Early newborn bathing practice and its determinants among postpartum women in Ethiopia: a systematic review and meta-analysis

This systematic review and meta-analysis study was done to determine the level of early bathing practice and its determinants among women in Ethiopia. The study is reported per Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guideline (Supplementary file 1). The study is registered at PROSPERO with the registration number CRD42023493338.

The study utilized PICOS criteria. The population consisted of postpartum women who had given birth; the exposures are factors associated with the level of early newborn bathing practice; the outcome of the study was early newborn bathing practice; the studies' design were observational studies that looked at early infant bathing practices; and the context of the studies was Ethiopia. Additionally, the studies considered in this systematic review and meta-analysis study were written in English and from inception to October 10, 2023. This systematic review and meta-analysis excluded case series/reports, reviews, commentary, and editorials.

We performed intensive literature searches on different database platforms such as Web of Sciences, CINAHL (EBOSCO), Science Direct, MEDLINE, EMBASE, Google Scholar, and PubMed and websites until October 10, 2023. We also tried to access different universities' institutional repository websites. Unpublished articles were also searched directly on Google web search.

The search was accomplished using a combination of Boolean logic operators (AND, OR, NOT), Medical Subject Headings (MeSH), and keywords. The search strategy for advanced PubMed includes "infant, newborn"[MeSH Terms] AND ("bathes"[All Fields] OR "bathings"[All Fields] OR "baths"[MeSH Terms] OR "baths"[All Fields] OR "bathe"[All Fields] OR "bathed"[All Fields] OR "bathing"[All Fields]) AND ("woman"[All Fields] OR "women"[MeSH Terms] OR "women"[All Fields] OR "woman"[All Fields] OR "women’s"[All Fields] OR "women's"[All Fields]) AND "Ethiopia"[MeSH Terms]. For other databases, keywords early newborn bathing AND Postpartum women AND Ethiopia were used.

All articles searched from the database and websites were consolidated, and exported to the reference management application (Endnote version X8). Then duplicate articles were manually removed. The titles and abstracts of the papers were then carefully evaluated. Two authors (AD, and AE) independently reviewed the full texts of the remaining publications to determine their eligibility based on predetermined inclusion and exclusion criteria (PICOS). The objectives, methodology, population, and significant findings (level of early newborn bathing practice, determinants, factors/predictors of early newborn practice) of the full-text studies written in English were then reviewed further. The two authors came to a logical agreement to handle any questions that developed during the extraction process, and the final agreement was finalized with the assistance of the authors (TG and AK). The PRISMA statement flow diagram is used to represent the entire research selection process.

After identifying articles that met the inclusion, the two authors (AK and AE) extracted the data independently using a Microsoft Excel 2016 sheet. Different variables of interest were considered in the extraction and presented in table of summary. The accuracy of the data extraction was checked by comparing the results produced by the two authors. The information used for meta-analysis from the included articles was extracted, which included the sample size, frequency of the occurrence of early newborn practice, measure of associations with 95% CI, and effect size.

The primary outcome of this meta-analysis was the level of early newborn bathing practice among postpartum women in Ethiopia. The early newborn bathing practice was defined as the practice of bathing a newborn 24 h after giving birth or before the WHO recommended time [17, 18] The secondary outcome of interest was identifying significant factors associated with Early newborn bathing practice among postpartum women in Ethiopia.

The risk of bias was assessed using the Newcastle Ottawa scale (NOS) which is a validated tool for assessing the quality of non-randomized studies (cross-sectional studies) [19]. The risk of bias assessment tool includes the following domains: selection domain including representativeness of the sample, non-respondents, ascertainment of the exposure (risk factor); comparability domain (the subjects in different outcome groups are comparable, based on the study design or analysis and confounding factors are controlled); outcome domain such as assessment of the outcome and statistical tests. The included studies' methodological validity and the quality of their conclusions were scrutinized. Two authors (AD, and AE) assessed and scored the quality of the study using NOS. The mean score of the authors was utilized to make the final decision. Based on their performance against each tool indicator, the included studies were categorized as high, moderate, or low quality. The quality score of the 6 studies ranges from 6 to 7, with most studies (7 studies) scoring 7. All of the six studies were considered of adequate quality for inclusion in the analyses.

Statistical analysis was conducted using STATA 18 software. The overall estimate of the level of early bathing and overall effect size of determinants with 95% CI were presented using forest plots. A meta-analysis of the level of early newborn bathing was done by random effect model using Der Simonian Liard method analysis to reduce the heterogeneity of the included studies. Since there was significant heterogeneity existed between included studies, subgroup analysis, sensitivity analysis, and meta-regression were performed to deal with such heterogeneity. Subgroup analyses were performed based on region and year of publication. Both bivariate and multivariate meta-regression were performed to determine and identify the source of heterogeneity. Sensitivity analysis was done using the leave-one-out meta-analysis method to identify the effect of a single study on the overall estimate and to identify the outliers. A meta-analysis of observational studies was undertaken based on the recommendations of the I2 statistic given by Higgins et al. (an I2 of 75/100% and above implying considerable heterogeneity). Additionally, a significant P value in the test of theta indicated heterogeneity between studies.

To look for potential publication bias, the researchers utilized Egger's Regression Test, trim fill analysis, and a visual evaluation of a funnel plot. Egger's test with a p-value less than 0.05 showed a small study effect on the estimate but an insignificant p-value indicates no small study effect. Trim fill analysis that showed no difference in observed and combination of observed and imputed effect indicated no publication bias. While we conducted a visual inspection of the funnel plot, the asymmetric presentation of the funnel plot indicated the presence of publication bias. Meta-analysis of factors associated with the level of early newborn bathing was analyzed using the random effect model by the Hedge method.

A total of 244 published and unpublished articles were found in legitimate databases and websites. Using ENDNOTE and visual inspection, 39 publications were removed from all identified studies due to duplication. The remaining 205 studies were then maintained and screened based on title and abstract. After being vetted based on titles and abstracts, 167 were removed. Thirty-eight articles were considered eligible for inclusion. Of those, 32 studies were removed due to reasons such as studies assessing essential newborn care practices, studies conducted outside Ethiopia, and studies assessed hypothermia. Finally, this systematic review and meta-analysis comprised six published observational studies that met the inclusion criteria (Fig. 1). Using the NOS criteria, all of the six studies yielded moderate and high-quality scores. Then all of the studies were considered eligible in this meta-analysis.

This systematic review and meta-analysis comprised six observational studies (4 cross-sectional and two mixed studies) that assess early newborn bathing practice and associated factors. The sample size of the included studies ranges from a minimum of 382 in a study by Getachew, G. et al., [18] to a maximum of 610 in studies done by Girma, Tsinuel [20]. The majority of the included studies were published after 2022 [5, 17, 18, 21]. This illustrated that early newborn bathing is a recent concept and a current health problem for newborns. Of the included studies, half of the studies (3) were conducted in the Oromia region [5, 20, 21], one study from Harari [6], one study from Afar [18], and one study from the central Ethiopia region [17]. Three of the studies were conducted at the hospital while the other three studies were conducted in community settings. A total of 2787 postpartum women were included in this systematic review and meta-analysis. A summary of the main characteristics of the studies included in this systematic review and meta-analysis is presented in Table 1.

All included studies were considered in a meta-analysis of the primary outcome of the study. The level of early newborn bathing practice ranges from 32% (95% CI: 28–37) to a maximum of 84% (95% CI: 80–88). The pooled level of early newborn bathing practice among postpartum women in Ethiopia was 55% [95% CI: 38–71]. There is significant heterogeneity among individual studies (i² = 98.76%), the P value of the test of theta is also significant (P = less than 0.0001) and the confidence interval does not even overlap (Fig. 2).

Based on subgroup analysis by region, the highest level of early newborn bathing practice among postpartum women in Ethiopia was among studies conducted in the Afar region which was 73% (95% CI: 69–77). The heterogeneity among the subgroups was significant as the test of group difference showed that there is significant heterogeneity (P value less than 0.0001). Based on subgroup analysis by the publication year, the highest level of early newborn bathing practice among postpartum women in Ethiopia was observed among studies conducted during and after 2021 which was 59% (95% CI: 35–71) (Fig. 3).

Of the six studies, three studies report the association between primary education and early bathing practices among women in Ethiopia. The overall pooled effect size showed that primary education is associated with early bathing practices among women in Ethiopia. Mothers who had primary education were 49% less likely to practice early newborn baths compared to mothers with no education (AOR = 0.51, 95% CI: 0.24, 0.78) (Fig. 5).

Of the total included studies, three studies examined the association between knowledge of hypothermia and early newborn bathing practice. The overall estimate showed that there was no statistically significant association between knowledge of hypothermia and early newborn bathing practice [OR = 2.33, 95%CI: -0.42–5.08] (Supplementary file 2). Similarly, three studies examined the association between antenatal care visits and early newborn bathing practices. The overall estimate showed that there was no statistically significant association between having antenatal care visits and early newborn bathing practice [OR = 1.53, 95%CI: 0.34–2.71] (Supplementary file 3).

Leave one out meta-analysis using the Der Simonian Liard method was done to assess the effect of a single study on the overall estimate and to identify outliers. As shown in the figure below when individual studies were removed and the overall prevalence re-estimated, there was no such significant deviation from the prior determined prevalence (Fig. 6). However, the confidence interval of individual studies overlaps each other.

The study's strength is that the publications were found in a variety of genuine databases, and most of them were recent publications. Another strength is that, to the best of the investigators' knowledge, it is the first SRM on the level of early newborn bathing practice and associated factors among postpartum women in Ethiopia. However, the majority of the included studies came from a few regions. During the article search, only the English language was taken into account, which may influence the overall estimate.