Why maternal continuum of care remains low in Northwest Ethiopia? A multilevel logistic regression analysis.

BACKGROUND: Non-adherence to the maternal continuum of care remains a significant challenge. Though early initiation and continuum of care are recommended for mothers' and newborns' well-being, there is a paucity of evidence that clarify this condition in resource-limited settings. This study aimed to assess the level of women's completion of the maternal continuum of care and factors affecting it in Northwest Ethiopia.
METHODS: A community-based cross-sectional study was conducted from October to November, 2020. Data were collected from 811 women who had a recent history of birth within the past one year. The random and fixed effects were reported using an adjusted odds ratio with a 95% confidence interval. The p-value of 0.05 was used to declare significantly associated factors with women's completion of the maternal continuum of care.
RESULTS: The study revealed that 6.9% (95%CI: 5.3-8.9%) of women were retained fully on the continuum of maternal care, while 7.89% of women did not receive any care from the existing healthcare system. Attending secondary and above education (AOR = 3.15; 95%CI: 1.25,7.89), membership in the women's development army (AOR = 2.91; 95%CI: 1.56,5.44); being insured (AOR = 2.59; 95%CI: 1.33,5.01); getting health education (AOR = 2.44; 95%CI: 1.33,4.45); short distance to health facility (AOR = 4.81; 95%CI: 1.55,14.95); and mass-media exposure (AOR = 2.39; 95%CI: 1.11,5.15) were significantly associated with maternal continuum of care.
CONCLUSIONS: The maternal continuum of care is low in rural northwest Ethiopia compared to findings from most resource-limited settings. Therefore, the existing health system should consider multilevel intervention strategies that focus on providing maternal health education, facilitating insurance mechanisms, encouraging women's participation in health clubs, and ensuring physical accessibility to healthcare facilities to be more effective in improving maternal health services.

Introduction

Every day, about 810 women die due to pregnancy and childbirth-related complications across the world [1]. Annually, 295,000 women die as a result of pregnancy and childbirth-related complications, of which 94% occur in low- and middle-income countries. Statistics, according to WHO 2019 report on trends in maternal mortality estimates from 2000–2017, show that the probability that a 15-year-old woman will die eventually from a maternal cause is estimated as 1 in every 90 women globally, while 1 in every 5400 and 45 in developed and developing countries, respectively [1]. Likewise, the life time risk of maternal death is recorded as 1 in every 37 in Sub-Saharan Africa [1]. Almost all pregnancy-related deaths are associated with complications during pregnancy and childbirth, and nearly 75% of these deaths could be avoided if women were kept on the maternal continuum of care [2]. According to the 2019 report, from global figures, approximately 14,000 maternal deaths and 1/55 lifetime risk of maternal mortality were reported in Ethiopia [1]. Despite recent progress, Ethiopia has continued to suffer from an unacceptable number of maternal deaths [3].

Improving maternal health has been a global public health priority [4, 5]. The maternal continuum of care through pregnancy to postpartum is a rallying call to make a woman’s life safe and joyful [6]. It is a proven intervention for maternal mortality reduction and is efficient for making a woman more productive in her life [7], as it links a woman to the maternal healthcare services provision centers such as home, community, and health facilities [6]. The World Health Organization (WHO) suggested at least four Antenatal Care (ANC) visits and one or more postnatal visits within two days of delivery to improve maternal health and birth outcome [8, 9].

ANC is a vital component of the maternal continuum of care for women and is believed to be the basis for women and newborns to survive and thrive [10, 11]. Early initiation and proper ANC follow-up reduce adverse pregnancy outcomes in both mother and baby since it promotes early identification of pregnancy-related complications and provides an opportunity for a woman to discuss with a healthcare provider about pregnancy and health behavior [12]. Healthcare facility delivery with the assistance of a Skilled Birth Attendant (SBA) is another strategy for reducing maternal morbidity and mortality [9], as births carried out in homes under unhygienic conditions could lead to a high incidence of maternal death [13]. Receiving Postnatal Care (PNC) from a healthcare provider at the recommended time with its appropriate content also prevents complications that could arise after childbirth and helps a mother to get comprehensive health education [9].

With regard to the health sector’s mission in Ethiopia, a three-tiered healthcare delivery system has been put in place to facilitate the realization of the health system’s goals; at the primary level of the healthcare delivery system, Health Extension Workers (HEWs) are designated for consultation and service provision in the areas of family planning, ANC, PNC, and child health services [14]. Moreover, the government’s initiatives and commitment towards the maternal continuum of care indicate there are identified intervention areas to strengthen maternal health programs [15]. Despite the improvement in maternal healthcare utilization, however, completion of the maternal continuum of care is very low [16] and improving maternal health remains a major challenge for the health system in Ethiopia [3].

According to past studies, nearly four out of every ten women did not receive ANC; seventeen percent of women in Ethiopia had their first prenatal care visit before the fourth month of pregnancy [17] with 56% of women giving birth in a health facility after receiving four or more ANC visits, and 17% received postnatal visits within 48 hours after birth [9].

According to the body of literature, both individual and community-level factors are influencing maternal healthcare utilization [18-20]. Though there was promising progress in maternal healthcare utilization, and it is a prioritized agenda in Ethiopia [21], the available evidence on women’s completion of the maternal continuum of care is inconclusive. Understanding the existing point of women’s completion of the maternal continuum of care and identifying contextual factors that limit women’s completion of the maternity care continuum helps to be effective in the implementation of maternal health programs. Considering the hierarchical structure of women with a child less than one year (level 1) nested within a health center (level 2), it is appropriate to assess the point of women’s retention on the maternal continuum of care. Therefore, this study aimed to determine the level of women’s completion of maternal continuum of care and identify individual and community-level factors affecting it in northwest Ethiopia.

Methods

Study design and setting

A community-based cross-sectional study was conducted in 11 clusters in Wogera and Gondar Zuriya districts in the central Gondar zone of Amhara National Regional State from October to November 2020. The districts are located 658km from Addis Ababa, the capital city of Ethiopia. There were 16 health centers and 88 health posts in the districts. The total population was 524, 907 (female = 260879 and male = 264028) at the time of the survey, of which 122,303 women in the reproductive age group (15–49) and 16745 were surviving infants (Central Gondar Zone Health Bureau report-unpublished).

Study population

All women in the study districts who gave birth within the past one year prior to data collection were the source population, and randomly selected women who were permanent residents of the selected kebeles and willing to participate in the study were the study population.

Sample size determination

For the single population proportion formula, the assumptions used were proportion of focused ANC p = 39.9%, proportion of SBA after completing at least 4 ANC visits p = 31.1%, and proportion of women retained in the continuum of maternal care p = 12.1% [22]. This formula considers 95% CI, margin of error = 5%, design effect of 2 and 10% non-response rate, the sample sizes were 811,724, and 360, respectively. Double population proportion formulas were considered in estimating different sample sizes, considering factors affecting maternal health service utilization. The sample size was computed using the STAT-CALC program of Epi-info version 7.0. software using the following assumptions: 5% level of significance (two-sided), 80% power, a 1:1 ratio of exposed to non-exposed maternal health service utilization, 49.7% of the outcome in the exposed group and 50.3% of the outcome in the unexposed group, and considering being a model household in the community as an exposure variable for good maternal health service utilization [15], a design effect of 2, and a 10% non-response rate, n = 559. Of the different sample sizes estimated, the largest sample size, 811, was obtained from a single population formula and considered in the study.

Sampling procedure

A two-stage sampling technique was employed to reach study participants. Since the districts have fifteen health centers, eleven health centers (clusters), kebeles at each cluster were selected randomly with systematic sampling. Households were selected from a sampling frame developed using a family folder of HEWs, which contained a list of all births for rural women. The sampling interval was obtained by dividing the source population by the total sample size estimated. The first woman was chosen from a sampling frame in the family folder using a simple random sampling technique, and the random selection was carried out using the random between functionality in Microsoft Excel. Using this technique, the required sample size was obtained, and a total of 811 eligible women were interviewed for the study and included in the analysis.

Study variables and measurement

Response variable

Women’s completion of the maternal continuum of care is categorized as "yes" if the women received at least 4 ANC visits, SBA, and PNC visits within 48 hours after birth, and "no" if otherwise.

Explanatory variables

Ranges of predictor variables were selected and grouped as individual and community level variables based on previous literature.

Individual level factors

Individual level variables were age of the respondents, marital status of the respondents, occupation, education status of respondents, education status of their partners, pregnancy intention, distance from home to facility, household wealth index, being a member of community-based health insurance, mass-media exposure, HEWs’ home visits during pregnancy, being a member of Women’s Development Army (WDA), and getting health education from a nearby health post during pregnancy.

Community level factors

Community and cluster level variables were aggregated from individual and cluster level data such as health facility readiness, the average distance from their village to a health facility, wealth status in the cluster, and mass-media exposure. Health facility readiness for service provision was assessed using the WHO service availability and readiness assessment guideline [23]. Aggregated clusters were categorized as low if the proportion of clusters’ readiness for service provision was 0–72% whereas high if the proportion was 73–100%. The household wealth index median value was three, and the community value was classified as low if the median value of the community is below three and high if otherwise (middle and above). The mass-media exposure status of the community was categorized as low if the proportion of exposure status was 0–17%, and whereas high if the proportion of mass media exposure status in the community was 18–100%. Besides, the walking distance from home to health facility was near if the aggregated proportion was 0–71%, and far if the aggregated proportion of walking distance was 72–100%.

Data collection instrument and procedures

The instrument was developed in English and translated into the local language (Amharic). Since experts’ views were sought for the psychometric properties of the face and content validity, experts with health management system expertise, health informatics professionals, midwifery professionals, gynecologist and obstetrician, and emergency officers at obstetrics and gynecology were invited to review the relevance of each question in the instrument. The tool was revised according to the experts’ views and was piloted out of the selected clusters and validated before data collection. The proportion of the rated item and scale content validity index was computed, and the experts’ rated item content validity index and scale level validity index were 0.98 and 0.81, respectively. Data collectors and supervisors were trained on the instrument. The household data were collected using an interviewer-administered questionnaire via face-to-face interviews.

Operational definition

Maternal continuum of care

It is the continuity of care throughout pregnancy (utilization of at least 4 ANC visits, SBA, and PNC within 48 hours after birth) [6].

Data management and statistical analysis

The data were entered into Epidata and analyzed with Sata14. For wealth index assessment, principal component analysis was employed to reduce data that measure the same construct together, and the data were recoded into binary variables. Variables with low frequencies were combined, and frequencies greater than 95% and less than 5% were excluded from the analysis. Under the component loading, there were twenty-one variables correlated with measuring the wealth index, such as having one’s own house, toilet, electricity, kitchen, charcoal, television, radio, modern bed, cotton, mobile phone, farming land; having animals in the house, such as cows, horse, donkey, goat, sheep, chickens; having a separate room for animals; having a beehive, and having a bank account number with deposited Birr (Ethiopian currency). The wealth quantile was categorized into five categories, such as poorest, poor, middle, richer, and richest.

Bivariable multilevel logistic regression was used for each factor against the maternal continuum of care without controlling the effect of other explanatory variables. For multivariable multilevel logistic regression analysis, factors at both individual and cluster levels with a p-value <0.25 were considered candidate variables. The fixed-effect of individual and cluster level variables was reported using the Crude Odds Ratio (COR) with a 95% confidence interval.

During model building, four models were built to estimate both the fixed effects of the individual and community-level factors and the random effects between-community variation on the women’s completion of the maternal continuum of care. Model I was built without any explanatory predictors to examine the random effect of cluster variation by using Intra-Class Correlation (ICC) to justify the application of multilevel analysis in this study. Therefore, random parameters in this model were used as a benchmark to compare parameters of successive models (Model II, Model III, and Model IV) by looking at the decline of the ICC value [24]. Model II was built to examine the contributions of individual-level factors. Model III was fitted to determine between-cluster variations. Finally, Model IV (individual and cluster-level factors model) was built by combining both individual and cluster-level factors simultaneously by controlling for the effect of other predictors. The measure of association was reported as Adjusted Odds Ratio (AOR) with a 95% CI. The p-value < 0.05 was used to identify factors significantly associated with women’s completion of the maternal continuum of care. The random effect was presented using ICC; thus, in all models, ICC and its change from the null model were examined. The Proportional Change in Variance (PCV) was computed with reference to the null model to examine the relative contribution of level II, III, and IV predictors in explaining the odds of women completing the maternal continuum of care.

To estimate the goodness-of-fit of the adjusted final model, the Akaike information criteria (AIC) and loglikelihood model were used in comparison with other models. The multicollinearity effect was checked by using the mean of the variation inflation factor (VIF) value at a cut-off point of 10, and it indicated that there was no multicollinearity effect among predictor variables [25]. The interaction effect of the variables was checked by creating a new variable, and the created new variable, or product term, became either statistically significant or not at p-value<0.05.

Ethical considerations

Verbal consent was obtained and participants were informed about the objective, importance of the study, procedure and duration, risk and discomfort, benefits of participating in the study, confidentiality, and the right to refuse or withdraw during data collection. Study approval and ethical clearance were obtained from the University of Gondar ethical review board (R.NO. V/P/RCS/05/2020). A formal letter of approval was taken from Amhara national regional health state bureau and central Gondar zonal health department. For participants age <18 verbal informed consent was taken from their parents and assent obtained from the minor/participant. And it was approved by the ethical review committee of the institute of public health on behalf of IRB of University of Gondar. After obtaining the relevant information, participants were counselled on the benefits of attending maternal health care services and the consequences of missing maternal health care services. The COVID-19 protocol was maintained throughout the study.

Results

Sociodemographic and reproductive characteristics of the study participants

A total of 811 women participated in the study, with a response rate of 100%. The minimum and maximum ages of the respondents were 15 years and 48 years, with a mean age of 28 years. The majority of the participants were married, 790(97.4%), and housewives, 801(98.8%). About two-thirds, 507(61.7%), of participants did not attend education. The proportion of women who intended the current pregnancy was 642 (79.2%). The proportion of women was nearly equal across the wealth quantiles at household level, with 163(20.1%) the poorest, 162(20.0%) poorer, 163 (20.1%) middle, 161(19.8%) richer, and 162(20.0%) the richest Table 1.

Table 1

Sociodemographic and reproductive characteristics of study participants in Northwest Ethiopia, 2020 (n = 811).

Variables	Category	Maternal continuum of care		(p-value)
		Yes(%)	No(%)
Age group of respondents (years)	15–24	16(7.0)	211(93.0)	0.935
	25–34	28(7.1)	366 (92.9)
	35 and above	12(6.3)	178(93.7)
Marital status	Married	54(6.8)	736(93.2)	0.632
	Single/divorced/widowed /separated	2(9.5)	19(90.5)
Occupation	Housewife	55(6.9)	746(93.1)	0.698
	Employee/laborer/merchant	1(10.0)	9(90.0)
Education level of respondents	Did not attend education	23(4.6)	477(95.4)	<0.001
	Primary education	17(7.9)	198(92.1)
	Secondary and above	16(16.7)	80(83.3)
Education level of husbands	Did not attend education	27(5.2)	496(94.8)	<0.001
	Primary education	15(7.3)	190(92.7)
	Secondary and above	14(16.9)	69(83.1)
Pregnancy intention	Yes	49(7.6)	593(92.4)	0.111
	No	7(4.1)	162(95.9)
Distance home to health facility	>5km	23(9.8)	212(90.2)	0.039
	< = 5km	33(5.7)	543(94.3)
Household wealth index	Poorest	11(6.7)	152(93.3)	0.446
	Poorer	6(3.7)	156(96.3)
	Middle	13(8.0)	150(92.0)
	Richer	12(7.5)	149(92.5)
	Richest	14(8.6)	148(91.4)
Being insured	Yes	41(10.3)	353(89.6)	< .000
	No	15(3.6)	402(96.4)
Membership in WDA	Yes	35(13.5)	225(86.5)	< .000
	No	21(3.8)	530(96.2)
HEWs home visiting	Yes	31(8.6)	330(91.4)	0.091
	No	25(5.6)	425(94.4)
Mass media exposure	Yes	15(10.7)	125(89.3)	0.051
	No	41(6.1)	630(93.9)
Getting health education at health post level	Yes	29(9.2)	285(90.8)	0.037
	No	27(5.4)	470(94.6)

Being a member of community-based health insurance and being a member of a WDA in the kebeles was 394 (48.6%) and 260 (32.1%), respectively. Study participants who were visited by HEWs during the current pregnancy were 361 (44.5%) and those who got health education from a nearby health post were 314 (38.7%) Table 1.

Community level characteristics of the study participants

Women’s retention in the maternal continuum of care was significantly associated with clusters (health centers) (chi-square = 28.19, p-value = 0.001). In this study, the wealth index at the community level indicated that 611 (75.3%) of the participants had high wealth status. The level of health facility readiness for ANC service provision was 619 (76.3%). The proportion of study participants who had high media exposure at the community level was 372 (45.9%). Among the participants, most respondents, 615(75.8%), had a walking distance of less than or equal to five kilometers from their home to a health facility. Bivariable multilevel logistic regression analysis indicated that media exposure in the community and distance from home to health facility were significantly associated factors with the maternal continuum of care Table 2.

Table 2

Community level characteristics of respondents in Northwest Ethiopia, 2020 (n = 811).

Variables	Category	Maternal continuum of care		(p-value)
		Yes(%)	No(%)
Health facilities	Ambageorgis	16(15.2)	89(84.8)	0.001
	Gedebiye	11(11.6)	84(88.4)
	Birra	3(5.0)	57(95.0)
	Tirgosgie	6(9.0)	61(91.0)
	Woybey	5(10.9)	41(89.1)
	Dergaj	3(7.1)	39(92.9)
	Miniziro	0(0.0)	104(100.0)
	Maksegnit	3(3.5)	82(96.5)
	Lamba	2(2.5)	78(97.5)
	Enfranz	4(4.3)	88(95.7)
	Abawarka	3(8.6)	32(91.4)
Wealth Index	High	45(8.2)	506(91.8)	0.039
	Low	11(4.2)	249(95.8)
Health facility readiness	High	48(7.8)	571(92.2)	0.087
	Low	8(4.2)	184(95.8)
Mass-media exposure	High	34(9.1)	338(90.9)	0.021
	Low	22(5.0)	417(95.0)
Walking distance from home to health facility	>5km	4(2.0)	192(98.0)	0.002
	< = 5km	52(8.5)	563(91.5)

Maternal continuum of care

In this study, women who attended at least 4 ANC visits, SBA, and PNC visits within 48 hours after birth were considered during the analysis. Any ANC visit: Women who received at least one antenatal care visit from the existing healthcare system. At least 4 ANC visits: Women who received at least 4 ANC visits. Retention on SBA: This is the continuity of care from at least 4 ANC visits to skilled delivery. Retention on PNC: This is about the continuity of care from ANC to SBA and PNC within the first 48 hours after delivery. The outcome variable was “yes” if a woman received ANC to SBA and PNC within the first 48 hours after birth and “no” if otherwise. Accordingly, the rates of any ANC visit during pregnancy, at least 4 ANC visits, and SBA after ANC4+ completion were 86.8% (95% CI: 84.3–89.1), 39.6% (95% CI: 36.2–43), and 29.1% (95% CI: 26–32.4), respectively. The proportion of women who were retained fully on the maternal continuum of care during pregnancy was 6.9% (95%CI: 5.3–8.9) Fig 1.

Fig 1

Maternal health service utilization in completion of maternal continuum of care from ANC to PNC care within 48hrs after birth in Northwest Ethiopia, 2020.

Multivariable multilevel logistic regression analysis of maternal continuum of care

During analysis, four mixed-effect regression models were built to predict the maternal continuum of care based on the different exposure variables at individual and community levels Table 3. The intercept-only model was run without any predictor to test the random effect between cluster variation on the maternal continuum of care. An estimate of ICC was 12% (95% CI: 3%, 35%), implying that 12% of the variation in the maternal continuum of care was due to cluster-level factors and 43.5% was due to differences across clusters, and this variation was significant (τ = 0.43, p 0.000).

Table 3

Multivariable multilevel logistic regression analysis of predictors of women’s completion of maternal continuum of care in Northwest Ethiopia, 2020 (n = 811).

Variables	Category	COR (95%CI)	Model I	Model II	Model III	Model IV
			(Null Model)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)
Women education level	Did not attend education	1		1		1
	Complete primary education	1.70(0.88,3.28)		1.45 (0.69,3.02)		1.54 (0.74,3.21)
	Complete secondary education and above	3.61(1.78,7.32)		2.93(1.18,7.28)		3.15(1.25,7.89)
Husband education level	Did not attend education	1		1		1
	Complete primary education	1.43(0.74,2.78)		0.93(0.43,1.96)		0.87(0.41,1.83)
	Complete secondary education and above	3.48(1.68,7.21)		1.63(0.63,4.18)		1.39(0.53,3.62)
Being a member in WDA	Yes	3.84(2.16, 6.83)		2.85(1.53,5.33)		2.91(1.56,5.44)
	No	1		1		1
HEWs home visiting	Yes	1.85(1.04,3.28)		1.91(1.03,3.57)		1.69(0.88,3.24)
	No	1		1
Being insured	Yes	3.07(1.65,5.71)		2.52(1.29,488)		2.59(1.33,5.01)
	No	1		1
Getting health education	Yes	1.87(1.06,3.29)		2.40(1.314.42)		2.44(1.33,4.45)
	No	1		1
Facility readiness	Yes	1.66(0.52,5.23)			0.78(0.27,2.26)	0.54(0.19,1.51)
	No	1			1	1
Distance from home to health facility	<=5km	4.37(1.23,15.29)			5.09(1.52,16.99)	4.81(1.55,14.95)
	>5km	1			1	1
Mass-media exposure	High	1.89(0.72,4.98)			2.20(0.96,5.03)	2.39(1.11,5.15)
	Low	1			1	1
Random effects	Variance		0.43	0.37	0.1	0.03
	ICC (%)		12	10	3	0.9
	PCV (%)		ref	14	76	93
	AIC		401.43	365.38	398.52	362.62
	-2Loglikelihood		397.44	345.38	388.52	336.62

At the individual level (Model II), predictors such as women with secondary and above education status, being a member of the WDA, being a member of community-based health insurance, getting health education at the health post level, distance from home to health facility, and husband’s secondary and above education level were statistically associated with women’s completion of the maternal continuum of care. ICC in Model II indicated that 10% of the difference in women’s completion of the maternal continuum of care was attributed to cluster variability.

The finding of this study in Model III showed that exposure to mass media and walking distance of less than or equal to five kilometers from home to a health facility were statistically significant in women’s completion of the maternal continuum of care.

Finally, after controlling all factors, the full model (Model IV) was developed, including individual and cluster-level factors simultaneously. The findings indicated that there was a substantial reduction in variance in predicting women’s completion of the maternal continuum of care. About 93% of women’s completion of the maternal continuum of care in the clusters was explained in the last model where the lowest Akaike information criteria and loglikelihood were observed (362.62 and 168.31, respectively). The odds of adhering to the maternal continuum of care were 3.15 times (AOR = 3.15; 95%CI: 1.25, 7.89) higher among women with secondary and above education as compared to their counterparts with lower-education. Women who were members of a WDA in the kebeles were 2.91 times (AOR = 2.91; 95%CI: 1.56, 5.44) more likely to complete the maternal continuum of care as compared to women who were not members of a WDA. Women insured in community-based health schemes were 2.59 times more likely to complete the maternal continuum of care than their counterparts (AOR = 2.59; 95%CI: 1.33, 5.01).

Our study found that women who got health education from a health post during the pregnancy period were 2.44 times (AOR = 2.44; 95%CI: 1.33, 4.45) more likely to complete the maternal continuum of care as compared to women who did not get health education from a nearby health post. Furthermore, women who walked less than or equal to five kilometers from home to health facility were 4.81 times (AOR = 4.81; 95%CI: 1.55,14.95) more likely to complete the maternal continuum of care than those who walked more than five kilometers. Finally, women who had been exposed to mass media were 2.39 times (AOR = 2.39; 95%CI: 1.11, 5.15) more likely to complete the maternal continuum of care as compared to their counterparts Table 3.

Discussion

We conducted a study on women’s completion of the maternal continuum of care among women in rural areas of Northwest Ethiopia. Our study showed that nearly one out of every fourteen women completes a maternal continuum of care, while one out of every thirteen women does not receive maternal healthcare from the existing health system. The final model of multivariable multilevel logistic regression analysis indicates that at the individual level, completing secondary and above education, being a member of a WDA, getting health education from a health post, and being a member of community-based health insurance; and at the community level, distance from home to health facility, and mass-media exposure were significant predictors of women’s completion of the maternal continuum of care.

According to our review of literature, the proportion of women completing the maternal continuum of care ranges from 6.4% [26] to 67.8% [27]. The current study was congruent with research studies conducted in Ghana [28, 29]. However, our finding was lower as compared to findings of studies conducted in other parts of Ethiopia [22, 27, 30–34], and elsewhere in Tanzania [35], and nine South Asian and sub-Saharan African countries [36].

The deflated proportion of women in the current study as compared to other studies could be attributed to the fact that our study considered women who had received at least 4 ANC visits, SBA at birth, and PNC visits within 48 hours after birth, whereas others included the post-partum period within six weeks after birth [28–30, 32, 34, 37]. The larger drop out in women’s completion of the maternal continuum of care puts women at higher risk of unwanted death, as women could miss continuity in maternal healthcare utilization. Another justification may be the low rate of early initiation and completion of ANC visits [38-40]. It is believed that early booking of ANC creates an opportunity for a woman to get pregnancy-related information such as birth preparedness, danger signs during pregnancy and the postpartum period, nutritional counseling, and complete content of focused ANC components and helps with early checkups for medical and obstetric conditions [32, 40–42]. Another possible justification could be that in the current study, the rate of late initiation of ANC (after 16 weeks of pregnancy) was nearly 50%, i.e., one in every two women did not receive an early ANC visit, as women have not been provided the aforementioned healthcare benefits of early initiation of ANC visits. Therefore, we suggest early booking of ANC and staying on the path of the maternal continuum of care to improve women’s and newborns’ well-being. The low point of women’s completion of the maternal continuum of care may be explained by the fact that though "home delivery free" is the program that has been launched by the government of Ethiopia, it is still one of the unresolved challenges of maternal health program indicators. For example, the Ethiopian demographic survey of 2016 [9] showed that 73% of births occur at home. In our study, among women who did not receive facility delivery and PNC visits, nearly two-thirds of women preferred Traditional Birth Attendants (TBA) to give birth and did not have awareness about the importance of PNC after birth. They perceived that they and their babies were safe and did not intend to go to a health facility. Another reason might be that health-seeking behavior, beliefs, and norms about the place of birth could contribute to a low rate of women’s completion of the maternal continuum of care. Literature states that the most common reasons for home delivery are family and relatives influence, usual practice, unexpected labor, and not being sick, depending on traditional healers and spiritual healing, such as prayer [43]. Women’s perception that pregnancy is not an illness and that attending ANC is useless could act as a barrier to utilizing healthcare services [44]. Moreover, unlike other studies [27, 30–32, 34, 36, 37, 45, 46] that reported higher proportions of the maternal continuum of care by using ordinary logistic regression, our study considered multilevel multivariable logistic regression models.

Regarding predictors, our finding revealed that the odds of women’s completion of secondary and above education had a significant association with women’s completion of the maternal continuum of care. This finding aligns with the findings of studies conducted in Ethiopia [22, 30, 31, 47] and abroad [29, 36, 37, 48, 49]. The possible justification may be that educated women could easily grasp pregnancy-related counseling during healthcare provision and information through mass media. Moreover, educated women may have a role in self-determination so that they can decide by themselves to seek healthcare from healthcare providers. In addition, educated women may not be influenced by beliefs and norms. A study conducted in rural Mali on the role of beliefs and norms for maternal healthcare utilization revealed that women were expected to obey their husbands during childbirth in order to have an easy delivery [50]. If the husband orders the woman to stay at home during labor, the woman should obey him and stay at home for home delivery.

In this study, women who were members of a WDA were more likely to complete the maternal continuum of care as compared to their counterparts. This finding was supported by studies done in Ethiopia [51, 52]. The similarity of our findings might be that an active woman in a WDA could get different benefits in the kebele, such as health education about birth preparedness and complication readiness, being linked to health facilities by HEWs at the time of referral, reducing delays that could be related to a pregnant woman, and enhancing harmonized relationships with HEWs, community, and healthcare facilities.

The odds of women completing the maternal continuum of care were higher among respondents who were members of community-based health insurance as compared to those who were not insured. Our finding was in concord with other studies conducted in Ethiopia [53, 54], Ghana [55], Manzi [56], Nigeria [57], and Tanzania [58]. Ethiopia has been implementing the Community Based Health Insurance (CBHI) scheme since 2011 with a vision of promoting health of poor rural residents [59, 60], as the CBHI package benefits a woman to get all family health services which is part of Ethiopia’s essential health package [61]. The association of CBHI and maternal continuum of care in the current study might be due to the fact that the existing CBHI scheme in Ethiopia provides maternal healthcare services for free to all women [62]. Due to the low economic conditions of the country, women could be less motivated to use maternal healthcare services if they were required to pay for them. Another possible explanation might be that community health insurance reform in Ethiopia could have had an influence on the health-seeking behavior of a woman [53]. Moreover, CBHI enrolment rate has been growing in Ethiopia, particularly the highest rate was recorded in Amhara region where the study was conducted [60].

The findings of this study showed that getting health education at the health post level made women more likely to complete the maternal continuum of care as compared to women who did not get health education. Our finding was consistent with a prior study [63] and a study done in Pakistan revealed that facility delivery was higher among women who got health education from community health workers when compared to those who did not get health education during the pregnancy period [64]. In Ethiopia, frontline HEWs are commonly tasked with providing maternal and child health services at primary healthcare units [65]. HEWs’ health message advocacy for maternity service utilization at the health post level could have an influence on the maternal healthcare continuum. This finding was reaffirmed by [66], as community health workers have a vital role in facilitating women’s completion of maternal continuum of care.

Regarding community level distance, the odds of completing the maternal continuum of care were higher in women who had a walking distance of less than or equal to five kilometers from home to reach health facility as compared to their counterparts. Studies conducted in other parts of Ethiopia [30, 31, 67] and elsewhere [28, 63, 68–70] supported this finding that less travel time to health facilities predicted completion of the maternal continuum of care. This could be because women who live in difficult-to-reach areas or are unable to access health facilities may face lack of transportation to health facilities, particularly at night [71] as many times labor begins at night [72] and most women take walks by foot or animal; lack of transportation and long distances are barriers for maternal healthcare utilization [73]. Another reason might be that women at remote locations could have less awareness and miss a chance to get pregnancy-related information as compared to women who live near a country/health facility. Another possible explanation might be the inaccessibility of health facilities, according to Berhan et al., as women pay around 4000 Birr for less than a hundred kilometers of travel in Ethiopia, which is unfair and too high in the country by any standard [74].

This study found that women who were exposed to the mass media were more likely to complete the maternal continuum of care than those women who were not exposed. This finding is in line with previous studies done in Ethiopia [29, 37]. According to a study in Pakistan, about 58% of women with weekly exposure to mass media gave birth in health institutions as compared to 36% of women with less frequent mass-media exposure [64]. The possible explanation for this is that the mass media has the potential to influence developing positive behavior towards maternal health service utilization because women can easily access maternal and child health-related information at home.

Strengths and limitations of the study

First, we carried out a community-based study to understand the current level of maternal continuum of care and its associated factors using individual and community level variables. Second, the sample size we used in this study was yielded after a thorough estimation of sample sizes using proportions of focused ANC visits, SBA, and continuum of maternal care. Third, the use of a simple random sampling technique in this study has contributed to the strength of the study. Fourth, we provided adequate training and close supervision during field activities. Fifth, each estimate was tested using the demand and supply side inquiries, which can be reflected as an actual problem of women’s completion of maternal continuum of care.

Recall bias may be the limitation of this study, as the participants might not remember the previous event other than recent events. Nevertheless, we attempted to specify questions related to the service given during prenatal, natal, and postnatal periods by probing. In our study, we did not exclude women who had previous pregnancy complications; this could have introduced social desirability bias in which respondents may have answered questions that they thought would lead to their being accepted. However, during the response, an open-ended approach was used to prevent the participants from simply agreeing or disagreeing. Owing to resource constraints, the study was limited in a small locality and this might make it to fall short of generalizability to a wider population. Follow-up studies covering larger segments of population and focusing on other contextual factors beyond the individual and community level variables could give better findings for the improvement of maternal continuum of care. Moreover, we tried to control confounding through maximizing potential variables to be included in the study and applying multivariable logistic regression during analysis stage.

Conclusions

Women’s completion of the maternal continuum of care was extremely low in the study area. There was a substantial reduction in community variation in the final model, indicating that there was a reduced representation of unobserved variables that explain the variation. The findings suggest that women’s education, being insured, being a member of a WDA, health education at the health post level, distance from home to health facility, and women’s exposure to mass media are contextual factors contributing to women’s completion of the maternal continuum of care. Therefore, the existing health system should consider multilevel intervention strategies that focus on providing maternal health education, facilitating insurance mechanisms, encouraging women’s participation in health clubs, and ensuring physical accessibility to healthcare facilities to be more effective in improving maternal health services.

Dataset of why maternal continuum of care remains low in Northwest Ethiopia? A multilevel analysis.

(DTA)

Click here for additional data file.

12 Jul 2022

Submitted filename: PONE-D-21-38409_reviewer_comments.pdf

Click here for additional data file.

Submitted filename: PONE-D-21-38409_reviewer comments.docx

Click here for additional data file.

14 Aug 2022

Date 20/07/2022

Rebuttal Letter

Subject: Question-by-question responses

Manuscript title: Why maternal continuum of care remains low in Northwest Ethiopia?

a multilevel logistic regression analysis

Manuscript number: PONE-D-21-38409

Authors: Tesfahun Hailemariam, Asmamaw Atnafu, Lemma Derseh Gezie, and Binyam Tilahun

Dear reviewers

Thank you for taking the time to review and comment upon our manuscript, (Manuscript number: PONE-D-21-38409, Why maternal continuum of care remains low in Northwest Ethiopia? a multilevel logistic regression analysis). We found the advice constructive and have incorporated the suggestions into our revised or cleaned manuscript. We have responded to each of the reviewers’ comments in our comment-by-comment responses below. It is our belief that the manuscript is substantially improved after all those comments and suggestions were addressed. We noted that the corrected texts at cleaned version may not exactly found on the page and line as mentioned by reviewers at the original commented manuscript because of the extensive revisions of each page and line. We want to inform you that the corrections made in the main document are indicated with blue color.

Thank you again for your thoughtful comments.

On behalf of the group of authors

Sincerely,

Tesfahun Hailemariam

Reviewers' comments to author:

The paper is clearly and logically written, well organized, and easy to follow. The introduction provides the reader useful background and context that sets up the rest of the paper well. The authors present robust and novel findings that address the factors that affect women’s completion of the maternal continuum of care in the Wogera and Gondar districts through a community-based cross-sectional study. The methods were clearly defined, and the results and discussions well outlined and supported with existing evidence and past studies in Ethiopia. The paper also did well to adjust for the design effect by assessing various levels of intraclass correlation coefficients using four different models.

The paper requires major and minor revisions including the revision of wording choice and consistency, the justification of the interview tool and sampling technique, addition of citations, and the expansion on the discussion of study limitations. Moreover, I would advise the author to complete a grammar revision of the paper to improve the flow and readability of the text and revise the structure of some of the sections within the paper.

Comment: Revision of wording choice and consistency

Response: Thank you. We have revised the paper for its word choice and consistency according to the comment given.

Comment: The justification of the interview tool and sampling technique

Response: Thank you so much. The interview tool was developed in English, after thorough literature review, and translated into the local language (Amharic). Since expert’s view was sought for psychometric properties of face and content validity, experts (Gynecology and obstetrics specialist (1), research and publication directorate (1), Health Informatics specialist (2), Midwifery specialist (2), Emergency and obstetric officer specialist (1) were invited to review the relevance of each question in the interview tool. The tool was revised according to expertise’s view and piloted before data collection was started. The proportion of rated item and scale content validity index was computed and experts rated item content validity index and scale level validity index were 0.98 and 0.81, respectively. For wealth index assessment, principal component analysis was employed to reduce data that measure the same construct together and the data was recoded into binary variables. Variables with low frequencies were combined together and frequency more than 95% and less than 5% were excluded from the analysis. Under the components loading, there were twenty-one variables correlated with measuring the wealth index such as having own house, toilet, electricity, kitchen, charcoal, television, radio, modern bed, cotton, mobile phone, farming land, having animals in the house such as plough, house, cows, horse donkey, goat, sheep, chicken, having separate room for animals, having beehive, having bank account number with deposited Birr (Ethiopian currency). Scale reliability coefficient and Kaiser-Meyer-Olkin measure of sampling adequacy of variables were 0.72 and the 0.80, respectively. Wealth quantile was categorized into five categories such as poorest, poor, middle, richer and richest. Data collectors and supervisors were trained on the instrument. The household data were collected, using an interviewer administrated questionnaire via face-to-face interviews.

Regarding to sampling technique, in this study, we used a multi-stage sampling technique to reach study participants. Wogera and Gondar zuriya districts (the districts where our study conducted) were selected randomly among the six transformation woredas in the central Gondar zone, Northwest Ethiopia. Specifically, the primary sampling units were kebeles with the respective health centers while the secondary sampling units were women who gave birth during the past 12 months. The health centers (or the respective kebeles) were selected with a simple random sampling technique. And to select sample participants, a sampling frame of mothers was developed using a list of eligible women from the community health information system family folder(pouch). Then the sampling interval was obtained by dividing the source population (8393) by the estimated total sample size (811). The first study participant (the 2nd mother) was selected by a simple random sampling technique from the first sampling interval, and all other mothers were selected systematically by taking every 10thmother in the frame. When two or more women were found at a single household level, one of them was selected randomly and included in the study. Seriously ill or women who were unable to speak were excluded from the study. For those women who were selected for an interview and were not available during data collection, we waited for them for three to five days to get them. If they were not still accessible, they would be considered non-respondent.

Comment: Addition of citations, and the expansion on the discussion of study limitations.

Response: Thank you so much. We added recent references and expanded study limitations in the discussion section.

Question: Moreover, I would advise the author to complete a grammar revision of the paper to improve the flow and readability of the text and revise the structure of some of the sections within the paper.

Response: Thank you so much. Critical refinements were made to improve further its English and revisions on the structure of the paper.

Please refer to the minor revisions for detailed recommendations.

Minor revisions:

1. The paper needs more discussion of study limitations which may include expanding on:

Comment: Selection bias: error due to random chance, bias from selecting the 11 health centers (clusters) and kebeles for the study; limited external validity: study only conducted in 2 districts in Ethiopia

Response: Thank you so much. In order to minimize the selection bias that is possible from selecting the 11 health centers (clusters) and kebeles for the study, we have applied probability sampling technique using the defined target population and sampling frame. Owing to resource restrictions, the study was limited in a small locality (Wogera and Gondar Zuriya districts) and this might make it to fall short of generalizability to a wider population. Longitudinal studies covering larger segments of populations and focusing on other contextual factors beyond the individual and community level variables would give better findings for the improvement of maternal continuum of care. We have included this limitation in the discussion section of limitation part. We feel that with its limitation the study could be interpreted/compared to other similar settings.

Question: Confounding bias: any other unknown or unmeasured confounding factors that were not controlled for

Response: Thank you so much. We have exhaustively reviewed literatures and included the potential variables in the study in order to minimize biases that could be arised due to confounding. The random effect we applied in this study could capture the effect of unaccounted variables beyond the individual variables and may quantify the bias estimate related with third variable. Moreover, the conventional alpha 0.05 significance wouldn’t report the errors in the estimates so that the estimate could be over or under estimated. The findings in our study were presented with adjusted odds ratio which reduces systematic and random error misclassification bias.

Minor revisions:

1. Paper Section: Overall

Suggestion: Keeping the word choice and labeling consistent

Specific example

Utilize relevant and defined acronyms consistently throughout paper:

Comment: Page 3, Line 17: antenatal care was defined as ANC. Suggest using “ANC” consistently throughout rest of the paper

Response: Thank you so much. We have used “ANC” consistently for antenatal care throughout the paper.

Comment: Suggest assigning acronym “TBA” to ‘Traditional birth attendants” earlier in the paper and standardizing use of “TBA” throughout the rest of paper

Response: Thank you so much. We have used “TBA” early in the paper for Traditional Birth Attendant and used the acronym throughout the paper according to the comment.

Comment: Suggest assigning acronym “SBA” to “Skilled birth attendants” early in the paper and standardizing use of “SBA” throughout rest of paper

Response: Thank you so much. We have used “SBA” for Skilled Birth Attendant early in the paper and we have applied it throughout the paper.

Comment: Suggest assigning acronym “WDA” to “Women’s Development Army” earlier in the paper, as Table 3 references “WDA” without the acronym being defined earlier and standardizing use of WDA in the rest of paper

Response: Thank you so much. We have abbreviated “WDA” early in the paper and used the acronym throughout the paper as per the comment.

Comment: Suggest assigning acronym “HEW” to “Health Extension Worker” earlier in the paper (before page 20) since Table 3 references ‘HEW” without the acronym being defined earlier and standardizing use of HEW throughout the rest of the paper

Response: Thank you so much again. We have abbreviated “HEWs” early in the paper and used the acronym throughout the paper.

Comment: Suggest assigning acronym “PNC” to postnatal care and standardizing use of PNC throughout the rest of the paper

Response: Thank you so much. We have used “PNC” early in the paper and used it throughout the paper as per the comment given.

Comment: Intraclass correlation was assigned acronym “ICC” on Page 9, Line 12. No need to define the acronym again on Page 9, Line 21. Suggest using “ICC” consistently throughout rest of the paper

Response: Thank you so much. We have made correction on the manuscript based on the comment, using ICC consistently throughout the paper.

Comment: Suggest revising “antenatal care four and above” to “at least 4 ANC visits”

o Page 5, Lines 23-24

o Page 8, Lines 12-13

o Page 13, Line 5

o Page 14, Line 1

o Page 18, Line 10

Response: Thank you so much. We have revised “antenatal care four and above” to “at least 4 ANC visits” consistently throughout the paper.

2. Abstract

Word choice

Comment: Page 2, Line 3: suggest replacing “intractable” to “significant”

Response: Thank you so much. We have replaced an “intractable” word to “significant”

3. Introduction

Wording clarification and consistency

Suggest revising the following sentences

Comment: Page 3, Line 4-6: clearly explain what the numbers represent (e.g., number of deaths, life-time risk). Also, please clarify which year these statistics were estimated in. Page 3, Line 10: suggest adding year of estimation.

Response: Thank you so much. We have explained the numbers presented at page 3 line 4-6 and the year in which the statistics were estimated.

Comment: Page 5, Line 3-4: please clarify what you mean by “women with less than one year” – what is the less than one year referring to?

Response: Thank you so much. We have made a clarification of the word “women with less than one year.” The author has made change on the paper as “women with a child less than one year”.

Wording choice and consistency:

Comment: Page 4, Line 1: suggest changing “skilled healthcare attendant” to “skilled birth attendant” (SBA) to align with wording in rest of paper

Response: Thank you so much. We have changed skilled healthcare attendant to “skilled birth attendant” (SBA) in page 4 line 1 and aligned it in all parts of the paper where SBA is mentioned.

Comment: Page 4, Line 11: suggest changing “besides” to “moreover”

Response: Thank you so much. We have changed the word “besides” to “moreover” as per you suggestion.

Comment: Page 4, Line 16: suggest revising “studies” to “past/recent studies”

Response: Thank you so much. We have revised the word “studies” to “past studies” according to the comment given.

4. Methods

Sample size clarification

Consider including a sample size calculation table in Appendix:

Comment: Suggest presenting sample size calculation matrix table under different scenarios (and assumptions) and justification of the chosen sample size

Response: Thank you so much. Our study was based on the following sample sizes estimated using different assumptions as presented in the following calculation matrix table.

Table 1: Sample size calculation for the study objective using single population proportion formula and considering factor affecting the outcome variable

No

Variables Study

In 95% CL 80% power OR Exposed (%) Unexposed (%) Design effect Ratio Sample size NR- rate Total

Sample size

Proportion of focused antenatal care (39.9%) (22) 95% - - - - 2 - 736.9 10% 811

Proportion of SBA after completing at least 4 ANC visits (31.1%) (22) 95% - - - - 2 - 658.5 10% 724

Proportion of women retained in the continuum of maternal care (12.1%) (22) 95% - - - - 2 - 326.8 10% 360

Associated factors with CoC Being a model family in the community (23) 95% 80% 2.13 49.7 50.3 2 1:1 254 10% 559

NB: Finally, after internal comparison, we have considered the largest sample size (811) and used it to community based cross-sectional study in order to assess women’s completion of maternity continuum of care.

Clarify age range

Comment: Page 5, line 14: Please clarify what the age range for reproductive age group is (e.g., 15-54) and explain why this range was chosen.

Response: Thank you so much. For this study, women of reproductive age group (15-49) were selected. Though the age range for reproduction is possible out of the mentioned age range, the reproductive age group in Ethiopia (15-49) was sought our attention in order to assess the prevalence and its associated factors of maternal continuum of care using both individual and community level variables. The current Ethiopian demographic health survey [1] also focuses the age range of women 15-49 as the majority of women fall in this age ranges. Other study also supports these evidences [2].

Clarify outcome variable

Comment: Page 9, Line 3: Suggest clarifying that the outcome variable is the “maternal continuum of care” (binary).

Response: Thank you so much. Maternal continuum of care contains three combined maternity care levels, first: women with antenatal care one to antenatal care four and above; Second: women who were completed antenatal care 4 and above to skilled birth attend, and third: women from skilled birth attendant to postnatal care visit within 48 hours after delivery. Hence, our study considered maternal continuum of care “1” if a woman received the three combined service together, and “0” otherwise.

5. Results

Revise Tables

Comment: Tables 1-2: Suggest removing column “Total (%)”: it is not contributing more useful information

Response: Thank you so much. We have removed table 1 and table 2 total (%) column as per the suggestion given in page 10 and 11, and 12 and 13 respectively.

Comment: Tables 1-3: Make sure that spacing is consistent in table 1 (e.g., N (%))

Response: Thank you. We have made consistent spacing in both “Yes” and “No” column in page 10 and 11.

Wording choice

Comment: Page 11, suggest revising “intendedness of pregnancy” to “pregnancy intention”

Response: Thank you so much. We have replaced “pregnancy intendedness” to “pregnancy intention”, page 11.

Wording choice and consistency:

Comment: Standardize capitalization of “Model I-IV” in Results section,

o Page 14, Line 22: Capitalize “model II”

o Page 15, Line 1: Capitalize “model III”

o Page 15, Line 4: Capitalize “model IV”

Response: Thank you so much. We have capitalized Model I-IV in result section in page 14 and 15, and throughout the paper as well.

Comment: Clarify “counterparts” when summarizing Odds Ratios and comparisons

Suggest revising “counterparts” to “lower-education counterparts” (Page 15, Line 11)

Response: Thank you so much. We have revised the word “counterparts” with “lower-education counterparts”.

Missing results

Comment: Page 14, Line 17: fill in the blank: what = 0.43?

Response: Thank you so much. We have filled the missing result, page 14, line 17

6. Discussion

Add citations

Comment: Page 17, Line 3-4: add citations for statistics “one of every fourteen” and “one out of eight”

Response: Thank you so much. In the first paragraph of the discussion section, we outlined our summary findings. One of the findings is maternal continuum of care as it was presented in Fig 1, i.e., 6.9% of women compete maternal continuum of care. This was explained in the discussion section in the form “one from every fourteen women”. The other finding indicates that from all the women who were surveyed at the study districts, about 12.5% of women didn’t get any antenatal care visit during pregnancy life time from existing health system. This was also explained in discussion section in the form of “one out of eight women” were not getting maternity service from existing health system. These findings were explained in detail in result section at “dropouts of women completion of maternal continuum of care” part, thus, we feel that it does not need citations as they are the findings of our/current study.

Comment: Page 19, Lines 16, 19-20: add citations for the community-based health insurance scheme/reform in Ethiopia

Response: Thank you so much. We added citations for the community-based health insurance scheme/reform in Ethiopia.

Wording choice and consistency and clarification

Wording choice and consistency:

Comment: Suggest revising Page 18, Lines 12-13 to “puts women at higher risk of unwanted death”

Response: Thank you so much. As per your suggestion, we have revised the word “risks women to unwanted death” to “puts women at higher risk of unwanted death”

Comment: Suggest revising Page 19, Lines 18-19 to “aligns”

Response: Thank you so much. We have revised the word “is in consonance” to “aligns” according to your comment.

Wording clarification:

Comment: Please clarify what “elsewhere” means (Page 19, Line 19)

Response: Thank you so much. In our study we used the word elsewhere to indicate countries where the primary studies were conducted (Ghana (Ref 30), South Asia and sub-Saharan Africa (Ref 38), Nepal (Ref 40), Cambodia(Ref 42), and Egypt (Ref 44).

Clarification needed

Comment: Please justify why Pakistan and other countries (e.g., Cambodia, Nepal, Lao PDR, Egypt) were included as comparisons of interest to the results of the study. Would suggest removing these comparisons and only keeping references in Ethiopia.

Response: Thank you so much. In this section we have attempted to compare utilization of maternal continuum of care at different countries which have different context. However, the variation of our finding from previous finding conducted in Africa and other countries were justified why is low or high using the existing evidences. The only thing comparing our finding in this section with studies conducted in some other countries like Cambodia, Egypt, Lao etc is that to show the existing evidence or what is known regarding to our findings. According to the comment given, we removed references and tried to keep the references in Ethiopia.

Strengths & limitation

Comment: Strengths: No strengths were outlined in the Discussion section

Response: Thank you so much. We have included the strengths section in the discussion part of the paper.

Comment: Limitations: to consider further limitations in the study

Response:

Comment: Consider selection bias: error due to random chance, bias from selecting the 11 health centers (clusters) and kebeles for the study;

Response: Thank you so much. In order to minimize the selection bias that is possible from selecting the 11 health centers (clusters) and kebeles for the study, we have applied probability sampling technique using the defined target population and sampling frame.

Comment: limited external validity: study only conducted in 2 districts in Ethiopia. Explain the relevance of the study beyond the context of Ethiopia.

Response: Thank you so much. Owing to resource restrictions, the study was limited in a small locality (Wogera and Gondar Zuriya districts) and this might make it to fall short of generalizability to a wider population. Longitudinal studies covering larger segments of populations and focusing on other contextual factors beyond the individual and community level variables would give better findings for the improvement of maternal continuum of care. We have included this limitation in the discussion section, at limitation part. We feel that with its limitation the study could be interpreted/compared to other similar settings.

Comment: Consider confounding bias: any other unknown or unmeasured confounding factors that were not controlled for

Response: Thank you so much. We have exhaustively reviewed literatures and used the potential variables that to be included in the study in order to minimize biases that could arise due to confounding. The random effect we applied in order to capture the effect of unaccounted variables, beyond the individual variables may quantify the bias estimate related with third variable. Moreover, it is obvious that the conventional alpha 0.05 significance wouldn’t report the errors in the estimates so that the estimate could be over or under estimated, therefore, the findings in our study were presented with adjusted odds ratio with systematic and random error misclassification bias.

Comment: Study instrument: clarify whether tool is standardized in Ethiopia / beyond; explain who the “expert reviewers” were

Response: Thank you so much. This section was elaborated in detail in the first page in similar comment section. The explanation of who the expert reviewers were elaborated at the instrument section of the paper. We applied both face and content validity of the tool in which expert judgment on what the instrument is looks like: wording, layout, clarity, comprehensiveness of the tool was checked in face validity and content validiy in which using content validity index. The mean proportion of experts rated item relevant is in a range (I-CVI)=0.98 and S-CVI/UA= 0.81, 98% of the total items are judged content valid. Accordingly, the tool was checked for its appropriateness and revision was also undertaken.

7. Grammar

Grammar

Comment: Please revise comments throughout manuscript (attached PDF).

Response: Thank you so much. The whole manuscript sections were revised and refinement was made. The cleaned version and manuscript with comment addressed in blue color were attached in the submission system.

Thank you again for the effort you made to review our manuscript!

1. Mini E, Demographic E. health survey 2019: key indicators report. The DHS Program ICF. 2019.

2. Yeshaw Y, Kebede SA, Liyew AM, et al. Determinants of overweight/obesity among

reproductive age group women in Ethiopia: multilevel analysis

of Ethiopian demographic and health survey. BMJ Open 2020;10:e034963. doi:10.1136/ bmjopen-2019-034963

Submitted filename: Response to reveiewers.docx

Click here for additional data file.

4 Sep 2022

Why maternal continuum of care remains low in Northwest Ethiopia? a multilevel logistic regression analysis

PONE-D-21-38409R1

Dear Dr. Tesfahun Hailemariam ,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Sebsibe Tadesse, PhD

Academic Editor

PLOS ONE

8 Sep 2022

PONE-D-21-38409R1

Why maternal continuum of care remains low in Northwest Ethiopia? a multilevel logistic regression analysis

Dear Dr. Hailemariam:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Sebsibe Tadesse

Academic Editor

PLOS ONE