Programme theory maps

Alison Turner; Abeda Mulla; Andrew Booth; Shiona Aldridge; Sharon Stevens; Mahmoda Begum; Anam Malik

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Turner A, Mulla A, Booth A, et al. The international knowledge base for new care models relevant to primary care-led integrated models: a realist synthesis. Southampton (UK): NIHR Journals Library; 2018 Jun. (Health Services and Delivery Research, No. 6.25.)

The international knowledge base for new care models relevant to primary care-led integrated models: a realist synthesis.

Show details

Contents

< Prev Next >

Chapter 7Programme theory maps

This chapter maps the evidence base pertaining to the remaining five programme theory components. These maps are intended to be indicative, rather than representing an exhaustive review of the literature base, providing a starting point for further exploration. The focus of the maps is on signposting the relevant wider literature bases, which offer valuable, transferable learning for the design, implementation and evaluation of population health-based models of care.

M1: collective responsibility improves quality and safety outcomes

As described in Chapter 3, integrated pathways feature as a focal point in MCP models, driving collective responsibility for quality and safety. Nationally, the Right Care programme is developing pathways to ‘support Local Health Economies to concentrate their improvement efforts on where there is greater opportunity to address variation and improve population health’ (contains public sector information licensed under the Open Government Licence v3.0).²⁹⁸ Related closely to place-based contracting driving new outcome frameworks (R2) and more collaborative working (R3), this programme theory component relates to the alignment of processes, people and technology:

If providers share collective responsibility for outcomes through standard integrated pathways, then providers will improve the quality and safety of care.

The evidence map for this component of the programme theory is derived from five ‘sets’ of related evidence (Table 5). Three of these evidence sets relate to context and two relate to potential mechanisms:

TABLE 5

Evidence map: M1 – collective responsibility improves quality and safety outcomes

care for long-term conditions and multimorbidity (context)
population-based approaches to primary and community care (context)
integrated and collaborative care (context)
quality improvement (mechanisms)
service development (mechanisms).

The three contextual evidence sets are primarily theoretical, whereas those relating to mechanisms are underpinned by varying levels of empirical and theoretical evidence.

Outcomes relating to the quadruple aim

There is some evidence to suggest that integrated care pathways could contribute to improved health outcomes. A disease-management programme in Germany has established disease-specific objectives with defined treatment goals and specific criteria for referrals to secondary care. The programme has an emphasis on patient education and self-management. Patients choose a physician, who co-ordinates their treatment; the framework clearly sets out when specialists should be involved in patient care to avoid gaps in care provision.³⁰³^,³⁰⁸ Further afield, the shared guidelines and standard procedures of the Mayo clinic are considered to contribute to lower complications and improved clinical outcomes.³⁰³^,³⁰⁹ An end-to-end pathway may enable the focus on prevention common to many of the MCP logic models; one example suggests that integrated pathways may contribute to a reduced incidence of falls, following knee arthroplasty.³⁰² Although it is difficult to attribute population health outcomes to pathways, theoretically, integrated pathways may help to stimulate an environment that encourages communication, collaboration and peer review, thus leading to improvements in quality and safety.³⁰³ Audit and feedback can strengthen the impact of integrated pathways, identifying what works in improving health outcomes and identifying problems that may indicate potential improvements or amendments to the pathway.³⁰³^,³⁰⁶

Integrated pathways, through reducing waste and duplication, could contribute to improved cost-effectiveness and lower unnecessary utilisation. The Nuka system of care, run by the Southcentral Foundation in Alaska, is credited with achieving improved quality for a lower cost, demonstrated by a decrease in unscheduled care alongside improved patient and staff satisfaction.¹³⁹ Integrated pathways have been used to enable a shift of care from hospital to community settings, demonstrated by the Veterans Home Administration, which measured fewer hospital bed-days and overall admissions.³⁰³^,³¹⁰^,³¹¹ Integrated pathways may also lay the foundations to introduce technology to support the delivery of care. Shared health records have a role in driving patient safety and adherence to guidance, supporting smoother discharge and transition between teams or services.³⁰¹ Telecare and telehealth can support more home-based care through remote monitoring and, in some contexts, has been found to contribute to reduced bed-days and increased patient satisfaction.³⁰³^,³¹² Opportunities exist to embed technology within pathways; for example, patient deterioration can trigger a responsive multidisciplinary package of care to avoid further decline, based on an agreed care plan facilitated by record-sharing.³⁰³ Some countries have initiated condition-specific approaches, such as disease-management programmes in Germany, Sweden’s chains of care and Scotland’s managed clinical networks, finding the most impact on patients with long-term conditions. All of these approaches involve elements of integrated care, recognising the importance of co-ordinating care and working across systems, and creating pathways across traditional ‘single-disease’ pathways.³⁰³

Unintended consequences

Integrated pathways are dependent on behaviour change and multidisciplinary working. However, behaviour can be unpredictable, with the risk that practice may not change in the way intended. This could lead to inconsistency, fragmentation and ‘disharmony’, ultimately affecting the safety and quality of care.³⁰³ Integrated pathways intricately link the success or failure of providers, supporting a population health-based approach and shared accountability¹⁰² for outcomes, as illustrated by ACOs such as Kaiser Permanente.³⁰³ There is, however, a risk that misaligned responsibilities between different professions, departments and organisations can prevent the effective integration of pathways,³⁰³^,³¹³ particularly beyond defined boundaries.³⁰³^,³¹⁴ The intention to collaborate and consolidate brings potential benefits of economies of scale, but could backfire if the collective organisations are perceived as a monopoly, hindering competition; nonetheless, Curry and Ham³⁰³ suggest that competition and integration need not be mutually exclusive:

. . . there is no inherent contradiction between integration and competition provided that patients are able to exercise choice either within or between integrated care organisations.
Curry and Ham³⁰³

It could be argued that integrated pathways encourage a ‘cookie-cutter’ approach to patient care, failing to consider the individuality of patients – their needs, preferences and case mix.²⁹⁹^,³⁰² The integrated pathway may also assume a level playing field in terms of infrastructure, which could, in reality, be strikingly different across settings.²⁹⁹ This highlights a need for strategies to achieve closer co-ordination of care.³⁰³

Mechanisms

A virtuous cycle, akin to the theory of the MCP model, could operate where clinicians and managers work more effectively across boundaries to attain financial and quality goals, which then improve cost-effectiveness, which can support investment in quality initiatives.³⁰² In such an environment, integrated care pathways can act as a focal point for clarifying and articulating the roles, responsibilities, processes and systems to underpin co-ordinated care, shared decision-making and care-planning.³⁰²^,³⁰³ This co-ordination can provide the stimulus for information-sharing, supporting patient-centred care. Where the design of integrated care pathways is based on engagement with patients and practitioners, patients may feel empowered and supported to maintain independence for longer.¹⁰²^,³⁰¹^,³⁰³ Such engagement can also be instrumental in enabling access to a more holistic approach to care.³⁰¹ From a patient perspective, an integrated pathway can translate into a single point of entry, a single assessment and a personalised care plan, delivered by a MDT (see Chapter 7, M2: multidisciplinary teams provide continuity for patients with long-term conditions/complex needs), which can help to reduce the burden on carers and address unmet needs.³⁰¹^,³⁰³ This co-ordinated standardisation can, in theory, address unwarranted variation, drive up quality and improve equity of access.³⁰² Patient satisfaction can be strengthened by the routine collection of patient-reported outcomes and regular patient feedback.³⁰³^,³⁰⁶

The process of developing integrated care pathways to incorporate shared values and goals can support the integration of teams and influence organisational culture and staff motivation, making it more likely that new processes will be embedded in routine practice.³⁰¹^,³⁰⁴ Closer working and standard pathways can help to identify those patients with more complex needs who need specialist secondary care, and those who are better managed in the community.³⁰³ Pathways informed by formal needs analysis – for example, Joint Strategic Needs Assessments – and community engagement can help to align the investment of resources within local communities,¹³⁹^,³⁰⁰^,³⁰⁵ factoring in the wider determinants of health and significant risk factors specific to local communities.¹⁰² This, in turn, can help to promote holistic and person-centred care and a sense of ‘shared ownership’, shared goals and incentives to address prevention.¹³⁹^,³⁰³

Early engagement of staff will help to ensure that pathways are comprehensive and realistic,²⁹⁹ and can ‘offer an opportunity for professionals to renew their implicit rules for operation and shape collective meanings and objectives’.³⁰² Affording more opportunities for communication across teams³⁰² can contribute to ‘an environment in which excellence in clinical care will flourish’.³⁰⁷ Standardisation facilitates monitoring and evaluation;³⁰² supported by transparent reporting, this can be a catalyst for quality improvement when larger teams are better able to exploit performance data for quality improvement, and to address wider issues, such as prevention and health promotion.³⁰³ Benchmarking, using national audits and other tools, can help to demonstrate the impacts of changes to pathways and to highlight issues relating to demand and capacity. A culture of personal professional responsibility and innovation among health-care professionals can be encouraged through more autonomy.³⁰¹^,³⁰⁴^,³⁰⁷ Multidisciplinary training and development, with a focus on communication and trust, can encourage more collaborative working.²⁹⁹^,³⁰¹^,³⁰⁴^,³⁰⁷ When individuals understand their contribution and that of others, pathways are more likely to be ‘owned’ by staff and successfully embedded in practice.¹³⁹^,³⁰² Problems can be identified and resolved collectively,³⁰¹ thus developing a learning culture and a ‘continuous quality improvement “mind-set” ’.³⁰⁷ Where a history of co-operation and trust pre-exists, this may be a smoother process; conversely, where relations are more adversarial, behavioural change is likely to require significantly more investment.¹³⁹^,³⁰³ Cultural barriers, such as competitive funding, poor relationships between senior leaders and risk aversion can stifle innovation and motivation.³⁰¹ Local barriers and facilitators need to be considered, for example, around data-sharing.¹⁰²

Although engagement is to be encouraged, there is evidence to suggest that the design and testing of integrated care pathways may be better managed by small focused teams, which are sufficiently agile to make the decisions needed to progress, but also sufficiently connected to ensure staff commitment and confidence.³⁰¹ Pathways alone are unlikely to deliver significant transformation – delivery may be contingent on new or expanded roles within primary and community settings.¹³⁹ In areas where there are long-standing recruitment issues, integrated pathways could be hindered by key staff shortages that, without workforce planning and development, could serve to undermine morale.³⁰² Multidisciplinary working will need to be supported by organisational development, to tackle cultural aspects, contractual incentives to influence behaviours and operational logistics, such as systems to enable shared data,¹⁰²^,³⁰¹ as demonstrated in Geisenger Health.³⁰³ Information systems, in particular, may help to facilitate timely and efficient information flow, supporting knowledge-sharing, the spread of good practice, consistent standards and communication between generalists and specialists.²⁹⁹^,³⁰³ Shared governance can promote care standards, patient safety and professional competencies;³⁰¹ conversely, instability can severely constrain the implementation of integrated pathways.¹⁰²^,²⁹⁹^,³⁰³

M2: multidisciplinary teams provide continuity for patients with long-term conditions/complex needs

Multidisciplinary working is core to the MCP model of care, underpinning care-planning and more proactive management of patients with long-term conditions and more complex needs, as described in Chapter 3. Specifically, this programme theory component relates to the impact on the continuity of care of MDTs:

If commissioners and providers train and fully engage staff in service transformation, then staff will drive the cultural change that underpins new ways of working.

Continuity denotes consistency and can be described from a patient perspective, entailing a continuous relationship, or from a provider perspective, entailing co-ordinated or seamless care.³¹⁵ It is a longitudinal concept, judged over time.³¹⁵ Informational continuity is an important element, dependent on shared records and information to ensure a full history.³¹⁶ Continuity is often described on a continuum with access, with suggestions that patients with long-term conditions prefer continuity of care, whereas patients with acute episodic needs prefer convenient access;³¹⁶ it is unlikely to be so simplistic in practice, and will be influenced by contextual factors. Continuity has been associated with reduced unnecessary utilisation and improved prevention,³¹⁶^,³¹⁷ so could be an important ingredient to achieving the quadruple aim through the use of MDTs.³¹⁸

The evidence map for this component of the programme theory is derived from five ‘sets’ of related evidence (Table 6). Three of these evidence sets relate to context and two to potential mechanisms:

TABLE 6

Evidence map: M2 – MDTs provide continuity for patients with long-term conditions/complex needs

management of long-term conditions (context)
population-based approaches to primary and community care (context)
integrated and collaborative care (context)
quality improvement in primary care (mechanisms)
team and organisational development (mechanisms).

The three contextual evidence sets are primarily theoretical, whereas those relating to mechanisms are underpinned by varying levels of empirical evidence.

Outcomes relating to the quadruple aim

There is evidence to suggest that an integrated approach to care, via MDTs, is associated with improved outcomes for patients with long-term conditions or complex needs;³⁴⁰ however, the majority of studies focus on single conditions – for example, reduced glycated haemoglobin values/improved glycaemic control in diabetes,³³⁵^,³⁴⁴ blood pressure control,³²⁸ treatment compliance,³²⁸^,³⁴⁴ reduced depression and anxiety³²⁴^,³²⁷^,³³⁴^,³⁵² – and it is unclear if benefits are sustained over time. As with any package of interventions, attribution is a challenge and, typically, MDTs are implemented alongside programmes of self-management, case management and enhanced pathways.³⁴⁰ It has been suggested that the ‘supportive network of professionals and peers’³⁵² contributes to improved health and well-being. Additional evidence suggests that MDTs can facilitate disease management and encourage self-management,³³⁴ reflecting the common features of MDT working, including case management,³⁵¹ care-planning,³⁵¹ the promotion of self-management and medication adherence.³²⁵^,³³⁴^,³³⁸^,³⁵³

Conversely, there seems to be little or no impact of MDT-working on patient-reported outcomes, such as activities of daily living,³³⁸^,³⁵³ functioning,³²² quality of life³²²^,³²⁵ and disability.³²⁵ There may be limited benefits for older people with multimorbidity or frailty.³⁵¹ The extensivist model (based on the Care More model in the USA)³²¹ is designed to provide more intensive support for older patients with very complex needs, and is being implemented as part of the MCP model in Fylde Coast. It can be challenging to measure the effectiveness of MDTs, notably the appropriate follow-up period within which to measure impact and the baseline (with a ‘ceiling effect’ of limited impact, whereby primary care was already well organised).³²²

Patients with multiple conditions may find it challenging to manage the complexity of their conditions and to prioritise needs, particularly within the context of deprivation or social isolation.³³⁶ It can be difficult to navigate the health and care system, and care packages developed by MDTs could be a critical element in improving support to patients and carers, as demonstrated in Torbay.³²³ Patient involvement may be facilitated by certain roles within MDTs; one qualitative study³⁵⁴ reported perceived improvements in care co-ordination and problem-solving, following the inclusion of a nurse practitioner within the MDT. Trust was also greater, possibly resulting from more time spent with patients and more opportunities for communication. The value of a MDT is in the opportunity for interdisciplinary diversity, with multidisciplinary participation emphasised as a key benefit of MDT meetings,³³⁷ supporting a range of social functions such as peer support, team-building and bonding,³³⁷ which can, in turn, have an impact on health outcomes.³⁵⁰

Several studies report reductions in service utilisation associated with integrated care via MDTs (hospital admission,³²⁶^,³³⁴^,³⁵⁵ A&E attendance,³³⁹ readmission³⁴¹ and length of stay)³²⁶ with a potentially higher impact in those ‘with at least moderate disease’.³²⁶ Few cost-effectiveness studies explore the effect of MDTs. Of those studies that do report costs, few have reported savings.³⁴⁴ One cost-effectiveness study estimated³³⁰ a mean cost per QALY of £14,248, arguing that collaborative care is expected to be less costly and more effective.

Unintended consequences

Integrated governance frameworks have been used to clarify roles, responsibilities and reporting arrangements,³⁴⁵ and the principle of MDTs seems to be valued by professionals.³³⁸ However, dissatisfaction and frustration can arise from perceptions of inequality and underinvestment in relationships,³³⁷^,³⁵⁶ potentially leading to issues regarding team processes and structure, available resources and sharing of information.³¹⁵^,³³⁷ This suggests that the reality of multidisciplinary working may diverge from the vision of increased co-ordination. Decision-making may even be impaired by the multidisciplinary approach, where the enabling cultural change (see Chapter 7, M3: engaged and trained staff expedite cultural change) has not occurred.

Attendance and participation at MDT meetings can be problematic as a result of workload pressures. An observational study,³⁴⁹ found that the number of participants ranged from 7 to 27, with an average of 14 participants; however, the mean number of people contributing was only three. There seem to be clear cultural differences across teams, with some dominated by relatively few people, and others demonstrating higher levels of participation. Medical dominance is perceived in some contexts to be a barrier to participants having an equal voice.³³⁷^,³⁴⁹ There may also be tensions from the underlying perspectives of different disciplines, which influence how practitioners perceive their role and relationship with patients.³³⁷ Cancer Research UK found a relatively low level of verbal contribution from nurses in meetings,³⁴⁹ which could impact patient care if the nurse has greater contact with patients and greater familiarity with patient preferences. The report suggests that 10–15% of MDT recommendations are not implemented, partly because such preferences were not taken into account.³⁴⁹ Contrary to this, Raine et al.³³⁷ did not find a link between discussions on preferences and the decisions made. Preferences may change over time, with differing opinions on how this knowledge should be incorporated into care-planning. Interestingly, Raine et al.³³⁷ found that those teams that were more medically dominated showed a stronger record of implementing decisions made. Implementation was found³³⁷ to be influenced by socioeconomic factors, which may have implications for the role of MDTs in building community resilience (see Chapter 6), and this is noted as an area that needs further investigation.

Significant variation in how MDTs operate reflects the different contexts within which teams are working. However, this could reflect a lack of clarity on best practice and a need for guidance.³³⁷ Delays in decision-making have been reported, as a result of a lack of availability of key staff and/or information,³³⁷^,³⁴⁹ with staff shortages being cited as having a strong effect on attendance. Notwithstanding attendance and participation barriers, the caseload of MDTs can have an impact on the quality of discussions. The Cancer Research UK study³⁴⁹ found that meetings lasted up to 5 hours; the mean length of patient discussions was 3.2 minutes, and over half of the discussions were < 2 minutes long.

Mechanisms

The professional composition of MDTs is variable and typically comprises three to eight members, with GPs and nurses most frequently represented. The inclusion of voluntary sector organisations (VSOs) in MDTs has recently increased, often to support self-management and more holistic care.³⁵⁷ Involving other public services, such as housing, enables a care plan that addresses the wider determinants of health, and supports patient-focused goals to tackle social isolation.³⁵⁸ Specialist nurse roles, with arguably more capacity than GPs, can help to support patient education³²⁹ by spending time with patients to personalise information. This could be of particular benefit to patients who may struggle with more generic approaches, for such reasons as language, learning difficulties and cognitive decline. The ‘drip feed’ of information can support a continuous relationship.³²⁹ For patients with multimorbidities, decisions on the trade-offs between the benefits and harms of treatments can be supported by specialist roles.³²⁹ Conversely, it has been argued that patients with very complex needs might need more support from GPs.³³¹

In bringing together multiple services, the role of care co-ordinator can be critical in tackling fragmentation,³⁵⁹ supporting relational continuity between the services and the patient; informational continuity between services and with the patient; ensuring accountability to the patient; and acting as the patient’s advocate to ensure patient-centred care. Nurses without advanced qualifications or advanced training have reported feeling ill equipped to take on the care co-ordination role, highlighting the importance of support (information systems, monitoring processes, guidelines), as well as training and development.³⁴⁷

Space and time for informal communication and shared communication methods are identified as enablers,³⁴⁶ underpinning access to, and the exchange of, essential information and ideas. MDT working can be facilitated by a shared understanding and vision of integrated care,³⁴⁰^,³⁴⁶ translated into clear, agreed objectives.³⁶⁰ Operationally, explicit processes³³⁸^,³⁴⁵^,³⁴⁶ can help to clarify expectations; for example, mandating attendance in job descriptions, regular feedback on team performance, provision of administrative support and sharing of information with other practitioners involved in pathways. An absence of such clarity can have an impact on the implementation of decisions.³³⁷ Organisational development can support the development of shared goals, values and outcome measures.²³⁹

To effectively manage multimorbidity, interventions may offer greater benefit if targeted at specific combinations of common conditions or specific problems facing comorbid patients,³³² highlighting the role of dedicated case managers.³²⁷^,³³³ Risk stratification and predictive modelling are often used to support case-finding,³⁶¹ and several recent studies present alternative methods to predict future utilisation.³¹⁹^,³²⁰ Sharing and access to patient records is critical in MDTs working across organisational boundaries to support clinical decisions and more holistic support.³²³^,³⁴³^,³⁴⁵ The use of technology is variable,³³⁷ challenging information-sharing within current systems. The experience of the Buurtzorg model suggests that technology can underpin timely communication, operational transparency, knowledge-sharing and quality improvement, through team access to audit and performance data.³⁴³^,³⁴⁵

System leadership,³⁴⁶^,³⁶² with its attributes of building a common vision and engaging with colleagues and partners to build commitment, may act as a trigger to develop the collaborative culture that is key to MDTs. Maintaining a shared vision that is clearly communicated, alongside an explicit commitment to quality improvement through benchmarking and peer review, are considered to be essential to creating the environment for integrated working.³³⁷

The vanguards seem to recognise that integrated working is unlikely to emerge simply by forming teams. Team development is needed at all levels of the system – not just at the executive level – and investment in organisational development is needed to underpin structural and cultural change.³⁶⁰^,³⁶³ Without explicit development, teams can lapse into ‘domain thinking’.³⁴⁸ Specific training and development can help to define a shared language and strengthen relationships.³⁴⁸ Development may involve coaching, leadership training, 360-degree feedback, personal development and action-learning sets.³⁶⁴

Although not a multidisciplinary model, the underlying philosophy of the Buurtzorg model,⁵⁵ with its emphasis on autonomy and empowerment, may offer valuable learning. Care is delivered by self-organising teams, typically comprising 8 to 12 nurses or health-care assistants. Teams are given the freedom to develop their own models for joint working with other professions. Value-based recruitment has created a virtuous cycle, in which the high levels of patient and staff satisfaction have attracted highly motivated individuals who perpetuate the cycle. The Buurtzorg mission is based on a shared sense of purpose, which ‘appeals to professional pride’, aimed at stimulating a culture of innovation and improvement. Teams are embedded within communities, developing detailed knowledge of their ‘clients’, their families and community resources; this informs problem-solving, and, in turn, helps to build the resilience of patients, families and communities.³⁴³

The mix of professional backgrounds in MDTs can enable ‘participants to transcend their own professional roles and routines, leading to learning and a more collaborative environment’.³⁴⁵ Support for reflection, innovation, open communication and supportive colleagues can create an environment in which participants feel part of the team, supporting a belief in interprofessional care.³⁴⁵ On a practical level, a nominated lead to co-ordinate and chair meetings, supported by administrative resource, can help to nurture the team as it develops,³⁴⁵^,³⁴⁸ with appropriate personal development. This need for development is emphasised by West et al.,³⁵⁶ who note that leadership skills are unlikely to emerge from clinical experience alone and will need nurturing through training and support.

M3: engaged and trained staff expedite cultural change

Closely related to the emphasis on multidisciplinary working is the aspiration for a systematic approach to engagement and development of the workforce. This is referenced earlier in Chapters 4–6, demonstrating engagement and development as important enablers to new care models:

If commissioners and providers train and fully engage staff in service transformation, then staff will drive the cultural change that underpins new ways of working.

The evidence map for this programme theory component is derived from five ‘sets’ of related evidence (Table 7). Two of these evidence sets relate to context and three to potential mechanisms:

TABLE 7

Evidence map: M3 – engaged and trained staff expedite cultural change

population-based approaches to primary and community care (context)
integrated and collaborative care (context)
quality improvement in primary and community care (mechanisms)
team and organisational development (mechanisms)
leadership development (mechanisms).

The two contextual evidence sets are primarily theoretical, whereas those relating to mechanisms are underpinned by varying levels of empirical evidence.

Outcomes relating to the quadruple aim

Findings from the UK National Staff Survey suggest that staff engagement is the best overall predictor of NHS organisations’ outcomes (care quality and financial performance; patient satisfaction; staff absenteeism; and health and well-being).³⁷⁴ The impact of staff experience on the quality and safety of care has been explored, following the Francis inquiry³⁸⁰ and the Berwick report,³⁸¹ with suggestions that strong team-working and clear aims among staff can contribute to patient outcomes.³⁸²^,³⁸³ Engagement has also been linked with organisational productivity, quality of patient care, lower mortality rates and better financial performance.³⁸³ In one study, practices demonstrating higher leadership scores were more likely to implement clinical improvements.³⁶⁸

Leadership development, through training and increasing workforce capabilities, is seen as a particularly critical element of workforce development. Staff on leadership programmes, such as the Leadership Challenge programme, have reported improved organisational impact, including raised awareness of costs and cost-savings, as well as improved patient care.³⁶⁴ The experience of staff is often linked to patient satisfaction.³⁵⁰^,³⁶⁰^,³⁷⁴^,³⁸⁴^,³⁸⁵ Staff views of leaders are strongly related to patients’ perceptions of the quality of care. Where leadership enables effective team-working, health-care delivery is more effective, and organisations exhibit higher levels of innovation and patient satisfaction.³⁷⁴ Team working, in particular, has been found to be a critical ingredient in high-quality care, particularly in relation to care for long-term conditions.³⁶⁵

Higher levels of support for staff from managers has been associated with higher numbers of patients reporting better care.³⁵⁰^,³⁷⁴ Greater engagement has also been linked to safer patient outcomes and significant associations with patient satisfaction, patient mortality and infection rates.³⁸⁵ When staff experience is positive, for example when they feel that they have ‘meaningful roles’, individual staff outcomes (e.g. job satisfaction) and organisational outcomes (e.g. absenteeism and staff turnover) have tended to improve.³⁵⁰ Conversely, higher turnover may exist where working environments are perceived as negative, because of bullying, harassment and aggression.³⁵⁰ A learning organisational culture, exemplified in many high-performing services, has been linked with organisational commitment, job satisfaction, innovation and performance.³⁶⁶ Linking with the theory components M2 and M4, learning organisations typically demonstrate an effective use of data to reflect and learn from performance at the individual, team and organisation levels.³⁷¹

Unintended consequences

The criticisms of leadership development programmes, referenced earlier, can result in a process-driven and task-focused approach to change, which is incongruous with the system leadership style advocated for complex change.³⁶²^,³⁷³ There is a risk that organisational development cannot keep pace with the skills development needed to design, deliver and evaluate service transformation. On a broader scale, the ability of formal education to keep pace has been questioned.³⁸⁶ For example, problem-based learning to develop emotional intelligence, strategic planning and organisational awareness will take time to establish.

An inevitable tension exists between the competing priorities of continuing to run existing services, while planning and commissioning new services. Focusing efforts on the development of selected cohorts could limit opportunities for ‘cultivating greater levels of understanding and communication networks across professional groups’,³⁷¹ and runs the risk of developing subcultures of unnecessary variation.³⁵⁰ Staff engagement is recognised as a critical enabler to service transformation,¹⁴⁶ but in an environment of ambiguity, volatility, complexity and uncertainty,³⁶² it can be challenging to maintain efforts in a meaningful way.³⁷⁵

Mechanisms

Organisational development could usefully focus on empowering teams to explore opportunities to improve and innovate. Positive staff experience has been associated with the ability to contribute towards service and quality improvement at work and strong communication.³⁵⁰ Clarity on the opportunities for involvement and engagement in change has been identified as ‘an important contextual element’.³⁷⁵ Multidisciplinary learning, potentially working via networks, can help to cement the vision of integrated working³⁶⁵^,³⁷⁸ and develop the adaptive skills needed to implement and evaluate service transformation. Leadership and quality improvement skills in primary care have been identified as a priority; team dynamics may also be influenced by value-based recruitment.³⁶⁵

Leadership development often focuses on processes, targeting more senior or medical roles, thereby missing the opportunity to develop skills more widely.³⁷¹ Various groups within health care, notably primary care staff, have not engaged as much with leadership development, possibly because of capacity issues.³⁶⁵ This points to an opportunity to engage more of the workforce in leadership development. A ‘learning organisation’ culture can be encouraged through processes and systems to support the sharing of knowledge, data and feedback.³⁶⁴ With leadership support, including time for reflection, teams can be trained and empowered to audit and benchmark their performance, analyse problems and share knowledge.³⁶⁴^–³⁶⁶ For clinicians, autonomy may be a particularly important factor, enabling more collaborative and innovative behaviours.³⁷⁹^,³⁸⁷ Evidence suggests that learning and development should be seen as an ongoing commitment; incremental change to encourage continuous learning and improvement³⁶⁴^,³⁷⁵ has been shown to sustain cultural change.

High levels of motivation are associated with higher levels of productivity, and tend to be reflected in positive relationships with patients/service users.³⁶⁷ Motivation is influenced by levels of engagement and the working environment; continuous high pressure and evidence of discrimination are associated with lower levels of motivation and patient satisfaction.³⁵⁰^,³⁸² A positive environment may be encouraged through clear goals³⁸² and a systematic approach to learning and development.³⁶⁶ The attributes of senior leaders are an important influence here, with a key role to play in inspiring collaborative working through fair-mindedness, the empowerment of staff, vision and organisational awareness:³⁷⁹^,³⁸⁶ ‘capacity building and effective leadership development implies social systems to help build commitments among members of a community’.³⁸⁶

Integrated care depends on a system perspective, and much has been written about system or collective leadership,³⁶²^,³⁶⁹^,³⁷⁰^,³⁷⁴ where there is a ‘notional dispersal and sharing of leadership’.³⁷⁹ Distributed leadership models, which allocate time and resources to staff engagement, can ‘create a shared sense of energy’,³⁸⁸^,³⁸⁹ activating ‘a learner’s sense of psychological safety’. Involving people in projects can help to build the credibility of the change programme, demonstrating the feasibility of change and encouraging uptake, spread and sustainability.³⁶⁶^,³⁷⁵

Interpersonal skills are a critical element in creating ‘the foundation for making good ideas into tangible improvements’;³⁷² this is especially the case with complex change, for which senior clinicians will need to influence others to contribute expertise and ideas. A longer-term view is needed to build the relationships that are needed to support complex change, and this will also help to develop insight to spot opportunities in an environment of ambiguity and uncertainty.³⁷² A key priority is to develop primary care staff, at all levels, to support quality improvement and innovation,³⁶⁵^,³⁷⁶^,³⁷⁷ with a clear role for mid-level managers in creating and sustaining change.³⁶⁸^,³⁷⁸

The pace and scale of change involved in new models has the potential to demotivate staff. Although evidence suggests that employing a range of staff can extend the current skill mix and improve working life, an accompanying risk of stress and anxiety highlights the importance of mutual support, which can be encouraged by team development.³⁶⁵^,³⁶⁷ It is important to staff that they feel that their roles make a difference to patient care, which increases their trust in their employer; a sense of shared values and a feeling of belonging can build mutual respect, which is particularly important in potentially stressful contexts. From these perceptions of trust and respect, motivation and satisfaction can increase, demonstrated by lower levels of staff turnover. Similarly, when staff feel like they are under less pressure, training and appraisals seem to increase in frequency. Personal experience is strongly affected by instances of negative behaviours – bullying, harassment, discrimination and aggression – particularly, when it is perceived that it is inadequately addressed, leading to low morale, and potentially increased absence, which affects patient satisfaction.³⁵⁰^,³⁶⁷

M4: system learning embeds and sustains transformational change

System learning is a thread that runs through several of the identified programme theory components, demonstrating its critical role in transformation. Evaluation and feedback loops are presented as underpinning new ways of working within enhanced primary care and MDTs, and informing strategic commissioning and place-based accountability:

If MCPs learn and adapt quickly using evaluation/monitoring loops and knowledge sharing, then MCPs will sustain transformational change.

System learning takes place at multiple levels, notably within MCPs, across the MCP vanguard sites, and, ultimately, through the lessons learnt, as they are shared within the overall health system. As ‘system-learning’ is a nebulous concept, Figure 19 is included to describe it, within the context of a logic model. The evidence map for this programme theory component is derived from six ‘sets’ of related evidence (Table 8). Three evidence sets relate to context and three to potential mechanisms:

FIGURE 19

Logic model for programme theory component M4: system learning embeds and sustains large-scale transformational change (context).

TABLE 8

Evidence map: M4 – system learning embeds and sustains transformational change

large-scale transformational change (context)
population-based approaches to primary care (context)
complex adaptive systems (context)
learning theory (mechanisms)
quality improvement (mechanisms)
audit and feedback (mechanisms).

The three contextual evidence sets are primarily theoretical, whereas those relating to mechanisms are underpinned by empirical evidence.

Outcomes relating to the quadruple aim

In contrast to other programme theory components, the achievement of transformational change does not map directly to the quadruple aim. Instead, it is best seen as an intermediate outcome by which the medium- to long-term outcomes articulated by the quadruple aim might be achieved. An alternative view might locate transformational change as a subsequent context, within which the achievement of outcomes might be demonstrated through the monitoring of outcomes and the creation of feedback mechanisms.

Population health

A key underpinning theory to US ACOs is that the allocation of a population-based budget set against the delivery of agreed outcomes will provide the incentive for providers to invest in activities to encourage better health, thus reducing utilisation.³⁹⁷ In their review of large-scale approaches to primary care, Pettigrew et al.³⁹⁰ examined the claim that population health management enables improvements in the quality of care and reduces unwarranted variation. They presented four published observational before-and-after studies in support of this claim, concluding that the results appear promising, demonstrating a step-wise improvement in most, but not all, areas.³⁹⁰ A qualitative study³⁹⁸ examined by the same authors³⁹⁰ evaluated quality and safety processes, and provided staff views on their job satisfaction and staff views on patient experience. Study results were equivocal, with evidence to suggest that cross-coverage and greater skill mix may inadvertently contribute to fragmentation of care and a loss of continuity of care. Pettigrew et al.³⁹⁰ observed that, collectively, these studies support the theory that large-scale general practice collaborations may provide a better environment for standardisation of care, greater community resilience, investment in technology and improved access to care.

Other large-scale population-health providers exist across the globe (see Ewbank et al.³⁹⁹ and Figure 16). Ribera Salud, a private service provider in Valencia, encapsulates the attitude of a population-health approach, focused on improving health outcomes for the whole population, while managing demand and achieving value for money.³⁹⁹ This incentivises Ribera Salud to intervene at the most effective moment.

Cost-effectiveness

A supplemental theory underpinning US ACOs is that receipt of a population-based budget incentivises providers to shift care from higher-cost hospital sites to community-based settings.³⁹⁷ Ribera Salud in Valencia was contracted to provide secondary care, extending the contract to primary care for financial reasons.⁴⁰⁰ Overall, Ribera Salud has delivered care improvements at a 26% lower cost than providers within the wider Valencia region.⁴⁰⁰

Pettigrew et al.³⁹⁰ highlighted a lack of evidence to identify the characteristics (e.g. size, governance) of large-scale general practice that are likely to contribute to sustainability. Arguments for large-scale primary care organisations tend to rest on improved financial sustainability. Having reviewed the literature, Pettigrew et al.³⁹⁰ concluded that further research is needed to evaluate the effects of general practice operating on a large-scale. In particular, such effects should be evaluated over the long term in relation to effects on the workforce, patient experience, clinical outcomes and costs.

Patient experience

A 15-month mixed-methods study showed that three-quarters of practices were now involved in bigger partnerships or federations,¹⁹⁰ but this had a minimal impact on performance against a range of quality indicators, and patient satisfaction had fallen.⁴⁰¹ Rosen et al.¹⁹⁰ recommends a phased introduction of alternative contract models for large-scale general practice organisations and integration with community providers, citing a lack of evidence that large-scale general practice leads to improvements in the quality and value of care.

Staff/provider experience

Achieving population-based care within a NHS context requires practices to organise into larger-scale practices. The Primary Care Workforce Commission acknowledges that larger-scale general practice collaborations could enable the delivery of a wider range of services, offer better opportunities for staff development and training and allow more effective relationships with commissioners, specialists, hospitals and social services.⁴⁰² Mechanisms for achieving such improvements relate to ‘investing in technology, strengthening clinical governance, standardising procedures, performance monitoring and benchmarking, peer review and feedback, spreading best practice, and having a population-based approach to services’.³⁹⁰

Unintended consequences

A counterargument to measurement activities is their counterproductive or unintended effects: ‘what gets measured influences behaviour’.⁸⁸^,⁴⁰³ Transformational processes are less easily measured than transactional activities, possibly contributing to a countertransference from the longer-term objectives to short-term measurable goals.⁸⁸ Those working in health care split their time between the ‘two jobs’ of health care: ‘to do their work and to improve it’.⁴⁰⁴ Many front-line staff face competing priorities, with associated undue effort in ‘feeding the beast’ through metrics and accounting for performance.⁴⁰⁵ Metrics and learning, the third of five key components of accountable care (population; outcomes; metrics and learning; payments and incentives; and co-ordinated delivery), require both the means (i.e. metrics and processes) by which to monitor outcomes and the feedback loops required to learn from unintended outcomes.⁴⁰⁶ All components are related in a closed-system feedback loop.⁴⁰⁷ This principle has been fundamental to developments within the vanguard sites.

Leaders need time to realise the benefits from new models of care, not being accountable simply for short-term performance. Alderwick and Ham⁴⁰⁸ comment that the time needed to implement large-scale change in the NHS and to demonstrate its impact is often dramatically underestimated.⁴⁰⁹ At the same time, expectations from new care models have been overestimated, requiring a more realistic appraisal from policy-makers and health service leaders. Feedback loops need to blend quantitative and qualitative data in order to ‘make sense of the transformation effort’.⁸⁸

Vize⁴¹⁰ highlights a further paradox, namely that, at a time when increasing time, money and effort are being invested in integrated care, the system is experiencing pressures in funding research that explores the impact of integrated care on ‘costs, outcomes, and patients’ experiences’. Even when such evaluation activities are forthcoming, their results are not always used to transform practice.

A rapid realist review⁸⁸ identified five ‘simple rules’ of large-scale transformational change. One rule involves establishing feedback loops: ‘Almost without exception, successful large-scale transformational efforts were recognized and sustained through the careful identification of measures and judicious disclosure of those measures to those both inside and outside the organization’.⁸⁸

Do feedback loops work under all circumstances and contexts? Best et al.⁸⁸ attributes success within transformational change to two factors: (1) leadership commitment to reporting performance throughout and outside the organisation, in order to effect transformation (a necessary accompaniment of Vanguard status);⁴¹¹ and (2) the quality of the supporting information infrastructure (see programme theory in Chapter 7, M5: proactive population health is dependent on shared and linked data) in order to effect the aggregation of data for audit and performance management.⁴¹²^,⁴¹³ The literature offers vigorous debate as to if and how NHS performance management systems are comparable to those of, for example, Kaiser Permanente.⁴¹⁴

Large-scale longitudinal case studies conducted across 12 health-care systems identified five interactive elements considered to be critical to successful transformational change. The elements include an impetus to transform, leadership commitment to quality and improvement initiatives that engage staff in meaningful problem-solving. Additional elements include alignment between organisational goals and resource allocation, and integration across traditional intraorganisational boundaries.⁴¹⁵

A systematic review⁴¹⁶ of organisational change identified 56 recent studies, 13 of which were in health care. Limited differences were found between health care and other industry studies, suggesting a future productive line of inquiry. Published research reveals that transformational change is multifactorial, attesting to the need to adopt a complexity lens for evaluation.⁴¹⁶ A mixed-methods study of transformational change in the NHS, located in north-east England similarly revealed ‘a need to develop new methods to understand how change occurs, or fails, in complex settings like the NHS’.⁴¹⁷ In-depth mixed-methods case studies are required to supplement evidence synthesis work, such as our own report.

Complex adaptive systems (context)

Feedback loops are equally prominent within complex adaptive system (CAS) frameworks that seek to explain how transformation occurs.³⁹⁴^,⁴¹⁸^,⁴¹⁹ A CAS is ‘a collection of individual agents with freedom to act in ways that are not always totally predictable, and whose actions are interconnected so that one agent’s actions changes the context for other agents’.⁴¹⁸ The influential Institute of Medicine report, Crossing the Quality Chasm,⁴²⁰ played a major role in recognising how a CAS’s lens can help in understanding how to improve and transform health systems. It reinforced the idea that health organisations operate within CASs,⁴²¹ an idea further developed in a series of papers in the British Medical Journal.⁴¹⁸^,⁴²² In focusing not on organisational elements themselves, but on the interdependency of those elements, the CAS approach seems particularly well suited to the complexities of evaluating new care models.⁴²³

Even though CASs are complex and unpredictable, they are open to the influence of simple rules that are sufficiently flexible to allow for adaptation. Rather than trying to manage the complex system through top-down large-scale transformation efforts, a CAS approach seeks to mobilise the natural creativity of health-care professionals to evolve, through system learning, locally owned solutions by which to achieve quality.⁴²⁴

Complex adaptive systems, such as that within which integrated care operates,⁴²⁵^,⁴²⁶ pose a further challenge with regard to evaluation; with so many different aspects to care, the system itself inevitably changes over time as practice develops.⁴¹⁰ Evaluators seek to measure a moving target.

Best et al.⁴²⁷ employed an action research methodology to explore transformation within the British Columbia health system, using a systems-thinking perspective. Requirements for managing large-scale clinical change included creating a context to prepare clinicians for health system transformation initiatives; promoting shared clinical leadership; strengthening knowledge management, strategic communications and opportunities for networking; and clearing pathways through the complexity of a multilevel, dynamic system. More specifically, Sturmberg and Lanham³¹⁸ identify future priorities for research into systems-thinking and complexity theory within general practice, ‘applying nonlinear dynamics and empirical modelling to patient care, and to organizing and developing local practices, engaging in community development, and influencing health care reform’.

Feedback loops can generate change or stability within a system. Where feedback loops generate change, two systems that initially are quite similar may develop significant differences over time. Even the same system, after the passage of time, may bear little resemblance to its previous configuration. Because the context for each CAS is unique, and each CAS is context dependent, each CAS is unique.⁴²⁸ The implications for planning are far-reaching, requiring that notions of ‘control’ over the process of change should be avoided, as should language that emphasises ‘overcoming resistance’.⁴²²^,⁴²⁹ Instead, efforts should be directed towards iterative planning and practice cycles that acknowledge that successful action is less about meeting targets and more about shifting the system’s behaviour through generic guidance and steering mechanisms. Changing the principles by which people carry out their work is considered to be more important than attaining seemingly arbitrary predefined targets.

Complex adaptive systems require constant monitoring and adaptation if transformational change is to be achieved. CASs challenge the linear logic models utilised by the vanguards; they are ‘self-organizing and constantly adapting to change; they are driven by interactions between systems components and governed by feedback; and they are nonlinear and often unpredictable, with changes on one part of the system producing unexpected changes in other parts’.⁴³⁰ Although the realist method offers one particular lens for exploring such complexity, further work is required to develop methods for the evaluation of real-world CASs.⁴³¹ The recognition of system complexity opens up the use of a diverse range of ‘systems-thinking tools’,⁴³² which might include knowledge synthesis, concept-mapping, social network analysis, programme budgeting and marginal analysis and system dynamics modelling.⁴¹⁹

Mechanisms

Learning theory (mechanisms)

The link between system-thinking and learning theory is explicit:

Systems thinking is an iterative learning process in which we replace a reductionist, narrow, short-run, static view of the world with a holistic, broad, long-term, dynamic view, reinventing our policies and institutions accordingly.
Reproduced from Sterman⁴²⁹

The challenge for 21st-century health services is to ‘aim not merely for change, improvement, and response, but for changeability, improvability, and responsiveness’.³⁹⁵ This requires a change from enabling not just competence, but also capability. Effective implementation requires a system-wide approach in which ‘the needed capabilities are combined, articulated and developed simultaneously in the same delivery system. Networks of providers will need to put energies and effort into the management of change and ensuring capabilities are embedded’.²²⁰ Commentators recognise that needed capabilities are often in short supply and investment in leadership and organisation development is required if NHS providers are to be ready to take on additional risks and responsibilities.⁴³³

System learning requires an environment within which individuals are able to develop sustainable abilities appropriate for a continuously evolving organisation³⁹⁵ and that enables organisations to capitalise on local knowledge to generate continuous improvements.⁴³⁴ Such a supportive learning environment enables an organisation to harness relationships and the skills and capacities of individuals in the system.⁸⁸ A learning culture, focused on developing the capability, capacity and confidence of staff and service users, provides an environment for the collaborative problem-identification and solving, experimentation and evaluation that is key to addressing systemic issues.⁴³⁵ The US ACO experience confirms that successful ACOs manufacture mechanisms and systems for learning (and then act on that learning).¹²⁶ Such learning systems must be agile in driving rapid quality improvements, as they empower staff and service users to address shortcomings in care.¹²⁶

‘Rapid-cycle’ evaluation is an identifying characteristic of many organisations that seek to achieve transformational change.⁴³⁶ In a study of transformational change in acute care, Greenhalgh et al.⁴³⁷ emphasise the criticality of ‘tools and techniques for quality improvement, including systematic data capture with rapid feedback loops (plan–do–study–act cycles)’. Rapid-learning cycles enable organisations, such as the ChenMed organisation, which offers high-quality primary care to older people in the USA, to test different ways of working across different contexts,¹¹⁷ tailoring approaches in the light of agile feedback. However, the evidence base for plan–do–study–act (PDSA) methods, notwithstanding their prominence in system transformation, modernisation and quality improvement, is equivocal at best. Few formal objective evaluations of their effectiveness have been conducted, and the achievements of PDSA approaches range from significant improvements in care and patient outcomes through to no improvement at all.³⁹⁶ A systematic review of PDSA cycles in quality improvement found that many inherent advantages of the PDSA method, such as the facility for iterative learning, and the ability to build up to larger cycles from small-scale changes and capitalise on learning from abandoned cycles, remain underexploited.³⁹⁶

In Saskatchewan, Canada, primary health-care teams were equipped with the tools and resources to allow them to understand, monitor and compare their own performance as part of the drive towards large-scale transformation.³⁹¹ Rather than reflecting a strong accountability structure, such tools place learning at the ‘forefront of the re-design process’;³⁹¹ building in the capacity for rapid learning is likely to be equally critical to the development of integrated care in the NHS.

The link between individual and organisational learning requires the identification and negotiation of shared mental models.⁴³⁸ Closely related is the concept of double-loop learning, which arises when such shared mental models enable action at an organisational level.⁴³⁹ Whereas single-loop learning corrects any defect or mismatch between desired and actual outcomes without challenging any underlying assumptions, double-loop learning effects a more profound, longer-lasting and, thus, transformational change by changing not just the defect but also reconfiguring the underlying decision-making rules.

Research and evaluation consistently emphasise the importance of organisational and system architecture in supporting high performance in health care and other sectors.⁴⁴⁰ Collins²⁰⁰ claims that how the vanguards work together as effective learning systems will prove to be critical to their success. Membership of the network of vanguards is predicated on a principle of ‘reciprocal benefit’;⁴⁴¹ sharing of data and current practice in turn opens up the prospect of accessing the data and practice of other organisations and, thereby, accelerating the lessons learnt.

Quality improvement (mechanisms)

Within large-scale primary care, stronger performance monitoring and feedback has been cited as a way of improving clinical quality and removing variation.³⁹⁰ Performance data on quality and cost are key to improvement at an overall practice and individual physician level. Rosen et al.¹⁹⁰ report that monitoring and improving quality represents a constant priority for the central management team, with significant resources allocated to identifying and addressing weak performance in individual practices and services. Dashboards and other performance management systems frequently feature in vanguard documentation. An internal modality report stated that the use of technology to centralise information across all sites and to improve monitoring systems had allowed leaders to hold partners to account for variations in quality and performance.¹⁹⁰ Fifteen indicators were used for monitoring the performance of the primary care organisations across four domains: (1) prescribing, (2) hospital activity, (3) the QOF and (4) patient satisfaction (the GP Patient Survey).¹⁹⁰

Variable results from quality improvement can be attributed, at least in part, to the fact that ‘improvement strategies typically treat practices as something . . . that can be independently isolated and “fixed”, and as such, optimally approach change through the implementation of sequential steps, or stages, to achieve a set objective.’³⁹³ Understanding primary care organisations as CASs would recognise their need for support in ‘achieving reflective, adaptive, and action-oriented approaches to managing the future’. This represents a different feedback culture from the current emphasis on ‘hard’ performance indicators. Sheaff et al.⁴⁴² describe ‘concertive control’, whereby members monitor each other’s work through peer pressure and peer review, using a combination of organisational culture and technical knowledge to implement collective decisions. Concertive control represents a natural evolution, as organisations move from hierarchical control to self-managing teams.⁴⁴³ Rosen et al.¹⁹⁰ contrast a ‘directive’ model of change, whereby executive staff both determine the need for change and then facilitate staff to undertake it with a ‘consultative’ model of change, in which peer review and peer pressure are prominent. The latter model corresponds to ‘concertive control’ – exemplified by federations in which individual practices kept their own separate contracts.⁴⁴²

Audit and feedback (mechanisms)

Evidence on the value of audit and feedback, although not exclusive to a primary care setting, is considered to be quite conclusive, albeit to modest effect. A Cochrane review⁴⁴⁴ of audit and feedback activity recorded ‘small but important’ improvements in professional practice. A cumulative meta-analysis by the same authors led them to question if allocating further resources to testing the effectiveness of audit would be worthwhile.⁴⁴⁵ At least 20 systematic reviews report data on the use of audit and feedback in primary care.⁴⁴⁶ Audit and feedback in primary care achieved only modest effects, with a median change of 1.3%. Many unanswered questions remain, for example on whether single or multiple component feedback mechanisms are more effective and exactly what the mechanisms are by which audit and feedback achieves its effect. A realist review of clinical audit by the University of Groningen⁴⁴⁷ suggests four mechanisms for audit: (1) those relating to continuous improvement, (2) the fact that the process is ‘bottom-up’, (3) the active involvement of professionals and (4) a miscellaneous category of other mechanisms.

M5: proactive population health is dependent on shared and linked data

Closely related to the new ways of working (R1 and M2) and system learning (M4) is the need for linked and shared data to inform decision-making and continuous improvement:

If MCP staff are not able to access shared/linked data, then patients will continue to experience fragmented care.

The evidence map (Table 9) for this programme theory component is derived from four ‘sets’ of related evidence. Three of these evidence sets relate to context and one to potential mechanisms:

TABLE 9

Evidence map: M5 – proactive population health is dependent on shared and linked data

population-based approaches to primary and community care (context)
evidence-informed commissioning/value-based health care (context)
integrated and collaborative care (context)
quality improvement in primary and community care (mechanisms).

The three contextual evidence sets are primarily theoretical, whereas those relating to mechanisms are underpinned by varying levels of empirical evidence.

Outcomes relating to the quadruple aim

The importance of information to assist with public health is well recognised. At the local level, data are used to target interventions and monitor population health. At the national level, information is used to make informed decisions about prioritisation, planning and commissioning. At the global level, information assists with the burden of disease and tackling global health threats. Routinely collected health data are not only used by public health agencies, but also for academic research and the development of technology. Large population-level data can help to accurately identify social inequities, and target specific groups and geographical and regional challenges, which can be addressed through evidence-based health interventions.⁴⁵⁴ The Symphony Project in south Somerset is an example of the use of extensive data sets for the purpose of improving the integration of care. Data on utilisation, costs and patient characteristics have been combined with local intelligence to identify groups that would benefit the most from proactive care.⁴⁵⁰ To this end, examples also exist in which segmentation analysis combined with ‘big data’ has divided patient populations into distinct groups, for needs-based targeting of care models and intervention programs.⁴⁵⁵

Southampton City CCG is an example of an organisation that has used data to assist with the reduction of unwarranted variation in health care. NHS RightCare provided data that highlighted the key priority areas for the reduction of variation, which could increase the value of health care. The CCG used these data to target areas of concern for spend and quality indicators, which subsequently led to the development of an implementation plan to improve the management of gastrointestinal patients.⁴⁵¹ A UK study suggests that information on variations in costs, activity and outcome could be better exploited, perhaps indicating a need for tools to help understand the causes of variation in health-care delivery.⁴⁵⁶ Despite the availability of data, however, data on disease surveillance, cause-specific mortality, intervention coverage and other vital statistics remain underused data sources,⁴⁵⁴ suggesting missed opportunities in discovery and innovation.

The potential for effective information-sharing to achieve cost-savings remains inconclusive. One systematic review concluded that there is sufficient evidence to suggest a reduction in costs through the use of electronic health records,⁴⁵⁷ whereas another study suggests that, although such cost-savings are achievable over a long period of time, there is a high degree of variation dependent on the practice setting.⁴⁵⁸ The high costs of implementation can also function as a barrier to the uptake of electronic health records.⁴⁵⁹ Such costs may also not be recovered.⁴⁶⁰ Support for the cost-effectiveness of shared electronic health records remains limited and conflicting.⁴⁵⁹

The government-anticipated ‘big data’ and the use of large data sets for the purposes of analysis have been promoted as potentially reducing costs, with an estimated £20B of benefit over a 5-year period for the central government.⁴⁴⁹ This estimate was based on the understanding that the cost-effectiveness of services is, to some degree, dependent upon data transparency and effective data-sharing.⁴⁴⁹ To this end, linked data have proved to be useful in exploring the utilisation and costs of services. Patient-level linked data have been used to explore the NHS and social care costs of the care pathway associated with elderly patients recovering from a fall in Torbay, Devon, UK. In this case, cost data patterns proved to be helpful in identifying and targeting patients to fall prevention services.⁴⁴⁸

Information-sharing can help to improve patient–doctor interaction, including through the use of telephone, e-mail and web-based consultations, with tools such as webGP and askmyGP enabling patients to share information with clinicians to support clinical decision-making.⁴⁶¹ Modality, a super-partnership in Birmingham and Sandwell, and one of the MCP vanguards, operates a remote hub to support triage, self-management and same-day consultations, enabling 65% of health concerns to be dealt with remotely.⁴⁶¹ A similar system, based at south Somerset’s Symphony, one of the PACS vanguards, has established a web platform to enable patients to access their care plan, messaging capabilities (such as Skype™; Microsoft Corporation, Redmond, WA, USA), test results, information about their condition and telehealth functionality. The platform enables a ‘single view of clinical activity between the [integrated] hub and practices’, directly linked with the GP system. From an information governance perspective, such systems can also clarify permissions for sharing information with carers and family members, which can be useful in the event of a crisis or emergency or when the patient is reluctant to engage with technology.⁴⁶² Mobile phone applications are increasingly being used⁴⁶³ and may help to support patient empowerment and shared decision-making.

An earlier HSDR programme study⁴⁵² explored continuity and co-ordination in primary care, including the impact of informational continuity. Up to half of the patients at all five sites studied described instances of information, regarding history, diagnosis and treatment, not being passed between providers. Patients reported correcting health worker omissions and out-of-date information, and having to repeat information to the next practitioner. In one instance, a district general hospital delayed action on an urgent GP referral, resulting in emergency admission of the patient 2 days later. Poor protocols for sharing data and unclear responsibilities were cited as contributing factors. Informational continuity seems to decline across traditional boundaries, with some patients with long-term conditions maintaining their own paper-based histories.

Effective communication, to share and link data, underpins multidisciplinary and collaborative working,¹⁷⁴^,⁴⁵³ and can range from a basic model through to a shared care record with e-referrals, e-prescribing and e-discharge. Data-sharing, enabled by telemedicine, allows GPs to access advice from specialists; examples include GPs working with dermatologists and ophthalmologists to share data and enable remote specialist consultation, with mixed results.¹⁷⁴ Informational continuity, through shared records across primary and secondary care, seems to show potential in improving decision support outside of specialist clinics.¹⁷⁴

Data-sharing is considered to be one of the key factors driving high-quality care for chronic conditions, with the potential to improve efficiency, drive up quality and reduce duplication.⁴⁵³^,⁴⁶⁴^,⁴⁶⁵ The chronic illness care model⁴⁶⁶^,⁴⁶⁷ encourages the use of information systems to facilitate data-sharing and communication across traditional boundaries, helping to translate knowledge into practice. One example demonstrates integrated medical records for frail elderly patients providing read–write access for all health professionals involved in care, as well as patients and carers.⁴⁵² Such systems can be built on to develop tools and dashboards to alert MDTs to patients at risk of admission⁴¹⁷^,⁴⁵² and identify potential bottlenecks and delays.⁴¹⁷ In north London, an information technology tool was developed and used to support information storage, sharing and analysis to support integrated care programmes.⁴⁶⁸ Information-sharing does, however, depend on data quality; GP systems and templates may enable standardised and complete data collection.⁴⁵³^,⁴⁶⁹

Unintended consequences

The push for data-sharing and health information exchange has given rise to concerns surrounding privacy, confidentiality and control of data about individuals,⁴⁷⁰ specifically, concerns around user trust, data privacy, transparency of data ownership and the potentially intrusive nature of data analytics. Poor communication and confusing messages can turn public and patient attitudes against the sharing of personal data, as demonstrated in the issues surrounding care data.⁴⁷⁰ The risks and benefits should be clearly articulated, as should information on data governance and anonymisation techniques.⁴⁷⁰ Data governance is likely to become an even more important issue, given the predicted rise in wearable technology and digital applications, which have the potential to link directly to records to support decision-making and remote monitoring.⁴⁶¹ Privacy terms and conditions are less clear, and the process for opting out is sometimes not explicit.⁴⁷⁰

Inevitable concerns persist in relation to sharing data with patients and carers, with the ‘digital divide’ highlighting the socioeconomic disparities in access to technology,⁴⁶² which can be exacerbated when patients with low educational attainment and health literacy rely on carers for information. NHSE ran a widening digital participation programme from July 2013 to April 2016, which provided digital health training to 221,941 people who were at a risk of being digitally excluded.⁴⁷¹

The focus on shared data and technology-driven data collection could unintentionally signal a shift away from the social contract that underpinned the traditional doctor–patient relationship,⁴⁷⁰ with criticisms regarding the ‘depersonalisation’ of care and risks to provider autonomy.⁴⁷² There is an additional risk that staff may feel that data collection is being driven by targets rather than patient-centred care.⁴¹⁷

Mechanisms

Facilitated by new ‘disruptive’ technologies enabling large data sets,⁴⁷⁰ real-time analytics can be a catalyst to drive population-based health care. Future iterations of clinical decision support systems are likely to enable data linking across sectors, enabling primary and secondary care clinicians to access data to inform decisions at key points in the pathway.⁴⁶¹ Currently, information exchange is constrained by such factors as local information governance protocols, incompatible information systems leading to duplication and rescanning, unreliable connectivity and inconsistent clinical coding.⁴⁵² Digital strategies should encompass a range of technologies to support data-sharing and co-ordinated multidisciplinary care; shared electronic records, used in combination with predictive analytics; telehealth; and patient-facing tools, such as portals.⁴⁶⁹ Consideration needs to be given to infrastructure (e.g. data storage, real-time analytics and integration of distributed data sets);⁴⁷⁰ capacity and skills;⁴⁷⁰ and financial sustainability.⁴⁵⁴^,⁴⁶⁴ Some form of central administration is needed to ‘monitor, mediate and facilitate data-sharing across multiple stakeholder organisations’⁴⁵⁴ and to manage protocols, processes and data specifications.⁴⁷⁰ At a national level, there is also a role to ensure interoperability between policies, systems, data standards and supporting market development,⁴⁵⁴^,⁴⁶¹ in order to avoid unnecessary variation⁴⁷² and conflicting policies.⁴⁶⁹^,⁴⁷³

The sustainability of data innovations is dependent on demonstrating benefits for patients, commissioners and practitioners.⁴⁷² Data-sharing needs to be supported by cultural changes, such as multidisciplinary working and support for shared decision-making.¹⁷⁴^,⁴⁶¹ Innovations are also likely to be more acceptable if patients and staff have been engaged, and communication on the legal and ethics implications has been clear.⁴⁷⁰

Both patients and health-care professionals will need support in developing the skills and knowledge to benefit from data-sharing.⁴⁷⁰ Training centres specialising in big data (large linked data from electronic health records and health data collected by personal wearable devices) and open data (the sharing of data sets) can help services to reap the benefits from data-sharing to address population health needs.⁴⁷⁰ In the absence of training, the ‘noise’ of large data sets may risk the misinterpretation and misuse of data.⁴⁷⁰

Numerous barriers to technology transformation have been identified, including the absence of standards, significant variation in workflows, inefficiency of health information technology, incompatible legacy systems, incomplete information flows and poor system design.⁴⁵²^–⁴⁵⁴^,⁴⁷² An absence of incentives for data-sharing, opportunity costs and disagreements around the use of data can also hinder effective information exchange.⁴⁵⁴ Poorly defined goals can result in uncertainty about what can be shared and which data need to be included,⁴⁶⁴ potentially leading to variation in the capability of organisations and services to share and act on data.⁴⁶⁴ Relationships based on trust and collaboration will be key to overcoming structural, cultural and ethics barriers.⁴⁵⁴

Chapter summary

The intention for these maps has not been to systematically cover all of the relevant literature. Indeed, it is clear that each of the evidence sets relating to the programme theory components holds the potential to sustain an in-depth analysis of empirical literature and the underpinning theory. The maps demonstrate the interdependencies between individual theory components. Structural developments, such as the introduction of MDTs and integrated pathways, are contingent on cultural change, which, in turn, needs to be stimulated through organisational development and system leadership behaviours. Staff autonomy and empowerment is identified as a critical thread in cultural change, associated with trust and collaboration. This shift to a more collaborative, population-based approach to health care can be hindered by misaligned systems and processes, and it is clear from the evidence base that shared data can offer the opportunity to improve the co-ordination and continuity of care for patients and to instigate improvement and innovation. System learning, within each vanguard and then across vanguards under the aegis of NHSE initiatives, can be demonstrated to be a key component of the New Care Models programme to a degree that has not typically been encountered with vanguard-type initiatives. Indeed, similarities can be detected with the Institute of Medicine’s Learning Health Care System model.⁴⁷⁴ In particular, feedback loops built into audit and formative evaluation could potentially trigger a range of different mechanisms, variously interpreted as responses to transformational change, learning or quality improvement resources.

Copyright © Queen’s Printer and Controller of HMSO 2018. This work was produced by Turner et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Bookshelf ID: NBK508137

Contents

< Prev Next >

PubReader
Print View
Cite this Page
Turner A, Mulla A, Booth A, et al. The international knowledge base for new care models relevant to primary care-led integrated models: a realist synthesis. Southampton (UK): NIHR Journals Library; 2018 Jun. (Health Services and Delivery Research, No. 6.25.) Chapter 7, Programme theory maps.
PDF version of this title (2.7M)

Programme theory maps - The international knowledge base for new care models rel...
Programme theory maps - The international knowledge base for new care models relevant to primary care-led integrated models: a realist synthesis
UDP-glucuronate 4-epimerase 1 [Musa acuminata AAA Group]
UDP-glucuronate 4-epimerase 1 [Musa acuminata AAA Group]
gi|2730182290|ref|XP_009417460.2|
Protein
Cyclin-dependent kinase B2-2 [Cucurbita argyrosperma subsp. argyrosperma]
Cyclin-dependent kinase B2-2 [Cucurbita argyrosperma subsp. argyrosperma]
gi|2053620620|gb|KAG7037976.1||gnl| DJN|SDJN02_01609
Protein
Chrozophora oblongifolia genes for 5.8S rRNA, ITS2, 28S rRNA, partial and comple...
Chrozophora oblongifolia genes for 5.8S rRNA, ITS2, 28S rRNA, partial and complete sequence
gi|1761012877|dbj|LC503611.1|
Nucleotide
Homo sapiens chromosome 10 paired-box protein PAX2 gene, complete cds; and PAX2 ...
Homo sapiens chromosome 10 paired-box protein PAX2 gene, complete cds; and PAX2 gene, complete sequence
gi|1036029750|gb|AH006910.2|
Nucleotide

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Bookshelf

The international knowledge base for new care models relevant to primary care-led integrated models: a realist synthesis.

Chapter 7Programme theory maps

M1: collective responsibility improves quality and safety outcomes

TABLE 5

Outcomes relating to the quadruple aim

Unintended consequences

Mechanisms

M2: multidisciplinary teams provide continuity for patients with long-term conditions/complex needs

TABLE 6

Outcomes relating to the quadruple aim

Unintended consequences

Mechanisms

M3: engaged and trained staff expedite cultural change

TABLE 7

Outcomes relating to the quadruple aim

Unintended consequences

Mechanisms

M4: system learning embeds and sustains transformational change

FIGURE 19

TABLE 8

Outcomes relating to the quadruple aim

Population health

Cost-effectiveness

Patient experience

Staff/provider experience

Unintended consequences

Complex adaptive systems (context)

Mechanisms

Learning theory (mechanisms)

Quality improvement (mechanisms)

Audit and feedback (mechanisms)

M5: proactive population health is dependent on shared and linked data

TABLE 9

Outcomes relating to the quadruple aim

Unintended consequences

Mechanisms

Chapter summary

Views

In this Page

Other titles in this collection

Recent Activity