Use of maternal and child health services in a community with large refugee population in Nairobi, Kenya: a cross-sectional survey


 Background

The population of urban refugees in the Eastleigh area of Nairobi, Kenya, is estimated at 100,000 to 150,000. Limited information on access to health care is available for this population. The health care a mother receives during pregnancy, at the time of delivery, and soon after delivery is important for the survival and well-being of both the mother and her child. The disadvantages in accessing health care and vulnerabilities of immigrant and refugee women are well documented in the literature. The purpose of our study was to assess the use of maternal and child health (MCH) services among the residents of Eastleigh and sources of services sought.
Methods

In July and August 2010, we surveyed households in Eastleigh. Households were chosen using a multistage cluster sampling design. A standard questionnaire on household demographics was administered to the household caretakers. If the household included a woman who delivered a baby or had a pregnancy that lasted at least 7 months within the past year, she responded to the MCH questionnaire separately.
Results

We gathered data from 673 households with 3,005 household members. There were 981 (32.6%) women of reproductive age (15–49 years), of whom 116 (11.8%) reported having a pregnancy or delivering in the past year. Of these 116 mothers, 104 (89.7%) made at least one antenatal care (ANC) visit, and 60 (63.2%) made at least four visits. Of 104 women who made at least one ANC visit, 23 (25.4%) first visited in the first trimester; 69 (64.2%) in the second trimester; and 13 (10.4%) in the third trimester. Compared with women who attended secondary school or higher, those with religious education only were twice as likely to get ANC in the early stages of pregnancy while those with no education were 3.33 times as likely. Fewer women (24%) sought delivery services at government health facilities than the 75% who delivered at private health facilities. Only three women delivered at home.
Conclusion

Use of MCH services was high among Eastleigh residents. Women actively sought care at numerous facilities, predominantly private ones. However, the quality of care in these private facilities requires further assessment.


Abstract Background
The population of urban refugees in the Eastleigh area of Nairobi, Kenya, is estimated at 100,000 to 150,000. Limited information on access to health care is available for this population. The health care a mother receives during pregnancy, at the time of delivery, and soon after delivery is important for the survival and well-being of both the mother and her child. The disadvantages in accessing health care and vulnerabilities of immigrant and refugee women are well documented in the literature. The purpose of our study was to assess the use of maternal and child health (MCH) services among the residents of Eastleigh and sources of services sought.

Methods
In July and August 2010, we surveyed households in Eastleigh. Households were chosen using a multistage cluster sampling design. A standard questionnaire on household demographics was administered to the household caretakers. If the household included a woman who delivered a baby or had a pregnancy that lasted at least 7 months within the past year, she responded to the MCH questionnaire separately.

Results
We gathered data from 673 households with 3,005 household members. There were 981 (32.6%) women of reproductive age (15-49 years), of whom 116 (11.8%) reported having a pregnancy or delivering in the past year. Of these 116 mothers, 104 (89.7%) made at least one antenatal care (ANC) visit, and 60 (63.2%) made at least four visits. Of 104 women who made at least one ANC visit, 23 (25.4%) rst visited in the rst trimester; 69 (64.2%) in the second trimester; and 13 (10.4%) in the third trimester. Compared with women who attended secondary school or higher, those with religious education only were twice as likely to get ANC in the early stages of pregnancy while those with no education were 3.33 times as likely.
Fewer women (24%) sought delivery services at government health facilities than the 75% who delivered at private health facilities. Only three women delivered at home.

Conclusion
Use of MCH services was high among Eastleigh residents. Women actively sought care at numerous facilities, predominantly private ones. However, the quality of care in these private facilities requires further assessment. 1 Background Page 3/15 According to estimates by the World Health Organization (WHO), 580,000 maternal deaths are recorded worldwide every year (1), 99% of which occur in developing countries. The incidence in sub-Saharan Africa is estimated at 640 maternal deaths for every 100,000 live births (2), most due to preventable causes; while the neonatal mortality rate is estimated at 38.8 per 1,000 live births (3). Access to and use of quality health care services are important determinants of maternal mortality (4). The health care a mother receives during pregnancy, at the time of delivery, and soon after delivery is important for the survival and well-being of both the mother and her child. The major objective of antenatal care (ANC) is to identify and address conditions associated with adverse outcomes (5). During an antenatal care visit, women are screened for complications and given advice on a range of issues, including place of delivery and referral hospitals. Information on antenatal care is of great value in identifying subgroups of women who do not use such services and in planning improvements to these services. Antenatal care is more bene cial in preventing adverse pregnancy outcomes when it is sought early in pregnancy and is continued through delivery (6). WHO recommends that a woman without complications should have at least four antenatal care visits, the rst of which should take place during the rst trimester (7). The quality of antenatal care is essential in detecting pregnancy related complications that can otherwise lead to maternal and child morbidity and mortality (8). Therefore, pregnant women should routinely receive information on the signs of complications and be tested for them during antenatal care visits.
Increasing the proportion of babies that are delivered in health facilities can be important in reducing health risks to both mother and baby. Proper medical attention and hygienic conditions during delivery can reduce the risks of complications and infection that can cause morbidity and mortality to either the mother or the baby (9).
Many maternal deaths occur during childbirth and the period immediately following the birth, with half of the deaths occurring within the rst 24 hours (10). Postnatal care is key to treating complications arising from delivery and providing the mother with important information on how to care for herself and her child. It is recommended that all women receive a check on their health within 2 days after delivery (11).

Urban refugees in Nairobi
According to recent statistics from the United Nations High Commissioner for Refugees, almost half of the world's 10.5 million refugees now live in cities and towns, compared to one-third who reside in camps (12). In Kenya, the total refugee population stood at 387,372 at the end of January 2010, of whom 272,712 (70.4%) lived in Dadaab (refugee camp in northeastern Kenya), 69,414 (17.9%) lived in Kakuma (refugee camp in northwestern Kenya), and the remaining 45,246 (11.7%)lived in Kenya's capital city of Nairobi. Uno cial estimates put the number of refugees in Nairobi at much higher gure (13) (14).
Most refugees in Nairobi live in Eastleigh. Eastleigh is a suburb on the eastern side of Nairobi inhabited largely by Somalis-both Kenyan nationals and migrants from Somalia (13) (15). The presence of this large Somali community has motivated new refugees from Somalia and from Kenya's refugee camps to settle in Eastleigh to seek better job opportunities and education (12)(13). Every day, ve to six buses arrive from Dadaab and Kakuma and several commuter buses operate between Eastleigh and Kenya towns bordering Somalia.
Little information is available regarding the health-seeking behavior of those living in Eastleigh and the services available to them. This survey was conducted to learn health-seeking behavior for maternal and child health services (MCH), identify barriers to MCH care, provide information to stakeholders working with urban refugees, and provide a foundation for future health surveillance activities.

Setting
We conducted a cross-sectional survey in July and August 2010 to assess use of MCH services in Section II of Eastleigh North division of Kamukunji district in Nairobi province. The area we surveyed is densely populated and has the highest concentration of health care facilities. This section of Eastleigh hosted largest number of urban refugees. The sole public facility is managed by the City Council of Nairobi in collaboration with the nonpro t German Society for Technical Cooperation (GTZ) and is situated at Fourth Street of Section II.

De nitions
A household was de ned as a group of people living under the same roof sharing the same cooking arrangements and under the responsibility of one head who was identi ed as the caretaker. A caretaker was de ned as a household member who was responsible for the health of the household members. A household member was de ned as someone who had slept within a compound, apartment, or room within the study area for at least 3 of the preceding 12 months, with the exception of infants, who were included if their mother met the de nition for household member

Criteria for inclusion of subjects
Household was considered eligible to be selected to participate if it was in Section II of Eastleigh north division and someone responded when approached for interviews. Any individual who met the de nition for household member was also included in the study.

Criteria for exclusion of subjects
Households in which the caretaker did not consent to participate were not surveyed. Similarly, households were not included if there was no response after the fourth attempt. Individuals living in hotels or other facilities that rent rooms by the night and visitors who stayed in the household less than 3 months were excluded from the survey and were not included in the sample size calculation.

Sampling
We used the standard sample size calculation formula for random samples in determining the sample size, which resulted in a total sample size of 785 households. We wanted to estimate maternal and child health-seeking behavior in women who delivered a baby or had a pregnancy that lasted at least 7 months within the past year and responded to a standard MCH questionnaire. We used a multistage random cluster sampling design. The rst stage clusters were blocks among Section II of Eastleigh North, the second stage clusters were plots within the blocks, and the third stage clusters were households within the plots.
Households were selected in advance and the interviewers were provided with a list. Every household was given four opportunities to be included, however, if at the fourth visit the inhabitants were still not available, the household was replaced with another household that had been randomly selected in advance.

MCH use survey
Interview teams gathered data over a 3-week period from July 29, 2010, through August 14, 2010. The data collection phase followed a 3-week training of 21 community interviewers and one week of mapping and piloting. Interviewers were grouped into teams of six, each supervised by a team leader. Community mobilizers arranged appointments in advance for interviews with caretakers. Four attempts were made to interview the household. Informed consent to participate in the survey was obtained from all interviewees.
Household caretakers answered questions on household demographics, and in certain circumstances such as when the mothers were not present, they answered questions for the women. A detailed household demographics form was administered to all consenting household caretakers, followed by a form identifying household members and screening for illnesses. The survey took between 15 and 60 minutes per household depending upon the number of people in the household and the number of mothers and deliveries.
A detailed questionnaire was verbally administered in both English and Somali to participating caretakers and mothers. If a caretaker or mother spoke another language, e.g., Oromo, Kiswahili, or Amharic, an interpreter was provided. The caretaker was asked if there was a woman in the household who had delivered a baby or had a pregnancy that lasted at least 7 months within the past year. If yes, that woman was administered the MCH questionnaire about use of antenatal, delivery, and postnatal services.
Mothers were asked whether the child who was delivered within the 12 months preceding the survey was vaccinated and whether vaccination cards were available. The interviewer copied the vaccination dates directly onto the questionnaire. When there was no vaccination card for the child, vaccination information was not recorded.

Ethical Review
The study protocol was reviewed and approved by the Scienti c Steering Committee and the Ethical Review Committee of KEMRI (Kenya Medical Research Institute). It was also reviewed by the US Centers for Disease Control and Prevention and determined to be non-research, meaning CDC Institutional Review Board review was not required.

Results
Data were entered from each questionnaire into a Microsoft Access 2007 database. Data then were reentered. The two data sets were compared and cleaned. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary North Carolina, USA). The analysis took into account clustering in the survey design, and estimates were weighted to account for sampling probability. We used descriptive statistics and cross-tabulations to describe the maternal health care use. Logistic regression was used to t models in bivariate and multivariable analysis to understand the factors associated with attending ANC in the early stages of pregnancy.   shows that the place of birth of the mother, caretaker's education, and mother's age were signi cant factors explaining the timing of ANC visits in the bivariate analysis. When these factors were tted in the multivariate analysis, non-Kenyans were threefold more likely to attend ANC in the early stages of pregnancy than Kenyans (OR 3.04 CI(1.94-4.75)). There was a signi cant correlation between the level of education and the timing of ANC visits. Mothers with more education were unlikely to attend ANC at early stages of pregnancy. Those with no education were 3.33 times more likely, while those with religious education were twofold more likely, to get ANC in the early stages of pregnancy than those with secondary school or higher education, which was statistically signi cant at p values of 0.0007 and 0.0023 respectively. Being young increased the likelihood of getting ANC in the early stages of pregnancy; those aged 15 to 24 years were 2.29 times more likely to attend ANC than those aged 30 years and above (

Discussion
We discovered that use of MCH services was high (89.7%) among Eastleigh residents. Women actively sought care at numerous facilities, predominantly in the private sector. ANC coverage for non-Kenyans was slightly higher than the coverage for the Kenya nationals (7). Interestingly, while Kenya-born mothers preferred the government facilities, mothers living in households whose caretaker was born in Somalia chose private facilities. Several hypotheses have been put forward to explain why migrant communities tend to avoid government-run facilities, including immigration status, stigma, and language barriers. In one study women reported a lack of trust in the available services due to health care worker attitude and service quality at these government facilities (16). It is important to highlight that this area is predominantly served by private health care providers. There was only one government hospital in the study site, and study participants, when asked about the facility where they sought care, gave the names of more than 30 private facilities in that small area. There is evidence that private facilities thrive in areas with few government regulations and are motivated by price competition. Generally the laws and agencies regulating Kenya's private health care sector are viewed to be inadequate (17). These facilities charge considerably higher fees than the government facilities. Previous studies in Kenya have also shown that many people working at such facilities lack training and may therefore incorrectly prescribe and dispense drugs (18). Such errors could mistreat life-threatening illnesses and contribute to the development of drug resistance.
A large Kenyan-Somali community lives in the former North Eastern Province of Kenya, bordering Somalia. Considering that more of the people we surveyed were Somalis, one might expect to see healthseeking behaviors that are similar to the Kenyan-Somali community of North Eastern Province. Use of maternal and child health services was higher in our survey participants compared to both the Kenyan ethnic Somali community and also the other non-Somali Kenyan nationals. There were higher rates of antenatal care use, hospital delivery, and postnatal care. Similarly, the rate of caesarean section was higher in the participants in our survey than the national rate of 6% (7).
Our study had two interesting ndings that were not consistent with expectations; the nding that migrant mothers had better health seeking behaviour compared to the Kenyan born mothers and the nding that lower/religious education was associated with receiving ANC earlier in pregnancy. As much as we might not have exact explanation for these two ndings, we think that the household wealth which we could not assess directly might be the main contributing factor to these both situations. Anecdotally, we were told that the migrant mothers receive remittances and are generally accepted to be in a better position economically than the local mothers. The Kenyan mothers are hustling to make ends meet hence they skip or delay health seeking while Somalia born mothers who might have not had an opportunity to attend the formal education provided in Kenya but had a religious education and also had the nancial capability to pay for the widely used private facilities in this area compared with local mothers who have some level of formal education but are struggling economically to attend the private ANC clinics. We have described in other parts of the paper that the mostly utilized health facilities are the private facilities and that the majority of our respondents were foreign born mothers. Further studies will need to carried out to conclusively ascertain the actual reason for these two speci c ndings.
This study has some limitations. Eastleigh was in the media on several occasions during the survey period because of allegations of money laundering and piracy, and there were several arrests. This environment fostered suspicion in the community toward anyone asking about money, nationality, immigration status, or other personal questions. Therefore, our study did not ask about income or refugee status. While we assume that many of the participants were urban refugees based on demographic data, we could not verify their status.
Data collected were based on the mothers' recall of events, including gestation dates and services received during the pregnancy, which might bias the results. However, we limited the questions on the most recent pregnancy to not longer than 12 months prior to the interview.
The ndings from this study indicate a high ANC utilization rate; however, fewer mothers (25.4%) are starting their ANC within the rst trimester as recommended by WHO. Also, this study has shown that migrant women, mothers born outside Kenya, have better health-seeking behaviors compared to their Kenya-born counterparts; however, they mostly seek these services at private, for-pro t facilities with varying quality of care.