The GREAT Network

The World Health Organization (WHO) uses a rigorous process for guideline development, based on the standards of the WHO Guideline Review Committee.[12] The Department of Reproductive Health and Research (RHR) at WHO issues recommendations on a range of reproductive health topics.[13] However, since guideline development and dissemination alone is not enough to change behaviour and improve outcomes,[14,15] there is a clear need for effective strategies to support implementation of these guidelines into practice globally. To meet this challenge, RHR partnered with the Knowledge Translation Program at St. Michael’s Hospital, University of Toronto, Canada to establish a collaboration on KT in maternal and perinatal health, called the GREAT Network (Guideline-driven, Research priorities, Evidence synthesis, Application of evidence, and Transfer of knowledge).[16]

The GREAT Network uses a systematic approach based on evidence and theory to support LMICs in the implementation of maternal and perinatal health guidelines. The Network’s activities are guided by the Knowledge-to-Action (KTA) cycle (Fig 1), a process model which presents the steps involved in bringing research to practice. The theoretical elements exist on two levels: first, the KTA was developed based on a review of over 30 behavior change theories and frameworks[17]; second, the process of selecting implementation strategies to address identified barriers and facilitators to change involves incorporating behaviour change theory, for example the theory of planned behaviour or the capability, opportunity, motivation—behaviour theory.[17] The KTA also emphasizes the need to consider evidence at each step along the process. The GREAT Network conducted a mixed-methods pilot study on the determinants of implementation of maternal health guidelines in Kosovo in 2012.[18]. The process was adapted and expanded into a multi-country, mixed methods research activity described below.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Knowledge-to-Action Cycle (Reproduced with permission from[38]).

Reprinted from Straus SE, Tetroe J, Graham I. Knowledge translation in health care: moving from evidence to practice. 2nd ed. BMJ Books, Wiley, 2013. under a CC BY license, with permission from Wiley, original copyright 2013.

https://doi.org/10.1371/journal.pone.0160020.g001

This project focused on the first steps of the action cycle of the KTA (Fig 1), namely: adaptation of knowledge to the local context (i.e. adapting WHO guidelines to each country setting), assessing barriers and facilitators to knowledge use, and selecting and tailoring implementation strategies to local contexts. The study utilized an integrated KT approach, where countries’ and relevant WHO maternal health guidelines were identified based on in-country needs. The premise of integrated KT is that end users should be involved in research and implementation throughout the process so that the activities and outcomes directly address their needs.[19] This article describes the methods, findings and lessons learned from these activities. In subsequent phases, local working groups will progress through the action cycle steps of tailoring, implementing, monitoring, evaluating and sustaining the use of guidelines in their local contexts.

Setting

Participating countries were identified through purposive sampling. Within each country, activities related to a specific guideline implementation challenge, identified as a national priority at the time of the activities. In Myanmar, the activity was conducted in the context of a Ministry of Health initiative to improve the coverage of basic maternal and newborn healthcare nationwide, with particular emphasis on task-shifting from midwives to auxiliary midwives (AMWs). Uganda, Tanzania and Ethiopia were identified as priority countries within the UN Commission on Life Saving Commodities for Women and Children, which aims to improve access to 13 essential commodities (including the maternal health commodities oxytocin, misoprostol and magnesium sulfate). In-country activities were conducted in Myanmar in June 2014, Uganda in August 2014, Tanzania in November 2014 and Ethiopia in May 2015.

Design

Myanmar

A total of 31 people completed the pre-workshop survey and 97% agreed that task shifting to AMWs is a priority in Myanmar. In the survey, the three most prevalent barriers to implementation of the task-shifting recommendations were: 1) the need for training, retraining and supportive supervision of AMWs; 2) the need for clear policy on AMW roles and responsibilities; and 3) community preferences for different cadres of health workers.

Forty-two participants attended the workshop, which included six AMWs and four MWs. Four FGDs were conducted, stratified by role and/or level of the healthcare system (AMW, midwives (MW), Ministry of Health staff at central and township level, and a mixed group of UN and non-governmental stakeholders and representatives of professional organizations) and focused on the feasibility of implementing the task-shifting recommendations in Myanmar.

Participants in FGDs discussed factors affecting implementation of AMW-specific task-shifting recommendations. At the level of the healthcare system, factors included: shortages of MWs and AMWs; lack of resources (including birth kits and medicines), the need for better accountability and monitoring of AMWs (specifically, a need for AMW supervision, regulatory oversight and improved data collection for births attended by AMWs) and a need for health policy changes (particularly those governing AMW and MW role descriptions). Issues at the level of the healthcare provider were common, including the role and capacity of AMWs and MWs (particularly, AMWs capacity to take on additional tasks and the need for MWs to provide better supervision for AMWs); the need for improved quality, regulation and monitoring of AMW and MW education and training and the willingness, buy-in and motivation of AMWs and MWs, as well as the relationships between these health cadres. Factors at the level of the patient/community included the existence of strong community support for and trust in AMWs (described as an important facilitator). Cultural practices and health-seeking behaviours were also identified, specifically the perceptions that many women are reluctant to seek hospital care (due to financial and cultural factors) and AMWs were less likely to refer women to hospital.

Respondents ranked the feasibility of implementing AMW-specific task-shifting interventions in Myanmar (top ranked recommendations are listed in Table 1). Identified barriers to implementation of these prioritized recommendations included: variable capabilities of AMWs to perform tasks appropriately; poor drug quality, supply, and availability to AMWs; need for additional monitoring of drug use (including inappropriate use) and performance supervision; and insufficient funds, materials and skilled trainers for training. More task-specific barriers included current policies that do not permit AMWs to use these interventions; a lack of AMW-specific national guidelines (on prevention of PPH, neonatal resuscitation and management of puerperal sepsis); and a lack of resuscitation equipment.

The workshop participants proposed that the primary targets for implementation efforts should be AMWs working in rural areas, whereas MWs and AMWs working in urban areas and interested in improving skills were secondary targets. Potential implementation strategies included: policy briefs (to support development of evidence-based health policies that are practical and feasible); registration and regulation of AMWs through a new or existing national organization (such as a new AMW council, or the existing National Nursing and Midwifery Council); development of AMW guidelines on obstetric emergency management; and standardization and improvement of AMW training, tailored to specific needs of rural-based AMWs. Incentives for training (e.g. honoraria) were proposed to encourage participation. The need for improvements in drug distribution (including better logistics) was also identified.

Uganda

A total of 16 people completed the pre-workshop survey; the majority of respondents were from Kampala and Mityana (37.5% and 25.0%, respectively). Prevention and treatment of PPH was identified by all respondents as the highest priority guideline for implementation in Uganda. Thirty four stakeholders participated in the two-day in-person workshop, with balanced representation from each stakeholder group, including 20 (58.9%) practicing healthcare providers.

Four FGDs were conducted to elicit perceptions around factors affecting implementation of the WHO guidelines in Uganda. One cross-cutting issue that was consistently highlighted as a barrier to health care provision in general (including guideline implementation) was lack of human resources. This included shortages of doctors and midwives, poor recruitment and retention rates (particularly in rural areas) and limitations on allowable numbers of health care workers that can be recruited at various health facility levels (irrespective of actual workload). These staffing pressures can limit effective implementation due to lack of time for routine care practices (such as labour monitoring and partograph completion).

Gaps in drug procurement, management and distribution processes as well as lack of essential equipment such as refrigerators for oxytocin storage, fetal scopes and partographs, were also described as barriers. It was noted that certain medicines (eg: magnesium sulphate) are not approved for use at lower levels of the health care system (such as health centres), which hinders their use, despite awareness of WHO recommendations. Concerns were also raised regarding safety and appropriate medicine use, including what cadres of healthcare providers are trained in their use. Examples cited included the growing practice of labour induction in outpatient settings, and anecdotal reports of combined use of oxytocin and misoprostol for labour augmentation, resulting in uterine rupture in some cases.

The need for better site-specific data on implementation of evidence-based practices was also highlighted. While quantitative data on selected outcomes are routinely collected and reported to the Ministry of Health, participants reported these are not fed back to facilities to support improvements in care. All FGDs described the need for improved accountability and monitoring of provider adherence to guidelines and professional standards. However, they also described that current government policies prohibit the implementation of some recommendations. Specifically, lack of consideration of available research evidence in policy decisions that govern how some medicines are distributed and used by various cadres was cited. Two focus groups discussed that the administration of misoprostol by community healthcare workers is currently not supported by policy in Uganda, despite studies in Uganda that have demonstrated safety and effectiveness of this approach (under the supervision midwives). Lack of up-to-date guidance on recommended practices was also highlighted, particularly around the use of misoprostol for prevention of PPH in health facility settings where drugs like oxytocin and ergometrine are also available. Recent improvements in misoprostol availability in facilities have not yet translated into their routine use; there have been instances of misoprostol expiring unused.

Factors affecting implementation were also identified at the provider level including: negative attitudes (for example around the use of the partograph); lack of buy-in from providers (particularly, their resistance to change); a lack of knowledge and skills (for example in administering magnesium sulphate and in interpreting the partograph); and the lack of opportunities for professional development for providers. The need for clear task shifting/sharing policies was highlighted, as the current role definitions of certain cadres were thought to hinder implementation. At the patient/community level, factors identified included traditional beliefs (eg: misconceptions around certain medical conditions, such as eclampsia); a lack of knowledge and awareness of certain health behaviours (such as a fear that mobility during labour can negatively affect outcomes); and the role of socioeconomic status in seeking and accessing healthcare in Uganda. The high value placed on companionship during labour in Uganda was cited as an important facilitator.

Respondents ranked three recommendations as the most feasible (Table 1). In the small group discussions, participants were asked to consider potential strategies to implement priority recommendations. Each group was assigned 1–2 guidelines to focus on during discussions; however, implementation strategies that were broadly applicable were also welcomed. Table 1 summarises the six implementation activities relevant to prioritized recommendations.

Tanzania

A total of 15 stakeholders completed in the pre-workshop survey. Survey respondents represented six different regions across Tanzania with Dar es Salaam the most highly represented region (50%). The survey respondents varied in terms of role and the level of the healthcare system in which they were situated. The majority of respondents (n = 11) selected the prevention and treatment of PPH guideline as the highest priority in Tanzania.

A total of 32 stakeholders participated in the in-person workshop. FGDs explored priority recommendations, as well as factors influencing their implementation. At the healthcare system level, participants identified a lack of access to resources (including drugs, equipment, supplies and human resources) and challenges to referring patients to higher-level facilities safely and quickly. Other factors included poor continuity of antenatal and postnatal care for women who cannot afford or access services, the need for improved monitoring and evaluation of medicines distribution, and the need for improved policies that support implementation. At the provider level, identified factors included: healthcare providers’ existing beliefs and attitudes regarding practices, and the need for their buy-in as part of implementation efforts; the need for improved knowledge, skills and self-efficacy amongst providers; and the need for continuous training, coaching, and professional development to ensure skill retention and adherence to guidelines. At the community level, identified factors included: women’s health-seeking behaviors and preference for care; the role of community champions in promoting care-seeking in early labour, and financial constraints affecting women’s abilities to seek and use care.

In the ranking exercise, five recommendations were deemed the most feasible to implement in the Tanzanian context (Table 1). In small groups, participants considered potential implementation strategies—a summary of these strategies is shown in Table 1.

Ethiopia

Activities in Ethiopia focused on WHO recommendations on prevention and treatment of postpartum haemorrhage. Fifty-three stakeholders completed the pre-workshop survey and nineteen stakeholders participated in the in-person workshop. The survey identified ‘use of uterotonics’ as the highest priority (66%) in clinical area within the guideline.

Findings from FGDs described issues at the level of the healthcare system, including: access to resources (a lack of drugs, supplies, personnel and facilities); challenges to drug procurement, distribution and management (including stock-outs, particularly in rural areas); data collection and monitoring (particularly, a lack of routine data collection and reporting on maternal deaths and a lack of facility-level audit and feedback to improve provider performance); the need for policies to mandate protocol adherence; need for incentives in facilities to allow more equitable distribution of human resources, and for more visible and accessible guidelines and protocols. Notably, while participants agreed readiness to change was important, there were mixed views regarding current readiness. At the Ministry of Health level, the level of readiness for health system changes related to guideline implementation was perceived to be high. However, at the provider level, some participants perceived resistance to change (i.e., lack of buy-in, low morale), while others disagreed, indicating that providers would recognize the importance of the issue and will support change regardless of frustrations related to low morale.

Several factors at the level of the healthcare provider that may affect guideline implementation were also identified. Some participants indicated that provider resistance to using available guidelines, lack of adherence, and lack of buy-in by providers can negatively affect use of guidelines. A lack of knowledge and skills was also identified. Participants described a lack of knowledge exchange opportunities between senior healthcare providers and newly trained providers, and a need for more training, supportive supervision and mentorship opportunities. Positive inter-professional dynamics and a clear understanding of one’s role were discussed as factors that may facilitate guideline implementation at the provider level. At the community level, factors identified included: preferences regarding childbirth location (home births in Ethiopia are common); a lack of knowledge and awareness amongst pregnant women about the risks of home birth (and the need for more and better education to address this) and the financial, economic and social barriers preventing women from accessing to healthcare services.

In the ranking exercise, five recommendations were deemed most feasible to implement in the Ethiopian context (see Table 1). Based on these findings, ten recommended activities were formulated by the group to guide next steps of guideline implementation in Ethiopia (see Table 1).

Key Findings

Despite the different prioritized recommendations and contexts, barriers identified across countries were often similar. Findings suggested that provider knowledge was not the primary barrier to guideline implementation; rather, they described that poor implementation was largely due to practical issues. Health system level factors, including health workforce shortages, and need for strengthened drug and equipment procurement, distribution and management systems, were consistently highlighted as limiting the capacity of providers to deliver high-quality care, despite their willingness to do so. All countries identified the need for health policies to support these implementation efforts, specifically the need to address outdated or restrictive health policies that were not in line with current evidence (for example, policies governing what tasks midwives are allowed to perform). However, the need to improve knowledge and skills of healthcare providers (through education, coaching, mentorship and other professional development activities) was consistently identified as an important determinant of implementation at the level of the health provider.

Community/individual level determinants of guideline implementation were also quite consistent across countries. More generic barriers to utilizing care (including geographical, economic and social barriers) can complicate timely use of effective interventions.[37] For example, the benefits of oxytocin for prevention and treatment of PPH are likely limited when women present in facilities a long time after childbirth. These factors highlight the need for guideline implementation efforts to explicitly consider and address factors affecting communities and individual women. The KT literature reflects a disproportionate focus on interventions directed at healthcare providers,[29] with comparatively less knowledge on the effectiveness of interventions directed at individual women and communities.

Strengths and Limitations

The strengths of this study include the use of a mixed-methods approach that incorporates theory and evidence. It has permitted a standardized assessment of barriers and facilitators to inform implementation efforts, as well as offering a roadmap for future directions. This approach could be applied to other guideline implementation challenges. However, it has not yet been demonstrated that this approach is superior to other approaches in terms of its effect on clinical outcomes (such as morbidity and mortality); further evaluation is required.

Our study had several limitations. First, we had variable numbers of participants and responses to the pre-workshop survey, largely due to technological difficulties. As the survey was circulated to large groups of people to maximize number of responses, the denominator (and response rate) is unknown. As the survey was anonymized, we are unable to assess what proportion of survey responses came from workshop participants. While we regard the survey as a useful resource to inform the workshop discussions, with experience we came to consider the diversity of participants represented within the focus groups as a more important determinant of success. While we aimed to engage a mixed group of stakeholders, community representatives and advocates were under-represented. More effort on how to best engage them is important (for example, considering separate community-specific research activities).

Although these activities were conducted in multiple countries, the scope of the activity was deliberately narrowed to specific priorities for implementation within each country, with consideration of broader contextual or systems factors that can affect or impede implementation. The findings cannot necessarily be generalized to other LMICs (with their own contexts and challenges), or to other implementation activities. However, other countries and organizations could adopt and adapt a similar methodological approach to implementation of maternal health guidelines (or other evidence-based guidelines) in their own settings.[16] Certain barriers (such as the need for training, and improved drug and equipment procurement) are likely relevant to a range of implementation efforts.

Lessons Learned

Several lessons were learned in executing these activities, which could be incorporated into adaptations of this methodology. In some countries, a large number of recommendations were initially considered. This requires additional time and effort and adds complexity to activities. Focused, sustained implementation efforts addressing few (rather than several) KT gaps are more likely to succeed in the longer term. Hence we advise narrowing the focus at the outset to as few recommendations as is practical.

Identifying and engaging local stakeholders as champions was critical—ideally, such stakeholders are well-resourced and motivated, have local networks and partnerships that can be leveraged to achieve project objectives, and can navigate local challenges. Such stakeholders should be engaged early and often, with a sincere emphasis on local leadership. Conversely, staff turnover of key individuals (such as within Ministries of Health) can endanger guideline implementation efforts—institutional ownership by the appropriate local entities (such as the Ministry of Health, WHO Country Office or university department) is therefore advised. Consideration should be given to (where relevant) aligning or combining guideline implementation efforts with existing local or national initiatives on improving quality of maternal healthcare. Effort should also be made to maintain ongoing engagement with workshop participants, such as sharing resources on coaching, technical support, funding or other opportunities.

In the FGDs, we noted a tendency to over-emphasize barriers; additional attention to considering facilitators is advised. We came to appreciate that the “right mix” of stakeholders in the workshops was essential. Including policymakers, researchers, clinicians and other stakeholders with sufficient seniority and power to make decisions was necessary to ensure buy-in for future implementation efforts. Representation from frontline clinicians (including doctors, midwives, and nurses) meant that the discussions focused on the practical reality of delivering care, which can often be missed in discussions by senior policymakers alone.

Next Steps

In this paper we have reported on activities relating to the first few steps of the KTA cycle. The GREAT Network partners continue to support activities in these countries related to subsequent steps—findings from these will be the subject of a separate, future publication. Evaluation of how this approach can improve guideline implementation (in terms of both process and outcome indicators) will also be required. For example, we recently conducted a process evaluation of the outcomes in Kosovo two years after the workshop.

Guideline implementation efforts in lower-income countries need to be more consistently documented and reported, so that all can learn from successes and failures in the field. The local working groups, with support from the GREAT Network, are developing and implementing activities based on these findings. For example in Myanmar, findings of this activity stimulated revisions of Ministry of Health policies regarding the scope of auxiliary midwifery practice and training, including administration of misoprostol for PPH prevention (prior to referral), management of puerperal sepsis with oral antibiotics, and performance of neonatal resuscitation. The workshop outputs and related discussions played a major role in influencing the policy changes. However, successful implementation of the policies and strategies identified in these activities will require strong leadership at multiple levels, sustained engagement of multiple stakeholder groups and mobilization of necessary financial, technical and human resources to support them.