‘How to know what you need to do’: A cross-country comparison of maternal health guidelines in Burkina Faso, Ghana and Tanzania

Background: Initiatives to raise the quality of care provided to mothers need to be given priority in Sub Saharan Africa (SSA). The promotion of clinical practice guidelines (CPGs) is a common strategy, but their implementation is often challenging, limiting their potential impact. Through a cross-country perspective, this study explored CPGs for maternal health in Burkina Faso, Ghana, and Tanzania. The objectives were to compare factors related to CPG use including their content compared with World Health Organization (WHO) guidelines, their format, and their development processes. Perceptions of their availability and use in practice were also explored. The overall purpose was to further the understanding of how to increase CPGs’ potential to improve quality of care for mothers in SSA.Methods: The study was a multiple case study design consisting of cross-country comparisons using document review and key informant interviews. A conceptual framework to aid analysis and discussion of results was developed, including selected domains related to guidelines’ implementability and use by health workers in practice in terms of usability, applicability, and adaptability.Results: The study revealed few significant differences in content between the national guidelines for maternal health and WHO recommendations. There were, however, marked variations in the format of CPGs between the three countries. Apart from the Ghanaian and one of the Tanzanian CPGs, the levels of both usability and applicability were assessed as low or medium. In all three countries, the use of CPGs by health workers in practice was perceived to be limited.Conclusion: Our cross-country study suggests that it is not poor quality of content or lack of evidence base that constitute the major barrier for CPGs to positively impact on quality improvement in maternal care in SSA. It rather emphasises the need to prioritise the format of guidelines to increase their usability and applicability and to consider these attributes together with implementation strategies as integral to their development processes. © 2012 Baker et al; licensee BioMed Central Ltd.

The study is a multiple case study design consisting of cross-country comparisons of three case studies [6,30,31] using: document review to compare the content of national CPGs with WHO recommendations and to assess the format of guidelines; and key informant interviews to explore knowledge of national maternal health CPGs, their development processes, relation to WHO recommendations, and perceptions of their availability and use by health workers in practice. The study was carried out in two West African countries, Burkina Faso and Ghana, and one East African country, Tanzania. In all countries, an intervention and a non-intervention district have been chosen where the QUALMAT project will be conducted in 2009 to 2014 in rural primary healthcare centres providing maternal and neonatal care. All three study countries suffer from a critical shortage of health workers [32,33], but variations are large with the density of nurses and midwives in Ghana five times that of Tanzania [33]. Skilled attendance at delivery is similar with national averages around 50% in all three countries [33]. Data collection was carried out between September 2009 and March 2010. National CPGs for maternal care with contents corresponding to the WHO guideline ‘Pregnancy Childbirth Postpartum and Newborn Care: A Guide for Essential Practice’ (PCPNC) sections B9-E, the sections relevant to the CDSS development, were identified and collated (see further explanation under document review). This was done with the help of local research partners and co-authors (MS, BK, JW, RM, SM), as well as key informants at the ministries of health and other organisations involved in the development of national guidelines for maternal health. Several rounds of crosschecking were done to make sure that all the relevant national maternal health CPGs had been identified. Semi-structured interviews were held with a total of 38 key informants in the three countries. The interview guide had three sections which covered information on: 1) national CPGs for maternal health; 2) motivation and incentive policies for human resources; and 3) current maternal health projects and initiatives in the study areas (see Additional file 1 for interview guide for topics one and three). Ten of the key informants worked in human resources and administration and were not asked questions regarding topics one and three. Only the 28 interviews covering information on CPGs for maternal health (Table ​(Table2)2) were therefore included in the study presented in this paper. The sampling was mainly purposive where an initial list of key informants [34] representing relevant government departments, international and national non-governmental organisations (NGOs) and with knowledge of maternal health guidelines, was prepared by the research teams in each country headed by RM, JW and MS. Each key-informant was also asked about other key stakeholders, but the majority of the ones mentioned were already on the initial lists. Because the objective was to identify national guidelines and policies, the majority of interviews took place at the central level, e.g., government departments and ministries as well as at national offices for NGOs and United Nations (UN) organisations. In Burkina Faso and Ghana, however, the research teams were based near the intervention and control districts, and therefore took the opportunity to interview a few key informants also at the regional and district level as well as at a few health facilities (Table ​(Table2)2) [35]. Interviews in Tanzania and Ghana were conducted in English and recorded and transcribed by the first author (UB) and two research assistant (MD and GH, see acknowledgements), respectively. In Burkina Faso, interviews were conducted by one of the co-authors (SM) in French and later translated into English. Key informant characteristics A conceptual framework describing the implementation process from development and production to health workers’ access to and use of CPGs was constructed to aid analysis and discussion of the results in this paper (Figure ​(Figure1).1). It emanated both from the document review and analysis of interviews, as well as from the adaptation of a recently published guidelines’ implementability framework by Gagliardi et al. [21]. While Gagliardi and colleagues present a comprehensive model with eight overarching domains related to guidelines’ implementability, our framework is purposefully limited with a focus on health workers as the users of guidelines. In this context, usability refers to the user-friendliness of a guideline and includes aspects such as ease of navigation and the format of evidence and recommendations such as tables and algorithms [21]. Applicability represents the degree to which the guideline can be used and applied by health workers for the management of individual patients [21]. Adaptability describes the availability of different versions of a guideline such as electronic formats or leaflets aimed at different users [21]. Conceptual framework for cross-country comparison of clinical practice guidelines (CPGs). The framework describes the process from development and production to health workers’ access to and use of CPGs. It emphasises three selected features (from Gagliardi et al., 2011 [21]) of CPGs related to their implementability and use by health workers in practice: usability, applicability, and adaptability. A recent paper [28] presents desirable features for a framework aimed at cross-country analysis. These include the need for such a framework to: offer insights into a broad variety of contexts; address the same problems in the same way whilst allowing for contextual differences; and use concepts sensitive to both quantitative and qualitative differences [28]. We believe that our framework considers all of these features. The cross-country comparison thus focuses on the description of contrasting and common features of the themes and categories emerging from the data analysis from each country, linking them to the elements of this framework. This is a methodology commonly applied for cross-country comparisons [6,30]. The WHO PCPNC guideline [36] is part of the WHO tool kit for the ‘Integrated Management of Pregnancy and Childbirth.’ It provides generic norms and standards to be adapted for skilled health service providers at the primary care level. It is a guide supporting clinical decision making for health workers engaged in antenatal care, delivery, postpartum, post-abortion, and emergency care, as well as care for newborns immediately after delivery and during the first week of life. The introduction to the guide states that ‘correct use of this guide should help reduce the high maternal and perinatal mortality rates […], thereby making pregnancy and childbirth safer’ [36]. This guideline is an internationally recognised tool on which national guidelines are commonly based, and was therefore chosen by the QUALMAT project as a gold standard on which to base the development of the CDSS and to be used for comparison with national guidelines. All identified guidelines with contents corresponding to the WHO PCPNC sections B9-E were included in the data analysis [36]. Each guideline was systematically compared [37] with the WHO PCPNC, which is divided into chapters A to N with different numbered sub-sections. For this, we used a data extraction checklist (see Additional file 2) of key contents in the sections relevant to the CDSS development (sections B9-E). The data extraction from national guidelines was done according to sub-elements of these sections, which included: emergency care in pregnancy; bleeding in early pregnancy; antenatal care; labour; delivery; and postpartum care (see Additional file 2 for full contents of data extraction check-list). Deficiencies and discrepancies between national CPGs and the WHO PCPNC were noted down in a grid. The format of guidelines was then examined to make an assessment of their usability and applicability, both attributes believed to influence their use in practice. It is also a primary objective of the CDSS in QUALMAT to ensure a high level of these attributes [21,38]. Usability and applicability of each guideline included were graded as high, medium, or low. Usability was assessed by: indexing and ease of navigation; format of text (narrative, check lists); and availability of treatment/management algorithms. A guideline was deemed as having a high degree of usability if it was easy to navigate and contained comprehensive checklists and algorithms. Guidelines with text in a narrative format, poor indexing, and no algorithms were deemed as having a low degree of usability, and guidelines falling between these categories as having a medium degree of usability. Applicability was assessed by: the availability of treatment/management algorithms to guide decision-making for individual patients; and availability of Partograph to monitor labour. Applicability was deemed high for guidelines containing clear treatment/management algorithms for different levels of care. Guidelines containing no algorithms but with a Partograph were deemed as having a medium degree of applicability, and guidelines without both algorithms and Partograph as having low applicability. The guideline review was primarily carried out by JE with additional input from UB and SM, and complemented by discussions with maternal health specialists in Burkina Faso and Ghana (see Acknowledgements). The qualitative analysis of interviews was carried out with the help of Nvivo9® software [39] and applied thematic content analysis [35,40]. Interview transcripts were initially sorted into content areas, and sections containing information on guidelines were further analysed to identify themes. These were partly framed by the initial interview guide and partly by the information emerging from analysis of the interview transcripts. Within each theme, further coding of information took place revealing several categories [35,40]. Data from each country were initially analysed independently. Themes and categories were then harmonised between the countries and in the last stage, a cross-country comparison of the results from interview data was performed. Analysis of interviews was carried out jointly by UB, JE, and GT. Verbal consent was obtained from all participating key informants who were informed that their participation was voluntary and that they could withdraw from the study at any time without stating a reason and without any implications. Ethical clearance was granted by: the Institutional Review Board at the Navrongo Health and Research Centre (Ethics Approval ID NHRCIRB 085) for Ghana; the Muhimbili University of Health and Allied Sciences Ethical Review Committee (ref no.MU/AEC/VOLXIII/96) for Tanzania; and the Ethics Committee for Health Research in Burkina for Burkina Faso.

• baker, ulrika
• tomson, göran
• somé, mathias
• kouyaté, bocar
• williams, john
• mpembeni, rose
• massawe, siriel
• blank, antje
• gustafsson, lars l.
• eriksen, jaran