Understanding new models of integrated care in developed countries: a systematic review

Relationship with health-care provider

A qualitative study from the UK,100 of the North West London Integrated Care Pilot, indicated mixed results regarding whether or not the intervention changed relationships between patients and providers. Although 54% of participants responded positively to a question regarding an improved relationship in the survey evaluation, the study reports that 54% of participants also responded that they had not experienced any changes. Two international qualitative studies273,279 report more-positive outcomes: that integration of a community worker into a team had improved patient knowledge and engagement,273 and that an ICP had improved patient involvement. 279

Patient understanding of treatment

One UK intervention study59 reported that ICPs were associated with improved patient understanding of treatment and reduced anxiety among patients.

Discharge planning

Two UK intervention studies59,63 reported that discharge planning had improved. The first of these59 evaluated the introduction of ICPs in hospitals and concluded that there was strong evidence that they improved discharge planning. Nursing discharge planning rose by 20% (p < 0.01) and there was a significant increase in general discharge planning, from 33% to 93% of patient records being examined (p < 0.01). The second study63 (which had a comparator group) found that discharge planning for older patients at high risk began within 2 working days of admission in most cases, compared with at the end of a stay or not at all prior to the addition of a social worker to a general practice. One systematic review191 also concluded that case conferences assisted discharge planning (although it reported mixed evidence regarding care planning for stroke patients).

Timeliness of care and/or patient flow

One UK intervention study49 reported that timeliness in terms of patient assessment, care packages being put in place and equipment being provided improved (across all of these elements) at three sites adopting a Kaiser Permanente^® (Oakland, CA, USA) approach. UK qualitative studies report positive perceptions that integrated care could co-ordinate services more effectively,98 improve the flow of patients out of hospitals,101 streamline services,109 lead to efficient and timely discharge,116 lead to more-rapid service delivery,121 increase co-ordination,124 provide continuous and pro-active care125 and improve efficiency. 146

Two systematic reviews suggested improved efficiency and reduced overlap following the introduction of ICPs184 or collaborative working. 153 Two international comparator studies214,238 supported this. The first study214 reported parent perceptions of greater satisfaction with care co-ordination (p = 0.058) compared with a control group, with parents reporting that services were more co-ordinated following the intervention (50% prior to the intervention and 80% post intervention; p = 0.003). A second study238 found that, compared with a control group, patients who received a care-co-ordinator intervention were more positive about the co-ordination of care (p = 0.001). However, the difference between groups in reporting improved care co-ordination was not significant. An international qualitative study282 similarly reported perceptions of a positive effect on care timeliness and efficiency. A non-comparative evaluation257 of a patient-centred medical home found that participants reported improved care co-ordination (p > 0.001).

Continuity of care

The international literature highlighted the increased continuity of care that integrated care may provide for patients. Bird et al. 198,199 described how a care co-ordination intervention provided continuity of engagement with patients, which allowed care facilitators to better understand the needs of patients and provide a point of contact. The authors of another study204 attributed frequent visits, continuity of care and ease of access to staff to the ability to deliver structured care at home, thereby reducing hospital usage. A MDT intervention in a psychiatric hospital was similarly described as providing continuity of care. 239 Continuity of care was highlighted as a key change by other international intervention studies. 225,227,238 One UK intervention study64 evaluating an ICP referred to increased continuity of care following the intervention.

Systematic reviews provide a more varied picture regarding whether or not continuity may improve as a result of integrated care. Myors et al. 184 described that continuity of care was achieved via a link worker or identified contact. However, another review160 reported patient perceptions of continuing issues in communication between agencies, resulting in continuity of care being broken. Another review175 concluded that there were few significant findings in favour of short-term programmes related to continuity of care, although the short-term nature of the programmes may have influenced this finding. A review of nurse-delivered care co-ordination interventions186 that reported staff perceptions found that they did not think that the programmes fostered increased communication or continuity of care.

A qualitative study from Australia267 provides a different perspective on continuity by distinguishing three different elements: information continuity, management continuity and relational continuity. The authors reported that continuity of information was important in effective care as it was perceived to reduce repetition and provide shared access to records.

Variance in practice and adherence to guidelines

Reduced variance following the introduction of ICPs was reported by two UK evaluation studies. 56,59 Kent et al. 59 found an absolute improvement of 35–40% in the recording of indicators from clinical guidelines and best practice. Another evaluation of an ICP70 described implementation as standardising decision-making. A general service redesign including the introduction of standardised documentation reportedly ensured consistency between staff. 139

One systematic review180 reported that adherence to guidelines improved in 14 out of 19 studies that were examined. Another184 concluded that pathways and guidelines helped to ensure consistency.

A UK qualitative study of a pathway in palliative care described greater consistency in care. 109 Another study,97 however, warned that pathways could constrain judgement. Greenhalgh et al. 110 found that common guidelines and protocols were difficult and time-consuming to achieve in practice.

Duplication

One UK qualitative study119 described a perception of reduced duplication following an integrated care initiative. Another129 reported the perception that working in a fragmented way causes duplication and waste. A systematic review184 reported that benefits of collaborative working were perceived to include reducing wastage and overlap in staff roles and duties.

Practitioner time and workload

One UK intervention study60 found evidence of a potentially positive reduction in staff time. This cross-sectional evaluation of MDT working reported that 68% of staff perceived that the average of 2 hours that was spent in meetings each week saved time later. The UK qualitative literature, however, provides few examples of perceptions of integrated care as being time-saving, although one study109 reported views regarding the use of the Liverpool Care Pathway in hospices. Staff described the pathway as time-saving as it streamlined documentation.

In one study,105 staff reported concerns regarding potential increases in workload that may result from a Pioneer programme. GPs also expressed concerns that an integrated telephone service may increase their workload. 113 Another study122 described staff becoming preoccupied with additional administrative burdens, time spent in meetings and completing care plans in an integrated care pilot. Team meetings were described as being long and time-consuming, with another paper,123 originating from the same integrated care pilot, suggesting that their frequency should be reduced. In a further paper134 from the same study, views were described as variable, with some staff perceiving meetings as valuable, whereas others expressed concern regarding time and cost. An evaluation of a new integrated GP role in cardiology137 found that this could distract from other GP work and that the associated administrative tasks had been underestimated. Workload had increased to the point that GPs used their own time for the new role. In another study,147 staff also reported increasing their working hours to meet patient needs and expectations from secondary care following the introduction of a new MDT model. Additional time required to enter data in unfamiliar systems was also described by staff. 149,150

One systematic review153 concurred with these reports of increased time being required, with improved documentation (ICPs) reportedly leading to the ‘burden of increased workload’.

In contrast, a non-comparator evaluation of a team-based approach in the USA243 reported that practitioners, when surveyed, reported that care co-ordination was time-saving (saving at least 30 minutes per month per patient). A qualitative study in the Netherlands,268 however, echoed the UK literature by reporting that staff perceived that collaborative initiatives were overly time-consuming.

Shared information

A key purpose of improving communication between staff is to enhance the exchange of patient information. One of the UK intervention studies64 included outcomes related to the transfer or sharing of information. This survey of staff found that 77% of staff perceived that accessibility of information had improved. The UK qualitative studies provide variance in reported perceptions of the effect of integrated care on the process of information sharing. One study109 found that staff described the value of knowing what other members of the team were doing following the introduction of ICPs. They also valued having clearly documented decisions that had been made by the family available for all staff to access. Another study125 of palliative care also described the sharing of written reports to aid collaboration. A third study143 evaluating pathways in palliative care described benefits from having a structured record, which was especially useful out of hours. It was mentioned, however, that pathways could be used too rigidly by some staff. 143

A UK qualitative study139 exploring a variety of initiatives reported that there had been limited progress in information sharing across different sites. Similarly, another study149 exploring the implementation of a pathway for frail elderly patients and rapid response team described information as being difficult to share because of there being different systems for health and social care. An evaluation of the North West London Integrated Care Programme151 reported difficulty in putting data sharing agreements in place and achieving operability between different health and social care systems. Staff working in integrated initiatives within community disability services reported concerns regarding the reliability and accessibility of care records, with some groups being unable to access NHS records. 107 Team meetings were perceived to be an important means of sharing information. 107

A systematic review of models of joint working160 (which included only UK studies) concluded that there could be difficulty in sharing information and getting access to information (because of incompatible systems). An international qualitative study267 reported issues regarding confidentiality, multiple and/or incompatible data systems and data ownership (governance). A study in Germany292 evaluating team meetings found that key information was not available in 4–5% of the cases discussed. A study by Berendsen et al. ,268 from the Netherlands, echoed the UK study discussed previously143 by describing staff views of integrated pathways as being too restrictive.

Accuracy/completeness of recording information

Three UK intervention studies59,64,70 (none with a comparator design) reported improvements in the accuracy and/or completeness of the recording of information such as initial assessments by staff following the introduction of ICPs in hospital or community settings, although it was noted that some elements of care remained poorly reported (such as disinfection). One UK qualitative study107 described concerns regarding shared care records being incomplete or out of date.

Systematic reviews of ICPs for stroke and nursing home patients at the end of life154,173 report an increase in the documentation of rehabilitation aims, communication with patients and carers, discharge notification plans and plans for ‘do not attempt resuscitation’. In contrast, a systematic review of teamworking171 highlighted gaps in communication and documentation. A non-comparator study from the Netherlands260 of the Liverpool Care Pathway reported that the documentation was more comprehensive than it was before the introduction of the pathway.

Shared knowledge

A perceived increase in knowledge sharing between staff was described in four UK qualitative studies. 106,107,123,132 Three of these papers107,123,132 explored multidisciplinary teamworking. A systematic review171 (also of multidisciplinary teamworking) supported this, concluding that the sharing of knowledge may improve team performance. A qualitative study from Canada279 concurred that staff perceived that there was greater exchange of knowledge (attributable to the use of ICPs).

Reconfigurations

The UK intervention studies describe several models of organisational reconfiguration. Examples of organisational change include the merging of services delivered in two hospitals;41 a merger between several partner organisations in primary care;45 a merger between an acute trust, a primary care trust (PCT) and social care services;43 integration between hospitals and the community;49 the formation of partnerships between general practices, which may also include nursing/care homes;46 new units or reconfigured wards;71 and services rearranged around health networks. 72 A non-comparator study that explored health-care reforms in Canada259 reported that there were few changes at a system level but that there were new forms of primary care organisations. This is echoed by a qualitative study from the USA281 that reported only small process changes following a reorganisation.

A qualitative case study from Sweden,263 which compared areas of successful integration with areas that were less successful, concluded that none of the successful cases had initiatives focused on management change; all had a focus on improving the quality of services.

Governance and accountability

There are few data directly related to the process of organisational change, beyond one UK intervention study,87 which described 66% of staff agreeing that two or more statutory organisations could share risks in an effective way, and that strong governance was important. The authors reported that the partnerships that appeared to be the most strengthened were those between health-care services, local authorities and voluntary organisations. The UK qualitative literature provides some insights into the organisational change process. One qualitative study103 described differences in the way that organisations described and perceived the process of joint commissioning. The authors reported that many sites that were investigated rejected the term ‘joint commissioning’, preferring terms such as ‘integrated commissioning’, and that staff struggled to describe how joint commissioning differed from joint working generally, and what it should achieve or what the outcomes should be. The process of putting formal structures in place could be seen as the outcome rather than service improvement.

Challenges regarding changing commissioning were reported in another study. 105 This exploration of Pioneer sites describes many as having an official or nominal lead organisation, with a balance that needed to be achieved between driving change and shared ownership. This study105 also reported that varied governance arrangements were operating, with many sites choosing to maintain existing arrangements, and a perception that new contracting models were needed. Another study141 described how new management structures were at an early stage and budgets were yet to be resolved in a reconfiguration of community services. Effective governance was described as a key aspect of integration, with a health and well-being board overseeing the process and a named person having responsibility for the allocation of a pooled budget. 139 In this study of an integrated care community programme,139 progress was limited in the area of financial and contracting mechanisms.

Further evidence regarding the challenges in changing organisational structures and processes is described by Holder et al. 151 This evaluation of the North West London Whole Systems Integrated Care programme describes initial difficulties in relation to data sharing and governance arrangements, payment systems and accountability arrangements, which had led to delayed implementation. It was highlighted that the programme of change was led by NHS commissioners, and the study authors151 reported that there could be personal agendas operating and confusion regarding alignment with wider plans.

Financial arrangements

One UK qualitative paper91 provides a detailed description of the management and financial arrangements in multiple Children’s Trust Pathfinders, including the formation of joint boards that are responsible for the co-ordination of services (29/31 trusts), a commissioning unit (one Pathfinder) or NHS and local authority managers occupying similar positions in a single organisation. Mechanisms for integrating accounting and budgeting varied between the different sites: via pooling budgets through legal contracts, making formal or informal agreements, or sharing information but keeping budgets separate. Another UK qualitative study139 described the pooling of budgets as a frequently highlighted area of tension. In contrast, another UK qualitative study116 reporting views of leaders of integrated care programmes described a perception of there being no debates over budgets. A further UK intervention study87 reported that there were challenges in moving any cost savings to and from budgets, with an inability to move money from health care to local authority budgets.

A systematic review165 of review-level evidence in relation to general health system financing, funding allocations, direct purchasing arrangements, organisation of service provision and health service integration concluded that most financial and organisational system-level reforms have inconclusive or negative effects on funding processes. A second systematic review182 of financial integration across health and social care highlighted the difficulties of achieving financial integration, with a failure to break down service boundaries or take overall control of service use from individual providers. The authors concluded that even when budgets are pooled widely, total resources remain limited. A study by Battersby et al. 195 evaluated a change to an outcomes-based funding system, and reported that 2 years was an insufficient amount of time in which to assess its effects. A further systematic review161 concluded that having a unified budget to cover joint working was important.

Focus of the initiative

Seven UK qualitative studies99,105,114,122,134,136,139 across a range of conditions and service contexts highlighted the influence of what the focus of the new model of care was (or was perceived to be), and in particular whether the driver was improved patient care, service reconfiguration or cost saving. Individual practitioner and organisation interests in partnership working may be perceived as being prioritised over service user needs, which reduces motivation for staff. 99 A focus of integration on delivering care more efficiently and effectively provides motivation and hope for the initiatives,105 and alignment of perspectives regarding the purpose of change is important in enabling change or the development of opposition from staff. 114 Initial enthusiasm for the vision of improving care and a willingness to be involved could change to antipathy and withdrawal of involvement if a focus on service efficiency and reorganisation or potential financial gains was perceived. 122 Tension could be created when managerial focus on financial savings contrasted with clinician focus on improved services. 136 One study139 reported the use of patient case studies to focus change on outcomes for patients, and to break down staff barriers.

Patient engagement in changes

The importance of patients being engaged in the development and implementation of new models of care was emphasised in seven UK qualitative studies,95,100,105,112,127,134,139 two systematic reviews185,199 and one international comparator study. 199 The need for engagement to be sustained over time from the development stage through to implementation has been highlighted,105,199 in addition to the potential for different community groups to have different preferences. 112 A systematic review157 concluded that the involvement of patients and families raises awareness of their priorities and including these priorities increases the perceived validity and sense of equity of changes. Another systematic review185 highlighted that patient/community engagement was a core element that is necessary for integrated governance across primary/secondary care. A qualitative study297 from the USA reported that patients were unsure about the meaning of the term ‘integration’.

Sharing of patient information

Three UK qualitative studies91,126,151 highlighted that concerns regarding the security and confidentiality of patient data could be a barrier to integrating services. This was a particular challenge in reconfigurations across multiple partners. 126 Ensuring that agreements on information sharing were put in place at an early stage was key to ensuring progress in integration. 151

Emotional response to change

Two studies88,196,197 described the need to pay attention to the emotional responses of staff when introducing change. The first of these (a UK qualitative study)88 described ‘painful feelings’ among staff in health and social care regarding reconfiguration of teams and roles, and the need to support them through periods of uncertainty. The second study (reported in two papers),196,197 from Canada, reported a lack of adjustment time for personnel as a particular barrier to an integrated care programme for older people.

Professional identity

A key area of uncertainty during change may be adjustment in professional identity. Seven UK qualitative studies89,90,94,113,128,145,146 with diverse settings and patient groups described professional identity as a key potential barrier to change. New models of care required the development of a shared group identity,89,90 and there could be concerns regarding a perceived erosion of roles and identities,94 which required trust and a sense of ownership to overcome. 113 There could also be a privileging of one professional discourse over another. 145

Professional role boundaries

A large number of studies referred to the challenges of existing professional role boundaries for new models of integrated care. Sixteen UK qualitative studies88,91,94,96–100,105,108,120,124,126,128,131,144 examining a variety of interventions, different patient groups and locations described how issues regarding professional roles had been encountered. The realignment of roles could be stressful and generate resistance, and needed time to be effective. 88,91 Adequate preparation and supervision were described as important,94,131 and some role changes had the potential to raise issues of accountability. 96 One study120 highlighted that resistance to change because of professional boundary changes was not always a barrier, as it could lead to greater professional engagement in the process. Another146 described initial wariness and friction subsiding over time.

A systematic review exploring teamworking193 confirmed that blurring and misunderstanding of professional roles can be common in integrated working, and are key barriers that could lead to professional conflict and intractable differences.

Stability of the workforce

Three UK qualitative studies94,118,126 described a lack of stability of the workforce as a potential barrier to integrated working. Vacancies, staff turnover and use of temporary staff could act as a barrier to continuity of care. 94 Macfarlane et al. 128 described a general climate of unwanted externally opposed change with frozen posts and job uncertainty, which led to challenges related to people accepting new roles and retention of staff. One international comparator study197 reported that resignations of key staff as a result of funding uncertainty was a key barrier to the implementation of a programme.

Employer

One qualitative study88 highlighted that staff in the same team having different employers could lead to staff having different employment practices and working to differing agendas and priorities. There could be a perceived dominance of health over social care, and hospital over community. 116

Power and hierarchies

Established hierarchies of power and status constituted a frequently reported barrier to the successful introduction of initiatives in the UK qualitative literature. 91,97,100,103,105,110,115,116,120,132,145,147,305 In particular, MDT meetings were often described as being dominated by medical professionals. 100,115 Individuals with less power were restricted in the changes that they were able to make. 110,120,145 New models of integrated working could, however, facilitate the removal of this barrier. An advantage of an ICP reported in one study97 was reportedly enabling midwives to deflect medical dominance and pressure. Integrated management was also reported as being helpful to ensure that all professions had equal respect and influence. 116 One systematic review193 supported the UK evidence, showing that the status of team members was an influencing factor in multidisciplinary working. A qualitative study from Italy295 described how a more balanced role of medicine was perceived to exist since integration reforms. The traditional distinction between management and professionalism had been eroded, with a new concept of professionalism that incorporated management. 295

History of co-operation

Five UK qualitative studies93,105,110,126,133 described history and working relationships between staff and different services as influencing the introduction of new models of integrated working. Previous and ongoing changes could prepare the ground for further change,93,105 and a history of collaboration was an enabler of integrated working. 110

Provision of training

The importance of providing adequate training for staff during the introduction of new models of working across different patient groups, types of intervention and contexts was frequently emphasised. Ten UK qualitative studies91,94,96,116,118,125,128,129,146,148 described training as being essential for changes in role or working practice. A systematic review162 reported that studies describing higher levels of integration tended to provide staff with ongoing support or training, rather than at periodic intervals. A second systematic review183 reported that education regarding why and how to use an integrated pathway increased confidence and openness among staff, which then improved communication and collaboration. A third systematic review185 described training as important to support new ways of working and in aligning cultures, and also as being required to address the need for rigorous evaluation and leadership.

Staff involvement and engagement

A lack of engagement in new models of care by staff was frequently reported in the UK qualitative literature. 100,101,111,113,114,118,120,122,126,127,139,142,147,151 Engagement among clinicians could be variable,100 with those outside hospitals reported as being less engaged. 100 GPs were reported to be particularly difficult to engage. 113,120,126,142,147

One UK intervention study77 compared outcomes following a diabetes mellitus integrated care initiative between fully engaged and less engaged general practices. In the more engaged practices, mean hospital admission rates reduced, with comparable changes not seen in the less engaged practices. International comparator studies196,197,205 have also noted difficulties in engaging GPs in interventions. One comparator study from Germany219 described a plan to introduce incentive payments for clinicians. In a UK qualitative study,134 it was reported that staff were concerned about the incentives offered to GPs to be involved in an integrated care pilot. A qualitative study from Spain282 highlighted that previous experiences are influential, finding that individuals who lacked experience of collaboration could have negative views of its usefulness.

Local leaders/champions

The identification of local leaders or champions to drive initiatives forward was described as an enabling factor in nine UK qualitative studies. 91,93,106,117,118,126,128,136,150 Leaders need to have enthusiasm;91 a strong commitment to implementing lasting change;126 an ability to take charge, and the time and skills to do so;106 oversee and motivate other staff;150 and both ‘hard’ and ‘soft’ skills in transformative change with ‘multiple’ and ‘flexible’ skills. The leader could be a senior member of staff with the support of a facilitator,118 or clinicians from primary or secondary care,136 and it was highlighted that additional training in change management for senior staff was beneficial. 128 The non-UK qualitative literature294 also echoed the importance of the management role in initiatives, with effective leadership described as ‘critical’.

One UK qualitative study117 described how champions appeared to be beneficial in the early stages of adoption of innovations. However, as projects were rolled out and new stakeholders became involved, there could be a perceived threat to the champions in terms of status and lack of exclusive rights to their work, which could create tensions between their supporters and other staff. The authors117 concluded that it should not be assumed that champions always have a positive effect; they may be useful at an early stage to lead the vision, but may then need to hand over the implementation to others. A second study128 also cautioned that although a clinical champion role could work well, it could also be problematic because of personal agendas and problems with back-filling time.

Effective leadership

Having strong and effective leadership was a commonly recurring theme in the UK qualitative literature across different settings, services and patient groups. Fourteen UK qualitative studies97,100,105,106,118,126–129,133,139,141,145,151 emphasised the importance of this factor in the successful implementation of new models of integrated care. Effective leadership included being an enthusiastic advocator but also challenging and debating with staff. 118 Leaders need to be clear about their own role, build informal networks and have buy-in from staff at a senior level and clear lines of accountability. 127,139 It has been reported that there should be an emphasis on leadership rather than management,129 which needs to be strong and fair. 133 The need for protected time for senior leadership to devote to specific projects was also emphasised. 139 A lack of leadership was reported to create tensions, questioning of legitimacy and other obstacles. 139 Changes in management staff adversely affected the management of change. 96

Five systematic reviews162,179,185,190,193 also emphasised the importance of leadership as an enabler of new models of care. One185 highlighted that managers need organisational support to develop the ability to make change happen. Another190 emphasised the need for leadership to have vision, and suggested that physicians should be integrated at all levels, including in leadership roles. A mixed-methods study from Australia269 reported that although survey data suggested a weak association between management and initiative implementation, qualitative interview data suggested that the commitment of proactive managers could influence the implementation of new initiatives. Non-UK qualitative studies288,289,294 also describe the critical importance of strong leadership.

Support for innovation

Three UK qualitative studies89,93,111 highlighted the beneficial staff impact of the promotion of initiatives (e.g. by winning awards or having good publicity)93 and developing a culture of development and implementation of innovative ideas. 89,111

Capacity and resources

The UK qualitative literature92,93,95,98,105,106,114,126–128,134,151 described how the level of funding available and the availability of resources/capacity within organisations could influence implementation and outcomes of varying types of new models of care. Limited funding was perceived to be a significant obstacle to new services98,114 and to achieving successful transition of care between services. 92 Severe financial restraints were reported by staff105 and successful outcomes were attributed to the availability of resources by staff participants in one study of an integrated care and support pioneer initiative. 106

Information technology and logistics

Frustration with information technology (IT) systems was frequently reported in the UK qualitative literature. 72,94,95,100,102,105,107,110,118,124,126,134,146,149 A lack of compatibility between systems used in different services, such as health and social care or general practices and hospitals, was a key barrier to integrated working. 94,95,102 In one study,100 over half of the clinicians (56%) reported that they were dissatisfied with IT systems.

Three systematic reviews179,185,190 echoed the importance of adequate IT systems. One185 described integrated IT as essential infrastructure that enables providers to focus and manage risk, provides accurate and detailed information to inform clinical decision-making, supports change management and allows tracking of high-risk patients and outcomes. A second190 concluded that other processes within integrated models of care are only possible with computerised information systems that are accessible in any location. International qualitative studies294,299 echoed the need for better IT to support co-operation.

Clarity of vision

In Focus of the initiative, we reported that the perceived target outcomes for new models of care (patient care vs. cost saving) could influence the views of staff regarding new initiatives. UK qualitative studies72,100,103 highlighted that the clear communication of aims and outcomes was required to ensure that staff fully understood the vision. Dickinson and Glasby103 described a lack of understanding regarding what joint working meant. Another study122 outlined a perception of ‘being dragged along’ by clinicians, and poor communication about the aims leading to a lack of involvement and ownership. This link between communication of the benefits of change and staff involvement was also made by another study. 126 Another study105 described the challenge of splitting the vision of integrated care into practical activities. The authors reported that staff working in some areas understood the vision less well than others, and that making changes that were meaningful to patients could be difficult. The authors also noted an absence of articulation of the implications of the vision for changing power relationships.

One systematic review185 supports the need for clarity of vision in service integration. Nicholson et al. 185 referred to a shared vision of optimal health care as getting stakeholders ‘to the table’. They suggested that having joint board members with directors on each other’s boards facilitates shared vision, trust and collaboration. A qualitative study from the USA288 similarly concluded that there is a need for a clear mission and vision, and a culture that reflects the values and norms of the organisation.

Organisational culture

Eleven UK qualitative papers91,93,99,110,111,126,127,133,134,139,151 referred to the influence of differing organisational cultures in acting as a potential barrier to implementation and outcomes of new models of integrated care in varying service settings. Barnett et al. 93 described organisational culture as ‘a critical factor in change’, in particular openness and having a fit between organisational values, beliefs and the innovation. Greenhalgh et al. 110 similarly described the compatibility of values across organisations as a key factor in enabling change. A 2-year follow-up of this study by the authors111 found evidence of stakeholders’ conflicting interests and values, which had led to lack of sustainability of the service transformation. Another study127 reported how a particular challenge to change was staff in different organisations working within different care models.

Another study133 exploring the merger of two NHS trusts described how, although there were differences (one was more entrepreneurial, team oriented and less hierarchical), there were key similarities in culture, such as the emphasis on staff development, working in partnership and overcoming adversity. The authors reported that there could be challenges in preserving desirable cultures in shifting landscapes, but that best practice and goodwill were important. 133 Other studies described the importance of establishing a culture of collaborative working, which required sharing of clear objectives and processes134 and the need for the alignment of perspectives to enable change,114 together with the need to lever existing partnerships and the requirement for perseverance. 139 Huby et al. 120 outline that good cross-boundary relationships and access to sources of power and political leverage are important for organisational change. In this study of reconfigurations, including new roles for nurses and GPs, a facilitator was the aligning of initiatives with existing organisational priorities. 120 In contrast, another study151 describes confusion regarding alignment with wider plans, with organisational fragmentation also slowing down implementation.

A qualitative study from Sweden263 described four organisational culture facilitators: (1) that development is perceived as desirable and appropriate, (2) that there is backing from management, (3) that there is trust between participating organisations and (4) that there is acceptance from the body of physicians in each organisation. A non-comparator intervention study from Australia248 described culture change as being key in any initiative. A non-comparator study from the USA253 found that motivational readiness and organisational climate for change were associated with higher odds of co-ordination. Kathol et al. 283 also described the need for a culture shift.

Systematic reviews echo the role of organisational culture in new models of care. Suter et al. 190 described the need for an organisational culture to match the vision of integrated care. This was echoed by Maslin-Prothero and Bennion,181 who reported the necessity for the development of a shared culture, including the purpose and vision of change. Boult et al. 158 highlight that ‘real world’ adoption of new models involves having to fit in with existing organisational cultures. A review of working practice between care home staff and health-care staff162 reported that different working cultures of staff acted as a barrier. A review of teamworking171 concluded that culture affected change. Organisational culture was also reported to influence the introduction of ICPs. 179

External threats and policy context

The UK qualitative literature has widely reported the influence of policies and decision-making outside the control of the organisations involved in new models of integrated care. For example, changes in the political landscape, such as devolution in Scotland, could change policies and priorities. 95 Commissioning mechanisms, severe financial restraint, national reorganisation (such as NHS foundation trusts and staff regrading) and conflicting government policies also had an impact on the delivery of initiatives. 105,114,122,128,133,151 One study114 described how the choice of model that was introduced was dictated by the funding stream. A systematic review of UK literature on integrated health and social care179 echoed the influence of national policy on the introduction of new models of care. A qualitative study from the USA265 similarly emphasised the influence of the external environment and need for policies to facilitate change.

Audit and evaluation

Three systematic reviews157,183,193 highlighted the importance of feedback loops in the introduction and sustainability of new models of integrated care. A review of ICPs183 concluded that positive feedback from a national audit on service improvements helped sustainability, and that audits that identified areas of need could facilitate change. A review of integrated governance systems185 emphasised the need for mechanisms of reporting and auditing to be in place. This was echoed in a review of the literature on health integration, which reported that performance monitoring systems should be in place to measure outcomes at different levels. A qualitative study from the UK101 described how having outcome targets in place for only one area of care (emergency services) meant that there was a lack of incentive for other services (such as community) to improve. Greenhalgh et al. 111 describe how the meaning and significance of measures of effectiveness were contested. Roberts et al. 139 highlight the importance of taking time out to monitor and reflect on progress. This was seen as important in maintaining focus and momentum. A qualitative study from Ireland280 highlighted that audit was an important tool in the implementation of new models of care.

Budgetary and financial factors

Roberts et al. 139 found slow progress in developing financial and contractual mechanisms at the different sites they studied. They reported that it was common for the sites to report significant challenges regarding this aspect of new models of care. Some cited a lack of capacity to lead the development of this area, and others found finance and contractual arrangements to be complex and challenging activities. Bespoke external expert help was helpful in enabling progress. Conversations regarding finance were reported to be extremely challenging, particularly when considering moving money from acute to community settings. Many left development of this aspect to later in the process.

Two systematic reviews185,190 provide further data regarding financial challenges. Suter et al. 190 described different service funding as a major barrier to integration, with financing mechanisms requiring pooling of resources and a population focus. Nicholson et al. 185 described the need for new financing and a change from an organisation focus to care for a whole population. A qualitative study from the Netherlands290 highlighted the need to develop incentives for financial collaboration. A US qualitative study294 reported the need for standard contracts and outcome measures to encourage integration and commissioning between providers, with a requirement to demonstrate improved patient experience and clinical outcomes in regulation of integrated care initiatives. The authors emphasised that the pricing strategy needs to incentivise integration, and bundled payments and local tariffs for particular conditions and pathways should be developed. 294

Non-UK intervention and qualitative studies283,284 discussed how payment systems could potentially act as barriers to, or enablers of, new models of integrated care. A comparator study from Germany219 described a plan to introduce incentive payments for clinicians. A non-comparator study from the USA252 recommended that GPs should be rewarded for interventions that avoid potentially preventable admissions/contacts. The authors highlighted that current payment schemes focus on physician effort, rather than patient characteristics or outcomes. A qualitative study from the USA265 described misalignment of reimbursement systems as a barrier to patient-centred medical homes. Another qualitative study from the USA291 reported that barriers to becoming part of an Accountable Care Organization were predominantly finance related, with legal and regulatory barriers also present.

A qualitative study examining various initiatives in Australia to improve co-ordination267 noted differences regarding the role of payment systems in different services. In one case study, financial incentives were not perceived as important for involving GPs, who were reported as being most interested in involvement for the benefit of patients. In the second case study, it was perceived that use of an electronic record had been hampered by a lack of incentive payment for doctors. In the third case study, it was reported that a lack of incentive payments made it difficult to attract GPs initially, but efficiency gains overcame this.

A qualitative study from the Netherlands296 reported that fee-for-service systems lead to inappropriate incentives for non-co-operation, an inability to influence the service, a service that is not patient-centred and inflexibility. The authors described the view among participants that integrated financing was incompatible with the existing financing system, was problematic for patients with comorbidity and lacked evidence. They also noted that integrated financing is challenging when there is a diverse patient population, varying care, differing requirements at points in the pathway of care and a lack of clarity regarding who has principal responsibility for care.

Governance

The need for strong accountability mechanisms and risk sharing across organisations was described by a UK qualitative study. 100 Governance was reported to be an issue in further UK qualitative studies,105,126,151 with effective governance being a key enabler of integration. 139 One study105 reported that governance arrangements varied in different Pioneer sites, with many choosing to maintain existing arrangements. A UK intervention study66 described having clear clinical governance as a key aspect of a new MDT initiative. Strong governance with clearly understood lines of responsibility and accountability was reported as important in another UK intervention study. 87

Three systematic reviews181,185,190 describe the importance of clear governance arrangements in new models of working. One185 recommended that goals and strategies are jointly agreed, with formal agreements put in place to manage deliverables, risk and process. Another190 concluded that a flatter and more responsive organisational structure facilitates integration, with a need for effect mechanisms of accountability and decision-making. A qualitative study from the USA274 described that having common objectives, contractual relationships, representation on the board of executives and having a large market share facilitated collaboration.

Clinician contacts

Two UK intervention studies,44,87 five non-UK comparator studies158,206,217,218,306 and one non-UK non-comparator study249 evaluated the number of clinician contacts as an outcome measure. A UK cluster RCT44 found that clinician contacts in the first 12 hours after admission increased following the introduction of an ICP for children with breathing difficulties in an ED (p = 0.0004). In contrast, a community intervention for older adults87 was reported to have reduced the number of physiotherapy or occupational therapy appointments and led to a 25% reduction in the number of visits to a practice nurse (p = 0.05).

Non-UK studies report a range of findings regarding the effect on clinician contacts, with three RCTs or cluster RCTs (two from the USA158,206 and one from Finland217) reporting a reduction in clinician visits (p = 0.03 to p < 0.001; 1% reduction) following the formation of new teams or standardisation of practice interventions for older adults. In contrast, one Canadian non-RCT study218 found an increase in health professional, voluntary and home help visits in frail older adults following a case management care plan and standardised assessment intervention. A Swedish RCT306 similarly reported higher resource use in terms of nurse visits among adults with heart failure following a team and key contact intervention, although the text and the data provided seem contradictory.

A non-comparator study from the USA249 evaluated an intervention comprising shared electronic records and a messaging system for patients and staff. The study highlights the need to consider system-wide effects resulting from new models of care. Total numbers of patient in-person contacts across all professions reduced by 26% over a 3-year period following introduction (p < 0.001) in both primary care (25% reduction) and specialist services (21% reduction). Numbers of ED attendances and urgent care visits increased, (p < 0.001) but not by as great an extent as the reduction in patient in-person contacts. However, the number of telephone contacts by patients also increased, resulting in an overall increase of 8% in patient contacts (p < 0.001).

General practitioner appointments

Three UK studies58,63,87 reported the effect of new models of integrated care on GP appointments (rather than all contacts or non-medical clinician contacts). Two studies63,87 reported a reduction, and one58 found no significant difference. Julian et al. 58 evaluated the effect of a GP-led ICP for women with menstrual problems via a small-scale non-RCT study, and found that there was no significant difference in the number of GP appointments (although it had a positive effect as there was a reduction in the number of hospital appointments).

The first58 of two studies reporting a reduction in the number of GP appointments had a comparator group design, and evaluated the placing of a social worker in a general practice, together with joint assessments and care packages for the care of older and at-risk patients. The number of GP consultations were reported to have reduced by 3% when comparing the project year to the previous year. A before-and-after evaluation of the Partnership for Older People Projects (POPPs)87 reported a 10% reduction in the number of GP appointments (p = 0.009), a 28% reduction in the number of telephone calls to GPs (p = 0.014) and a 25% reduction in the number of visits to a practice nurse (p = 0.05) for the case co-ordination type of interventions that had been put in place in some sites.

Length of stay

A large number of UK and international studies evaluated the impact of new models of care on length of stay in hospital. Of the UK studies, five34,44,61,66,85 reported that length of stay was not significantly different, one69 reported that it increased and 1434,36,40,46,49,54,59,63,65,68,71,72,81,87 reported that it had reduced. One additional study40 found contrasting results, with a reduced length of stay for elective surgery and emergency medical patients (p < 0.001), but longer stays for surgical, medical and paediatric emergency cases.

One study reporting no significant difference66 evaluated the introduction of a children’s community outreach team with an integrated pathway and extended hours. The other three studies evaluated ICPs for children with breathing difficulties,44 gynaecological patients61 and patients with diabetic ketoacidosis. 85 One study69 reported an increase in length of stay, but it has very limited reporting and is of very weak quality.

Of the 14 UK intervention studies reporting a reduction in length of stay, four36,54,63,81 are of a higher-quality controlled before-and-after design; three54,63,81 of these were carried out in populations of older adults in community settings. The first of these81 evaluated eight POPPs with older adults in the community (generic workers, integrated teams and out-of-hours services). It reported that the integrated health and social care team project reduced the number of bed-days following emergency admissions (no detailed data). This was in contrast to a signposting service that appeared to increase the number of bed-days. The second study54 evaluated a case manager role in general practice and found a shorter length of stay in intervention practices than in the control (adjusted relative risk 0.896, relative risk reduction 10.41%, 95% credible limit 0.820 to 0.979; p = 0.015). The third study81 evaluated having a social worker located in a general practice, with joint assessment and packages of care in conjunction with a district nurse. The study found that there was a significant difference in length of stay between patients from the project practice and patients from other practices (p = 0.02; mean 31% reduction). The fourth UK comparator study36 was a cost-effectiveness analysis of a hospital at home service, which found that the mean length of stay in the intervention group was 3.3 days [standard deviation (SD) 3.9 days], compared with 10.4 days (SD 7.7 days) prior to the introduction of the initiative.

The 10 other studies34,40,46,49,59,65,68,71,72,87 also reporting a reduced length of stay were uncontrolled before-and-after studies. They were carried out in a range of populations and settings encompassing general hospital patients,59 children in hospital,68 older adults in hospital and community settings,49,65,87 older adults and/or people with diabetes mellitus,72 hospital mental health services,71 urgent care,40 patients at the end of life,34 and adult patients with complex needs. 46

The intervention elements associated with a reduced length of stay also varied, including MDTs and care managers,46 whole service redesign with multiple elements,40,49,71,72 ICPs59,65,68 and multiple types of intervention, including case management and discharge planning. 87 Two studies59,65 highlighted a difference in the effectiveness of ICPs for different groups of patients. The first of these59 found a reduction in length of stay associated with only about half of the conditions in a hospital where ICPs had been introduced, and the second65 concluded that pathways in a hospital setting were more effective for the older and male patients among those undergoing hip or knee arthroplasty.

Ten systematic reviews158,168,169,171,172,175,177,180,182,191 included evaluation of length of stay outcomes. Six158,168,171,172,180,182 concluded that new models of integrated care led to reduced length of stay, and four169,175,177,191 concluded that findings were inconclusive or that there was no clear impact. The reviews reporting uncertain effects examined studies of older adults in services crossing both acute and community settings,177,191 hospital services for people with brain injury175 and hospital mental health services. 169 The interventions comprised case management and the formation of MDTs. All four studies169,175,177,191 are from Australia and meet just over half of the quality criteria for systematic reviews. The systematic reviews, which concluded that length of stay was reduced with new models of integrated care, included a range of settings and patient groups, including older adults,168,171,172 those with any condition182 and the chronically ill. 180 They included case management interventions in hospital or community settings,168 MDTs in hospitals and community settings,171 financial integration182 and any multidisciplinary intervention. 180 This group of systematic reviews includes one168 that met all but two quality criteria.

The international comparator studies echo the diversity with regard to length-of-stay outcomes, with two studies, the first in older adults in a community setting198 and the second with a range of patients in a hospital medicosurgery unit,216 suggesting a significant reduction of 19.2% in the number of bed-days (p = 0.001) and length of stay (p = 0.31), respectively. In contrast, two studies (the first in an emergency unit for older adults,221 the second in a psychiatric hospital239) reported no significant difference in length of stay or the number of bed-days in hospital. The first of these evaluations221 found that both the intervention arm and the control arm had a reduced hospital length of stay, and although the intervention group reduced to a greater extent, this difference was not significant. One of the four studies in this group239 used randomised allocation to a study arm. The findings are from studies carried out in Australia, the USA and Switzerland. The two studies finding a significant reduction in length of stay198,216 evaluated co-ordinated assessment and management and case management interventions. The two studies finding no significant difference221,239 evaluated a multielement co-ordinated case management intervention and a MDT intervention.

Three international non-comparator studies243,261,307 report divergent findings regarding effect, with two243,307 indicating reduced length of stay and one261 indicating increased length of stay in a nursing home. This study261 is the only one that used length of stay in a nursing home rather than hospital length of stay as a measure.

Unscheduled admissions

Most evidence regarding the effect of new models of integrated care on unscheduled and emergency admissions comes from the UK intervention studies. Four UK studies,46,48,78,80 including two with a comparator design,46,78 indicated no effect on the number of unscheduled admissions. 46,48,78,80 Four studies reported in five documents43,45,74,81,82 indicated a significantly increased number of admissions. Three of these74,81,82 were of a higher-quality comparator design. The Steventon et al. 81 study found that only the intermediate care intervention seemed to be associated with an increase; the health and social care team projects appeared to reduce the number of emergency admissions for a particular high-risk subgroup. Stokes et al. 82 reported that the increase was only 0.01 admissions per patient per month, an effect size of 0.03, which is of limited clinical significance.

Intervention studies from the UK have reported a reduction in the number of unscheduled admissions of 22%,37 a 16.3% decrease in the number of emergency medical admissions and a 3.9% decrease in the number of emergency surgical admissions,40 17 admissions for 13 patients,47 a 2% reduction for multiple admitters,54,66 a reduction from 935 to 840 admissions in a year for COPD patients,72 and a significant effect size following a service reconfiguration compared with other hospitals (p = 0.0002). 39 None of these studies was of comparator design.

The patient populations and locations for interventions in this group of studies were diverse, including children in primary care,66 older adults,47 people with COPD,72 urgent mental health care in a hospital,78 and mixed patients in acute and community settings. 74 There was no clear pattern to the effect of different intervention elements. Case manager, MDT and staff relocation interventions were in the no-effect group; MDT and large-scale service reorganisation interventions were present in the group indicating an increase; and MDT, ICPs and role change interventions were in the group of studies reporting reduced admissions.

Findings from UK qualitative studies were also equivocal, with one reporting no perceived difference100 and one reporting a perceived reduction in length of stay. 106

Three systematic reviews168,177,191 together provide unclear evidence, with one suggesting no difference,168 one suggesting that the effect was unclear177 and another suggesting a reduction. 191 All three reviews examined interventions for older adults and encompassed case management in hospital and the community,168 and community integrated care initiatives generally. 177,191 The single international comparator study236 reported that the risk of having at least one unplanned hospital admission was lower in the intervention group after adjustment (odds ratio 0.39). The single international non-comparative study243 outlined that unplanned charges (arising from unscheduled hospital admissions or ED interventions) decreased by 51% at 12 months and a cumulative 64% by 24 months.

Non-emergency and scheduled admissions

Eight UK studies reported in nine papers,39,41,45,46,63,71,74,77^. 81 including two studies reported in three papers45,63,74 with comparator designs, indicated a reduction in admissions. Reductions were of 1% in the number of inpatient admissions in gynaecology;41 4% in the number of hospital admissions for a range of conditions following service reconfiguration (most of the reduction was reportedly in the case management interventions);45,74 15% in the number of admissions for older patients comparing general practices (no reduction pre–post intervention) following relocation of a social worker to a general practice;63 31% in the number of admissions for a subgroup of patients from early-engaging practices in a reconfiguration intervention;83 there was a decrease in standard admissions ratio;39 a decrease of the number of inpatients from 91% to 35%;71 a reduction from 19% more admissions than comparable areas to 8.7% more than comparable areas following a multielement initiative;77 and a reduction in the number of excess bed-days from 520 per 1000 patients in 2007/8 to 232 in 2008/9. 46

Two evaluations of ICPs44,80 reported no significant effect on admission rates (1) for children with asthma (comparator design)44 or (2) for admissions for fractures/falls in the elderly. 80 One of these studies80 also found indication of an increase in the number of admissions for patients with diabetes mellitus (p = 0.014) compared with before the intervention.

Another study43 reported a slight increase in the number of admissions to a community hospital following a hospital merger. Stokes et al. 82 similarly reported a very small increase (of 0.01 admissions per patient per month, effect size 0.03) following a MDT intervention. Another study of older adults81 also reported a slight increase of 0.64 additional admissions over a 12-month period compared with a control group for a MDT intervention with a support worker and an out-of-hours service. A study by Wilberforce et al. ,86 also in older adults, found that more-successful (higher-integration) teams had higher admission rates, and Clarkson et al. ,42 another study in older adults, found that integrating services could increase rates of care home admissions (hazard ratio 0.73; p = 0.12).

Systematic reviews have reported reductions in numbers of admissions of 24–31% among studies of ICPs,173 reductions in numbers of nursing home admissions in four of five studies and in hospital admissions in two of three studies of case management,177 reduced hospital admission in 10 of 17 studies of multidisciplinary interventions,180 and reduced admission in studies of case conferences. 191 In contrast, one review of patient-centred medical homes170 concluded that there was no significant effect on admissions.

The evidence from the international literature is mixed, with six comparator studies201,206,210,211,228,236 evaluating a range of intervention types indicating no significant effect on admissions (all but one201 of these studies was in older adults). Nine studies195,199,200,205,207,219,221,222,306 reported a reduction in the number of admissions for patients with a range of conditions; patients included older adults, general groups and those with chronic conditions. One of these studies195 found that, although admissions rates in one intervention area reduced, in another area they increased. Three international non-comparator studies echo the mixed picture, with two indicating reductions in numbers of admissions243,251 and one indicating no significant effect. 261

Readmission

Studies report conflicting findings regarding the effect on readmission rates, with three UK studies reported in four papers36,38,39,54 (including one study with a comparator group54) indicating no significant effect on (1) older at-risk patients,54 (2) patients originally admitted as emergencies38,39 or (3) patients with COPD following a hospital at home intervention, service reconfiguration or a case management intervention. 36 Two studies indicated an increase in readmissions,69,82 although one of these69 was of very poor quality. The other82 found only a slight increase in 30-day readmission (effect size 0.03). One UK study of a community outreach team for acutely ill children66 reported a reduction in the number of readmissions of 17.3%.

International studies carried out in populations of older patients, those with chronic conditions and those in acute care support the lack of evidence of a reduction in numbers of readmissions, with all five international comparator studies (reported in eight papers201–203,209,210,216,218,221) suggesting no statistically significant effect. In contrast, two international reviews171,180 suggested reduced numbers of readmissions, although a third182 concluded that there was conflicting evidence.

Accident and emergency department attendance

We identified three non-comparator design UK studies66,74,87 that reported accident and emergency (A&E) (ED) attendance outcomes. Two66,87 suggested a reduction of 5% and 69% for children and older adults, respectively, following varied interventions. The other74 reported no change (p = 0.04). One review170 concurred that interventions may lead to a reduction for patients with co-occurring conditions. A meta-analysis was conducted in this study, and found that patient-centred medical homes reduced the number of emergency visits [risk ratio 0.81, 95% confidence interval (CI) 0.67 to 0.98]. A second review197 found that although A&E attendance fell by 10%, this was not statistically significant.

Five of the international comparator studies also reported that A&E attendance reduced following interventions by 20.8%,199 by 57%,200 by 12.2 per 1000 patients annually,207 from 705 to 396 attendances (p = 0.03)210 and from 20.7% to 3% of patients visiting annually (p = 0.02). 237 The interventions included multiple elements of integration, an Accountable Care Organization model and a home-based team (they were delivered to older adult patients in all but two studies199,200). One non-comparator international study244 supported a reduction in A&E attendance in a general population.

Five studies reported in seven papers194,201–203,217,228,239 evaluated MDTs or integrated home care and found no significant effect on A&E attendance. Only one of the studies239 was not carried out in older adult patient groups, but in patients with a psychiatric diagnosis instead. Two studies reporting case management interventions in older adults found that the number of visits to an ED increased (odds ratio 1.28; p = 0.02;211 46% in the control group vs. 32% in the intervention group; p < 0.001218). The studies report opposing trends, with the first study211 finding an increase in the second year but not the first, and the second study218 finding a greater increase in the first year of the intervention.

Outpatient appointments

The UK evidence reports a reduction in outpatient appointments in all four studies outlined in five papers45,58,67,74,81 of 20–40% (one74 at the individual patient level only and not when analysed by general practice). The initiatives evaluated included a variety of interventions for a range of conditions. In contrast, one international comparator study of an ICP for older adults found no reduction in outpatient clinic use. This finding was supported by a non-comparator study of all-inclusive care for older adults,261 which similarly found no significant effect.

Prescribing rates

Few studies examined the effect of new models of integrated care on rates of prescribing. Among UK studies, no difference in prescribing rates is reported in one evaluation of an ICP,62 whereas a reduction is reported in another. 85 A systematic review158 found that pharmacy involvement in MDTs led to ‘improved prescribing’. ‘Improved prescribing’ was also reported in an international non-comparator study. 245

Perceived quality

The outcome of ‘quality of care’ presented challenges in analysis, as a result of the varied ways in which the construct was defined and considered. Four UK intervention studies45,64,73,148 reported on staff perceptions of quality of care following service redesign, case management or integrated pathway interventions in hospital or primary care settings for older adults;45 general caseloads;73,148 or patients with clostridium difficile infection. 64 All four studies reported that staff-perceived quality of care had improved. One of these64 reported that 91% of staff perceived improved patient care and treatment following the introduction of a pathway. In other studies, it was found that 54% of staff perceived that patient care had improved,45 ‘more than half’ of staff reported improved services148 and 91% of staff reported that improved relationships would have or had improved patient care. 73

Four systematic reviews163,165,182,186 and four international comparator studies reported in seven papers201–203,213,215,220,235 reported the effect on quality of care in terms of staff or patient perceptions. A review of reviews of organisational and/or financial interventions165 concluded that there was some evidence of benefit to quality of patient care from interventions including MDTs, case management and provision of primary care in or alongside EDs. The authors highlighted the wide range of measures that were used as indicators of ‘quality of care’, and commented that some studies did not specify what aspect of quality was measured. 165 The outcomes considered included, most commonly, patient satisfaction, but also aspects that we defined as process improvements (such as continuity or compliance with guidelines), together with clinical outcomes. Another review163 concluded that there was moderate evidence that comprehensive care approaches could improve patient perceptions of improved care quality. Peikes et al. 186 reported that physicians perceived that quality of care had improved following a care co-ordination and lifestyle programme. Mason et al. 182 examined the effect of financial integration on quality of care. Studies in this review typically assessed perceived quality using surveys of staff, patients or carers (including aspects that we have identified separately, such as satisfaction and access to care), reporting mixed findings.

One international comparator study reported in three papers201–203 supported the finding that quality of care was perceived by patients (older with chronic conditions) to have improved. One study evaluating MDTs and shared training in Norway226 provided contrary findings, reporting no difference in staff-perceived service quality.

Quality standards

Three studies evaluated quality not in terms of perceived quality of care but in terms of whether or not recommended procedures (such as vaccination, screening, prescribing or testing) had been carried out. One215 found a significant improvement in procedures from before to after the collaborative team intervention (p = 0.0013), although the authors concluded that the intervention was not cost-effective. Another study235 evaluated three general practices that were Accountable Care Organizations. The authors reported percentage point differences between the three practices and comparison groups of 0.7, 2.6 and 4.7, respectively, and that ‘the three practices were superior to their comparison group peers on all care quality measures’ but they did not provide analysis of statistical significance. They noted that the before–after differences for the practices were not statistically significant, and that in fact the performance of one practice had declined over the time period. Fagan et al. 213 evaluated the effect of pay-for-performance incentives for care co-ordination. They concluded that there was either no significant difference between the intervention and comparison groups in rates of screening/testing for diabetes mellitus, or that the improvement was significantly lower for the intervention sites.

Patient satisfaction

A large body of literature reports a positive effect on patient satisfaction following the introduction of integrated care initiatives. Nine UK studies35,37,43,46,51,58,66,75,83 across a range of conditions and services (including diabetes mellitus, complex needs, children’s services, complex and vulnerable patients in primary care, and hospital gynaecology) and following a variety of interventions (including whole-service redesign, MDTs and ICPs) reported increased levels of patient satisfaction. Only two studies74,78 highlighted less positive outcomes. The first of these74 noted that following a case management intervention, older patients reported that they were less able to see the GP or nurse that they preferred. A second study evaluating the siting of a psychiatric nurse in an A&E department78 found no difference in patient satisfaction levels.

Four UK qualitative studies121,130,138,146 support the finding that interventions including MDTs and service redesign result in increases in patient satisfaction. One qualitative study,108 however, reports more mixed views of a mental health integrated partnership, and another119 describes how satisfaction may have been higher with health services but not with social care services.

The systematic reviews support an increase in patient satisfaction, with 118,88,160,163,164,170,177,180,188,189,192 concluding that the evidence suggests a positive effect. Four international comparator studies197,214,228,237 similarly report increased satisfaction among older, acute and paediatric patient populations following service integration, case management and patient-centred medical home interventions. One of these studies197 highlighted that although patients reported greater satisfaction, this was not shared by carers. This was echoed by another study of carers,223 which found no impact on satisfaction with care and support services. Another study209 reports a lack of significant difference following home-based integrated care (p = 0.31) and a final study of an integrated care protocol to prevent overuse of acute services212 reports that patients’ satisfaction was at the same level as prior to the intervention.

Meeting patient preferences

Three UK studies of MDTs or ICPs in end-of-life/frail older patients34,53,144 highlighted that the interventions were able to increase the fulfilment of patient wishes to die outside a hospital setting. The percentage of deaths in a hospital reduced from 15% to 8% in one study,34 the percentage of home deaths increased from 19% to 23% in another53 and the percentage of home deaths was 55% in the third study,144 compared with a national average of 33% (however, the requested hospice placement had only been achieved for 34% of patients in this study).

Time spent in accident and emergency departments

One UK study44 reported no difference in time spent in A&E for children with asthma following the introduction of an ICP compared with standard care. In contrast, an Australian comparator study221 found that older adults spent around 45 minutes less in the ED following a new co-ordinated service, although this is of borderline significance (p = 0.0575; a reduction from 496.3 minutes to 435.7 minutes).

Numbers of incidents and complaints

One UK non-comparator study71 reported that the frequency of aggressive incidents among inpatients with mental health and behaviour problems was reduced following the introduction of an integrated assessment and treatment service. Another UK before-and-after study39 reported that numbers of formal complaints, incident reports and patients leaving before treatment declined following service redesign and the establishment of an integrated emergency assessment unit.

Length of waiting time

One UK study55 found that following the introduction of an ICP in urology, the time between referral and contact and the time from referral to treatment, both reduced. ICPs were also found to reduce waiting times during the management of inpatients with diabetes mellitus, including reducing time before giving intravenous infusions. 85 The reconfiguration and integration of gynaecological services was found to reduce waiting times for scans (p = 0.0054) and getting blood test results, although not for blood testing. 41 Improvements in waiting times for assessment by a social worker were reported in a comparator study evaluating the relocation of a social worker to a general practice. 63 Waiting time for assessment/diagnosis of autism and related disorders in children was reduced from 2 years to 5 or 6 months following the introduction of a multiagency team with a filtering panel and joint assessment. 75

Five qualitative studies104,119,131,146,147 supported the finding of a reduction in waiting time. There were perceived reductions in waiting time for allocation to a social worker, admission to a ward, waiting for assessment and waiting for services generally.

In contrast to these positive reports, one systematic review172 found that the length of wait for surgery for older patients with hip fracture was reduced in only one of three studies analysed. A review of UK literature that was carried out in 2010181 evaluated speed of delivery following the introduction of integrated health and social care teams. The 18 included papers were reviews, qualitative studies, surveys and one non-RCT. The authors concluded that the speed of response may be quicker, and services may be more responsive, as a result of introducing MDTs. 181 However, the data to support this conclusion are limited, and the quality of the review was assessed as weak (meeting only one of the quality criteria).

One international comparator study196 reported a reduction in waiting times for nursing home placements among older adults (p < 0.05), and another203 reported no effect on waiting times for appointments among patients with chronic conditions. The two non-comparator studies258,308 are similarly divergent, with one indicating a reduction in waiting time for cancer testing258 and another suggesting an increase in referral delay. 308

Identification of unmet needs

There was no UK evidence from intervention studies in relation to the outcome of the identification of unmet needs. One UK qualitative study106 reported perceptions of increased unmet needs being identified and more referrals being made in care homes, and another142 reported the potential for increased referrals being caused by unmet needs being identified. One US systematic review of studies evaluating the Program of Research to Integrate Services for the Maintenance of Autonomy for older adults188 found that two of the 45 studies reported decreased prevalence of unmet needs in the community (p < 0.001). A comparator study,214 also from the USA, evaluating a medical home/case co-ordination intervention for children with special needs reported that GPs perceived that the programme had identified unmet child and family needs.

Access to services

Five included (non-comparator) UK intervention studies49,55,73,86,87 reported that access to services in the community and/or specialists/intermediate care had increased. These studies evaluated MDTs, general service redesign or integration of hospital and community services. One UK qualitative study135 reported perceptions among staff that colocation of health and social care services for older people had enabled access. Two systematic reviews154,182 also reported that access to services had improved. The first of these154 reported increased access to clinical assessments and interventions following the introduction of an ICP for stroke patients, and the second182 concluded that there is largely positive evidence that interventions that integrate financial or resource flows across both health and social care can improve access to care. A systematic review of UK health and social care integrated interventions161 described the key to successful integration as being the ability of patients to access care at home.

Two international comparator studies reported in four papers201–204 supported the finding that MDT and multicomponent home care unit interventions improve access to services (for patients with chronic conditions and severe heart failure). Two further comparator studies95,204 describe the use of domiciliary services increasing as a result of increased access,195 and the ease of access to staff as being an important enabler of care delivery at home. 204 A non-comparator study evaluating a wide-scale integration model245 similarly reports improved access to services for patients. Another comparator study evaluating a patient-centred medical home in the USA257 also reported that patient access to care had improved.

Staff working experience

None of our included UK studies reported staff working experience outcomes. Two systematic reviews (of patient-centred medical homes and MDT interventions)170,181 concluded that integrated models can improve staff working experience. The first170 reported that working experience improved to a small to moderate extent, and the second181 reported that initiatives increased staff job satisfaction. However, an international comparator study from Norway226 found no effect on staff exhaustion or work engagement. A positive effect on job satisfaction was indicated by a qualitative study from Spain. 282