Selection of the studies

All articles related to prevalence and determinant of compliance with IFAS were collected from different sources. Since the year of publication for each articles were not limited, all articles published up to February 25, 2020 were incorporated for their eligibility in the review. Then, quantitative cross-sectional study design was used to make the work more clear and meaningful. However, articles published in qualitative methods were excluded because of the nature of the review and analysis chosen to be used in this paper. Then to have a deep understanding of each article, all authors read the title and abstract part independently. To avoid biases, all eligible articles were screened and selected after all individual’s full reading of the abstract section of each article. Then the divergence of the work was managed to enhance the reliability and validity of the review and analysis based on pre-set inclusion criteria.

Data extraction and quality assessment

In the process of making any review and analysis, the participation of more individuals has a significant value in data extraction. This is because of the principle of one plus one is greater than one. That is to mean scrutinized and condensed extracted data was obtained from more individual if they are participated equally on the extraction of a given raw data. To obey this rule and make the data more readable and meaningful, all authors were engaged in extraction. To do so data extraction template, which included author’s name, year of publication, study location, sample size, odds ratio, confidence intervals and P-value, were prepared before the extraction of data was carried out. However, after the extraction of the data by each individual independently, we made a cross checked and compared the results very carefully. To make the work to address the target objective, all of us discussed and came to consensus on little partiality observed between us during the work. Thus, Modified Newcastle-Ottawa quality assessment scale for cross-sectional studies was used to assess the quality of the studies in terms of its inclusion. The total score for the modified Newcastle–Ottawa scale for cross-sectional studies used was nine (9) stars as a maximum for the overall scale with the minimum of zero, and a study was considered to be a high quality if 7 was achieved from 9 and medium if 5 was achieved from 9 [46].

Operational definitions

Statistically analysis

The extracted data was copied to Microsoft excel to be exported to review manager version 5.3 and the compressive meta-analysis version 2 software for careful analysis. Accordingly, statistical description related to IFAS and its determinants were performed. To evaluate the existence of statistical heterogeneity, the Publications bias was tested by funnel plot and empirically through Egger’s regression test. Moreover, the degree of trustworthiness was contemplated. Consequently, the heterogeneity of studies was computed using the I-squared statistic. In this process, 25% was signified as low, 50% moderate and 75% as high heterogeneity score. Also subgroup analysis was executed by the study sub-region and study type (Community based and / or facility based). Therefore, the effect of particular predictor’s variables which consist of: Counseling on IFAS, knowledge of IFA supplementation, knowledge of anemia, fourth visit for ANC and early registration for ANC were estimated and the result of the Meta-analysis revealed forest pilot and Odd Ratio (OR) with 95% of CI.

Studies selection

Based on the objectives set for this work, we identified different studies related to the prevalence and determinants of IFA supplementation for the inclusion in meta-analysis before directly move to the other detail part of this paper. Accordingly, we found 1156 completed studies published on international journals. From these, 1113 of them were excluded for they were not satisfying the criterion of inclusion set in this study. However, 43 articles were scrutinized and carefully chosen from those studies for their eligibility. Out of these 20 studies were rejected due to their poor statistical reports and defect of data observed in each of them. Finally, only 20 studies were added in this analysis for their neatness and clear justification (Fig 1).

Fig 1

Flow diagram of the studies included in the meta-analysis.

Characteristics of included studies

Twenty-three Cross-Sectional studies from different countries of Africa were included in the meta-analysis [23–45] (Table 1). Out of those four of them (17.4%) were from Kenya, two (8.7%) from Malawi, two (8.7%) from Nigeria, one (4.7%) from Mozambique, one (4.3%) from Senegal, one (4.3%) from South Africa, eleven (47.8%) from Ethiopia and one (4.3%) from Uganda. Among those the highest sample size was observed in studies conducted in Malawi [23] which was equal 10750 and the lowest was found in South Africa, 57 [39]. The mean age of the respondents was 27 years. Out of twenty-three studies incorporated in this review and analysis, though eighteen studies [25, 27–33, 35–38, 40–45] were conducted on facility based, five of them [23, 24, 26, 34, 39] were done on community based work. However, all studies considered definition of good adherence toward IFA as a pregnant woman who received ≥65% of the total recommended IFA supplementation per month, but the reverse of this is true for non-adherence types [1]. Lastly, (Table 2) displays assessment of quality of studies included in systematic review and Meta-analysis.

Table 1

Characteristics of studies included in systematic review of prevalence compliance with IFAS among pregnant women in Sub-Saharan Africa.

Author	Region	Study design	Study type	Sample Size	Prevalence in %	Compliance IFAS	No compliance IFAS
Abinet Arega Sadore et al, 2015	Ethiopia	Cross sectional	Community based	296	39.2	116	180
Titilayo A. et al, 2016	Malawi	Cross sectional	Community based	10750	37.2	3999	6751
Chikakuda A. et al, 2018	Malawi	Cross sectional	Facility based	213	18.3	39	174
Bekele Taye et al, 2015	Ethiopia	Cross sectional	Community based	628	20.4	128	500
BI Nwaru et al, 2014	Mozambique	Cross sectional	Facility based	4326	79	3418	908
Demis et al, 2019	Ethiopia	Cross sectional	Facility based	422	43.1	182	240
Agegnehu G. et al, 2018	Ethiopia	Cross sectional	Facility based	418	28.7	120	298
Gebremariam et al, 2019	Ethiopia	Cross sectional	Facility based	241	40	96	145
Getachew et al, 2018	Ethiopia	Cross sectional	Facility based	320	64.7	207	113
Dessie G. et al, 2018	Ethiopia	Cross sectional	Facility based	348	19	66	282
Juma M et al, 2015	Kenya	Cross sectional	Facility based	352	18.3	64	288
K. Niang et al, 2017	Senegal	Cross sectional	Community based	1442	51	735	707
Kamau et al, 2018	Kenya	Cross sectional	Facility based	364	33.7	123	241
Kiwanuka et al, 2017	Uganda	Cross sectional	Facility based	370	11.6	43	327
Lucy Nyandia Gathigi, 2011	Kenya	Cross sectional	Facility based	264	10.6	28	236
LYNETTE AOKO DINGA, 2013	Kenya	Cross sectional	Facility based	200	24.5	49	151
Mbhenyane et al, 2017	South Africa	Cross sectional	Community based	57	90	51	6
Molla et al, 2019	Ethiopia	Cross sectional	Facility based	348	52.9	184	164
Niguse and Murugan, 2018	Ethiopia	Cross sectional	Facility based	296	59.8	177	119
Onyeneho et al, 2016	Nigeria	Cross sectional	Facility based	1500	33	495	1005
Shewasinad S, et al, 2017	Ethiopia	Cross sectional	Facility based	326	70.6	230	96
Tarekegn et al, 2019	Ethiopia	Cross sectional	Facility based	395	28.0	111	284
Ugwu, et al, 2012	Nigeria	Cross sectional	Facility based	396	65.9	261	135

Table 2

Assessment of quality of studies included in systematic review of prevalence compliance with IFAS among pregnant women in Sub-Saharan Africa.

Studies	Representativeness of the sample			Sample size			Non-respondents	Ascertainment of the exposure (risk factor)			Confounding factors are controlled		Ass. of the outcome	Statistical test				Total
	All subjects or random sampling*	Non random sampling*	No description of sampling strategy	Justified and satisfactory*	Not Justified	Satisfactory*	Unsatisfactory	Validated measurement tool**	non Validated measure or tool is available*	No description of Validated measurement tool	Study controls important factor*	Study controls Additional factor*	Independent blind assessment**	Record linkage**	Self-report*	No description	statistical test used to analyze the data is clearly described*
Abinet Arega Sadore et al, 2015	*			*		*		*					**					6
Titilayo A.et al, 2016	*			*		*		*			*		**					7
Chikakuda A. et al, 2018	*			*		*		**			*		**					8
Bekele Taye et al, 2015	*			*		*		**					*				*	7
BI Nwaru et al, 2014	*			*				**			*		*				*	7
Demis et al, 2019	*			*		*		**			*		**				*	9
Agegnehu G. et al, 2018	*			*				**			*		**				*	8
Gebremariam et al, 2019	*			*		*		*			*		**				*	8
Getachew et al, 2018	*			*		*		*			*		*				*	7
Dessie G. et al, 2018	*			*		*		*			*		**				*	8
Juma M et al, 2015	*					*		*			*		**				*	7
K. Niang et al, 2017	*			*		*		**			*		**				*	9
Kamau et al, 2018	*			*		*		**			*		**				*	9
Kiwanuka et al, 2017	*			*		*		**			*		**					8
Lucy Nyandia Gathigi, 2011	*			*		*		**			*		**				*	9
LYNETTE AOKO DINGA, 2013	*			*				**			*		*				*	7
Mbhenyane et al, 2017	*			*		*		**			*		**					8
Molla et al, 2019	*			*		*		**			*		*				*	8
Niguse and Murugan, 2018	*			*		*		**			*		**				*	9
Onyeneho et al, 2016	*					*		*			*		**				*	6
Shewasinad S, et al, 2017	*			*		*		**			*		**				*	9
Tarekegn et al, 2019	*			*		*		**			*		**					8
Ugwu, et al, 2012	*			*		*		**			*		**				*	9

Compliance with Iron and Folic-Acid Supplementation

From the analysis made, we can understand that the lowest prevalence of compliance (10.6%) was observed in Kenya [38], but the highest compliance, which is equal to (79%) was observed in study conducted in Mozambique [27]. On the other hand, the pooled compliance with Iron and Folic-Acid Supplementation (IFAS) amongst pregnant women in SSA was 39.2% (95% CI = 0.308–0.483) (Fig 2). Moreover, the heterogeneity test displayed is equal to I² = 99.27% and the statistical suggestion for heterogeneity is P<0.000). Hence, the random-effect analysis was the secondhand of the analysis. Therefore, the Bag’s and Egger’s test for publication bias indicated that there was no statistical suggestion for publication bias. That is p-value = 0.428 and 0.575 respectively (See S1 Fig).

Fig 2

Forest plot displaying association of counselling on Iron and Folic Acid Supplementation with compliance with IFAS among pregnant women in Sub-Saharan Africa.

Subgroup analysis

A subgroup analysis was done by classifying studies based on corresponding sub-regional location in Sub-Saharan Africa in order to compute and relate the prevalence of compliance with IFA supplementation focusing on athwart various participants’ characteristics. Based on this, the lowest prevalence of compliance with IFA supplementation in pregnancy was documented in Eastern Africa, (34.2%) (CI: 0.264, 0.430) and the highest prevalence of compliance with IFA supplementation was recognized in Western Sub-Saharan, (49.3% (CI: 0.203, 0.788). However, a greater prevalence of compliance, which is equal to (44.3%%) was detected in studies conducted at facility level than community level (CI: 0.337, 0.554) (Table 3).

Table 3

Subgroup analysis of prevalence compliance with IFAS among pregnant women in Sub-Saharan Africa.

Subgroup	No. of included studies	Prevalence(95%CI)	Heterogeneity Statistics	Tau Squared	P value	I2
By Sub- region
Eastern Africa	17	34.2(0.264,0.430)	787.551	0.594	<0.000	97.968
Southern Africa	4	58.3(0.277,0.836)	1995.397	1.693	<0.000	99.850
Western Africa	2	49.3(0.203,0.788)	131.117	6.927	<0.000	99.927
Overall					<0.000
By study type
Facility based	18	37(0.256,.0.501)	2333.651	0.696	<0.000	99.272
Community based	5	44.3(0.337,0.554)	218.345	0.230	<0.000	98.16

Counseling on Iron and Folic Acid Supplementation

Out of twenty-three chosen studies conducted on the area of the key concern and included in the meta-analysis, in nine of them [25, 28, 30, 31, 34, 35, 37, 38, 45] it was documented that counseling on IFA supplementation was associated with compliance to IFA supplementation in pregnancy. Moreover, the result from meta-analysis also revealed that pregnant women who had received counseling on IFAS during pregnancy were 1.96 times more likely than those who had adherence to IFAS as compared to those who had not received counseling, [OR: 1.96, 95% CI (1.76-, 5.93)]. Thus, the heterogeneity test revealed I² = 92% and the statistical evidence of this is P<0.00001). From this we can understand that the random-effect analysis was the secondhand one. Thus, the Bag’s and Egger’s test for publication bias indicated that there is no statistical evidence of Publication bias. That is their p-value = 0.284 and = 0.754 respectively (Fig 3).

Fig 3

Forest plot displaying association of counseling on Iron and Folic Acid Supplementation with compliance with IFAS among pregnant women in Sub-Saharan Africa.

Knowledge of Iron Folic Acid Supplementation (IFAS)

The association of lack of knowledge on Iron Folic-Acid Supplementation and risk of developing noncompliance to IFAS during pregnancy was stated in nine chosen studies [25, 26, 28, 30, 32, 35, 37, 38, 40]. Thus, the result of meta-analysis from those exhibited that women who had knowledge of IFA supplementation were 2.71 times more likely to have compliance to IFA Supplementation compared to those who had no knowledge of IFA Supplementation [OR: 2.71, 95% CI (1.33, 5.54)]. Therefore, the heterogeneity test indicated (I2 = 94%) and statistical evidence of this heterogeneity was P<0.0000). Hence, the random- analysis was the secondhand one again. Finally, the Bag’s and Egger’s test for publication bias indicated the absence of statistical evidence of Publication bias. That is their p-values are equal to 0.465 and 0.786 respectively (Fig 4).

Fig 4

Forest plot displaying association of knowledge of Iron Folic-Acid Supplementation (IFAS) with compliance with IFAS among pregnant women in Sub-Saharan Africa.

Knowledge of anemia during pregnancy

The important analysis was focus on the association of lack of knowledge on Iron Folic-Acid Supplementation and risk of developing noncompliance to IFAS during pregnancy. This was stated in seven different studies [25, 26, 31, 37, 38, 40, 45]. The result of meta-analysis from those paper showed that women who had knowledge about anemia were 5.42 times more likely to have compliance to IFAS in their course of pregnancy as compared to those who had no knowledge of anemia, [OR5.42, 95% CI (1.52, 19.43)]. The heterogeneity test indicated that (I² is equal to 97% and the statistical evidence of this is P<0.0000). Hence, the random-effect analysis was secondhand one, but the Bag’s and Egger’s test for publication bias indicated that there is no statistical evidence of Publication bias, which is equal to p-value = 0.283 and 0.682 respectively (Fig 5).

Fig 5

Forest plot displaying association of knowledge of Anaemia during pregnancy with compliance with IFAS among pregnant women in Sub-Saharan Africa.

Fourth visit for Antenatal Care (ANC)

The association between lack of fourth visit for antenatal care and risk of developing noncompliance to IFAS during pregnancy was stated in eight studies [25, 28, 30, 31, 36, 37, 40, 43]. Accordingly, the result from meta-analysis from those revealed that women who had fourth visit for ANC in pregnancy were 1.54 times more likely to have compliance to IFAS during pregnancy as compared to those had no fourth visit for ANC [OR 1.54, 95% CI (0.66,3.58.43)]. However, the heterogeneity test indicated I² = 93% and statistical evidence of this was <0.00001). Thus, random-effect analysis was the secondhand one. On the other hand, the Bag’s and Egger’s test for publication bias indicated that there was no statistical evidence of Publication bias. That is their p-value is equal to 0.980 and 0.982 consecutively (Fig 6).

Fig 6

Forest plot displaying association of Fourth visit for antenatal care (ANC) with compliance with IFAS among pregnant women in Sub-Saharan Africa.

Strength and limitation

In this review, an extensive exploration method and more than one reviewer had taken part in all courses of review process. To do so, PRISMA guideline was carefully tracked throughout the review procedure. However, the analysis has its own defects because of a number of factors. For instance, firstly compliance to IFA supplementation was defined in various studies in different way. Secondly, study lacks representativeness since no data was found from some of Sub-Saharan African counties. Thirdly, some studies have been omitted due to their poor statistical reports, their small sample size, their inadequate data and their qualitative nature of the studies. Finally, since only the cross sectional studies were involved in the analysis, the outcome variable may possibly be affected by confounding variable. Therefore, those limitations might affect the overall prevalence of compliance to IFA supplementation in pregnancy in Sub-Saharan Africa.