Postnatal care utilization within 48 h after birth and its determinants in northwest Ethiopia: a multilevel logistic regression analysis

Abstract

Postnatal care immediately after birth is a crucial component to save lives of mothers and the newborns. A paucity of evidence indicated that women’s unwillingness to receive care after birth is a remained challenge in resource-limited settings in general and in Ethiopia in particular. This study aimed to assess the utilization of postnatal care within 48 h after birth and its determinants in northwest Ethiopia. A community-based cross-sectional study was conducted from October to November 2020. A total of 811 women who had birth within the last one year were involved in the study. Both random and fixed effects were reported using an adjusted odds ratio with a 95% confidence interval and p value of 0.05. The study revealed that 13.3% (95% CI 10.9–15.7%) of women attended postnatal care visits within 48 h after birth. Attending secondary and above education (AOR = 2.39; 95% CI 1.18, 4.84); health extension workers home visiting during pregnancy (AOR = 1.74; 95% CI 1.08, 2.82); birth order (AOR = 2.07; 95% CI 1.03, 4.14); place of delivery (AOR = 7.49; 95% CI 3.73, 15.06); and short distance to health facility (AOR = 4.19; 95% CI 1.64, 10.70) were significantly associated with postnatal care utilization within 48 h after birth. Postnatal care utilization within 48 h is low in rural northwest Ethiopia compared to findings from most resource-limited settings. The existing health system should consider multilevel intervention strategies focusing on maternal health education, engaging community health workers in health promotion, and ensuring physical accessibility of healthcare facilities to be more effective in improving postnatal care utilization within 48 h.

Introduction

Nowadays, unacceptably high maternal deaths during and following pregnancy is a global burden. Worldwide, more than two women die from 1000 live births¹. Every day, 380 pregnancy- and childbirth related maternal deaths occur in LMICs². Sub-Saharan Africa (SSA) is the only region with very high MMR, estimated at 542 per 100,000 live births, which is more than two folds of maternal deaths from the global figure, i.e., more than five women die from every 1000 live births¹. Almost 75% of these deaths are directly associated with pregnancy and childbirth-related causes³, mainly during labor, childbirth, the first day and week after delivery⁴.

Nearly 40% of women develop complications following delivery, and almost 15% of them encounter potentially life-threatening complications^5,6 and most of these complications are preventable¹. Though a proper maternal healthcare utilization approach is recommended to avert preventable maternal deaths⁷, the poor maternal healthcare utilization during pregnancy, childbirth⁸ and the poor antenatal and postnatal care attendance⁹ and limited access to health services and the dearth of trained health professionals¹ accounts for a substantial number of preventable maternal deaths.

The Sustainable Development Goals (SDGs) considered women's well-being as a prime area of intervention¹⁰ and launched a target to reduce global maternal deaths to less than 70 per 100,000 live births by 2030¹⁰. Ethiopia has been striving to increase maternal health service coverage by investing on its health infrastructure expansion and health workforce development¹¹, considering them as building blocks of the health system¹²; the ultimate goal is to promote health status of its population. Though the government of Ethiopia is investing much to improve women health through designing and implementing policies, guidelines, and strategies inspiring to promote maternal health conditions and achieve SDGs targets^10,13, women attending adequate and timely healthcare utilization remains a significant challenge in Ethiopia¹⁴.

Maternal mortality in Ethiopia is still intolerably high³ and among the highest maternal death rate in the world¹⁵ which could be linked with low maternal healthcare utilization and can be averted if women receive adequate and timely healthcare utilization³. The available evidence showed that 44% of women who had four or more ANC visits gave birth at home¹⁶. Similarly, among women who delivered at home, only 8% received postnatal care (PNC) within 42 days after delivery¹⁶. There is an unwillingness among women to attend birth at a health facility and receive PNC services after completion of four or more ANC visits¹⁶.

A large proportion of maternal deaths occur during the first 48 h after birth¹⁷. Thus, receiving PNC from a healthcare provider at the recommended time with its appropriate contents also prevents complications that could arise after childbirth¹⁵. In Ethiopia, 34% of women receive a PNC visit in the first two days after delivery. About 71% of rural women do not receive a postnatal check-up within 2 days of delivery, whereas 52% of urban women do not receive early postnatal check-up¹⁷. According to the body of literature, women education, place of delivery, ANC attendance¹⁸, getting advice from healthcare provider¹⁹, having many children, mass media exposure, and accessibility of healthcare²⁰ were some of factors influence maternal early postnatal care utilization. Moreover, both individual and community-level factors are influencing maternal early postnatal care utilization^21–23.

Though there is promising progress in maternal healthcare utilization, and a prioritized agenda to improve early postnatal care utilization in Ethiopia, the available evidence on PNC visit within 48 h is inconclusive²⁴. There are limited studies on the postnatal care utilization within 48 h particularly in the study area²⁵. Although previous work in Ethiopia has paid attention on investigating factors determining postnatal care utilization within 48 h, there is a gap in considering community-level variables that could hinder women from getting postnatal care within 48 h after delivery²⁶. Despite postnatal care utilization within 48 h is critical indicator to improve quality of life and reduce neonatal and maternal deaths, the attention given to consider hierarchical level of variables were minimal²⁵.

Understanding the existing level of postnatal care utilization within 48 h and identifying contextual factors that limit women from attending postnatal care utilization within 48 h helps to be effective in the implementation of maternal health program. 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 predict the point of women attending postnatal care utilization within 48 h. Therefore, this study aimed to determine the level of women’s attending PNC visit within 48 h after birth and identify individual and community-level factors affecting it in northwest Ethiopia.

Methods

Study design and setting

In this study we applied a community-based cross-sectional study design. Eleven clusters were considered from Wogera and Gondar Zuriya districts in central Gondar zone northwest Ethiopia from October to November 2020. The districts away 658 km from Addis Ababa which is a central city of Ethiopia²⁷. In the districts, there were sixteen health centers and eighty eight health posts²⁷. According to the report from central health bureau, a total population of the zone was 524, 907 (female = 260,879 and male = 264,028) and the reproductive age group (15–49) and surviving infants were 122,303 and 16,745 surviving, respectively²⁷.

Study population

Source population of the study were all women in the study area who gave birth within the last one year. Permanent residents of the selected districts, randomly selected, and had willingness to participant 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 all maternal services (any ANC visit and ANC4 + and postnatal care visit within 48 h) p = 12.1%²⁸. 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 utilization²⁹, 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 multi-stage sampling technique was used in this study. We randomly selected 11 clusters and included all the health posts found under the selected clusters to get the required number of women in the study. Random selection was applied at each study cluster to reach out the study participants. Households considered in the study were selected from a sampling frame using a family folder at health post level that contain lists of women who gave birth in the district. The sampling interval was estimated using source population and sample size estimated in the study. Accordingly, systematic sampling was carried out and the first woman was selected using simple random sampling approach. Thus, a total of 811 study participants were obtained and included in the study.

Study variables and measurement

Response variable

Postnatal care utilization within 48 h is categorized as “Yes” if the women received PNC visit within 48 h after birth and “No” if otherwise²⁷.

Explanatory variables

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

Individual level factors

Individuals level variables included in this study were: age of the respondents, marital status, occupation of the respondents, educational status, pregnancy intendedness, distance, and community based health insurance status, home visiting by health extension workers, involvement in Women's Development Army (WDA), health education from a nearby health post, frequency of ANC visits, and place of delivery.

Community level factors

Community and cluster level variables were aggregated from individual and cluster level data such as community level health education from nearby health posts. Aggregated clusters were categorized as low if the proportion of health education was 0–38% whereas high if the proportion was 39–100%. Community based health insurance was categorized as low if the proportion of CBHI status was 0–48%, whereas high if the proportion of CBHI status in the community was 49–100%. Finally, 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%. Aggregation of data from individual level to cluster was based on the data point at individual level.

Data collection instrument and procedures

Data collection tool was developed through thorough literature review. The tool was developed in English and translated into Amharic which is a local language of the study district. Face and content validity were carried out by involving experts in the area and enriched the tool though collecting insights from experts. The tool was piloted in different clusters rather than the study clusters and validated to its level of standard before commencing the data collection. Accordingly, the item content validity index and scale level validity index of the tool were found to be 0.98 and 0.81, respectively. Intensive training on the data collection tool was given to data collectors and supervisors. The survey data was obtained through face-to-face interviews approach using an interviewer-administered questionnaire.

Data management and statistical analysis

The data were entered into Epidata and analyzed with Sata14. Bivariable multilevel logistic regression was used for each factor against the PNC visit within 48 h without controlling the effect of other explanatory variables. Factors at both individual and cluster levels with a p value < 0.25 were considered as candidate variables for multivariable multilevel logistic regression analysis³⁰. 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 attending PNC visit within 48 h. 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 value³¹. 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 attending PNC visit within 48 h after birth. The random effect was presented using ICC; thus, in all models, ICC and its change from the null model were examined. 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 variables³². 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. Consent was obtained from all subjects and/or their legal guardian(s). 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. All methods were carried out in accordance with relevant guidelines and regulations.

Ethics approval and consent to participate

The ethical review board of the University of Gondar and the ethical review committee of the Amhara regional health bureau's research and technology transfer office gave the permission and ethical clearance to the study. Permission to conduct the research was sought at all levels of government administration. Each study participant gave their informed consent. To prevent revealing personal identification, data was collected using the codes provided. To avoid unwanted access, the data were stored at the University of Gondar's repository.

Results

Sociodemographic and reproductive characteristics of the study participants

In this study, a total of 811 participants were involved with a response rate of 100%. The minimum and maximum ages of the respondents were 15 years and 48 years, the mean age of the study participants was 28 years ± 6.4SD. Most of the participants were married, 790 (97.4%), and housewives, 801(98.8%). About two-thirds, 500 (61.7%), of participants had no history of attending education. Being a member of community-based health insurance and distance less than or equal to five kilometers were 394 (48.6%) and 576 (61.6%), respectively. 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%) and 97 (12%) of women who gave birth at health facility were attended PNC visit within 48 h after birth (Table 1).

Table 1.

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

Variables	Category	Postnatal care utilization within 48 h		(p value)
		Yes (%)	No (%)
Age group of respondents (years)	15–24	34 (4.2)	193 (23.8)	0.408
	25–34	46 (5.7)	348 (42.9)
	35 and above	28 (3.5)	162 (20.0)
Marital status	Married	106 (13.1)	684 (84.3)	0.604
	Single/divorced/widowed/separated	2 (0.2)	19 (2.3)
Occupation	Housewife	107 (13.2)	694 (85.6)	0.756
	Employee/laborer/merchant	1 (0.1)	9 (1.1)
Education level of respondents	Did not attend education	57 (7.0)	443 (54.6)	0.004
	Primary education	28 (3.5)	187 (23.1)
	Secondary and above	23 (2.8)	73 (9.0)
Pregnancy intention	Yes	90 (11.2)	552 (68.1)	0.252
	No	18 (2.2)	151 (18.6)
Distance home to health facility	> 5 km	40 (4.9)	195 (24.0)	0.047
	≤ 5 km	68 (8.4)	508 (62.2)
Being insured	Yes	57 (7.0)	337 (41.6)	0.349
	No	51 (6.3)	366 (45.1)
Membership in WDA	Yes	46 (5.7)	214 (26.4)	0.012
	No	62 (7.6)	489 (60.3)
HEWs home visiting	Yes	63 (7.8)	298 (36.7)	0.002
	No	45 (5.5)	405 (49.9)
Health education from nearby HPs	Yes	52 (6.4)	262 (32.3)	0.031
	No	56 (6.9)	441 (54.4)
Parity	1	28 (3.5)	149 (18.4)	0.052
	2–4	36 (4.4)	322 (39.7)
	≥ 5	22 (5.4)	232 (28.6)
Frequency of ANC visit	Less than four	50 (6.2)	440 (54.3)	0.001
	Four and above	58 (7.2)	263 (32.4)
Facility delivery	Yes	97 (12.0)	411 (50.7)	< 0.001
	No	11 (1.4)	292 (36.0)

Community level characteristics of the study participants

Women’s attending PNC visit within 48 h after birth was significantly associated with clusters (health centers) (chi-square = 36.14, p value = 0.001). In this study, community-based health insurance at the community level indicated that 394 (48.6%) of the participants had high CBHI status. The level of getting health education from nearby health post at community level was 259 (31.9%). Among the participants, most respondents, 615 (75.8%), had a walking distance of less than five kilometers from their home to a health facility (Table 2).

Table 2.

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

Variables	Category	Postnatal care utilization within 48 h		(p value)
		Yes (%)	No (%)
Health facilities	Ambageorgis	27 (3.3)	78 (9.6)	< 0.001
	Gedebiye	13 (1.6)	82 (10.1)
	Birra	7 (0.9)	53 (6.5)
	Tirgosgie	11 (1.4)	56 (6.9)
	Woybey	12 (1.5)	34 (4.2)
	Dergaj	5 (0.6)	37 (4.6)
	Miniziro	3 (0.4)	101 (12.5)
	Maksegnit	6 (0.7)	79 (9.7)
	Lamba	10 (1.2)	70 (8.6)
	Enfranz	8 (1.0)	84 (10.4)
	Abawarka	6 (0.7)	29 (3.6)
Community based health insurance	High	65 (8.0)	329 (40.6)	0.010
	Low	43 (5.3)	374 (46.1)
Health education from HPs	High	23 (2.8)	236 (29.1)	0.011
	Low	85 (10.5)	467 (57.6)
Walking distance from home to health facility	> 5 km	11 (1.4)	185 (22.8)	< 0.001
	≤ 5 km	97 (12.0)	518 (63.9)

Postnatal care utilization in northwest Ethiopia

Of the participants, 19.9% (95% CI 17.1–22.6) of women utilized postnatal care within six weeks (42 days) after giving birth. The proportion of postnatal care utilization within 48 h after delivery was 13.3% (95% CI 10.9–15.7%) (Fig. 1).

Figure 1.

Maternal postnatal care utilization in northwest Ethiopia, 2020.

Multivariable multilevel logistic regression analysis of women’s receiving PNC utilization within 48 h after birth

During analysis, four mixed-effect regression models were built to predict PNC visit within 48 h after birth based on 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 variations on maternal PNC visit within 48 h. An estimate of ICC was 8% (95% CI 2%, 23%), implying that 8% of the variation in attending PNC visit within 48 h was due to cluster-level factors and 28% was due to differences across clusters, and this variation was significant (τ = 0.28, p 0.000).

Table 3.

Multivariable multilevel logistic regression analysis of predictors of PNC visit within 48 h after birth in northwest Ethiopia, 2020 (n = 811).

Variables	Category	COR (95% CI)	Model I (Null Model)	Model II AOR (95% CI)	Model III AOR (95% CI)	Model IV AOR (95% CI)
Women education level	Did not attend education	1		1		1
	Complete primary education	1.13 (0.69, 1.84)		1.19 (0.66, 2.16)		1.16 (0.64, 2.10)
	Complete secondary education and above	2.19 (1.24, 3.85)		2.36 (1.17, 4.78)		2.39 (1.18, 4.84)
Being a member in WDA	Yes	1.71 (1.11, 2.61)		1.33 (0.83, 2.13)		1.26 (0.79, 2.03)
	No	1		1		1
HEWs home visiting	Yes	2.08 (1.34, 3.21)		1.75 (1.08, 2.84)		1.74 (1.08, 2.82)
	No	1		1
Parity	1	1		1		1
	2–4	0.62 (0.36, 1.07)		1.07 (0.58, 1.96)		1.07 (0.58, 1.97)
	5 and above	1.11 (0.65, 1.89)		2.12 (1.05, 4.27)		2.07 (1.03, 4, 14)
ANC visit frequency	Less than four	1		1		1
	Four and above	1.68 (1.09, 2.57)		0.94 (0.59, 1.51)		0.92 (0.57, 1.48)
Health Facility delivery	Yes	8.20 (4.21, 15.96)		7.84 (3.89, 15.82)		7.49 (3.73, 15.06)
	No	1		1		1
Community level CBHI	High	1.60 (0.79, 3.24)			1.68 (1.06, 2.66)	1.74 (0.89, 3.39)
	Low	1			1	1
Community level education	High	0.54 (0.24, 1.18)			0.66 (0.39, 1.12)	0.81 (0.38, 1.76)
	Low	1			1	1
Community level distance	≤ 5 km	3.13 (1.35, 7.24)			2.97 (1.49, 5.91)	4.19 (1.64, 10.70)
	> 5 km	1			1	1
Random effects	Variance		0.28	0.54	0.02	0.14
	ICC (%)		8	14	0.4	4.0
	AIC		626.81	567.18	621.26	563.55
	− 2Loglikelihood		622.81	546.44	611.23	537.54

At the individual level (Model II), predictors such as women with secondary and above education status, parity, health extension workers home visiting, and place of delivery (health facility) were statistically associated with women’s attending PNC visit within 48 h after birth. ICC in Model II indicated that 14% of the difference in women's attending postnatal care utilization within 48 h was attributed to cluster variability. The finding of this study in Model III showed that walking distances of less than or equal to five kilometers from home to a health facility and support from community-based health insurance were statistically significant predictors of women attending PNC visit within 48 h after birth.

Finally, after controlling all factors, the full model (Model IV) was developed, including individual and cluster-level factors simultaneously. The last model showed a substantial reduction of ICC where the lowest Akaike information criteria and loglikelihood were observed (563.55 and 537.54, respectively).

The odds of attending PNC visit within 48 h after delivery were 2.39 times (AOR = 2.39; 95% CI 1.18, 4.84) higher among women with secondary and above education as compared to their counterparts with lower education. The odds of attending PNC visit within 48 h after delivery was 1.74 times (AOR = 1.74; 95% CI 1.08, 2.82) higher among women who had received HEWs home visit during pregnancy as compared to women who were not received HEWs home visit during pregnancy. Women who have parity of 5 and above were 2.07 times more likely to attend PNC visits within 48 h after birth than their counterparts (AOR = 2.07; 95% CI 1.03, 4.14).

Our study found that women who gave birth at health facility were 7.49 times (AOR = 7.49; 95% CI 3.73, 15.06) more likely to attend PNC visit within 48 h after birth as compared to women who did not give birth at health facility. Furthermore, women who walked less than or equal to five kilometers from home to health facility were 4.19 times (AOR = 4.19; 95% CI 1.64, 10.70) more likely to attend PNC visit within 48 h than those who walked more than five kilometers (Table 3).

Discussion

The study showed that nearly one out of every eight women attends PNC visit within 48 h after birth. This implies that a substantial number of women do not receive postnatal care utilization within 48 h from healthcare facilities in rural northwest Ethiopia. The final model of multivariable multilevel logistic regression analysis indicates that at the individual level, completing secondary and above education, home visit by HEWs during pregnancy, having parity of 5 and above, and health facility delivery were predictors of women’s attending PNC visit within 48 h after birth and at the community level, distance from home to health facility was a significant predictor of women’s attending PNC visit within 48 h after birth.

According to our review of literature, the proportion of women attending postnatal care utilization within 48 h ranges from 6.3% in Ethiopia³³ to 58% in Kingdom of Cambodia³⁴. The current study was congruent with research studies conducted in Mundri East Country, in South Sudan 11.4%³⁵ and in Ethiopia 11.7%¹⁸. However, our finding was lower as compared to findings of studies conducted in other parts of Ethiopia 45.5%¹⁹, 23.3%³⁶, and 29.7%³⁷ and elsewhere in Nepal 43.2%³⁸, in Kingdom of Cambodia 58% ³⁴, and in India 45%³⁹.

However, the proportion of women attending postnatal care utilization within 48 h in the current study was lower as compared to studies conducted elsewhere as it could be attributed because of the time differences of studies reported and the difference in health system structure, accessibility of healthcare facilities and social and cultural influences of different study settings^18,40. The low remained postnatal care utilization within 48 h utilization puts women at higher risk of unwanted death, as women and the newborn could miss the opportunities of receiving healthcare from trained professionals. Studies showed that lack of postnatal care utilization within 48 h risks women and newborn to a life threating complications and postnatal care utilization within 48 h could improve women’s and newborns’ well-being and prevent unwanted maternal and neonatal deaths^19,36.

Regarding predictors, the odds of attending PNC visit within 48 h after delivery were higher among women with secondary and above education as compared to their counterparts with lower education. This finding was consistent with the findings of studies conducted in Ethiopia^18,19 and abroad^41–43. The possible explanation might be that educated women could easily grasp pregnancy-related counseling during healthcare provision and information and they may have updated information on the benefits of getting adequate and timely maternal healthcare and its complication to herself and the newborn as well⁴⁴. The other reason might be educate women could participate actively in active in different social and economic program in the community that could influence them to receive adequate and timely maternal healthcare services¹⁸. Moreover, educated women may have a role in self-determination so that they can decide by themselves to seek healthcare from healthcare providers as they are less likely to be influenced by beliefs and norms that could affect their health conditions⁴⁵. Therefore, education is a proxy indicator to ensure improvement of maternal healthcare utilization.

The findings of this study showed that women who were visited by HEWs during pregnancy were more likely to attend postnatal care utilization within 48 h as compared to women who did not get home visit by HEWs during pregnancy period. Our finding was consistent with a prior studies in Ethiopia^46,47 and abroad^48,49. HEWs in Ethiopia are commonly tasked with the provision of maternal health services such as health education and communication, identification of pregnant women, and creation of demand for health services in the community⁵⁰. Women who receive home visits by HEWs have a chance of attending maternity care⁵¹. HEWs’ daily engagement in the provision of health services at the household and community level may boost the health-seeking behavior of a woman. The HEP implementation in Ethiopia has positively impacted community health in general and brought various improvements to maternal and child health programs in particular⁵². HEWs' health message advocacy for maternity service utilization at the health post level could have an influence on the maternal healthcare utilization. This finding was reaffirmed by Gilmore and McAuliffe⁵³, as community health workers have a vital role in facilitating maternal healthcare utilization. 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 period⁵⁴. According to studies^55,56, capacitating community health workers (CHWs) increases accessibility of essential health services, particularly in low- and middle-income countries. Hence, HEP based counseling by a HEWs enabling women more likely to seek postnatal care utilization within 48 h and engaging HEWs more in the community activities and optimizing HEP could enhance early postnatal care utilization.

The odds of women attending early PNC were higher among respondents who had parity of five and above as compared to those who had parity of one. Our finding was in concord with other studies conducted in Ethiopia³⁶. The possible justification could be that previous pregnancy could have shared exposure to seek postnatal care visit³⁶. However, the current finding was in contrary with previous study showing that mothers having firs pregnancy enable them to seek PNC service on time and those who had multiple pregnancies could be less attentive to facility care⁵⁷. The difference could be that the sociocultural differences where participants involved in those studies could attribute for the differences between the studies as it has ability to affect the decision-making power of women for seeking healthcare services.

Our study showed that the odds of attending postnatal care utilization within 48 h were higher among women who gave last child at health facility as compared to women who did not give birth at health facility. The finding was consistent with findings in Ethiopia^58,59 and elsewhere^43,60,61. The association could be women who gave birth in health facility have greater opportunities to get health education and counselling related to postnatal care service immediately after delivery from healthcare providers who assisted them during delivery. Moreover, women who gave birth at facility also have chance to access and learn more about the benefits, types, and availability of PNC service. This could help women to receive postnatal care utilization within 48 h service before discharging from the facility. Hence, postnatal care counseling during labor, or at time of delivery is a good opportunity to educate mothers about benefits of PNC and improve the level of PNC utilization. Moreover, enhancing facility delivery also could provide a good coverage of postnatal care utilization within 48 h. Home delivery free program in Ethiopia is a promising program and need to be capitalized in order to make every woman opportune to facility based maternity care as women who gave birth at home could not have a chance to get access to PNC service counselling and care for themselves and the newborns²⁴.

Regarding community level distance, the odds of attending PNC visit 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. This finding is consistent with the studies conducted in Ethiopia^62–64 and elsewhere⁶⁵ This could be because women who live in difficult-to-reach areas or are unable to access health facilities and may face lack of transportation to health facilities especially during emergency^27,66. Moreover, those at remote area could have less awareness and miss a chance to get information related to maternal healthcare utilization²⁷. This indicate that the existence of different geographic variations are bringing disparities in utilization of postnatal care utilization within 48 h. Hence, improving physical accessibility of health facility could help to improve PNC utilization and make women to get the benefits of PNC care from healthcare providers for herself and newborn⁶⁷.

In this study, we assessed community level postnatal care utilization and factors associated using both community and individual level variables. The authors have conducted rigorous sample size estimation and applied simple random sampling technique to reach out the study participants. Moreover, the adequate training and close supervision provided during the study period was an added value. Furthermore, the demand and supply side inquiries were considered during the estimation of the parameters to reflect the outcome of interest.

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 postnatal period by probing. To minimize the selection bias that is possible from selecting the 11 health centers (clusters) and kebeles, the smallest administrative unit, we have applied probability sampling technique using the defined target population and sampling frame. Moreover, reducing bias related with confounding, we tried to control for it through maximizing potential variables to be included in the study and applying multivariable logistic regression during analysis stage.

Conclusions

Postnatal care utilization within 48 h was low in the northwest Ethiopia. Community variation was substantially reduced in the final model which indicates that there was a reduced representation of unobserved variables that explain the variation. The findings suggest that completing secondary and above education, home visiting by HEWs during pregnancy, having parity of 5 and above, health facility delivery, and distance from home to health facility were predictors of postnatal care utilization within 48 h. Therefore, the existing health system should consider multilevel intervention strategies focusing maternal health education, engaging community health workers in health promotion, and ensuring physical accessibility of healthcare facilities to enhance postnatal care utilization within 48 h at community level.

Abbreviations

ANC

Antenatal care

WDA

Women Development Army

SBA

Skilled birth attendants

AOR

Adjusted odds ratio

LMICs

Low and middle income countries

MMR

Maternal mortality rate

SDGs

Sustainable development goals

PNC

Postnatal care visit

CBHI

Community based health insurance

COR

Crude odds ratio

ICC

Intra-class correlation

VIF

Variance inflation factor

Author contributions

Conceptualization: T.H., A.A., L.D.G., J.J., J.K., B.T. Data curation: T.H., A.A., L.D.G., J.J., J.K., B.T. Formal analysis: T.H., A.A., L.D.G., J.J., J.K., B.T. Investigation: T.H., A.A., L.D.G., J.J., J.K., B.T. Methodology: T.H., A.A., L.D.G., J.J., J.K., B.T. Project administration: T.H., A.A., L.D.G., J.J., J.K., B.T. Resources: T.H., A.A., L.D.G., J.J., J.K., B.T. Software: T.H. Supervision: T.H., A.A., L.D.G., J.J., J.K., B.T. Validation: T.H, Writing—original draft: T.H. Writing—review & editing: T.H., A.A., L.D.G., J.J., J.K., B.T.

Data availability

All relevant data are within the manuscript. The data upon which these findings were developed can also be available upon request from the corresponding author.

Competing interests

The authors declare no competing interests.