Executive summary
Background: Across the developed world, we are witnessing an increasing emphasis on the need for more closely coordinated forms of health and social care provision. Integrated care pathways (ICPs) have emerged as a response to this aspiration and are believed by many to address the factors which contribute to service integration. ICPs map out a patient's journey, providing coordination of services for users. They aim to have: 'the right people, doing the right things, in the right order, at the right time, in the right place, with the right outcome'. The value for ICPs in supporting the delivery of care across organisational boundaries, providing greater consistency in practice, improving service continuity and increasing collaboration has been advocated by many. However, there is little evidence to support their use, and the need for systematic evaluations in order to measure their effectiveness has been widely identified. A recent Cochrane review assessed the effects of ICPs on functional outcome, process of care, quality of life and hospitalisation costs of inpatients with acute stroke, but did not specifically focus on service integration or its derivatives. To the best of our knowledge, no such systematic review of the literature exists.
Objectives:
* To systematically review all high-quality studies which have evaluated the impact of care pathway technologies on 'service integration' and its derivatives in stroke care
* To examine how elements of service integration are defined in such studies
* To examine the type of evidence utilised to measure service integration
* To analyse the weight of evidence used to support claims about the effectiveness of ICPs on improving service integration
* To produce recommendations for ICP developers, users and evaluators.
Inclusion criteria: Types of participants
The review focused on the care of adult patients who had suffered a stroke. It included the full spectrum of services - acute care, rehabilitation and long-term support - in hospital and community settings.
Types of intervention(s)/phenomena of interest
Integrated care pathways were the intervention of interest, defined for the purpose of this review as 'a multidisciplinary tool to improve the quality and efficiency of evidence based care and is used as a communication tool between professionals to manage and standardise the outcome orientated care'
Here 'multidisciplinary' is taken to refer to the involvement of two or more disciplines.
Types of outcomes
'Service integration' was the outcome of interest however, this was defined and measured in the selected studies.
Types of studies
This review was concerned with how 'service integration' was defined in evaluations of ICPs; the type of evidence utilised in measuring the impact of the intervention and the weight of evidence to support the effectiveness of care pathway technologies on 'service integration'. Studies that made an explicit link between ICPs and service integration were included in the review. Evidence generated from randomised controlled trials, quasi-experimental, qualitative and health economics research was sought. The search was limited to publications after 1980, coinciding with the emergence of ICPs in the healthcare context. Assessment for inclusion of foreign papers was based on the English-language abstract, where available. These were included only if an English translation was available.
Exclusion criteria:
* This review excluded studies that:
* focused only on a single aspect of stroke care (e.g. dysphasia)
* evaluated ICPs as part of a wider program of service development
* did not make an explicit link between ICPs and service integration
* did not meet the definition of ICP utilised for the purposes of the review
* focused exclusively on the outcomes of variance analysis
Search strategy: In order to avoid replication, the Joanna Briggs Institute for Evidence Based Nursing and Midwifery Database and the Cochrane Library were searched to establish that no systematic reviews existed and none were in progress. A three-stage search strategy was then used to identify both published and unpublished studies (see Appendix III).
Data collection: Our search strategy located 2123 papers, of which 39 were retrieved for further evaluation. We critically appraised seven papers, representing five studies. These were all evaluation studies and, as is typical in this field, comprised a range of study designs and data collection methods. Owing to the diversity of the study types included in the review, we developed a single-appraisal checklist and data-extraction tool which could be applied to all research designs.32 The tool drew on the Joanna Briggs Institute (JBI) appraisal checklists for experimental studies and interpretive and critical research, and also incorporated specific information and issues which were relevant for our purposes (see Appendix VI). This extends the thinking outlined in Lyne et al.31 in which, drawing on Campbell and Stanley's classic paper, the case is made for developing an appraisal tool which is applicable to all types of evaluation, irrespective of study design.
In assessing the quality of the papers, we were sympathetic to the methodological challenges of evaluating complex interventions such as ICPs. We were also cognisant of the very real constraints in which service evaluations are frequently undertaken in healthcare contexts. In accordance with the aims of this particular review, we have included studies, which are methodologically weaker than is typical of many systematic reviews because, in our view, in the absence of stronger evidence, they yield useful information.
Data synthesis: Given the heterogeneity of the included studies, meta-analysis and/or qualitative synthesis was not possible. A narrative summary of the study findings is presented.
Results:
1. ICPs can be effective in ensuring that patients receive relevant clinical interventions and/or assessments in a timely manner, although these improvements may reflect better documentation rather than actual changes in practice.
2. ICPs can be effective in improving the documentation of rehabilitation goals, documentation of communication with patients, carers (diagnosis, prognosis and follow-up arrangements) and documentation of notification of primary care physicians of discharge. However, this can create additional burdens of work for staff.
3. Early studies of ICP-managed care in the acute stroke context have demonstrated reduced length of stay without any associated adverse effects on discharge destination, morbidity or mortality. These effects do not reach statistical significance, however, and may reflect wider changes in service provision and a general trend towards reduced length of hospital stay. While later studies in the acute and rehabilitation contexts do not reveal any significant reduction in length of stay, they do report greater documented use of certain clinical interventions and assessments, suggesting that ICPs can be effective in mobilising hospital resources around the patient.
4. ICPs implemented in the context of acute stroke care can be effective in reducing the occurrence of urinary tract infections, although we do not know whether this can be attributed to improved service integration.
5. ICP management in stroke rehabilitation may not be flexible enough to meet diverse patient needs and can result in insufficient attention to higher-level functioning and carer needs influencing perceptions of quality of life.
6. ICP management may assist in clarifying role boundaries and a shared understanding of the work, but this can result in some members of the disciplinary team perceiving that their contribution is not appropriately reflected in the documentation.
7. There is some evidence that ICPs may be effective in changing professional behaviours in the desired direction where there is scope for improvement, but in situations in which multidisciplinary working is effective, their positive effects may be limited. Furthermore, it is far from clear what the active ingredients of ICPs actually are. Kwan et al. suggest that it was the process of ICP development that had most impact on behaviours rather than the use of the artefact per se.20
8. None of the studies assessed the balance of costs and benefits of ICP use. Therefore, we do not know whether the costs of ICP development and implementation are justified by any of the reported benefits.
Conclusions: Implications for practice
There is some evidence that ICPs may support certain elements of service integration in the context of stroke care. This seems to be as a result of their ability to support the timely implementation of clinical interventions and the mobilisation of resources around the patient without incurring additional increases in length of stay. ICPs appear to be most successful in improving service coordination in the acute stroke context where patient care trajectories are predictable. Their value in the context of rehabilitation settings in which recovery pathways are more variable is less clear. There is some evidence that ICPs may be effective in bringing about behavioural changes in contexts where deficiencies in service provision have been identified. Their value in contexts where inter-professional working is well established is less clear. While earlier before and after studies show a reduction in length of stay in ICP-managed care, this may reflect wider healthcare trends, and the failure of later studies to demonstrate further reductions suggests that there may be limits as to how far this can continue to be reduced. There is some evidence to suggest that ICPs bring about improvements in documentation, but we do not know how far documented practice reflects actual practice. It is unclear how ICPs have their effects and the relative importance of the process of development and the artefact in use. As none of the studies reviewed included an economic evaluation, moreover, it remains unclear whether the benefits of ICPs justify the costs of their implementation.
Recommendation 1: In the absence of stronger evidence and given the costs of their development, service providers in the context of stroke provision are advised to restrict ICP use to those areas of the care pathway which are predictable and can be standardised, and/or to areas of service provision where there are clearly identified deficiencies in existing care provision.
Implications for research
Recommendation 1: Primary research is necessary in order to identify the active ingredients of ICPs, their interrelationships and the theories which underpin them.
Recommendation 2: Evaluations of ICPs need to be underpinned by clarity as to the purposes of the intervention.
Recommendation 3: Evaluations of ICPs must include theoretically informed strategic choice of outcome measures which takes into account the perspective of all relevant stakeholders.
Recommendation 4: Evaluations of ICPs should include theoretically informed process outcomes in order that the reasons for behavioural change or its absence are understood.
Recommendation 5: Evaluations of ICPs should provide adequate information on the 'control'.
Recommendation 6: Evaluations of ICPs should provide adequate information on the local context, taking care to identify critical success factors.
Recommendation 7: It is unlikely that ICPs will work for all purposes and in all contexts. Researchers should aim to produce realistic evaluations of ICPs which seek to develop an explanation (and therefore a theory) about how the intervention in question works in particular situations/contexts, by exploring the relationship between context, mechanism and outcome.