Factors associated with utilization of motorcycle ambulances by pregnant women in rural eastern Uganda: A cross-sectional study

Background: Evidence suggests that use of motorcycle ambulances can help to improve health facility deliveries; however, few studies have explored the motivators for and barriers to their usage. We explored the factors associated with utilization of motorcycle ambulances by pregnant women in eastern Uganda. Methods: This was a cross-sectional, mixed-methods study conducted among 391 women who delivered at four health facilities supplied with motorcycle ambulances in Mbale district, eastern Uganda, between April and May 2014. Quantitative data were collected on socio-demographic and economic characteristics, pregnancy and delivery history, and community and health facility factors associated with utilization of motorcycle ambulances using semi-structured questionnaires. Qualitative data were collected on the knowledge and attitudes towards using motorcycle ambulances by pregnant women through six focus group discussions. Using STATA v.12, we computed the characteristics of women using motorcycle ambulances and used a logistic regression model to assess the correlates of utilization of motorcycle ambulances. Qualitative data were analyzed manually using a master sheet analysis tool. Results: Of the 391 women, 189 (48.3%) reported that they had ever utilized motorcycle ambulances. Of these, 94.7% were currently married or living together with a partner while 50.8% earned less than 50,000 Uganda shillings (US $20) per month. Factors independently associated with use of motorcycle ambulances were: older age of the mother (≥35years vs ≤24years; adjusted Odds Ratio (aOR)=4.3, 95% CI: 2.03, 9.13), sharing a birth plan with the husband (aOR=2.5, 95% CI: 1.19, 5.26), husband participating in the decision to use the ambulance (aOR =3.22, 95% CI: 1.92, 5.38), and having discussed the use of the ambulance with a traditional birth attendant (TBA) before using it (aOR=3.12, 95% CI: 1.88, 5.19). Qualitative findings indicated that community members were aware of what motorcycle ambulances were meant for and appreciated their role in taking pregnant women to health facilities. Conclusion: The use of motorcycle ambulances was associated with older age of the mother, male participation in birth preparedness, and consultations with TBAs. These findings suggest a need for interventions to involve men in reproductive health as well as efforts to reach women younger than 35years of age. This was a cross-sectional, mixed-methods study that was conducted in Mbale district, eastern Uganda, using quantitative and qualitative data collection methods. The study was carried out among women who delivered in four health facilities that had motorcycle ambulances in 2013. The district has a population of 492,804 (2014 census estimates) with about 99,546 women in child bearing age. Mbale is a rural district with majority of the people living in the villages. It has 61 health facilities including one public regional referral hospital, two private hospitals, four health centers (HC) IVs, 23 HC IIIs and 34 HC IIs. Each health center IV serves as a referral center for obstetric cases from the lower-level health centers and the communities. Because of its hilly nature, transport from the rural parts of the district is difficult especially during the rainy season. Many people in the community are peasants with limited ability to afford, let alone, use private means of transport. In addition, cultural beliefs and the value attached to the placenta by women after birth [27] continue to hamper women’s willingness to deliver at health facilities. As a result, fewer women deliver at health facilities, with many opting to deliver at home or at the traditional birth attendants’ place. In view of these challenges, Mbale district introduced the use of motorcycle ambulances with the objective of increasing health facility deliveries which would in return help to reduce the high maternal mortality ratio, estimated at 567 per 100,000 live births in Mbale district (J Waniaye, personal communication). In 2010, Mbale district partnered with a local non-governmental organization; Partnership Overseas Networking Trust (PONT), to provide four motorcycle ambulances to four health facilities, i.e., Wanale HC III, Makhonje III, Busiu HC IV, and Namanyonyi HC III. Each health facility was assigned one well trained driver to man the ambulance. All ambulance services were provided free of charge; riders were provided with a monthly stipend and airtime for communication purposes from the district. About 30 traditional birth referral attendants (TBRA) were trained to work with the ambulance drivers so that pregnant women identified as needing emergency obstetric care could be referred to any of the participating health facilities using the ambulances. It was the responsibility of the TBRA in each village to call the facility where the ambulance was stationed and/or the ambulance drivers whenever there was need for referral. Communities were sensitized about the importance of motorcycle ambulances through community outreaches. TBRAs were provided with cell phones and their contacts were made known to village health team members to contact the drivers whenever a referral was necessary. Ambulance drivers didn’t have skills to conduct a delivery but TBRAs were trained to identify mothers with complications, call for an ambulance and escort the mother to the health facility while in transit on a motorcycle ambulance as needed (F Chemuto, personal communication). We obtained a representative number of women from each facility through stratified sampling procedures using a probability proportional to size strategy. A sample size of 391 women was estimated using Kish Leslie (1965) formula, assuming prevalence of use of the motorcycle ambulances at 20 %, a precision of 5 %, a standard normal deviation at 1.96 and an adjustment for non-response of 2 % [28]. Using maternity registers at each health facility, we obtained information (names, address, and telephone contact where available) on all women who delivered at each facility between January and December 2013, and generated numbered lists of all eligible women. Women were eligible to participate in the study if they delivered at any of the participating health facilities between January and December 2013; lived within the catchment area of the health facility supplied with a motorcycle ambulance; and were recorded in the maternity register. Using an online table of random numbers generator [29], specific numbers were selected and the names of women corresponding to these numbers were selected to participate in this study. This was done for each of the participating health facilities until the sample size was obtained. Women below 18 years included in this study were emancipated minors who were able to provide informed consent as guided by the national guidelines [30]. Qualitative interview participants were both men and women living in the catchment area of the health facilities supplied with motorcycle ambulances. They were all selected purposively. Women were selected to participate in the study if they had delivered in the previous two years at the participating health facilities but used other means of transport to access delivery services. On the other hand, men were selected to participate in the study as long as they had wives who were still in the reproductive age group (15–49 years). Participants were identified with the assistance of village health team (VHT) Chairpersons and the Community Development Officer (CDO). To facilitate the identification process, we informed the VHT Chairperson or the CDO about the enrolment criteria (as described above), and whenever we approved a household, either the VHT chairperson or CDO helped the research team to identify and verify the individuals that were eventually invited for the qualitative interviews. Quantitative data were collected between April and May 2014. Interviewer-administered, semi-structured questionnaires (See Additional File 1) were administered to consenting women who delivered at each of the four participating health facilities between January and December 2013. Data were collected on respondents’ characteristics, community factors, and health service factors. Respondents’ characteristics included; socio-demographic (age of the respondents, marital status, occupation, tribe, religion, level of education) and economic (monthly income, ownership of a phone) characteristics, pregnancy and delivery history (number of antenatal visits attended during that pregnancy, history of any obstetric complication) and motorcycle ambulance use-related characteristics. Community factors included; women not wanting to be seen in public while in labour and the cultural expectation that pregnant women should not sit on a motorcycle ambulance. Health service factors included; availability of obstetric drugs and supplies for delivery in the health facility, whether or not women believed that ambulance drivers have obstetric skills, health workers’ and drivers’ attitudes towards pregnant women in labour, whether or not women discussed with traditional birth attendants on whether to use or not to use a motorcycle ambulance during the course of her previous pregnancy and distance of the women’s home from the health facilities. Data were collected by four trained research assistants fluent in the local languages (Lumasaba and Luganda) and English. Women were traced and interviewed at home. At least two repeat visits were made to locate and interview all the eligible women. Those who were not found at home at the second attempt to locate were replaced with other women randomly selected from the maternity register that was used to select the initial cohort of women. All completed questionnaires were edited in the field to ensure completeness and accuracy; returned to a central office location on a daily basis and entered into Epi-data version 3.02. Qualitative data were collected using focus group discussion (FGD) guides (See Additional File 2). Data were collected on level of awareness, attitudes towards the ambulances, barriers and facilitators to using the ambulances, and suggestions for improvement on the functionality of the ambulances by a team of research assistants comprising a moderator and a note-taker. Interviews were conducted at the health facilities where the ambulances are stationed, and lasted for a period of one and a half hours on average. All data were audio-recorded with permission from the study participants. Six FGDs (3 for males and 3 for females) were conducted among 28 women and 29 men; with each FGD comprising of 8–10 participants. Male discussants were within the age group of 20–50 while females were within 20–45 years. Men and females were separated in these discussions to allow freedom of speech from all participants. FDGs were conducted in Luganda/Lumasaba/English concurrently for better understanding and all participants consented before the discussions. The primary outcome, utilization of motorcycle ambulances, was defined as a binary variable with Yes = 1 indicating that a pregnant woman used the ambulance to go to the health facility at the time of delivery and No = 2 indicating that a pregnant woman used other forms of transport to go to the facility. In this study, the term married was used to mean all forms of marriage; married in church/mosque, traditionally married or civil marriage. The term “living together” was used to mean all women who might have not been married in any of the above described forms of marriage but were living together with their spouses. The variable “occupation” was categorized into peasant, student, civil servant, business woman and others. We used the term “peasant” to refer to a woman who was engaged in subsistence farming, with no expected monthly salary. Women who said they earn some money during a typical month were asked to state how much money they earn per month, and the figures presented are based on self-reports. At the time of analysis, all quantitative data were transferred from EpiData to STATA statistical software version 12. We computed descriptive statistics to obtain characteristics of the study respondents. Categorical variables were expressed using frequencies and percentages while continuous variables were expressed as mean (standard deviation). A Chi Square test was used to assess for differences in proportions among women who used and those who didn’t use the ambulances. A logistic regression model was used to identify factors that were statistically associated with the primary outcome both at the bivariate and multivariable levels. All factors at the bivariate level that had a p-value of less than 0.2 were selected for inclusion into the multivariable logistic regression model. A p-value less than 0.05 was considered significant at the multivariable analysis level. Qualitative data were transcribed verbatim and entered into a Microsoft Word processing document for analysis. Data were analyzed manually using a master sheet analysis tool. All transcripts were printed out and read through by the first author to identify emerging issues based on a priori themes. These themes included knowledge about the existence of the motorcycle ambulances, attitudes towards use of motorcycle ambulances, barriers and facilitators to usage, and suggestions for improvement their role. Relevant quotations were identified and used to support each theme during the reporting process. This study was approved by the Institutional Review Board (IRB) of Makerere University School of Public Health (MakSPH) before conducting the study. Permission was also obtained from the district health officer (Mbale). Written informed consent was obtained from the respondents participating in the study. Data collection tools were coded and no names of respondents appeared anywhere for confidential purposes. The study was adherent to the STROBE criteria as outlined in Additional file 3.