Understanding quality improvement collaboratives in Ethiopia
A health worker supports a mother in Ethiopia.
Quality Improvement Collaboratives (QIC) are a common approach to bridging the quality of care gap in health facilities but little is known about implementation realities, especially in low-income settings. Implementation often occurs with little consideration of the mechanisms of change or the influence of context, and this may explain why QICs have had varied impact. A better understanding of QIC processes would help determine if they are suitable for all contexts and what adaptations are needed.
We conducted research on a large-scale maternal and neonatal health QIC intervention in Ethiopia implemented by the Ministry of Health and the Institute for Healthcare Improvement. We conducted:
- Research to understand the mechanisms through which the QIC worked and the influence of local context, in six health facilities and using qualitative methods.
- Research to determine whether QICs change the knowledge and motivation of health workers, using a quantitative before-and-after comparison.
Research exploring the collaborative element of QICs in a traditionally hierarchical setting, the health system drivers of data falsification and factors influencing motivation.
QIC mechanisms and the influence of context: We found that between-facility learning sessions increased awareness and focus on quality, and that participants learnt from experts about quality improvement (QI) methods and from each other about their implementation. They were motivated to achieve more when good results were publicly acknowledged or by seeing peers performing well. The shame of poor achievement inhibited participation for some; and some lower cadres felt alienated by the language and content of learning sessions. Within facilities, new structures and processes were created, although these were often fragile. QI team members learnt from, and were both supported and motivated by mentors, to whom they felt accountable. Infrequent mentor visits or visits by less skilled mentors were discouraging and impacted QI functioning.
In some facilities those outside of the QI team felt alienated by not being invited to learning sessions, a lack of knowledge transfer and not being visited and supported by mentors.
These mechanisms were more prominent, and QI more functional, in facilities with strong and supportive leadership, a culture of teamwork and openness and with fewer structural or resource challenges. These contexts worked together to form a positive environment for QI which enabled quality to become a priority; QI was seen as relevant because the foundations of health care were met. Where they existed, cultures of teamwork, shared goals and an active approach to solving problems meant that staff were flexible in their roles, more able to implement change ideas and that QI structures and processes were more stable as they were internally driven and owned. Teamwork and engaged leadership also meant that QI knowledge was transferred outside of the QI team, which enabled a culture of improvement to develop, reduced the impact of staff turnover and meant that those external to the QI team felt part of the process rather than alienated by it.
QICs effect on knowledge and motivation of health workers:
QIC improved knowledge of what history questions to ask a woman at the first ANC visit, what examinations and investigations should be performed and what drugs should be provided. Motivation was generally high and there was no evidence that it was influenced by QIC. Demotivating factors were beyond the influence of the QIC e.g. low salaries and lack of equipment.
The work confirms the need to carefully consider context in the planning and implementation of QICs
Facilities that more successfully implement QIC have an enabling environment and are most likely to already have characteristics that foster quality
In facilities with a less enabling environment, QIC structures and processes may be fragile and externally driven, with frequent visits by high quality mentors required. This has implications for implementation as well as for impact and sustainability
- In health facilities where the essential foundations for health care are not met, staff may struggle to see how QICs can improve quality, and their ability to deliver quality care and to engage with QICs may be reduced. Implementers need to consider whether QIC can effectively function in such settings and what additional support is needed
- Collaboration and support for the facility QI team from others in the facility is not automatic and is influenced by existing workplace cultures. Without broad buy-in QIC structures and processes can be alienating and poorly understood
- QICs can improve knowledge but may not influence motivation levels where these are largely driven by factors external to the QICs
Investigating the quality of care network in Ethiopia
Under the leadership of WHO, the Quality of Care Network (QCN) was established in 2017 in 11 low- and middle-income countries to improve maternal, newborn, and child health. In Ethiopia, the goals of QCN were aligned with government policy and national-level engagement was strong. There was a good fit between QCN activities and past experiences of quality improvement in the country, and to a large extent the Network was institutionalised within the health system. However, two key limitations were apparent: (i) more engagement was needed at the sub-national level; and (ii) government financial commitment to continue activities beyond the initial implementation period was lacking. As a result, it is not likely that QCN would be sustained in its current format, although some characteristics of the Network were likely to be carried forward.
The Quality of Care Network aimed to strengthen health system leadership for quality of care, accelerate and sustain quality improvement actions, facilitate learning, and develop mechanisms for accountability. The goal was, by 2022, to halve maternal and newborn mortality in health facilities in Network countries, as well as stillbirths, and to improve the experience of facility-based care. Network countries were Bangladesh, Cote d’Ivoire, Ethiopia, Ghana, India, Kenya, Malawi, Nigeria, Sierra Leone, the United Republic of Tanzania and Uganda. The vision was that the Network would be embedded within these member countries and continued beyond the initial implementation period.
Relatively little is known about how Networks such as QCN operate in low- and middle-income countries. To understand how QCN operated, the IDEAS project joined a UCL-led evaluation of the Network in Bangladesh, Malawi and Uganda, adding in the experience of Ethiopia.
Research questions included:
- Which capacities were available to enable Network functioning?
- Which actions were taken to sustain QCN beyond the initial implementation period?
- What influenced the quality of QCN data in health facilities?
Capacity for Network functioning
Individual, organizational and system-level capacities all played an important role in shaping implementation success in network countries, and these levels were inter-linked. Across all levels, actions that enabled leadership, motivated and trained staff, and created a positive culture of data use were critical – from the policy making arena to the day-to-day frontline practice. Some characteristics of QCN actively supported these levels, for example shared learning forums for continuous learning, a focus on data and tracking progress, and emphasising the importance of coordinated efforts towards a common goal. However, inadequate health system financing and capacity hampered network functioning, especially in the face of external shocks.
Actions taken to sustain QCN
Although vulnerabilities were observed, there was evidence that actions were taken to institutionalize QCN within country health systems, to motivate micro-level actors, plan opportunities for reflection and adaptation from the outset, and to support strong government ownership. But financial uncertainty was not pro-actively managed, community ownership not always fostered, and actions were least strong at the sub-national level.
Overall, evidence suggested that the QCN model would not be sustained in its original format, largely because of financial vulnerability and insufficient time to embed the innovation at the sub-national level. However, the efforts made to institutionalize QCN in existing systems meant that some characteristics may be carried forward within broader government quality improvement initiatives.
Quality of QCN data
QCN placed considerable emphasis on the importance of good quality data for learning and tracking progress. This emphasis had a positive effect on perceptions about data and data use for decision-making, with actors across the health system describing the potential power of data. However, in reality there were only limited improvements in the quality of data being generated. New data points were introduced to registers but not all data points were integrated with the routine health information system, causing some duplication of effort. Facilities also continued to lack the skills or resources needed to routinely record accurate data.
Understanding quality improvement in Lagos State, Nigeria
Quality improvement collaboratives are increasingly popular in low- and middle-income settings, often being implemented on a large scale. However, relatively little is known about the influence of context on implementation; consequently, new implementers may not benefit from the knowledge of what worked, what was adapted, and why. Working with partners in Lagos State, Nigeria, we studied the implementation of a complex quality improvement initiative that was driven by government and supported by a non-governmental organisation. Evidence revealed considerable influence of internal and external contextual factors, necessitating adaptations throughout implementation.
Since 2015 the Lagos State Ministry of Health, the Primary Health Care Board and managers of private facilities implemented the Nigeria Healthcare Quality Initiative (NHQI). NHQI was a QI intervention using a modified collaborative learning approach, guided by the principles described by the Institute for Healthcare Improvement, but contextualised to the Lagos health system. Three facility types were enrolled: public primary healthcare centres (PHCs), public secondary hospitals and private facilities. A broad set of shared change concepts was defined but it was anticipated that the complex needs of the different facility types would result in numerous local adaptations.
Through a qualitative multiple-case study, in which we defined a case as a collaborative of each facility type, we sought to understand whether and how NHQI implementation differed between facilities in Lagos State. We examined:
- What adaptations were made to NHQI to enable implementation.
- How implementation differences were influenced by contextual factors relating to facility type, health system and the stakeholders.
The theory of change for NHQI stated three core activities (collaboratives, measurement, and capacity building), each of which needed to be adapted to suit the implementation context in Lagos.
- Establishment of collaboratives needed to be adapted according to the characteristics of each facility type, taking into consideration the size of the collaborative, the number and geographical spread of the facilities within that collaborative; the governance structures of facilities and where power lies, including political power; and the level of service, reflecting whether predominantly preventive, promotive or curative services were provided.
- Measurement activities for facilities entailed identification of indicators to track performance, tools to measure accountability and QI readiness, and routine analysis of data for decision-making. Public hospitals could respond relatively easily to this activity but PHCs found it more difficult, and private hospitals had a weaker accountability mechanism for transparency in tracking data.
- Capacity Building activities needed to be designed for different levels: training for the state QI team and medical directors of private facilities focussed on governance; while facility QI team training focussed on QI skills and activities. In reality, plans for capacity building had to be adapted frequently because of high staff turnover.
Despite a common theory of change, implementation of the initiative needed to be adapted to accommodate the local needs, priorities and organisational culture of each facility type. Of note, in public facilities, the local governance structure could be adapted to facilitate QI coordination, but similar adaptations to governance were not possible for private facilities. Our findings underscore the importance of taking account of prevailing political commitment, the adaptability of available governance structures and the characteristics of facility types when planning QI implementation.
Our findings reveal that some QI priorities were common across facility types, often driven by the external context of health system leadership and external stakeholders. But many priorities were shaped by facility-level context such as the availability of clinical subject experts, available time and capacity of the facility QI teams, facility culture, the availability of data, and available finances.
The public hospitals and, to some extent, private facilities, but not PHCs, focused on complication management to enhance better health outcomes. This focus was explained by the level of care expected of public hospitals, a relatively high prevalence of maternal complications, availability of external support to conduct maternal death reviews, and availability of specialist doctors. Conversely, change concepts relating to tools, including utilities such as water and power supply, were mainly present in PHCs.
Individual leaders from government and NGOs play critical roles in influencing QI priorities at the facility level through coordination, support, mentorship and coaching to strengthen capacity of facility QI team and staff. These leaders, or sometimes the lack of them, were found to be important in Lagos State where facility teams with least QI capacity prioritised many easy tasks, such as quality of meals, while failing to address problems that could have greater impact on health outcomes.
All Understanding Quality Improvement related outputs can be found under Resources by selecting that theme from Research Areas.