Application of machine learning methods for predicting childhood anaemia: Analysis of Ethiopian Demographic Health Survey of 2016

Childhood anaemia is a public health problem in Ethiopia. Machine learning (ML) is a growing in medicine field to predict diseases. Diagnosis of childhood anaemia is resource intensive. The aim of this study is to apply machine learning (ML) algorithm to predict childhood anaemia using socio-demographic, economic, and maternal and child related variables. The study used data from 2016 Ethiopian demographic health survey (EDHS). We used Python software version 3.11 to apply and test ML algorithms through logistic regression, Random Forest (RF), Decision Tree, and K-Nearest Neighbours (KNN). We evaluated the performance of each of the ML algorithms using discrimination and calibration parameters. The predictive performance of the algorithms was between 60% and 66%. The logistic regression model was the best predictive model of ML with accuracy (66%), sensitivity (82%), specificity (42%), and AUC (69%), followed by RF with accuracy (64%), sensitivity (79%), specificity (42%), and AUC (63%). The logistic regression and the RF models of ML showed poorest family, child age category between 6 and 23 months, uneducated mother, unemployed mother, and stunting as high importance predictors of childhood anaemia. Applying logistic regression and RF models of ML can detect combinations of predictors of childhood anaemia that can be used in primary health care professionals.

Willingness to pay for Social Health Insurance and associated factors among Public Civil Servants in Ethiopia: A systematic review and meta-analysis

BACKGROUND

The provision of equitable and accessible healthcare is one of the goals of universal health coverage. However, due to high out-of-pocket payments, people in the world lack sufficient health services, especially in developing countries. Thus, many low and middle-income countries introduced different prepayment mechanisms to reduce large out-of-pocket payments and overcome financial barriers to accessing health care. Though many studies were conducted on willingness to pay for social health insurance in Ethiopia, there is no aggregated data at the national level. Therefore, this systematic review and meta-analysis aimed to estimate the pooled magnitude of willingness to pay for social health insurance and its associated factors among public servants in Ethiopia.

METHOD

Studies conducted before June 1, 2022, were retrieved from electronic databases (PubMed/Medline, Science Direct, African Journals Online, Google Scholar, and Web of Science) as well as from Universities' digital repositories. Data were extracted using a data extraction format prepared in Microsoft Excel and the analysis was performed using STATA 16 statistical software. The quality of the included studies was assessed using the Newcastle-Ottawa Scale for cross-sectional studies. To evaluate publication bias, a funnel plot, and Egger's regression test were utilized. The study's heterogeneity was determined using Cochrane Q test statistics and the I2 test. To determine the pooled effect size, odds ratio, and 95% confidence intervals across studies, the DerSimonian and Laird random-effects model was used. Subgroup analysis was conducted by region, sample size, and publication year. The influence of a single study on the whole estimate was determined via sensitivity analysis.

RESULT

To estimate the pooled magnitude of willingness to pay for the Social Health insurance scheme in Ethiopia, twenty articles with a total of 8744 participants were included in the review. The pooled magnitude of willingness to pay for Social Health Insurance in Ethiopia was 49.62% (95% CI: 36.41-62.82). Monthly salary (OR = 6.52; 95% CI:3.67,11.58), having the degree and above educational status (OR = 5.52; 95%CI:4.42,7.17), large family size(OR = 3.69; 95% CI:1.10,12.36), having the difficulty of paying the bill(OR = 3.24; 95%CI: 1.51, 6.96), good quality of services(OR = 4.20; 95%CI:1.97, 8.95), having favourable attitude (OR = 5.28; 95%CI:1.45, 19.18) and awareness of social health insurance scheme (OR = 3.09;95% CI:2.12,4.48) were statistically associated with willingness to pay for Social health insurance scheme.

CONCLUSIONS

In this review, the magnitude of willingness to pay for Social Health insurance was low among public Civil servants in Ethiopia. Willingness to pay for Social Health Insurance was significantly associated with monthly salary, educational status, family size, the difficulty of paying medical bills, quality of healthcare services, awareness, and attitude towards the Social Health Insurance program. Hence, it's recommended to conduct awareness creation through on-the-job training about Social Health Insurance benefit packages and principles to improve the willingness to pay among public servants.

Lesson Learned During a Pandemic: A Phenomenological Study of Pregnant Women's Experience in Resource-Limited Setting

Introduction

Pandemics like COVID-19 has strained healthcare systems globally, particularly in resource-limited settings where pre-existing healthcare gaps and constraints amplify the challenges faced by pregnant women. Having this lesson on pregnant women's experiences during COVID-19 offers invaluable insights into potential strategies and interventions needed to mitigate the impact of similar outbreaks in the future. Therefore, this phenomenological study aimed to uncover the experiences of pregnant women during the COVID-19 pandemic in Dilla town, Gedeo zone, southern Ethiopia.

Method and Materials

A descriptive phenomenological design was employed at Dilla University teaching hospital. A Purposive sampling technique was used and point of data saturation was declared at participant 16th. An in-depth interview was conducted in a private room after they received antenatal care services. The interviews were audiotaped, transcribed, and analysed using the open code software version 4.03. The interviews were transcribed verbatim to provide tick descriptions of data. Thematic analysis was employed using an inductive approach. Direct quotes from participants were included without any grammar editing to preserve their intended meaning.

Results

Our study found five predominant themes: mental health consequences of the COVID-19 pandemic, health impact of COVID-19, concerns over the economic impact of the COVID-19 pandemic, COVID-19 influence on social engagements, and challenge on getting reliable information.

Conclusion and Recommendation

This study identified that pregnant women experienced various challenges during the pandemic which includes psychological problems, health impact of COVID-19, economic problems, lack of reliable information and limited social support. Hence, implementing a customized approach by integrating psychosocial interventions with prenatal care services and conducting intensive health education sessions and economic empowerment are beneficial as mitigation strategies to better prepare for future pandemics.

Spatial patterns and spatially-varying factors associated with childhood acute respiratory infection: data from Ethiopian demographic and health surveys (2005, 2011, and 2016)

BACKGROUND

In Ethiopia, acute respiratory infections (ARIs) are a leading cause of morbidity and mortality among children under five years. Geographically linked data analysis using nationally representative data is crucial to map spatial patterns of ARIs and identify spatially-varying factors of ARI. Therefore, this study aimed to investigate spatial patterns and spatially-varying factors of ARI in Ethiopia.

METHODS

Secondary data from the Ethiopian Demographic Health Survey (EDHS) of 2005, 2011, and 2016 were used. Kuldorff's spatial scan statistic using the Bernoulli model was used to identify spatial clusters with high or low ARI. Hot spot analysis was conducted using Getis-OrdGi statistics. Eigenvector spatial filtering regression model was carried out to identify spatial predictors of ARI.

RESULTS

Acute respiratory infection spatially clustered in 2011 and 2016 surveys year (Moran's I:-0.011621-0.334486). The magnitude of ARI decreased from 12.6% (95%, CI: 0.113-0.138) in 2005 to 6.6% (95% CI: 0.055-0.077) in 2016. Across the three surveys, clusters with a high prevalence of ARI were observed in the North part of Ethiopia. The spatial regression analysis revealed that the spatial patterns of ARI was significantly associated with using biomass fuel for cooking and children not initiating breastfeeding within 1-hour of birth. This correlation is strong in the Northern and some areas in the Western part of the country.

CONCLUSION

Overall there has been a considerable decrease in ARI, but this decline in ARI varied in some regions and districts between surveys. Biomass fuel and early initiation of breastfeeding were independent predictors of ARI. There is a need to prioritize children living in regions and districts with high ARI.

Food taboo practices and associated factors among pregnant women in Ethiopia: a systematic review and meta-analysis

Food taboos have a negative impact on pregnant women and their fetuses by preventing them from consuming vital foods. Previous research found that pregnant women avoided certain foods during their pregnancy for a variety of reasons. This review aimed to determine the pooled prevalence of food taboo practices and associated factors in Ethiopia. In compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline, we searched the literature using PubMed/MEDLINE, AJOL (African Journal Online), HINARI, Science Direct, Google Scholar, and Google electronic databases. The random-effects model was used to estimate the pooled prevalence of food taboo and its determinants at a 95% confidence interval with their respective odds ratios. The pooled food taboo practice among Ethiopian pregnant women was 34.22% (95% CI 25.47-42.96), and after adjustment for publication bias with the trim-and-fill analysis, the pooled food taboo practice of pregnant women was changed to 21.31% (95% CI: 10.85-31.67%). Having less than a secondary education level (OR = 3.57; 95% CI 1.43-8.89), having no ANC follow-up (OR = 4.35; 95% CI 1.12-16.94), and being a rural resident (OR = 3.08; 95% CI 1.14-8.28) were the significant factors. Dairy products, some fruits, green leafy vegetables, meat, and honey are among the taboo foods. The most frequently stated reasons for this taboo practice were: fear of producing a big fetus, which is difficult during delivery; attachment to the fetus's body or head; and fear of fetal abnormality.

Prevalence of wasting and associated factors among children aged 2-5 years, southern Ethiopia: a community-based cross-sectional study

BACKGROUND

Wasting (acute malnutrition) is the most serious form of malnutrition for children in the near term. Malnutrition has a variety of causes, all of which are interconnected and hierarchically related. The purpose of this study was to assess the prevalence of wasting and its associated determinants among children under the age of five in the Wonago district, Gedeo zone, southern Ethiopia.

METHODS

Community based cross-sectional study was conducted from October 1 to 30, 2021 using a systematic random sampling technique. Data were entered using Epidata manager and STATA v.20 software was used for analysis. Descriptive statistics were reported to describe the study population. To identify associated factors of wasting, bivariate and multivariate logistic regression analysis were fitted. Variables having p-value < 0.05 were declared statistically significant predictors of wasting.

RESULTS

A total of 390 respondents participated with a response rate of 92.6%. The prevalence of wasting among children aged 2-5 years in Wonago district was 36.4% (95% CI: 31.76-41.32). Moderate household food insecurity (AOR = 0.35, 95%CI: 0.14-0.83), history of recurrent illness (AOR = 0.15, 95%CI: 0.26-0.84), and duration of breastfeeding greater than 2 years (AOR = 0.15, 95%CI: 0.26-0.84) were significantly associated with wasting.

CONCLUSION

Almost one-third of the children were wasted. Household food insecurity, breastfeeding, and recurrent illness were significantly associated with wasting among children aged 24-59 months. It is recommended that interventions be designed with food security, disease prevention, and breastfeeding awareness in mind and put the spotlight on food and nutrition policy to insure children's nutritional status.

Iodine concentration level of iodized dietary salt and its associated factors: a systematic review and meta-analysis

Background

Iodine deficiency disorder (IDD) is a major public health problem in Ethiopia. The availability of adequate iodized dietary salt at the household level is immensely important. Hence, this review aimed to estimate the pooled prevalence of adequate iodine concentration level of iodized dietary salt at the household level and its associated factors in Ethiopia.

Methods

We searched the literature using electronic databases (PubMed/Medline, Google Scholar, Science Direct, and Embase) and gray literature from January 9, 2022, to February 25, 2022. The rapid test kit was used to measure the adequacy of iodine level of dietary salt. The quality of studies was assessed using Joanna Briggs Institute critical appraisal tool. Heterogeneity between studies was checked using I2 test statistics and publication bias was checked using funnel plot and Egger's statistical test at a 5% significance level. A random-effects model was employed to estimate the pooled prevalence of the outcome variable and its determinants in Ethiopia.

Results

The search identified 149 studies of which 18 studies were included with a total of 10,556 participants. The pooled prevalence of adequate iodine levels of iodized salt in Ethiopia was 44.37% (95% CI: 35.85-52.88). Women who had formal education (adjusted odds ratio (AOR) = 1.99 (95% CI: 1.47-2.48)), good knowledge of women (AOR = 2.14, 95% CI: 1.36-3.36), packed iodized salt (AOR = 3.85 (95% CI: 1.88-7.87)) and storage of iodized salt at home for less than 2 months (AOR = 2.66 (95% CI: 2.11-3.35) were the significant factors.

Conclusion

This review suggests that the pooled prevalence of adequate iodine levels was low. Our finding highlights the need for considering the educational status, knowledge, and duration of salt storage to enhance the prevalence of adequate levels of iodized salt at the national level.

Diabetes mellitus and its association with central obesity, and overweight/obesity among adults in Ethiopia. A systematic review and meta-analysis

Background

Nowadays, diabetes mellitus is a serious public health problem in Ethiopia that has a profound impact on the health care system. However, no systematic synthesis and meta-analysis has been performed to depict the national prevalence. Hence, we authors aimed to assess the pooled prevalence of diabetes mellitus and its association with central obesity, overweight/obesity among adults in Ethiopia.

Methods

We did a systematic review and meta-analysis of 15 eligible studies on the national prevalence of DM and its association with central obesity, and overweight/obesity among adults in Ethiopia. We searched PubMed/Medline, Science Direct, Embase, and Google Scholar, from August 01 up to October 28, 2021, in accordance with PRISMA guidelines. Joanna Briggs Institute (JBI) critical appraisal tool was used to assess the quality of studies. Analysis was done using STATA version 14 software. Heterogeneity was checked using the I-squared test, and the publication bias was examined by funnel plot and eggers test. Moreover, Sensitivity analysis was done to check the influence of small studies on the outcome. The trim and fill analysis was performed to estimate the potentially missing articles because of publication bias.

Result

Total of 15 studies that met the inclusion criteria were included and the pooled prevalence of diabetes mellitus of the Federal Democratic Republic of Ethiopia was 6.26 (95%CI: 4.74–7.78). In the subgroup analysis, the prevalence of diabetes mellitus among the studies conducted in 2017 and before was 4.56 (95%CI: 2.98–6.14) but in studies done after 2017 was 7.55(95%CI: 4.69–10.41). The burden of diabetes mellitus was 5.79 times higher among those adults who had central obesity (OR = 5.79; 95%CI; 3.14–10.70), 5.70 times higher among adults who had overweight/obesity (OR = 5.70, 95%CI: 3.35–9.70).

Conclusion

The national prevalence of diabetes mellitus among adults in Ethiopia was higher and associated with central obesity, and overweight/obesity. Hence, the government of Ethiopia and stakeholders should give attention to strengthen the current health system regarding non-communicable diseases like diabetes mellitus and obesity/overweight.

Health related quality of life and its association with social support among people living with HIV/AIDS receiving antiretroviral therapy in Ethiopia: a systematic review and meta-analysis

Background

People living with HIV/AIDS (PLWHA) are frequently confronted with severe social issues such as rejection, abandonment, criticism, and stigma. This would negatively affect their quality of life. Several studies have been conducted so far to assess factors affecting the health-related quality of life among people living with HIV/AIDS who are on antiretroviral therapy (ART) in Ethiopia. However, to our knowledge, there is no previous study that has summarized the results of the studies that investigated health-related quality of life (HRQOL) among PLWHA in Ethiopia. Therefore, the purpose of this review was to estimate the pooled prevalence of HRQOL and its association with social support among people living with HIV/AIDS (PLWHA) on ART in Ethiopia.

Methods

A systematic search was carried out using several electronic databases (PubMed, Science Direct, Web of Science, and Cochrane electronic), Google Scholar, Google, and a manual search of the literature on health-related quality of life among people living with HIV/AIDS who are on ART. A Microsoft Excel data extraction sheet was used to extract pertinent data from an individual study. To assess the heterogeneity of primary articles, the Cochrane Q test statistics and the I2 test were carried out, and a random effects meta-analysis was used to estimate the pooled prevalence of HRQOL.

Result

Out of the 493 articles reviewed, ten with a total of 3257 study participants were eligible for meta-analysis. The pooled prevalence of HRQOL among people living with HIV/AIDS who are on antiretroviral therapy in Ethiopia was 45.27%. We found that strong perceived social support was significantly associated with higher levels of subjectively perceived HRQOL. PLWHA who were on ART and had good social support were four times more likely to report higher HRQOL when compared to their counterparts [AOR = 4.01, 95% CI 3.07–5.23].

Conclusion

A substantial number of PLWHA had poor HRQOL in Ethiopia. Social support was significantly associated with HRQOL among people living with HIV/AIDS. Hence, it’s recommended to encourage suitable intervention at every follow-up visit, and psycho-social support is also warranted to improve the quality of life.

Antibiotics non-adherence and its associated factors among households in southern Ethiopia

Objectives:

This study aimed at assessing the prevalence of antibiotics non-adherence and its associated factor among households in southern Ethiopia.

Methods:

A community-based cross-sectional study was conducted among 323 randomly selected households in Wenago town. To collect the data, structured questionnaire was used. Categorical variables were represented by frequency and percentage. For continuous variables, the mean value and standard deviation were used. Bivariate and multivariate logistic regression analyses were used to identify factors related to antibiotic non-adherence. Finally, for significant factors with p-values less than 0.05, the adjusted odds ratio with 95% confidence interval was calculated and evaluated.

Results:

The prevalence of antibiotic non-adherence in the household was 194 (60.1%) (95% confidence interval = 55.1–65.6). Remission of symptoms (63%) is one of the top reasons for antibiotic non-adherence in the home. Male sex (adjusted odds ratio = 1.77, 95% confidence interval = 1.03–3.08), lower educational status (adjusted odds ratio = 3.42, 95% confidence interval = 1.51–7.75; adjusted odds ratio = 2.37, confidence interval = 1.12–5.02), poor attitude toward antibiotics use (adjusted odds ratio = 1.89; 95% confidence interval = 1.23–3.04), poor knowledge about antibiotics use (adjusted odds ratio = 1.34; 95% confidence interval = 1.11–2.39), and no-prescription information from pharmacy (adjusted odds ratio = 2.02, 95% confidence interval = 1.09–3.72) were all associated with non-adherence. While no medication discomfort (adjusted odds ratio = 0.31, 95% confidence interval = 0.178–0.56) had a negative effect on non-adherence.

Conclusion:

In this study, antibiotic non-adherence was considerably high among the participants. Being male, lower educational status, poor attitude, poor knowledge, no-prescription information from pharmacy/druggist, and medication discomfort were related with antibiotic non-adherence. As a result, community service providers must provide relevant prescription information as well as appropriate counseling to antibiotic non-adherent patients.

Community-based new born care practice and its associated factors among women who give birth at home in Ethiopia: cross-sectional study

BACKGROUND

In Ethiopia, perinatal mortality rate was 33 per 1000 pregnancies and 64.4% of this death was occurred within the first 7 days of neonatal life. Moreover, more than 2.1% of new born babies were died within their first seven days of life in Ethiopia. Majority of neonatal deaths are preventable by applying an effective and lifesaving interventions. However, little is known about newborn care practice at the community level.

METHODS

A community-based cross-sectional study design was used. Multi-stage sampling techniques were used to get a total of 540 mothers who gave birth at home within the past six months from their kebeles in Ethiopia. Data was collected by using face-to-face interview with structured questionnaires. Then the data was coded, cleaned, and entered into Epidemiological data version 3.1 and exported to statistical package for social science version 23.0 for analyses. Bivariate and multivariable logistic regression model was used to identify statistically significant associations between dependent and independent variables. The odds ratio at 95% confidence interval with P-Value 0.05% was computed.

RESULTS

A total of 540 women were participated with a response rate of 98.3%. Only 208 (44%) of the women had good practice towards essential newborn care. Head of households [AOR (95%CI) 2.7417 (1.80-4.25)], type of birth attendant [AOR (95%CI) 3.962 (3.329-7.171)] and bad obstetrical history [AOR (95%CI) 3.151 (2.209-4.969)] were significantly associated with maternal newborn care practice.

CONCLUSION

Less than half of the mothers had good newborn care practice. In this study, head of household, type of birth attendant, and bad obstetrical history were significantly associated with maternal newborn care practice. Therefore, Ministry of Women and Woreda women and Child offices needs to promote the socioeconomic empowerment of women to increase the practice of essential newborn care practices.

Minimum dietary diversity and its associated factors among infants and young children in Ethiopia: evidence from Ethiopian Demographic and Health Survey (2016)

Background

Adequate infant and young child feeding during the first 1000 days of life is very essential to improve child health, survival, growth, and development through minimum dietary diversity (MDD). Hence, this study aimed to assess MDD and its multi-level factors among infants and young children aged 6–23 months in Ethiopia.

Methods

Ethiopian Demographic and Health Survey (EDHS-2016) data was used to identify both individual and community-level factors of dietary diversity. Weighted samples of 2,962 children were eligible and a multi-level regression model was used for the analysis. Finally, factors with a P-value of <0.05 were considered statistically significant.

Results

The prevalence of MDD among children in Ethiopia was 12.09%. According to this study, factors such as having a mother who attended higher education (AOR = 3.09, (95% CI; _1.67–5.71)), being a female household head (AOR = 0.62, (95% CI; _0.40–0.95)), having a mother's agricultural occupation (AOR = 1.89, (95% CI; _1.10–3.23)) and living in the household in the richest wealth index were significantly associated at the individual level. At the community level, children living in rural areas (AOR = 0.62, 95% CI; 0.39–0.98) were significant risk factors for MDD (AOR = 0.62, 95% CI; 0.39–0.98).

Conclusion

The educational and occupational status of the mother, wealth index, and region were significantly associated with MDD. Hence, strengthening of the existing nutritional intervention is helpful to increase diversified food consumption among children.