Level of completing continuum maternity care and the associated factors in the pastoralist and agrarian districts of the Ari and South Omo zones, Ethiopia, 2024: a comparative study

Abstract

Background

Completing continuous maternity care (COCMC) is crucial for preventing morbidity and mortality, but a low proportion of mothers use it. There is a paucity of evidence about the level of COCMC and the associated factors in agro-pastoralist communities.

Objective

To identify and compare the level of COCMC and associated factors among pastoralist and agrarian communities in Ari and South Omo zones.

Design

A community-based comparative cross-sectional study was conducted from March to May 2024.

Setting

A comparative study on COCMC was conducted in Ari and South Omo zones of South Ethiopia.

Participants

A multistage sampling technique was employed to select 632 women who gave birth 6 months before data collection.

Outcome measures

The outcome variables were the level of COCMC and its associated factors. The results of outcome variables were presented with proportions and adjusted ORs (AORs) with the corresponding 95% CIs. Data were collected with a standard questionnaire via the Kobo toolbox and analysed with the R software. Due to sparse data about the proportion of COCMC, a penalised likelihood (Firth logistic regression) model was employed. A p value <0.05 was considered a cut-off for statistical significance.

Results

A total of 632 women participated in this study giving the response rate of 97.4%. The overall level of COCMC was 46 (7.3%; 95% CI 5.3% to 9.3%), 27 (7.5%) among pastoralists and 19 (7.0%) among agrarian districts, the difference was not statistically significant (p value=0.84). The COCMC was associated with living in urban areas (AOR: 2.76, 95% CI 1.17 to 6.34), women’s ability to pass a decision on household resources (AOR: 2.57, 95% CI 1.17 to 5.55), being a government employee (AOR: 8.71, 95% CI 1.38 to 50.94), middle level household economy (AOR: 3.81, 95% CI 1.16 to 15.83) and receiving no home-to-home healthcare (AOR: 0.33, 95% CI 0.14 to 0.70).

Conclusion and recommendations

The level of COCMC is low in the study area. Therefore, equitable interventions targeting both communities, enhancing maternal decision-making power to use resources for maternal services, focusing on the housewives and those engaged in daily work and strengthening integrated home-based healthcare are all recommended.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The research used a short 6-month reference period to minimise memory errors, applied current WHO and Ethiopian Ministry of Health standards for antenatal care visits and compared maternal health between agrarian and pastoralist groups.

• Wide CIs suggest the sample size was insufficient, meaning the results should be interpreted with caution.

• Self-reported data likely introduced recall bias regarding the number of services used.

• Including urban residents in the study might have hidden the true differences between the two target communities.

• While the findings offer insights for similar agro-ecological regions in Ethiopia, further studies are required to reaffirm the evidence.

Background

Every mother is recommended to conduct at least eight antenatal care contacts (ANC8⁺), give birth with the assistance of a skilled birth-attendant (SBA) and conduct postnatal care (PNC) within the first 2 days after childbirth to prevent and manage complications that may arise during pregnancy, childbirth and the postnatal period to reduce maternal and newborn morbidity, mortality and unfavourable birth outcomes.^1–3 It was planned to reduce maternal mortality to less than 70 per 100 000 live births (LBs) by 2030, and maternal health and well-being were put as a core development agenda.⁴ Regardless of this commitment, in 2020, 287 000 women died during and following pregnancy and childbirth from the most common preventable causes.⁵

Every year, more than 303 000 women die from preventable causes associated with pregnancy and childbirth, and developing countries account for 99% of global maternal deaths. Sub-Saharan Africa and southern Asia account for more than 85% of the burden.^6–8 Although a significant reduction has been reported over time, Ethiopia is one of the countries with the highest rates of maternal mortality: 412/100 000 LBs in the 5 years preceding the 2016 Ethiopian demographic and health survey (EDHS)⁹ and 401/100 000 LB in EDHS 2019.¹⁰ The situation is unacceptably high and is left behind the target set to be achieved by 2020,^{10 11} and which calls for the strengthening of integrated and uninterrupted maternal healthcare services.^{3 12}

Completing continuous maternity care (COCMC) is a form of maternal healthcare using ANC8⁺ contacts, giving childbirth with the assistance of SBAs and conducting early PNC visits continuously.^{12 13} All of these services are associated with a reduction of maternal mortality and morbidity.^{2 6 14} More than 95% of mothers continually obtain all three services in developed regions, whereas a lower proportion of their counterparts in developing countries use these services.^{8 15 16}

In the 5 years preceding 2019 EDHS, only 43%, 48% and 35% of women used fourth ANC⁺, SBA and early PNC, respectively,¹⁷ and were found to be below half the level of COCMC recommended by the WHO.^{14 18} In the Southern Nations, Nationalities and Peoples’ Region, antenatal care (ANC), SBA and early PNC were found to be 69.4%, 50.1% and 32%, respectively.¹⁹ The rates of ANC, SBA and PNC services utilisation in the South Omo Zone are 24%, 33.7% and 47.9%, respectively,²⁰ and are associated with maternal literacy, living in urban areas and living in economically better-off households.¹⁰ Studies carried out in Ethiopia have shown that 11.2%–37.2% of mothers completed maternity healthcare on a continuum path, and their completion is positively correlated with increased maternal education, early ANC initiation, knowing the maternal danger signs, having a birth preparedness and complication readiness plans, having planned pregnancies and having decision-making autonomy.^21–27 Service utilisation is left behind the global 2030 targets: 90% coverage of four or more ANC contacts, 90% of births attended by skilled personnel and 80% early routine PNC (within 2 days).¹

There is a lack of robust and consistent evidence on the extent and factors associated with the completion of continuous maternity care in Ethiopia and Ari and South Omo Zones. The previous studies used four visits to define ANC uptake, which does not correctly estimate the level of COCMC according to the current WHO and the Ethiopian Ministry of Health recommendations. Most of the studies used 12 months of reference period after childbirth which might introduce recall bias. Moreover, almost all of the studies were conducted in agrarian districts, and little is known about pastoralist communities to guide context-specific interventions and ensure equity. No previous study examined and compared the level of completion of continuous maternity care among pastoralist and agrarian communities using the current WHO-recommended ANC8⁺ schedule to define COCMC.

Factors such as maternal age, number of children ever born to a woman, religion, mother’s educational level, mother’s occupation, marital status, family income, media exposure, current work status, time taken to reach health facility, cultural influence, place of residence, home-to-home visit by health workers, decision-making power to use household resources for maternal health service utilisation, and knowledge and attitude towards maternal healthcare are hypothesised to be correlated with completing continuum maternity care. Despite this, the consistent evidence on the factors associated with the COCMC in the Ari and South Omo Zones is lacking. Therefore, this study aimed to determine the level of completion of continuous maternity care and associated factors among mothers who gave birth within 6 months before data collection in pastoralist and agrarian communities of the Ari and South Omo zones in southern Ethiopia.

Method

Study area

The study was conducted in the South Omo and Ari zones of Southern Ethiopia. The residents were agrarians, semi-agrarians and pastoralists with different cultural backgrounds. According to the 2014 Ethiopian fiscal year Central Statistical Agency estimates, the zones had a total population of 821 405, with a male-to-female ratio of one. There were 11 woredas in the zones, the majority of the inhabitants relied mainly on animal rearing, and the rest were farmers growing crops (maize, lentils and animals (cattle, goat and sheep). With respect to access to health services, one zonal hospital, three district hospitals together with health centres and health posts served the population.

Study design and period

A community-based comparative cross-sectional study was carried out from March to May 2024.

Source and study populations

The source and study population were women of reproductive age who gave birth within 6 months before the data collection period in the study area.

Inclusion and exclusion criteria

All women of reproductive age (aged 15–49 years) who gave birth to a baby within 6 months before the data collection period were included, whereas women with missing observations of the outcome variables were excluded.

Sampling procedure

The South Ari district was randomly selected from the four agrarian districts, and the Male (semi-pastoralist) and Benatsemay (pastoralist) districts were randomly selected from the South Omo Zone via the simple random sampling (SRS). The catchment areas of health centres were selected at random from each district. In this context, the catchment areas of the Keyafer, Alduba and Shala health centres were selected from the Benatsemay district. The Koybe and Damiker health centre catchments were selected at random from the Male district. The catchment areas of Gazer Hospital, Getamer, Shekamer and Meytser health centres were randomly selected from the South Ari district. The list of women with babies aged <6 months was prepared via collaborating with health extension workers and health development armies. Health extension workers were blinded about the particular topic and the target population (whether woman or babies) of our study to reduce/avoid in-system bias. Finally, the study participants were selected from lists of mothers who gave birth within 6 months preceding the data collection period by SRS on the basis of size of the target population.

Sample size determination

The sample size for the current study was determined via the double population formula with an estimated completion of a continuum of maternity care of 12.1% in West Gojam²¹ and 9.5% in emerging regions of Ethiopia (Afar, Somali, Gambela and Benishangul Gumuz),²⁸ 95% confidence level, 5% precision (d), using a design effect of two and adding nonresponse rate of 10% as follows:

$${\displaystyle {\displaystyle \mathrm{n}=\frac{{\left({\mathrm{Z}}_{1-\frac{\alpha }{2}}\right)}^2\left(\left({\mathrm{p}}_1(1-{\mathrm{p}}_1)\right)+\left({\mathrm{p}}_2(1-{\mathrm{p}}_2)\right)\right)}{{0.05}^2}}}$$

$${\displaystyle {\displaystyle \mathrm{n}=\frac{{1.96}^2\ast \left((0.121\ast 0.879)+(0.095\ast 0.905)\right)}{{0.05}^2}=295}}$$

$${\displaystyle {\displaystyle =295\ast 2=590}}$$

After adding the 10% non-response rate, a total sample size became 649. The sample size calculated for determining the factors with the power of 80% was found to be lower than 649, thus the sample size calculated for the first objective was ultimately used to determine both objectives.

Data collection procedures and quality control measures

We collected data via an interviewer-administered and tested questionnaire via the Kobo toolbox. The questionnaire was first developed in English, then translated into Amharic and back-translated into English by two language experts to check its consistency. Pretesting was conducted in a Baytimal catchment area. The questionnaire covered socio-demographic and economic aspects, accessibility and utilisation of health services, behavioural factors, women’s decision-making power, husbands’ involvement in maternal service utilisation and questions to assess mothers’ attitudes. The data collectors and supervisors were trained on the objective of the study, data collection procedures, data collection tools, how to contact respondents, how to ensure data quality, confidentiality and respondents’ rights before the data collection date. The completeness and consistency of the questionnaires were checked daily by the responsible investigator and appropriate measurements were taken. Data cleaning was continued until the descriptive statistical analyses were completed.

Variables of the study

The dependent variable was the COCMC. The independent variables were the mother’s age, mother’s educational level, religion, mother’s occupation, marital status, family income, media exposure, current work status, time taken to reach health facility, cultural influence, place of residence, home to home visit by health workers, decision-making power to use household resources for maternal health service utilisation and feeling towards continuum maternity care. The explanatory variables were identified through reading previous similar studies.

Operational definition

Completion of a continuum of maternity healthcare

A woman had completed a continuum of maternity healthcare if she had ANC8⁺ visits in her recent pregnancy, a recent delivery was assisted by an SBA at a health facility and had PNC within 2 days after recent childbirth.²⁹

ANC visit

Having ANC8⁺ visits from conception to childbirth.^{2 3}

Health worker

Health extension or other health professionals.

Data analysis

The data collected via the Kobo toolbox were exported to a statistical package for social sciences −26 then to R software for analysis. The wealth index was generated through principal component analysis (PCA) of household assets, housing conditions and facilities. An asset owned by more than 95% or less than 5% of households was excluded during the data reduction process. Categorical variables with multiple options were transformed into binary categories (1 and 0)³⁰. In this regard, 16 variables were used to generate PCA (availability of latrine, ownership of table with chairs, availability of electricity, ownership of radio or television, use of charcoal for cooking, source of water (improved or unimproved), having a bank account, ownership of a motorcycle, ownership of telephone, ownership of bed with mattress, cash receipt from the safety net and other support, household enrolment status in community-based health insurance, private or living in charge-free home, the floor of the building (mud or other), roof (durable or other) and wall of the dwelling.¹⁰

Variables with too high (above 0.9) or too low (below 0.1) correlation during PCA were removed. The having a bank account and cash receipt from safety net and other financial sources had high correlation (0.92) so that was removed. The Kaiser-Meyer-Olkin Measure of Sampling Adequacy was 0.83 (above 0.6), and Bartlett’s test of sphericity was <0.0001 (significant). Approximately 61% of the total variance was explained by the first five components, which had initial eigenvalues >1. The first component was considered a wealth index, as it accounted for the majority of the variance.

The internal consistency of variables (maternal perceived decision-making power to use maternal healthcare) measured with the Likert scale was tested with a Cronbach’s alpha value ≥0.7. In this regard, a variable with the highest mean score among others (husbands must be involved in decision-making) had a reliability coefficient of 0.901. The ANC received by a woman was dichotomised into less than eight and eight or more contacts. A childbirth that took place at hospitals, health centres and clinics was considered SBA, and PNC was considered if a woman received services within the first 2 days of childbirth.

A multicollinearity test was carried out to assess interdependence among predictor variables using a variable inflation factor (VIF) of <10 or a tolerance of >0.1: no variables were found to be interdependent, and the highest VIF was 2.228. To overcome the sparse outcome data (only 46 COCMC from 632 participants) that do not allow more than four independent variables in a standard logistic regression, the Firth penalised logistic regression model was conducted to identify the associated factors. The overall model was statistically significant according to the penalised likelihood ratio test (χ² (at the degree of freedom=16)=50.08, p<0.001, n=632). The discriminative power of the final model was evaluated using the area under the receiver operating characteristic; 0.74 (95% CI 0.66 to 0.81) at the p value below 0.001.

The findings of the descriptive analysis are presented with proportions with 95% CIs, and the multivariable Firth penalised logistic regression outputs are presented with AORs with their respective 95% CIs at a statistical significance level <0.05.

Patient and public involvement

The finding of this study was disseminated via validation workshop and conference presentation at Jinka University where representatives of the public and private sectors, the public and academia participated.

Results

Sociodemographic characteristics of the respondents

632 mothers who gave birth within 6 months preceding the data collection participated, and the response rate was 97.4%. Approximately 43% were from the Ari Zone, and the rest were from the Male and Benatsemay districts of the South Omo Zone; 29.0% were from the Male district and 28.3% were from the Benatsemay district. The median age of the respondents was 27 (IQR=7) years. About half (49.8%) of the respondents reported that they could at least read and write. Three-fourths of the respondents were from male-headed households, and about 6 in 10, 372 (58.9%), were 25–35 years old (table 1).

Table 1. Sociodemographic characteristics of the study participants, Ari and South Omo zones, 2024 (n=632).

Characteristics	Frequency	%
Districts
South Ari	270	42.7
Male	183	29.0
Benatsemay	179	28.3
Kebele types
Urban	75	11.9
Agrarian	226	35.8
Pastoralist	331	52.4
Sex of the head of the households
Male	471	74.5
Female	161	25.5
Educational status
Illiterate	317	50.2
Grade 1–8	131	20.7
Grade 9–12	132	20.9
Above 12th grade	52	8.2
Age of the respondents
<25 years	183	29.0
25–34 years	372	58.9
≥35 years	77	12.2
Maternal occupation
Housewife	454	71.8
Self-business	75	11.9
Government	55	8.7
Others (non-governmental, daily labour and support)	50	7.91
Wealth Index
The poorest	137	21.7
Poor	118	18.7
Middle	125	19.8
Rich	127	20.1
The richest	125	19.8

Slightly more than half, 326 (51.6%) of the partners of the respondents were literate, and the median academic level attained was 10th (IQR=5) grade. The median number of children ever born to a woman was 4 (IQR=2).

Information on maternal health services and sources

About 17% of the respondents had never heard of any type of maternal health service; 27 (4.3%), 22 (3.5%) and 57 (9%) had not heard of ANC, delivery and PNC services, respectively. The dominant sources of information on maternal health services were health workers (health extension workers and health professionals); 479 (75.8%). Family members, friends, schools, mass media and churches served as sources of information on maternal health services for 340 (53.8%), 313 (49.5%), 236 (37.3%), 229 (36.2%) and 197 (31.2%), respectively.

More than one-third (37%) did not know that a woman can obtain maternal health services at hospitals. About one in four of the respondents (23.9%) did not know that maternal health services can be obtained at health centres, and 391 (61.9%) and 316 (50%) knew that a woman can receive maternal health services at health posts and home-to-home visits, respectively. Fewer than one-third of the respondents (32%) mentioned that a woman can receive maternal health services at private health facilities, and 9 (1.4%) reported that a woman can obtain care in traditional places.

Concerning women’s feelings about receiving maternal health services given by health workers, 16 (2.5%) felt that ANC by health workers was not important, 29 (4.5%) felt neutral about the importance of ANC given by health workers and 587 (93.0%) felt that ANC given by health workers was important. About 92% (580) felt that delivery assisted by health workers was important, and 28 (4.5%) were neutral regarding the importance of delivery assisted by health workers. 562 of the respondents felt that obtaining PNC by health workers was important, whereas the remaining 48 (7.6%) felt neutral, and 22 (3.5%) felt that the PNC health service was not important.

Maternal decision-making power and completion of continuum maternity care

About one-fourth of the study participants, 148 (23.4%), perceived that the husband was the only decision-maker for a woman to conduct an ANC visit while 70 (11%) were neutral and 414 (65.5%) disagreed on who should decide for a mother to use services. Concerning passing of the decision on a mother to give childbirth at a health facility, 57 (9%), 162 (25.6%) and 413 (65.4%) were neutral, agreed and disagreed, respectively, that husbands are the only persons to make decisions for a woman to give birth at health facilities. Concerning decision-making, a mother receiving PNC, 88 (14%), 144 (22.7%) and 400 (56.3%) were neutral, agreed and disagreed, respectively, on husbands being the only persons to make decisions. Regarding partner involvement in decision-making, more than half of the respondents (52.7%) agreed that the husband must be involved, whereas 248 (39.2%) disagreed and the remaining were neutral. More than three-fourths, 485 (76.7%), agreed that using maternal health services does not violate religious commandments. About 8 in a 100 (7.9%) agreed that using maternal care violates religious commandments, while 61 (9.7%) were neutral on this regard. More than three-fourths (82.4%) disagreed that using maternal healthcare violates cultural values, 50 (7.9%) agreed, while the rest were neutral. About 1 in 10 (9.6%) preferred traditional healers to health facilities. The majority (83.3%) agreed that maternal healthcare received at healthcare facilities was better than care given by traditional caregivers.

Maternal health service utilisation

The majority, 553 (87.5%) of the respondents had at least one ANC visit. Almost three-fourths (72.5%) had four or fewer ANC visits, and only 50 (7.9%) had eight or more contacts (table 2).

Table 2. ANC contacts conducted by the respondents in Ari and South Omo zones, 2024 (n=632).

Antenatal care visit frequency	Frequency (n=632)	Per cent
No visits	79	12.5
Once	8	1.2
Two times	41	6.5
Three times	127	20.1
Four times	215	34.0
Five times	63	10.0
Six times	39	6.2
Seven times	10	1.6
Eight or more times	50	7.9

ANC, antenatal care.

About 8% and 4.4% (95% CI difference 3.4 to 3.6) of women who gave birth within 6 months preceding the data collection had ANC8⁺ visits in pastoralist and agrarian districts, respectively.

The majority of them gave birth at health facilities; 196 (31%) at hospitals and 352 (55.7%) at health centres. The remaining 84 women gave birth to their recent baby at home, and the reasons mentioned for doing so were sudden onset of labour (4.3%), having no transport access when labour happened (3.3%), the labour was easy (1.9%), traditional birth attendants assisted (1.4%) and other reasons (1.1%). More than two-thirds of the respondents received PNC via home-to-home visit by health workers or themselves visiting health facilities within 45 days of childbirth. Only 128 (20.3%) of the women who gave birth within 6 months of data collection obtained PNC within the first 3 days of childbirth.

Only 7.3% (95% CI 5.3% to 9.3%); 7.5% and 7.0%, pastoralist and agrarian women respectively completed the continuous maternity care.

Penalised likelihood (Firth logistic regression) analysis of the level of completion of continuum maternity care

12 variables (maternal age, maternal education, marital status, maternal occupation, residence, household living standard (wealth index), maternal decision-making power to use domestic resources for maternal healthcare, maternal feelings about ANC, delivery and PNC by health workers, domestic spousal violence, access to diverse maternal health information sources, home-to-home visits conducted by health workers and a feeling that using maternal health service violates traditional commandments) were included in the penalised likelihood (Firth logistic regression) model. Five variables (place of residence, maternal decision-making on household resources, household economy, current occupation (being a government employee) and receiving frequent home-to-home visits by health workers) were statistically significantly associated with COCMC.

The odds of completing continuum maternity care among women living in urban areas was 2.76 (95% CI 1.17 to 6.34) times as high as the pastoralist mothers. Mothers who had the power to pass a decision on using domestic resources for maternal health services were 2.57 (95% CI 1.17 to 5.55) times more likely to COCMC than their counterparts. Compared with the mothers from the poorest households, those from households with middle level economy had 3.81 (95% CI 1.16 to 15.83) increased odds of completing maternity care in a continuum manner. Government employed women had 8.71 (95% CI 1.38 to 50.94) times higher odds of completing continuum maternity care compared with the housewives. The mothers who did not get healthcare by health workers by repeated home-to-home visits had 68% (0.32 (95% CI 0.13 to 0.71)) reduced odds of COCMC than those who got frequent visits (table 3).

Table 3. Firth penalised logistic regression analysis of factors associated with the level of COCMC, in the Ari and South Omo zones, 2024 (n=632).

Variable category	Completed CMC		COR (95% CI)	AOR (95% CI)
	No (n=586)	Yes (n=46)
Age of respondents
≤24 years	170	13	1	1
25–34 years	341	31	1.16 (0.61 to 2.33)	1.53 (0.74 to 3.31)
≥35 years	75	2	0.42 (0.08 to 1.42)	0.42 (0.07 to 1.66)
Maternal education
Cannot write or read	302	15	1	1
Grade 1–8	122	9	1.51 (0.63 to 3.43)	1.69 (0.58 to 4.80)
Grade 9–12	114	18	3.15 (1.55 to 6.48)	2.04 (0.69 to 6.09)
Above grade 12	48	4	1.81 (0.54 to 5.03)	0.16 (0.02 to 1.28)
Marital status
Married	538	44	1	1
Single, widow and divorced	48	2	0.51 (0.12 to 2.16)	0.38 (0.07 to 1.43)
Current occupation
House wife	428	26	1	1
Private business	63	12	3.12 (1.51 to 6.53)	2.04 (0.76 to 5.37)
Government employee	48	7	2.40 (0.98 to 5.82)	8.7 (1.38 to 50.94)*
Daily labour	47	1	0.35 (0.04 to 2.64)	0.41 (0.04 to 2.01)
Residence
Urban	60	15	4.33 (2.07 to 8.97)	2.76 (1.17 to 6.34)*
Agrarian	214	13	1.06 (0.51 to 2.18)	0.74 (0.28 to 1.84)
Pastoralist	312	18	1	1
Household economy
The poorest	134	3	1	1
Poor	113	5	1.98 (0.46 to 8.45)	1.81 (0.45 to 8.28)
Middle	113	12	4.74 (1.31 to 17.22)	3.81 (1.16 to 15.83)*
Rich	117	10	3.81 (1.03 to 14.20)	1.48 (0.36 to 7.00)
The richest	109	16	6.56 (1.86 to 23.08)	3.46 (0.85 to 17.00)
Mother could decide to use household resources for maternal healthcare
No	473	30	1	1
Yes	113	16	2.25 (1.17 to 4.20)	2.57 (1.17 to 5.55)*
Health workers did home-to-home visits
No	282	10	0.32 (0.13 to 0.66)	0.33 (0.14 to 0.70)*
Yes	304	36	1	1
The mother feels that ANC, SBA and PNC are all important
No	77	3	0.53 (0.14 to 1.41)	1.03 (0.25 to 3.22)
Yes	509	43	1	1
The mother had diverse sources of information on maternal healthcare
No	250	14	0.59 (0.31 to 1.12)	1.27 (0.54 to 2.98)
Yes	336	32	1	1
Using maternal health service violates traditional commandments
Disagree	485	36	1	1
Agree	101	10	1.33 (0.64 to 2.78)	1.68 (0.73 to 3.67)
The mother faced spousal violence
No	217	24	1.85 (1.02 to 3.38)	1.23 (0.61 to 2.43)
Yes	369	22	1	1

^*The variables are statistically significantly associated with the dependent variable.

ANC, antenatal care; AOR, adjusted OR; COCMC, completing continuous maternity care; COR, crude OR; PNC, postnatal care; SBA, skilled birth-attendant.

Discussion

The present study was conducted in the Ari and South Omo zones, southern Ethiopia, to examine and compare the level of COCMC among mothers who gave birth within 6 months prior to the data collection period in the agrarian and pastoralist zones. No statistically significant difference was found between the two zones (p value=0.84), and the overall level was 7.3% (7.5% in the pastoralist zone and 7.0% in the agrarian zone). The pooled COCMC was associated with maternal decision-making power regarding the use of domestic resources for maternal health services, residence in urban areas, household economy, current occupation of the women and home visits by health workers to provide integrated health services to mothers.

Our study revealed that the overall level of COCMC was 7.3% (95% CI 5.3% to 9.3%) in the study area. The level of COCMC was lower than reported in the studies conducted in the West Gojam zone; 12.1%,²¹ multilevel analysis of Mini Ethiopian Demographic and Health Survey; 12.9%, rural Ethiopia³¹ and Arbaminch Zuria district; 9.7%.³² These inconsistencies may be due to previous studies considering ANC completion as having four or more visits, which resulted in a higher COCMC rate. However, the completion rate of continuous maternity care remains low in Ethiopia, indicating that the study area and the country as a whole are unlikely to achieve the target set by the Sustainable Development Goal if the current situation persists. Therefore, there is a need to strengthen maternal healthcare activities nationwide. Rural and pastoralist communities require special attention to ensure completion of the continuum of maternity care. Achieving a good level of COCMC requires integrating health services, such as preconception care and early identification of pregnant women, and linking them with other health services, including community health days and various programme campaigns. Only 50 (7.9%) of the current study participants completed their eighth ANC visit. This low ANC8⁺ rate may be due to weaknesses in early identification of pregnant women and referral to the service.

The study revealed that women who had lived in urban areas were 2.76 (95% CI 1.17 to 6.34) times more likely to complete continuous maternity care than the pastoralist women were. This study is in agreement with a study conducted in Northwest Ethiopia³³ and a systematic review and meta-analysis.^{34 35} This implies that women in urban areas access health facilities, transportation services and health information more easily than their rural and pastoral counterparts. This calls for improved access to healthcare and equity in healthcare utilisation among all community groups and special consideration of the rural and pastoralist segments of the communities.

A woman from economically better-off households had increased odds of completing continuum maternity care. Compared with the poorest counterparts, the women from households with middle level economy 3.81 (95% CI 1.16 to 15.83) increased odds of completion of care. This is consistent with the evidence revealed by the systematic review and meta-analysis in Africa.^{10 34} It indicates that economically advanced women can easily decide on domestic and personal resources for maternal care. They can easily manage transport and other costs related to maternal healthcare utilisation.

Compared with the housewives, the government employed women had 8.71 (95% CI 1.38 to 50.94) higher odds of completing continuum maternity care. This might be because the government employees were more educated, close to health information and they had access to health services. The majority of the government employees live in the urban areas where access to information, healthcare workers and health services improved. This implies that the equitable service delivery is limited in the study area.

Mothers who could decide to use domestic resources for maternal healthcare had 2.57 (95% CI 1.17 to 5.55) times higher odds of COCMC compared with those with less decision-making power. This finding is inconsistent with studies conducted in Sub-Saharan Africa and South Asia assessing the continuum of care pathway for maternal health³⁶ and North West Ethiopia.³³ The more autonomous a mother is, the greater the likelihood of completing the care pathway. Thus, empowering women to make decisions about household resources and fostering supportive households is essential.

The odds of COCMC among the women who had not received home-to-home integrated healthcare by health workers were 68% (0.33 (95% CI 0.14 to 0.70)) reduced than those among mothers who had received frequent home-to-home maternal healthcare. The finding is in accordance with the findings revealed by a study carried out with the national health extension programme assessment survey; women who were exposed to the health extension programmes and had improved awareness were more likely to complete the continuum of care.³¹ Home-to-home provision of maternal health services saves the time of women, reduces costs related to transportation to health facilities and provides comfort as it is rendered in their own homes.

Limitations and strengths of this study

As a limitation, the wider CIs indicate inadequacy of the sample size and non-significant associations might be reflections of low power rather than true lack of association. The outcome variables were self-reported so that recall bias might be introduced. Thus, caution is needed when interpreting the findings of this study. We call for further investigations with the same objective and variable definitions to substantiate the evidence generated by our study. Nevertheless, these findings offer critical insights into maternal health disparities that are likely applicable to other pastoralist and agrarian populations of Ethiopia and other similar settings. The identification and listing of target population was done in collaboration and health development armies, as a result, there might be incomplete registration therefore precaution regarding a selection bias is required. The number and type of services used by the woman was self-reported therefore the recall bias likely led to an underestimation or overestimation of the true COCMC level, therefore, precaution is needed while interpreting. The quality of services and attitudes of providers should have been investigated. The inclusion of urban residents from each category might mask the real difference in COCMC between pastoralist and agrarian communities. As a strength, this study used a shorter reference period (6 months) to measure the level of COCMC, defined ANC visits according to the current WHO and Ethiopian Ministry of Health recommended ANC schedule and compared the level of COCMC among agrarian and pastoralist communities.

Conclusion and recommendations

The level of COCMC in the study area was very low and did not meet the standard recommended by the Ethiopian MOH and the WHO. To improve this, we recommend strengthening home-to-home maternal healthcare through the existing health system and periodic campaigns, empowering women to make decisions on domestic resources for maternal healthcare utilisation through partner engagement, education and training and intensifying the early identification of pregnant women and linking them with ANC in rural and pastoral communities. Enhancing household economy via income-generating activities and economic support is recommended to ensure completing the care among the poorest segments. Home-to-home visits to deliver comprehensive care, community-based health services provision improving the functionality of health development team is recommended to ensure healthcare utilisation by the housewives.

Data availability statement

Data are available in a public, open access repository.