Menu
Menu
Menu
Menu
Background: The Missed Opportunities for Maternal and Infant Health (MOMI) project, which aimed at upgrading maternal and infant postpartum care (PPC), implemented a package of interventions including the integration of maternal PPC in infant immunization services in 12 health facilities in Kaya Health district in Burkina Faso from 2013 to 2015. This paper assesses the coverage and the quality of combined mother-infant PPC in reproductive, maternal, newborn and child health services (RMNCH). Methods: We conducted a mixed methods study with cross-sectional surveys before and after the intervention in the Kaya health and demographic surveillance system. On the quantitative side, two household surveys were performed in 2012 (N = 757) and in 2014 (N = 754) among mothers within one year postpartum. The analysis examines the result of the intervention by the date of delivery at three key time points in the PPC schedule: the first 48 h, days 6-10 and during weeks 6-8 and beyond. On the qualitative side, in depth interviews, focus group discussions and observations were conducted in four health facilities in 2012 and 2015. They involved mothers in the postpartum period, facility and community health workers, and other stakeholders. We performed a descriptive analysis and a two-sample test of proportions of the quantitative data. The qualitative data were recorded, transcribed and analysed along the themes relevant for the intervention. Results: The findings show that the WHO guidelines, in terms of content and improvement of maternal PPC, were followed for physical examinations and consultations. They also show a significant increase in the coverage of maternal PPC services from 50% (372/752) before the intervention to 81% (544/672) one year after the start of the intervention. However, more women were assessed at days 6-10 than at later visits. Integration of maternal PPC was low, with little improvements in history taking and physical examination of mothers in immunization services. While health workers are polyvalent, difficulties in restructuring and organizing services hindered the integration. Conclusion: Unless a comprehensive strategy of integration within RMNCH services is implemented to address the primary health care challenges within the health system, integration will not yield the desired results.
The study sites were MOMI project sites in the Kaya health district in the Centre-Nord health region of Burkina Faso. The Kaya district has 564,867 inhabitants and 52 primary health facilities [21]. The MOMI project was implemented in 12 health facilities, including the Kaya Health and demographic surveillance system (Kaya HDSS). The MOMI baseline assessment showed that skilled health personnel in the Kaya health district attended 77% of births. Within 7 days after delivery, only 19% of women received maternal PPC, while all newborns received the BCG vaccination. Maternal PPC was poorly integrated within child immunisation and family planning (FP) services [3]. The identifiable factors hindering maternal PPC were the poor organisation of services and quality of care, issues related to understaffing and high staff turnover, a lack of knowledge about maternal health in the community, and traditional beliefs and practices [3]. The MOMI baseline assessment was carried out by local stakeholders and founded on the identification of problems and causes, then using system thinking approaches to specifically review RMNCH policies [3]. Based on this assessment, a tailored package of interventions to strengthen postpartum care delivery was selected and implemented, including: (1) upgrading the quality of immediate PPC provided at the health facilities, with a focus on the detection and management of postpartum haemorrhage and sepsis, (2) supporting mothers and infants during the PP period with the help of female community health workers (CHWs)1 and (3) the integration of maternal PPC in infant immunization services [3]. While infant PPC is the overarching anchor for the integration of maternal and child PPC, integration could also happen the other way around, i.e. infant PPC integrated in maternal PP services. Integrated mother-infant PPC in infant immunization services was provided within the health system organization on three occasions [22]. Firstly, at days 6–10, newborns are immunized against the Bacillus Calmette-Guérin virus (BCG) and receive the 0 dose of the poliomyelitis vaccine if they did not receive it before. During this visit, the mother is examined for signs of PP haemorrhage, sepsis, and anaemia. She receives information on the prevention, management and treatment of PP disorders for herself, and on the detection of warning signs in newborns. Secondly, at weeks 6–8, the infant is immunized against diphtheria, tetanus, pertussis, poliomyelitis, haemophilus influenza type B, rotavirus and pneumococcal using dose 1 conjugate vaccine/ (Penta + Hep 1). At the same time, maternal PPC focuses on the detection, management and treatment of postpartum anaemia, and on advising and providing postpartum family planning. The mother is also informed about the detection of warning signs in infants. Thirdly, at month 9, the infant is vaccinated against measles and yellow fever. The content of maternal PPC is the same as at the weeks 6–8 visit. We were not able to assess PPC at month 9 due to the lack of information in the registers and in the household surveys about this visit [23]. Instead, on top of the intervention at days 6–10, we consider the integrated maternal PPC visit at days 11–41 – between days 6–10 and weeks 6–8 –, and days 42–90, which can be seen as a prolonged weeks 6–8. For the child, we also check for PPC at days 0–5 and days 11–60. Moreover, as part of the strategy to improve immediate PPC, the newborn should receive the BCG vaccine during the first 48 h after delivery and before leaving the health facility. Therefore, we added the first 48 h to the analysis in order to consider immunization sessions provided during this period and the immediate PPC. The intervention implementation included several activities. Working sessions were performed in each facility with facility health workers (FHWs) in July 2013 in order to understand the existing level of integration. At the same time, FHWs and managers of primary health facilities, maternity wards and the Expanded Program of Immunization (EPI) were invited to workshops where the MOMI project and interventions were presented. In September 2013, FHWs received training on PPC, including the integration of maternal and infant care. Effective implementation started in October 2013. Staff training (several rounds) continued in December 2013 and in March 2015. Protocols and checklists to support the integration of services were developed, distributed and explained to FHWs from January to February 2014. Quarterly supervision of MOMI activities using a guide was performed from October 2013 to December 2015. As the study is conducted through the MOMI project, which is a case study, the study design follows its global configuration [3]. This present paper was designed around the comparison of the situation prior to and after 1 year of intervention, using a cross-sectional mixed methodology. The survey before the intervention data (including both quantitative and qualitative data) collection took place from December 2012 to May 2013. Another round of quantitative data collection occurred from August to December 2014, within 1 year after the start of the intervention. Qualitative interviews were conducted in 2015 in the framework of MOMI’s final evaluation and were compared with the data collected before the intervention. The quantitative surveys aimed to investigate the coverage and the content of the pair mother infant PPC and the integration in infant clinics, notably immunization services. The household surveys were carried out before and after 1 year of intervention in the Kaya HDSS, located in the Centre-Nord region of Burkina Faso. The Kaya HDSS contains 7 urban and 18 rural zones within a radius of 20 km, comprising 7 of the 12 primary health facilities included in the MOMI project [24]. The Kaya HDSS 2012–2013 routine household data collection – over a period of 6 months – covered 10,629 households, including 16,801 women of childbearing age, 326 pregnant women and 800 infants. The number of pregnant women is low, probably due to under declaration in the first trimester of pregnancy. We performed a multistage sampling technique, which included data extraction and systematic sampling with a random start -per health facility and village. The data were extracted from Kaya HDSS, pregnancy and delivery –for mothers– and birth records –for infants. The Kaya HDSS data collection and monitoring method is described in detail in Kouanda et al. [24]. We briefly identified all women who had given birth or who were in the last months of pregnancy throughout the Kaya HDSS. For the survey before the intervention, we selected a systematic sample of 840 mothers in their first-year of PP with an accuracy of 100% from the afore-mentioned database. Eight enumerators performed the data collection using personal digital assistants –small mobile/tablet handheld devices that provide computing and information storage and retrieval capabilities– from December 2012 to January 2013. The survey was specific to the Kaya HDSS. For the separate survey after 1 year of intervention, we selected a systematic sample of 880 mothers in their first-year of PP out of the 2014–2015 Kaya HDSS household newborn and pregnant women database, following the same procedure as in the survey before the intervention. Ten enumerators carried out the data collection using personal digital assistants and tablets within the Kaya HDSS routine data collection from August 2014 to February 2015. One sample of mothers in their one-year PP were interviewed before (757 out of 840: 90%) and another one after 1 year of intervention (754 out of 880: 86%) (Table 1). After applying exclusion criteria, there were 732 mothers in the survey before 2012–2013 and 705 in the survey after 2014. Furthermore, thirty three women assessed during the second survey were found to have delivery dates that preceded the intervention and were moved to the sample before the intervention. The survey is not a longitudinal survey, and the mothers are therefore not the same in the two samples. All the surveys were preceded by staff training, pre-testing, and pilot surveys. Data included by surveys We used the same questionnaire on PPC for mother and infant before and after 1 year of intervention, which we complemented with additional information from the Kaya HDSS routine data, such as individual characteristics and household socio-economic characteristics. The latter, including durable assets, housing characteristics and ownership, access to utilities and infrastructure, was used to calculate socio economic quintiles. These allowed us to assess the comparability of the two samples of mothers before and after 1 year of intervention [10]. The questionnaire developed by the researchers included queries about pregnancy outcomes, frequency and content of visits to PP services for maternal and infant immunization. The timing of the interventions that entered the analysis was for maternal PPC: first 48 h, days 6–10, days 11–44 and days 45–90; for infants, it was days 0–5, days 6–10 and days 11–60. We used cross-tables to study the proportion of mothers who had PP visit(s) during the first 48 h, at days 6–10, at days 11–44 and at days 45–90; and those whose infants had PP visit(s) at days 0–5, days 6–10 and days 11–60. The content of the PPC of the mother-infant pair (history taking, physical exam, consultations) targeting warning signs, detection and management of three main disorders in Burkina Faso (haemorrhage, sepsis, anaemia) and on the integration of services, were used as independent variables. We analysed the data using the infant’s date of delivery to determine whether the mother and infant were likely to have benefited from the intervention or not. This resulted in two periods (Table (Table1):1): 1) before the implementation of the intervention (in 2011–2012 and from January to June 2013), and 2) one-year after the implementation (from July to December 2013 and all of 2014). Frequencies and cross-tabulations were used at the bivariate level to analyse the distribution of the dependent variables by the content of visits. We ran a Chi-Square Test of Pearson as well as a two-sample test of proportions on the before-after linked data in order to determine whether there were differences in maternal and child PPC use before and after the implementation of the intervention. We performed descriptive and optimal scaling methods for multivariate categorical data analysis using Stata Statistical Software: Release 15 (College Station, TX: StataCorp LLC). The qualitative studies explored the changes in PPC with the implementation of the interventions from the point of view of the main stakeholders. It aimed to explain the numbers provided by the quantitative surveys and to complement the evaluation. Qualitative study recruitment was based on data saturation in four primary health facilities areas (two rural and two urban) before and during the MOMI project final evaluation study. In addition, the purposive sampling of four contrasting cases by level of implementation (high, low) was done during the evaluation study [25]. The number of CHWs and FHWs before and after did not affect the findings. Before the intervention, data were collected in January 2013 (A). The final evaluation data were collected in July–August 2015 (B). The data collection involved in depth interviews with health workers in the facilities (A, B) and communities (A, B), key informants (A, B) and women in postpartum period (B) (Table 2). Before the intervention, focus group discussions were conducted with women who had experienced childbirth and with mothers in the PP period. Those were not reproduced in the MOMI final evaluation survey whereby fewer women were included in A than in B. No respondents were interviewed before as well as after the intervention: both panels are composed of different individuals. Interview participants in the qualitative studies before and after the intervention aFour focus group discussions were conducted with 42 mothers (in groups of 8–12 women) in PP period (n = 22) and women with childbirth experience (n = 20) Furthermore, during the final evaluation phase, we used direct observation by following FHWs while they were delivering services such as routine PPC for mothers and infants, in infant clinics and in family planning services. Sixteen observations reports were written by three MOMI researchers in the four health facilities over a period of 2 weeks. The data were collected by qualitative research assistants using a guide developed for that purpose. Interviews with women focused on their individual experiences and explored their views on the need for PPC and on the barriers to or facilitators for receiving PPC. With FHWs, CHWs and other stakeholders, the interviews were centred around their views on the importance and provision of PPC, and contextual factors. The data were recorded and transcribed using NVIVO software. The content of the interviews were analysed along both surveys themes. The coding system was developed and implemented by one qualitative researcher for each survey and the consistency was checked a posteriori by the qualitative team members. The interviews were analysed to determine the opinions of participants on the overall changes in PPC (baseline status, changes in PPC, MOMI project contribution, enablers of PPC uptake) and on the integration of PPC (integrated services and barriers to integrate mother PPC into infant immunization services).
N/A
