Setting
According to the most recent national survey in the DRC conducted in 2018, 82% of women received at least one ANC consultation during their most recent pregnancy, and 43% received 4 or more ANC consultations [24]. Further, 85% of women had a skilled attendant at birth and 82% gave birth in a health facility. Coverage of outpatient PNC consultations reached 12% for newborns and 7% for mothers [24]. The maternal mortality ratio is high and estimated at 473/100,000 live births [25]. Due to ongoing conflicts, maternal healthcare coverage is unstable and deteriorating in several regions, hampering overall progress in maternal health outcomes in the country [26, 27].
In urban areas such as Lubumbashi, the health system is characterized by an increasing number of health facilities in the poorly regulated private sector. This has led to excessive commercialization of health care, including care related to maternal and child health [28, 29]. More than 90% of the population pays out of pocket to access basic and emergency obstetric and neonatal care, with the consequences of catastrophic expenditure for households and exacerbation of poverty [30].
The first case of COVID-19 in the DRC was identified on the 10th of March 2020[17]. The government immediately introduced a series of public health measures aimed at reducing virus transmission including the closure of bars, restaurants and schools which was followed by a declaration of a state of emergency, closure of international borders and restricting travel in and out of Kinshasa on 24 March 2020 [17]. The government implemented mandatory use of face mask and physical distancing on 20 April 2020. Mass testing for COVID-19 is highly recommended by the government, although not sufficiently available [27]. A total of 263,991 COVID-19 vaccine doses were administered as of end 2021, which below the WHO target to vaccinate 10% of the population by September 2021 [21, 28] .
This study took place in Lubumbashi city, the third-largest city in the DRC, with an area of 747 km² and an estimated population of over 2 million inhabitants in 2016 and an average density of 2,807 inhabitants per km². The city is located in Katanga Province and subdivided into eleven urban health zones (zones de santé – ZS), each with a General Reference Hospital (Hopital General de Reference - HGR) and, on average, 15 Health Areas (Aires de Santé - AS). Each health zone can have up to 35 health centers (Formations Sanitaires – FOSAs), including public and private polyclinics and health centers (Centres de Santé - CS) [28]. The 11th health zone, Kowe, is nested within the Kamalondo and Kampemba ZS, resulting in 10 different ZS to be included in this study. At the time of data collection, Lubumbashi city was experiencing the third wave of COVID-19: the first was in April 2020, the second in December 2020, and the third between April and October 2021. Until mid-January 2022, the provincial Ministry of Public Health, Hygiene and Prevention (MoPHH) had notified 4,555 cases of COVID-19 of which 109 died (2.39%)[33]. The response to the pandemic has varied over time. At the beginning of the pandemic, the response was centralized at the MoPHH. However, during the second wave the response was decentralized to the Central Offices of Health Zones (Bureau Central des Zone de Santé - BCZS) because of difficulties in coordinating notification and monitoring of cases. At the time of data collection, each ZS had a mobile team that carried out community investigations of suspected cases for the purpose of epidemiological surveillance.
Design
This cross-sectional study was conducted to assess maternal healthcare utilisation among women who were pregnant during the COVID-19 pandemic in Lubumbashi. This study used data collected in a household survey conducted in May 2021. Households were eligible for inclusion in the study if at least one woman residing within the household had been pregnant between March 2020 and May 2021. If a household was eligible, the head of the household and all women that had been pregnant within the given period were interviewed.
Population and sampling
We included all women of reproductive age (15–49 years) living in sampled households who were pregnant between March 2020 and the survey in May 2021, if they were present at the time of data collection and agreed to participate in the study. Among women who had more than one pregnancy during the study period, we collected information about all pregnancies, but for this paper we analysed information about the most recent pregnancy only. Households were selected using stratified random sampling with probability proportional to population size per ZS. Within each ZS, the avenues were randomly selected, followed by the street, then the households by means of a random walk starting from a single entry point of the street. A minimum sample size of 600 women was needed to test our main hypothesis (difference in the proportion of women using antenatal, childbirth and postnatal care) with a conservative estimate of 50% and a 95% confidence interval of 46%-54%.
Data Collection
The questionnaire included questions on women’s sociodemographic characteristics; knowledge, attitudes and practices regarding COVID-19; and a module capturing women’s use of maternal healthcare services. The questionnaire was tested and adapted by conducting a pilot study involving 89 households (not included in the analysis). The complete questionnaire in French and English can be found in Supplementary Material 1; the questions used for this analysis are set in bold. The data was collected using KoboToolbox on tablets by 10 trained enumerators.
Definitions
Use of maternal healthcare services was examined along the continuum of maternal health care with different denominators for each section. Complete care was defined as all women with a livebirth receiving at least four ANC consultations, having a skilled birth attendant during childbirth and receiving at least one outpatient PNC consultation. Women were asked if they received care for each component of care along the continuum and if not, for the reason. Reasons for not receiving each care component were formulated as open-ended questions and allowed for the recoding of multiple responses, which were categorised by the enumerators under predetermined categories or noted in free-text format. While the WHO recommends 8 ANC contacts since 2016 [29], we decided to keep the previous recommendation of four ANC consultations[30] as a threshold for complete ANC, taking into consideration that the new model is not yet fully implemented in the DRC. WHO recommends PNC at day three, between day 7 and day 14 and at 6 weeks after giving birth[31]. In line with these recommendations, women who had a child younger than two weeks at the time of survey and stayed in a health facility more than three days after birth were excluded from the estimates of PNC. Women who reported they received PNC for themselves and their baby (regardless of the timing) were defined as “receiving PNC”.
Women were also asked how much they spent in total for care received during pregnancy (= all ANC consultations) and for care received during childbirth. Women could answer either in the local currency CF (Congolese Frank) or US dollars. If the amount was given in US dollars, the exchange rate of the first of May 2021 was used to calculate the amount in CF. Furthermore, women were asked about care elements received during ANC consultations: blood pressure measurement, taking a urine sample, drawing a blood sample, measuring weight and height.
Additional questions collected information about women’s sociodemographic characteristics, including age, marital status, education, and occupation. Questions regarding women’s attitudes towards mitigation measures during the COVID-19 pandemic and COVID-19 vaccination were open-ended. Responses were categorised by the enumerators under predetermined categories or written open-text format. Open-text answers were recoded by the data analyst (AG) as follows: if women answered that COVID-19 had “a huge impact on earning daily bread and/or education of the children”, the answer was coded as “huge impact on daily life”. If the answer did not indicate clearly a positive or negative attitude, the answer was recoded as “no opinion”. The four final response categories to this question (Attitude regarding COVID-19 measures taken by authorities - Table 1) were mutually exclusive. The same approach was used for handling the data from the question regarding women’s attitude toward COVID-19 vaccination. The following responses regarding women’s attitude towards the COVID-19 vaccine were coded as “dangerous/conspiracy ideas/ineffective”: the vaccines are dangerous, it’s only for business, it’s a way to kill the black people and Africans, it is way to control the world population, the virus does not exist, and similar responses. If women said they wanted to wait before getting vaccinated or had doubts/concerns the answer was coded as “doubts/concerns”. If women did not express a clearly positive or negative attitude or the answer was not clear, the data was recoded as “no opinion”.
Analysis
The data were exported to SPSS, and checked for completeness, cleaned, coded and analysed. Descriptive statistics (percentages and their associated 95% confidence intervals, means and standard deviation) were used for to describe the characteristics of women included in the study.
We examined factors associated with the use of the full continuum of maternal healthcare by building a logistic regression model. The Akaike information criterion (AIC) was used for selecting the best fitted model[32]. Associated factors included women’s sociodemographic characteristics (mother’s age, newborn’s age, marital status, education level, occupation, ZS) and their attitudes to the COVID-19 restriction measures and vaccination. All categorical variables were recoded as dichotomous. Education was recoded into less than 4 years of secondary education versus four or more. Occupation was recoded into currently in employment or studying versus not. Marital status was recoded as currently married or cohabiting versus not. Attitudes regarding COVID-19 measures and vaccination were recoded into positive or not (including neutral). The age of the infant in days and age of the mother in years were added as continuous covariates. P-values of less than 0.05 were considered to have significant association between the outcome and the explanatory variables.
Missing data
Due to the automated skip-patterns and required answers in KoboToolbox, missing data was minimal for most variables with the exception of the question “how much did you pay for antenatal care (in total)?”, where 48 missing values were observed due to an initial incorrect skip pattern in the questionnaire. For descriptive analysis missing values are reported as such and in logistic regression analysis, pairwise deletion was applied, meaning that cases with missing values for any of the component variables of continuum of maternal care or independent variables in the model were excluded. This means that 28 out of 317 (8.83%) women with a livebirth were excluded from all analyses regarding completing the maternal health care continuum due to missing data regarding ANC, skilled birth attendance or PNC consultations for mother and baby.