioeconomic determinants of malaria in Ugandan children: An interpretable machine learning approach for public health policy#646452
• High Prevalence of Artemisinin-Resistant Plasmodium falciparum , Southeastern Sudan#422602
• 'Poverty is the biggest barrier': stakeholder perspectives on social determinants of malaria in Wakiso District, Uganda.#1054384