Contribution to the literature
-
Quality improvement collaboratives are a widely used approach. However, solid evidence of their effectiveness is limited and research suggests that achievement of results is highly contextual.
-
Previous research on the role of context in quality improvement collaboratives has not explored the dynamic relationship between context, mechanisms and outcomes. We systematically explore these through a review of peer-reviewed and grey literature.
-
Understanding contextual factors influencing intended quality improvement collaborative outcomes and the mechanisms of change can aid implementation design and evaluation. This systematic review offers a novel programme theory to unpack the complexity of quality improvement collaboratives.
Background
Methods
Clarifying scope of the review
Search strategy
Screening
Data collection
Synthesis and reporting of results
Results
Search results
No. | Author (ref) | Year | Country | Collaborative name | Topic | Study aim | Health setting | No. facilities (individuals) in study | Study design | Published | Focus |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | Amarasingham et al. | 2007 | USA | Keystone Intensive Care Units Project | Central line associated bloodstream infection | Assess correlation between automation and usability of clinical information systems and clinical outcomes. | Intensive care unit | 19 (19) | Uncontrolled before-after | Peer-reviewed | Context |
2 | Ament et al. | 2014 | Netherlands | ERAS (Enhanced Recovery after surgery) | Colonic surgery | Explore strategies for sustaining ERAS | Hospitals | 10 (18) | Qualitative | Peer-reviewed | Mechanism |
3 | Baker et al. | 2018 | Tanzania | EQUIP (Expanded Quality Management using Information Power) | Maternal and newborn health | Investigate how different components of a QIC were understood and experienced by health workers, and contributed to its mechanisms of effect | District hospital, health centre and dispensaries | 13(16) | Qualitative | Peer-reviewed | Mechanism |
4 | Benn et al. | 2009 | UK | Safer Patient Initiative | Patient safety | Understand participants’ perception of impact of the pilot programme | NHS Health Trusts | 4 | Mixed methods: cross-sectional and qualitative | Peer-reviewed | Mechanism and implementation |
5 | Benn et al. | 2012 | UK | Safer Patient Initiative | Patient safety | Analyse impact of intervention of safety culture and climate and role of contextual and programme factors in changes. | NHS Health Trusts | 19 [2 merged in 1] (284) | Uncontrolled before-after | Peer-reviewed | Context and implementation |
6 | Burnett et al. | 2009 | UK | Safer Patient Initiative | Patient safety | Analyse perceptions of organisational readiness and its relationship with intervention impact | NHS Health Trusts | 4 (41) | Mixed methods: cross-sectional and qualitative | Peer-reviewed | Context |
7 | Carlhed et al. | 2006 | Sweden | Quality Improvement in Coronary Care | Acute myocardial infarction (AMI) | Evaluate effect of QIC on adherence to AMI guidelines | Hospitals | 19 + 19 controls | Non-randomised controlled before and after | Peer-reviewed | Context |
8 | Carter et al. | 2014 | UK | Stroke 90:10 | Stroke | Explain processes and outcomes of the QIC intervention | Hospitals | 11(32) | Qualitative | Peer-reviewed | Mechanism |
9 | Colbourn et al. | 2013 | Malawi | MaiKhanda | Maternal and newborn health | Evaluate impact and processes of change | Hospitals and health centres | 9 and 29 | Mixed methods: cross-sectional and qualitative | Grey | Context, mechanism and implementation |
10 | Catsambas et al. | 2008 | LMICs various | 35 collaboratives funded by USAID between 2002 - 2007 | Various: Maternal and newborn health, nutrition, HIV/AIDS | Document and evaluate the implementation and results of the Quality Assurance Project | Hospitals and health centres | N/A | External review - multiple projects | Grey | Context, mechanism & implementation |
11 | Dainty et al. | 2013 | Canada | Ontario Intensive Care Units Best Practice Project | Evidence-based care practices in Intensive Care Units | Understand staff perspectives on QIC and hypothesise theoretical constructs that might explain the effect of collaboration | Hospitals | 12 (32) | Qualitative | Peer-reviewed | Mechanism |
12 | Dixon-Woods et al. | 2011 | USA | Keystone ICU Project | Central line associated bloodstream infection | Develop an ex-post theory of the project | Intensive Care Units | n/a | Case description | Peer-reviewed | Mechanism |
13 | Duckers et al. | 2009 | Netherlands | Better Faster | Patient safety | Test whether consensus on perceived leadership support among physicians influences the relation between physician’s perception and participation. | Hospitals | 8 (864) | Cross-sectional | Peer-reviewed | Context |
14 | Duckers M. et al. | 2009 | Netherlands | Better Faster | Patient safety | Assess relations between conditions for successful implementation, applied changes, perceived success and actual outcomes. | Hospitals | 23 (237) | Cross-sectional | Peer-reviewed | Context, mechanism and implementation |
15 | Duckers M. et al. | 2011 | Netherlands | Better Faster | Patient safety | Describe how the first group of hospitals sustained and disseminated improvements | Hospitals | 8 (8) | Qualitative | Peer-reviewed | Mechanism |
16 | Duckers M. et al. | 2014 | Netherlands | Better Faster | Patient safety | Test whether perceived average project success at QIC level explains dissemination of projects. | Hospitals | 16 (84 out of 148) | Cross-sectional | Peer-reviewed | Mechanism |
17 | Feldman-Winter et al. | 2016 | USA | Best Fed Beginnings | Breastfeeding | Describe collaborative and present lessons learned from implementation. | Hospitals | 89(89) | Case description | Peer-reviewed | Mechanism and implementation |
18 | Horbar et al. | 2003 | USA | Vermont Oxford Network Newborn Intensive Care Units /Q 2000 | Quality and safety of neonatal intensive care | Describe collaborative and present implementation strategy. | Hospitals | Case description | Peer-reviewed | Context, mechanism and implementation | |
19 | Jaribu et al. | 2016 | Tanzania | INSIST | Maternal and newborn health | Describe health workers’ perceptions of a QIC intervention | Health centres and dispensaries | 11 (15) | Qualitative | Peer-reviewed | Mechanism |
20 | Linnander et al. | 2016 | Ethiopia | Ethiopian Hospital Alliance for Quality | Patient satisfaction with hospital care | Analyse impact of QIC | Hospitals | 68 | Cross-sectional and uncontrolled before - after | Peer-reviewed | Context and implementation |
21 | Marquez et al. | 2014 | 38 LMICs | Health Care Improvement Project | various | Document and evaluate the implementation and results of the Health Care Improvement project | various | N/A | External review - multiple projects | Grey | Context, mechanism and implementation |
22 | McInnes et al. | 2007 | USA | HIV collaborative under HRSA/HAB | HIV/AIDS | Assess whether participation in QIC changes care processes, systems and organisation of outpatient HIV clinics | HIV clinics | 52 (104) Intervention and 35 (90) Controls from non QIC sites. | Non-randomised controlled before and after | Peer-reviewed | Context |
23 | Mills and Weeks | 2004 | USA | 5 Veteran Health Association collaboratives between 1999 - 2001 | Various | To identify the organisational, interpersonal and systemic characteristics of successful improvement teams | Hospitals | 134 medical QITs in 5 BTS collaboratives | Uncontrolled before – after | Peer-reviewed | Context and implementation |
24 | Nembhard | 2008 | USA | 4 collaboratives supported by IHI | Efficiency in primary care; complications in ICUs; reducing adverse drug events; reducing surgical site infections | Understand participants’ views of the relative helpfulness of various features of QICs | Hospitals | 53 teams (217) | Mixed methods: cross-sectional and qualitative | Peer-reviewed | Mechanism |
25 | Nembhard | 2012 | USA | 4 collaboratives supported by IHI | as above | Study the use of interorganizational learning activities as an explanation of mixed performance among collaborative participants | Hospitals | 52 teams (48 hospitals) | Cross-sectional | Peer-reviewed | Mechanism |
26 | Osibo et al. | 2017 | Nigeria | Lafiyan Jikin Mata | HIV/AIDS | Discuss lessons learned from QIC implementation and analyse effect of QIC activities on process indicators. | Hospitals and PHC centres | 32 (16 intervention + 16 controls) | Mixed methods: UBA and qualitative | Peer-reviewed | Mechanism and implementation |
27 | Parand et al. | 2012 | UK | Safer Patient Initiative | Patient safety | Identify strategies to facilitate the sustainability of the QIC | NHS Health Trusts | 20 (35) | Qualitative | Peer-reviewed | Mechanism and implementation |
28 | Pinto et al. | 2011 | UK | Safer Patient Initiative | Patient safety | Evaluate influence of various factors on the perceived impact of QIC | NHS Health Trusts | 20 (635) | Cross-sectional | Peer-reviewed | Mechanism |
29 | Rahimzai et al. | 2014 | Afghanistan | Maternal and Newborn Health Facility Demonstration Improvement Collaborative | Maternal and newborn health | Document implementation and describe results of a QIC project | Provincial hospitals, health centres and posts in provinces + large referral hospitals in Kabul | Participating facilities in “Demonstration wave”: 25 in provinces and 6 in Kabul: Wave 1–2: additional 6 facilities. | Case description | Peer-reviewed | Mechanism and implementation |
30 | Schouten et al. | 2008 | Netherlands | Stroke Collaborative I | Stroke | Explore effects of QIC and determinants of success | Stroke services | 23 | Cross-sectional and before - after with reference group | Peer-reviewed | Context |
31 | Sodzi-Tettey et al. | 2013 | Ghana | Project Fives Alive! | Maternal and newborn health | Document implementation, describe results and lessons learned of a QIC project | Hospitals (district and regional) and health centres | N/A | Case description | Grey | Context, mechanism and implementation |
32 | Stone et al. | 2016 | USA | California Perinatal Quality Care Collaborative | Breastfeeding | Assess factors that that affect sustained improvement following participation. | NICUs | 6 (n/s) | Qualitative | Peer-reviewed | Mechanism |
Characteristics of included studies
Focus | Total | Country setting | Internal or independent programme review | Before and after (controlled or uncontrolled) | Qualitative | Cross-sectional | Mixed methods | |
---|---|---|---|---|---|---|---|---|
High-income | Low or middle income | |||||||
Mechanism | 12 | 10 | 2 | 1 | 0 | 7 | 3 | 1 |
Context | 6 | 6 | 0 | 0 | 3 | 0 | 2 | 1 |
Context and implementation | 3 | 2 | 1 | 0 | 2 | 0 | 1 | 0 |
Implementation and mechanism | 5 | 3 | 2 | 2 | 0 | 1 | 0 | 2 |
All | 6 | 2 | 4 | 4 | 0 | 0 | 1 | 1 |
Total | 32 | 23 | 9 | 7 | 5 | 8 | 7 | 5 |
Context
Category | No. of studies | Evidence synthesis | Quality of evidence (ref.) | ||||
---|---|---|---|---|---|---|---|
Relationship with outcome | Relationship with mechanism | Quantitative and mixed methods | Qualitative and review | ||||
1 Healthcare setting in which a QI intervention is introduced | |||||||
Facility size | N = 1 | No | No evidence that hospital size is associated with improvement in outcome. | - | Not discussed. | Medium [42] | |
Base line performance | N = 1 | Yes | Lower base line performance of hospitals is positively associated with magnitude of outcome improvement. | Yes | Lower base line performance is positively associated with active participation in QIC. | Medium [43] | |
Voluntary or compulsory participation | N = 1 | No | No evidence of differences in outcomes. | - | Not discussed. | High [44] | |
Factors related to health facility readiness | N = 5 | Yes/No | Inconclusive evidence of association between programme pre-conditions (staff, resources, usability of health information system systems, measurement data availability and senior level commitment to target) and outcomes. | Yes | Bottom up leadership style may foster more positive perceptions of organisational readiness for change. Limited clinical skills, poor staff morale and few resources negatively associated with outcomes. | Low [48] | |
2 Project-specific contextual factors | |||||||
External support | N = 6 | Yes | Quality, appropriateness and intensity of quality improvement support positively associated with perceived improvement in outcomes. | Yes | The number of ideas tested by quality improvement teams partly mediates the association between external support and perceived improvement. | ||
Quality improvement team characteristics | N = 4 | Yes | Inclusion of opinion leader, team functionality and previous knowledge or experience of quality improvement is positively associated with outcome. | - | Not discussed | ||
3 Wider organisational context and external environment | |||||||
Leadership characteristics | N = 5 | Yes | Supportive leadership is positively associated with perceived improvement in outcomes. | Yes | Supportive leadership may motivate physicians to implement quality improvement and may enable active testing of ideas by quality improvement teams. Lack of supportive leadership may demotivate and stall quality improvement team efforts. | ||
Health system alignment | N = 4 | - | Not discussed. | Yes | Alignment with national priorities, national-level quality strategy, and incentives systems is essential to enable leadership support. | Medium [46] |
In what kind of facility setting may QICs work (or not)?
What defines an enabling environment for QICs?
Mechanisms of change
How may engagement in QICs influence health workers and the organisational context to promote better adherence to evidence-based practices?
Themes (No. studies) | Evidence synthesis | Quality of Evidence [ref.] | ||||
---|---|---|---|---|---|---|
Description of relationship QIC component–mechanism–outcome | Contextual enablers of mechanism (or barriers) | Quantitative and mixed methods | Qualitative and review | |||
QIC component | Mechanism of change | Outcome | ||||
Health professionals -knowledge, skills & problem solving (N = 4) | Use of continuous quality improvement approach | • Refreshed knowledge • Reinforced confidence and skills in improvement topic area • Facilitated a problem-solving approach | Change in clinical practice enabled | • Quality and appropriateness (mix of clinical and quality improvement expertise) of mentoring • Leadership and work culture open to bottom up discussion and reflection • Health workers participating in quality improvement interventions have adequate clinical competences (or a complementary clinical skills training programme is accessible) | Medium [46] | |
Health professionals engagement, attitude and motivation (N = 8) | Formulating shared goals Alignment with national priorities and fit with existing practices Use of run-charts to visualise progress Dissemination of success stories Credibility of change package | • Increased motivation, by reframing improvement topic as desirable, urgent and achievable • Removed resistance to use of data • Increased Commitment to change | Increased engagement in QIC—may lead to increased success | • Intensity of mentoring to increase data literacy and use for decision-making, particularly in LMICs • Supportive leadership • Barrier: competing programmes and initiatives. | ||
Organisational climate (N = 4) | General QIC approach | • Facilitated teamwork and multi-professional collaboration within and across departments • Facilitated bottom up dialogue and discussion | • Quality and intensity of mentoring • Wider use of improvement tools beyond unit of focus | High [60] | ||
Leadership (N = 2) | General QIC approach | • Enhanced leadership engagement • Decentralised/shared leadership promoted through encouraging bottom up problem solving | Staff morale boosted | • Previous success with quality improvement • Alignment with institutional responsibilities and participatory working culture | ||
Organisational structures, processes and systems (N = 5) | Process mapping | • Definition of standard care processes facilitated | New expectations on performance generated | • Previous success with quality improvement • Alignment with institutional responsibilities and priorities • Complementary approach (beyond QIC activities) to institutionalise new ways of working e.g. incorporation in induction or staff training; performance management frameworks for accountability at the level of health workers and/or organisation | ||
Organisational culture (N = 3) | General QIC approach | • Development of habits for improvement facilitated | Normalisation of new practices | • Leadership open to new practices • Health system enabling decentralised innovation |
What is it about collaboration with other hospitals that may lead to better outcomes?
Themes (No. studies) | Evidence synthesis | Quality of Evidence [ref.] | ||||
---|---|---|---|---|---|---|
Description of relationship QIC component–mechanism–outcome | Contextual enablers of mechanism (or barriers) | Quantitative and mixed methods | Qualitative and review | |||
QIC component | Mechanism of change | Outcome | ||||
Shared community of practice (N = 7) | Collaboration with other hospitals | • Sense of community reinforced or created • Increased motivation, by supporting reframing of improvement topic as desirable, urgent and achievable | Health workers motivated and empowered to take action towards common goal | • Settings where a community of practice amongst clinicians exists or can be developed • Barrier: external pressures on hospitals incentivising competition v. collaboration. | ||
Platform for capacity building (N = 5) | Collaboration with other hospitals | • Platform to refine skills for improvement provided • Definition of standard care processes facilitated | • Settings with quality-focused HR systems, e.g. incorporating quality objectives in professional development and performance appraisals • Barrier: high performing hospitals have less to gain from collaboration • Barrier: Collaboration can be undermined by free-riding (not all facilities contribute equally) and social loafing (leaving it to others to support low performing hospitals) | |||
Demonstration role (N = 3) | Collaboration with other hospitals | • Feasibility of improving outcome of focus is demonstrated | Increased engagement in QIC | • Supportive leadership • External support to disseminate success stories • Barrier: Large hospitals may have less to gain from collaboration | Medium [71] | |
Friendly competition (N = 6) | Collaboration with other hospitals | • Reputational gain from improvement (or conversely risk of non-improvement) at individual and organisational level achieved. • Access to others’ data and benchmarking for internal gains enabled. | Normative pressures to conform (change practice and improve) created. | • Open sharing of data on mutual performance • Alignment with institutional priorities (lack of which contributes to perception that collaboration is stressful and time-consuming) • Geographically dense professional network • Non-hierarchical teams facilitating decentralised decision making • Barrier: competition for financial incentives linked to quality criteria |