Selection and identification of articles
A total of 614 with (3413 participants) were retrieved through electronic databases and other sources. The title and abstract of every study was screened and duplicated or irrelevant articles were removed using EndNote Version7 software. Doing these, 146 articles were removed due to duplication and the remaining 468 articles, 460 were removed due to their title and abstracts was not in-line with our inclusion criteria( did not correctly report the correct outcome, studies conducted outside of Ethiopia). Finally, a total of only 8 articles were included for this systematic review and meta-analysis (Fig.1).
Description of the included studies
We included a total of 8 crossectional observational studies with 3413 participants in the final systematic review and meta-analysis. The smallest sample size was 233 in the study conducted from Southern Nations and nationalities region of Ethiopia[26]. Whereas, the largest sample size, was 824 from the study conducted in the Amhara region, Ethiopia. The detailed characteristics of the included studies were presented (Table1). In the current systematic review and meta-analysis, four regions and one administrative town in the country were included. Three studies were from Tigray [27-29], two were from Oromia region [23, 30].
Table1: Descriptive summary of eight studies included in the final systematic review and meta-analysis overall male involvement in birth preparedness and complication readiness in Ethiopia
Author
|
Year
|
Design
|
Region
|
Sample
|
r(pwo)
|
Proportion(p)
|
NOS
Score
|
Destaw et al.[24]
|
2014
|
Crossectional
|
Addis Ababa
|
403
|
264
|
65.5
|
5
|
Demessie et al.[23]
|
2016
|
Crossectional
|
Oromia
|
374
|
224
|
50.8
|
6
|
Tantu et al.[30]
|
2018
|
Crossectional
|
Oromia
|
421
|
127
|
30.2
|
7
|
Baraki et al. [31]
|
2019
|
Crossectional
|
Tigray
|
399
|
187
|
46.9
|
7
|
Weldearegay et al.[27]
|
2015
|
Crossectional
|
Tigray
|
376
|
227
|
60.4
|
6
|
Mersha et al.[32]
|
2016
|
Crossectional
|
Amhara
|
824
|
82
|
10
|
6
|
Gebrehiwot et al.[28]
|
2013
|
Crossectional
|
Tigray
|
376
|
101
|
26.9
|
7
|
Paulos et al.[26]
|
2020
|
Crossectional
|
SNNPR
|
233
|
72
|
30.9
|
6
|
Male involvement in birth preparedness and complication readiness
Overall involvement in birth preparedness and complication readiness
From a total of 8 studies [23, 24, 26-28, 30-32] crossectional observational studies, all studies provided information on overall involvement of males in birth preparedness and complication readiness. As stated in the forest plot (Fig.2), the pooled estimate of overall male involvement was 40.17% [95%CI (24.01-56.33)]. High heterogeneity was observed (I-squared= 99.2%), however there was no an evidence to show publication bias using Egger’s statistical test (p-value=0.258) (see additional file 1: Table S1). For reducing the potential random variations between the included studies, we conducted a sub-group analysis by region where the primary studies conducted. According to the sub-group analysis the highest proportion of male involvement was observed from Addis Ababa65.5% [95%CI (60.86-70.14)], followed by Tigray 44.71% [95%CI (25.29-64.13)] (Fig.3).
A Meta regression analysis was done using year of study and sample size as covariate to explore the possible source of heterogeneity (see additional file 1: Table S2). But, there was no significant statistical evidence that show the presence of heterogeneity, which explained by using these two variables
In addition, sensitivity analysis was conducted by using a random effect model to identify the effect of single study on overall pooled estimate. The analysis showed that there was no evidence of single study influence on overall pooled estimate (see additional file 2: Fig S1)
Male involvement in three main domains (Antenatal, delivery, and post-partum)
In this systematic review and meta-analysis, we would assess male involvement using three domains such as involvement during antenatal care (ANC), during delivery, and post-partum period. Accordingly, eight studies [23, 24, 26-28, 30-32] which reported antenatal care involvement were included in the pooled analysis of male involvement during ANC. In this meta-analysis, the proportions of pooled male involvement during antenatal care were found to be 41.14% [95% CI (23.68-58.60)] (Fig.4). According to the sub-group analysis the highest proportion of male involvement during ANC was observed from Addis Ababa 73.2% [95%CI (68.88-77.52)] followed by Tigray 44.71% [95%CI (25.29-64.13)] (Additional file2: Fig S4). Whereas, male involvement during delivery and post-partum period was found unfortunately similar with the pooled proportion of 45.56% [95%CI (28.82-62.13)] (Figs.5&6). High heterogeneity was observed (I-squared= 99.3%, p≤0.001), for reducing the potential random variations between the included studies, we conducted a sub-group analysis by region where the primary studies conducted (Additional file2: Figs S 5&6). We found Six studies [23, 27, 28, 30-32] that describe about male involvement during delivery and post-partum periods from three regions of Ethiopia (Oromia, Amhara, and Tigray). Similarly there was high heterogeneity for delivery and post-partum involvements (I-squared= 98.9%, p≤ 0.0001). Having this in mind we have done sub-group analysis by region and we found that the highest proportion of male involvement during delivery and post-partum period was reported from Tigray region of Ethiopia 51.09%[95%CI(23.55-78.64)] followed by 47.23%[95%CI(14.60-79.87)] in Oromia region of Ethiopia(Additional file2: Figs S12). We have also done the statistical tests of Egger’s and Begg’s to assess publication bias for the included studies in the Meta-analysis of male involvement during delivery and post-partum period (see additional file:Tables S3&S5). There was no evidence of bias (p-value>0.069) (see additional file:Table S5). We have performed meta-regression by using year of study and sample size as a covariate (see additional file:Table S6).
Factors associated with male involvement in birth preparedness and complication readiness (see Additional File 3: Figures S1-S6)
Among the variables included in the meta-analysis husband’s level of education was found to be statistically significant in three studies [23, 30, 32]. However, the pooled odds ratio found to be border-line statistically significant with [OR=1.27, 95% CI (0.26-6.30)] (see Additional File 3: Figure S1). Distance from health facilities was also statistically significant in three primary studies [26, 30, 32]. But, it was also border-line significant [OR=0.75, 95%CI (0.10-5.83)] (see Additional File 3: Figure S2). Moreover, husband’s knowledge of at least one danger sign during pregnancy was statistically significant in four included primary articles [23, 27, 30, 32]. The pooled odds ratio was found to be border line significant [OR=3.18, 95%CI [(0.54-18.66)] with male involvement for birth preparedness and complication readiness (see Additional File 3: Figure S3). The husband’s knowledge of danger signs during delivery was also found to be significant in two primary studies[27, 32]. However, the pooled odds ratio was found to be borderline significant [OR=4.60, 95%CI (0.06-375.06)] predictors of male involvement in birth preparedness and complication readiness (see Additional File 3: Figure S4). Similarly husband’s knowledge about post-partum danger signs was found to be [OR=2.50, 95%CI (0.06-104.65)] with involvement in birth preparedness and complication readiness (see Additional File 3: Figure S5). Husband’s attendance of at least one ANC visit with their wives was found to be [OR=3.20, 95% CI (1.97-5.19)] statistically significant predictor of male involvement (see Additional File 3: Figure S6).