Factors influencing deliveries at health facilities in a rural Maasai Community in Magadi sub-County, Kenya

The essence of the “Boma” model is that it recognizes the Maasai homestead, which is known as a “Boma” in Swahili, as the ideal interface between the semi-migratory community and MNCH health care delivery. It focuses on developing linkages between the community (level 1 of the health care system) and the formal health care system, particularly level two (dispensary), level three (health centre) and level four (county hospital) of the health care system. Project objectives included 1) to increase utilization of MNCH services provided by skilled health professionals in line with ministry of health guidelines; 2) to improve case management of childhood illnesses at community and health facility; and 3) to strengthen the community health information system and link it to the health facility information system.

The “Boma” model achieved a number of significant successes in increasing utilization of MNCH services provided by skilled health professionals in line with guidelines from the Ministry of Health. The proportion of pregnant women who attended four ANC visits increased from 4% in 2007 to 32% in 2010 and 59% by 2012 [10]. The 2012 end-term evaluation for the project revealed that more mothers were initiating ANC early in the first trimester (29% compared to 22% at baseline) [10]. Skilled attendant deliveries improved from 14% to 24%; however, this achievement is below the national average of 61% [7].

This was a cross sectional study using mixed methods approach to assess factors that influence health facility delivery. Quantitative research methods were used to assess coverage of facility births by various equity-related characteristics, while qualitative research methods identified motivators and barriers to health facility delivery.

This study was conducted in Entasopia community, which is located in Magadi sub-county of Kajiado County, Kenya. The area is one of the arid and semi-arid lands (ASALs) in Kenya, which are classified by the Kenya government as disadvantaged with respect to equitable distribution of national resources, infrastructure and access to essential social services including healthcare. ASAL regions are home to nomadic and semi-nomadic communities such as the Maasai and Turkana who have poor MNCH indicators. The poor MNCH indicators among these communities are attributed to a complexity of factors including inadequate, ill-equipped and poorly staffed health facilities; long distances to health facilities; migratory lifestyles; conservative cultural practices; and gender biases [10]. Magadi County was chosen as a study location because initially the “Boma” model was implemented there prior to extending it to Samburu County. The county has eight community units (Oldonyonkie, Shompole, Olkeri, Olkiramatian, Oloika, Pakase, and Entasopia, which contains two community units). Entasopia was the only community unit that had a relatively better quality health facility with 24-h obstetric care, and since the purpose of the case study was to understand why some women use and others do not use health facilities for delivery beyond lack of geographic access, it was decided to limit the case study to this unit. Data was collected from September 22 to October 3, 2014.

Households were used as the unit of selection. All households within community health units in the study site were mapped and given a unique number which gave rise to sampling frame. Systematic sampling technique was used to sample 50% of the total households. In the sampled households, all women aged between18–49 years and had given birth in the last 24 months were interviewed.

A survey tool was administered to women who had delivered their last child in the last 24 months (Additional file 1). Data were collected on paper forms and entered into two electronic spreadsheets using Microsoft Access. Ten percent of the questionnaires were randomly selected and rechecked for correct entry against the two spreadsheets, and the spreadsheet that had the lowest error percentage (1.8%) was selected for final analysis. Prior to the analysis, further data cleaning was carried out using frequency distributions and cross-tabulations of the dataset. The data were also coded and the required composite variables created. Finally, all personal identifiable characteristics such as name and date of birth were removed from the database. All analyses were conducted using SPSS version 20.

Five quantitative variables were involved; age, parity, antenatal care visits, distance from the health facility and wealth index. Antenatal care visits and distance from the health facility variables were each grouped into two categories (Antenatal care visits; fewer than four/ four or more, distance from the health facility; less than 3kms/ more than 3kms). A wealth index was developed and grouped into wealth quintiles (lowest, low, medium, high and highest). Wealth quintiles were calculated with standard techniques based on principal component analysis on the reported ownership of household possessions, main sources of income, lighting, and cooking fuel. Characteristics that did not show variation across households (e.g. ownership of mobile phones or type of toilets) were not included in principal component analysis. The variables included in the development of the wealth quintiles for the analysis included: total livestock, head of household occupation, woman occupation, ownership of (bicycle, motorcycle, radio, TV, foam mattress, gas cooker, furniture, solar, other). Variables with very few respondents (e.g. ownership of car, refrigerator and sewing machine), were merged to ‘other’ category. Age and parity were not grouped during analysis.

As the target subjects of the “Boma” model interventions, we considered women who had delivered their youngest child in the last 2 years. We also sought the perspectives of homestead heads as they had been targeted by the “Boma” model as entry points to influence decision making on health issues within the household. CHVs, the provincial administrators (chiefs), health care workers, village elders and TBAs also played a pivotal role in the implementation of the “Boma” model either as links to the health facility or service providers, and their views were also explored.

Two study villages, one within a short walking distance to the health centre and therefore with optimal access and one approximately more than 5 km away but where access to the health facility was still possible, were chosen for the qualitative portion of the study. Within each village, four women who delivered in a health facility and four women of similar parity who delivered at home were interviewed using a semi-structured in-depth interview guide. The key decision-influencers of women, that is, husband/mother/mother-in-law, for two women who delivered in a health facility and two who did not were also interviewed to better understand the decision-making process. In addition, the village elder and the traditional birth attendants in each village were interviewed. Finally, three focus group discussions with CHVs, the provincial administrators (Chiefs), and health care workers were conducted.

The maternity register at the Entasopia Health Center was used to identify women residing in the two villages and who had delivered there within the last 2 years. CHVs attached to the respective villages guided the interviewers to the “Bomas” of the identified women, as well as to the Bomas of women who had delivered at home. Since the CHVs were well known by the respondents, they assisted in introducing the research assistants who then introduced the study and sought informed consent prior to conducting the interviews. Table 1 outlines the type and number of respondents interviewed and the qualitative data collection method used.

Four research assistants who carried out the interviews underwent four-day training in conducting in-depth interviews and focus group discussions, including note taking, debriefing, transcribing and translating (Additional files 2, 3, 4, 5, 6, 7, 8 and 9). They all spoke the local Maasai dialect and used interview guides translated into Maasai to carry out the interviews. All interviews were tape-recorded and at the end of each day, the recordings were transferred to a password-protected study computer and backed up on an external hard drive. Debriefing sessions to reflect on each interview were held at the end of each day and written on a debriefing template. Debriefing notes and notes taken during the interviews were typed out and saved on the study computer. All interviews were transcribed verbatim and translated into English where necessary.

Following transcription of all interviews data were analyzed using a thematic analysis approach by two researchers (SK and PI) [11]. In the development of themes we adopted both a deductive and inductive approach. In the initial analysis a deductive approach was used to identify pre-set themes from the interview guides (topic guides). Transcripts were then read extensively by SK and PI using an inductive approach to identify emerging themes from the data. The emerging themes were those ideas or concepts that came up in the data and were different from the pre-set themes. SK coded all transcripts and PI coded a sub-set of the transcripts (approximately 20%) to ensure inter-coder reliability. SK and PI discussed any inconsistencies until mutual agreement was reached. The relationship between the quotes and codes was then explored to answer the research questions on motivators and barriers to health facility delivery. NVivo 10 software was used by both researchers throughout the analysis process.

Of the 200 Maasai-speaking women whose data were analyzed, mean age was 25.8 years, 137 (69%) had no formal education and 78 (39%) delivered in a health facility. Table 2 presents the socio-demographic characteristics of women by place of delivery and factors influencing health facility delivery. At bivariate analysis, level of education (p < 0.001), age (p = 0.05), marital status (p = 0.001), occupation (p = 0.015), parity (p < 0.001) and wealth quintiles (p = 0.001) were significantly associated with health facility delivery. By contrast, neither religion nor ANC was significantly associated with health facility delivery. After adjusting for other factors, belonging to the highest wealth quintile [aOR 4.9 (95% CI 1.5, 16.5)], currently not married [aOR 2.4 (95% CI 1.1, 5.4)], and living near the health facility [aOR 2.2 (95% CI 1.1, 4.4)] were positively associated with health facility delivery. High parity [aOR 0.7 (95% CI 0.5, 0.9)] was negatively associated with health facility delivery.

These results are organized into barriers and motivators to health facility delivery.

There are three groups of factors that appeared to hinder health facility delivery: Individual/family/community factors; geographic factors and health facility factors.

It appeared that most women were not the final decision-makers on the place of birth. The head of the household who is the eldest person in the “Boma” had the ultimate decision-making power. For example:

“…as the head of the “Boma” I must always give a go ahead of anything that need to be done at my home so generally the mother can suggest but I am the final decision maker” (FGD with CHVs)

“…she is mine and I will say where she will give birth” - Key decision-influencer for a woman who recently delivered in a health facility.

Most participants mentioned they did not make any birth plans when they discovered they were pregnant since in the Maasai community giving birth is seen as a natural occurrence and it is a taboo to make any birth plans.

“According to Maasai, you can’t choose for the baby the place you want to deliver; you just wait for the event to happen because the child can just come out at any place so there is no issue of planning.” A woman who delivered at home from a village far from the health facility

As a result, the decision on place of birth was made at the onset of labor unless a woman had previous health complications or had previously undergone a caesarean operation. Women were more likely to deliver at home if they had previously had a home delivery without experiencing any complications, thereby viewing facility-based delivery as unnecessary.

“It is because I have delivered all my children there (home), even this one because the TBA who assisted me is trained by Amref and so it is equal to a hospital delivery.” - A woman who lives near a hospital but delivered at home

Misconceptions about the services provided at the health facility, as well as the perceived quality of care also hindered women from delivering at the health facility. Participants believed that in some health facilities, women undergo unnecessary cesarean section and thus women avoid giving birth in these facilities. For example:

“Right now in…hospital, seven women were told that they will have to undergo operation because they could not deliver in the normal way. Those women came and delivered at home without any complications. Therefore that advice is frightening women”. - Key decision influencer

The gender of the health care providers also played an important role in influencing the location of birth, as some women reported not being comfortable being attended to by a male nurse. For example:

“When you turn to social factors, sometimes women don’t want to go to hospital because the attendant isn’t a woman, and don’t want to be attended to by a male, because she may not want everybody to see her private parts”- FGD with chiefs

Some women prefer home delivery due to subjection to unfamiliar birthing position, such as lying on the back compared to squatting. Home deliveries for such women ensure they retain desired birth practices.

“Others prefer home delivery because they say we don’t want to be told to lie on the back because it is uncomfortable. So such traditions matter”. – TBA who delivers women at home

Separate focus group discussions with CHVs and chiefs revealed that women may not deliver at health facility due to fear of disrespect and/or abuse by health care providers. Verbal abuse in form of shouting or rudeness, and physical abuse in form of pinching were some of the examples cited.

“…all the women would be delivering at hospital, it’s the issue of knowing who is the attendant at the hospital, may be Mr. (name) is the attendant and people know him as a rude person, so people opt elsewhere” – FGD with Chiefs

“There was another one who said that the reason why she doesn’t like going to deliver in the hospital is because there was a day when she was giving birth and a nurse pinched her with a pair of scissor so that she can push the baby, that was very painful to her and that discouraged her from delivering at the hospital.” - FGD with CHVs

It appeared that some health care providers are known to be good and others have bad attitude towards the mothers.

“They are not all the same, some are good and others are not…They treat people in a good and polite way but those who are not good shout at patients and do not show respect to them.” - A woman who lives near the health facility and delivered at the health facility

Accessibility can be a barrier to health facility delivery due to geographical distance to the health facility, availability of means of transport, and poor terrain that increases the time taken to reach the health facility and thus causing the fear of delivering on the way to the hospital. For example:

“Those who come from far have more challenges than others. There are places that even the ambulance can’t access…they have to be carried on motorbikes which are very risky for a mother in labor” - FGD with heath care providers

“Due to the distance to the health facility women are forced to deliver at home.” - IDI with TBA who delivers women at home in a village far from the health facility

Apart from health facilities being far, maternity services are not provided at night at some facilities, hence when a woman goes into labor at night she might prefer to obtain the services of a TBA whether or not she lives near a health facility.

“I didn’t choose I went into labor at night and who is at the hospital at night?”- Woman from a village near a health facility but delivered at home

In an attempt to overcome such barriers, some participants who live far from the health facility relocated and stayed with relatives who lived near a health facility or used other means of transport such as motorbikes which were deemed much better compared to walking to a health facility.

Delay in receiving appropriate care due to negative attitude of health workers and lack of essential drugs and supplies was cited as a hindrance to health facility delivery. A key decision influencer for a woman who lives near a health facility but delivered at home also stated:

“In our hospital, doctors neglect mothers a lot. You may call for assistance till the mother gives birth alone and she is in hospital.”

In some cases the local administration had to be involved for health workers to provide maternity services at night.

“The doctors now serve as they wish; there is no working at night nowadays unlike before. It is now a difficult issue. A chief has to be called to convince a doctor to attend to a patient…” FGD with Chiefs

A CHV gave an example of unskilled birth attendance at the hospital when she escorted her daughter to deliver at the health facility.

“There was a certain day when I brought my daughter here to deliver but the nurse refused because there were no gloves so I had to assist her myself.”

There are three groups of factors that appeared to motivate women to deliver in the health facility: individual, health system and geographic factors.

Individual factors include the status of the woman’s health or birth experience in choosing the health facility as the birth location.

Having an existing health condition, such as high blood pressure, iron deficiency, or generally being sick, played a big role in motivating them to deliver in a health facility.

“I think that’s what (health status) my husband considered because I became very sick and the nurses said that my blood level went down” – Woman who lived near a health facility and delivered in a health facility

Women who previously planned to deliver at home, but who experienced prolonged labor were also likely to deliver in a health facility. A woman who lived near a health facility and had planned to deliver at home gives this example:

“It was just because I had prolonged labor otherwise I could have given birth at home.”

Some women preferred to deliver in the health facility because they were “used to giving birth there”.

Physical factors include the role of geographic distance in selection of a birth location; living near a health facility increased the likelihood of utilizing a health facility during delivery. Ease of walking to a health facility without having to worry about availability of transport was described as a facilitating factor.

“I knew the hospital is close by so I would just go.” – Woman who lives near a health facility and delivered in a health facility

These include statements about the quality of care, availability of drugs and other commodities, and working closely with TBAs.

Availability of skilled personnel who can conduct life-saving interventions such as caesarean section, blood transfusion, labor induction and infusion was considered as a reason to deliver in a health facility.

“At the hospital one can be helped in two ways because sometimes a woman is unable to deliver in the usual way and if the hospital is unable to help her deliver, she will be operated on and the baby removed, and once she delivers the doctors are close so that they serve her in any way necessary for the person’s health.”- Key decision influencer

Availability of birth notification, drugs and other commodities given to women after delivering, such as diapers, towels, basins and mosquito nets, also motivated women to deliver in a health facility.

“When you give birth at the hospital…for that child who has been born, he/she has that document (birth notification) that day he/she is discharged. That is one advantage of giving birth at the hospital.” – Key decision influencer

“After I delivered, the placenta got retained and the nurses pressed my stomach, gave me two injections and it came out. So this was good because if I were at home I couldn’t have received this good care.” Woman who lives far from the hospital but delivered in a health facility

TBAs that have been integrated into the existing health system accompanied women to the health facility, assisted midwives in comforting women during labor, and helped reduce the language barrier between the health workers and Maasai women. During the focus group discussion with health workers, the example below was given:

“In this community, I think they have lived for long with the TBAs such that they have come to like them as part of them. So you never get a mother who comes to deliver without a TBA. So when you go to the maternity ward with that TBA, that mother will be very comfortable and won’t fear anything because the TBA is continually talking to them and instructing them on what to do. But when you go alone and leave the TBA outside you will get a hard time because there could be a communication barrier. You could be talking to them but they don’t get it, so unless they are there (TBAs) that is when you get the best delivery.”

Health education/advice received during antenatal clinics, home visits or seminars encouraged women to deliver in a health facility.

“We teach them the danger signs of pregnancies that they should be observing during pregnancy; we also tell them the importance of hospital delivery, and to give them nets and teach them how to use them for prevention of malaria during pregnancy.” – FGD with health workers

“I was the one who said let us go to hospital because whenever we go for clinic (ANC visit) we are usually advised that when we start to feel labor we should go to hospital” - Woman who lived far from a health facility and delivered in a health facility

In some cases women would describe health facility delivery as a “rule” that had been put in place by the health care providers whereby they were compelled to follow it.

The strength of this study is that it was conducted in an area that had received interventions to improve MNCH outcomes and that despite these interventions skilled birth attendance was still below the national average, hence this study documents factors that influence health facility delivery in order to better understand the context and inform design of programs to increase use of skilled care at delivery in marginalized communities.

An additional limitation was that the interviews were done several months after delivery. There is a possibility of recall bias and thus misinterpretation of the previous birth experience on the part of the mothers who participated in this study. Efforts to address this bias included use of local language and reference to the last born child when conducting interviews.