The results of this study show that the four phases of the Development Model of Integrated Care are confirmed empirically in integrated care practice by the participating integrated care coordinators. To underpin this conclusion we did multiple analyses in a four step approach.
Firstly, respondents confirmed the presented phases and all four phases were chosen by integrated care coordinators for the three patient groups. No phases or important phase characteristics were missed. Secondly, almost all the respondents stated that they had been through the previous phase, illustrating a certain change in development over time. A third result which underpins our conclusion is that elements that were related to earlier phases of development were also implemented earlier in time for stroke and AMI practices. This absence in the case of dementia services could be explained by the fact that they are substantially more recent and started only in 2007, with greater external pressure and time urgency.
Also, we found a relationship between the numbers of implemented and planned elements and the phase of development. In earlier development phases integrated care services had more plans for the future and this number of plans decreases over time. Corresponding with that, services that are in further development phases (in either way of assessment) do have more implemented elements of integrated care. These findings in integrated care services support the empirical validation of the DMIC which was based on the literature and experts in the field of integrated care. Although the phases of the DMIC were confirmed, the developmental process of integrated care services seems not to be linear and predictable. Some respondents mentioned that they were ‘in between’ phases, recognised aspects of two (following) phases or mentioned a fall-back. Phases can overlap or run into each other or there can be a relapse to earlier phases. There are no obvious or strict boundaries between phases. This makes it clear that the phases need to be seen as conceptual presentations, but can be helpful for evaluating and guiding integrated care development. These findings correspond with the related literature about organisational development and life cycles. Multiple authors describe life cycles with four to six phases , which was also the result in our study. The recent literature about networks also stresses the non-linear aspects of development, the complexity of collaboration and possibility of obstacles and fall backs .
The general applicability
Although the characteristics of the three groups of integrated care services differ on multiple aspects, the development phases appeal to all of them. This raises the question about the general applicability of the development phases and the question if they are useful for multiple or all kind of integrated care services. In our study, the stroke services can be seen as the ‘oldest’ of the three groups and are also the most developed in terms of number of implemented elements. About two thirds of them are in the third or fourth phase of development. The dementia services’ development is comparable with the AMI services, although the latter have existed for longer. It is remarkable that the dementia services have already experienced such a fast development and implemented such a large number of elements. The recent attention to dementia at client, professional and policy level in the Netherlands, initiatives like the National Dementia Improvement Programme and the development of a method for purchasing integrated dementia care, may have contributed to this. Financial preconditions like integrated budgets are not available for stroke and AMI services. The analyses of phase transitions show that next to CEO and higher management commitment, this condition is seen as the most important factor for proceeding to the next phase. The availability of a coordinator, a multidisciplinary care pathway, case management and clear agreements about roles, tasks, goals and ambitions are, regardless of setting, crucial elements that can speed up or hinder development. The recognition of the phases in all these different services, points out the question if the development phases would also fit in an international context. However we have not studied this, first steps to use the model in a Canadian context are now being undertaken and seem positive. Over the past decade the integration of care has gained increasing attention from managers, health care workers, policy makers and researchers in a large number of countries. There is a worldwide interest to better understand integration, implementing integrated care and stimulating development, regardless of systems or legislation. Whereas the Development Model of Integrated care is (also) based on the international literature and does not focus on a specific (patient) group, it is an interesting suggestion to further research the applicability in other countries.
Assessment of development phases
Although representatives of stroke, AMI and dementia services felt able to position their practice in one of the phases, the comparisons with the calculated phases based on the model are interesting. The self-assessed scores overlap for about one third with the calculated phases, which are based on the present elements as indicated by the coordinators themselves. The 7-out-of ten rule seems to fit the best with the self-assessed scores (highest kappa, significant p-value). When the calculations methods are more inflexible (eight out of ten or higher), the number of services that seem to overestimate their development rises, indicating that these rules may be too strict. Reasons for the low scores between self-assessment and calculated scores could be that the self-assessment scores are merely based on the integrated care coordinator, whose ability to assess therefore is an important factor. Coordinators may vary in their ability to assess, have different roles or involvement and their judgement is possibly influenced by multiple factors. Multiple studies from the fields of psychology and auditing show that people’s judgement about current situations are influenced by earlier experiences, perceptions about the history and the future, recent failures or successes and their situation compared to others [24, 25]. It is possible that these factors also play a role in this study. Our analyses show that increasing the available time per week for coordinators has a positive effect on the overlap between the coordinators’ and model’s phase assessment, which may be a manifestation of a more complete role. Recommended important next steps are therefore involving more key persons per integrated care service. When doing so, consensus among partners about present or future elements can be analysed and also the presence or absence of consensus about the assessed development phase.
Our study has some limitations. Although the response rates were high, the number of participants per patient group differed. AMI services were only represented by nine out of the twelve, but this is because the number of hospitals with interventional capacities and therefore the number of services is limited. For stroke and dementia, diagnoses and treatment can be initiated in almost every hospital. Further, the knowledge of the integrated care coordinator representing the integrated care service was important for the quality of the data. To optimise this, a number of respondents also consulted their partners in the care services before completing the questionnaire. To ensure that the right respondents took part, we explained the criteria for participation in personal contact with the respondents or even visited them. Nevertheless, it would be interesting to invite more respondents from each integrated care service to add additional perspectives and calculate consensus scores. In current studies in diabetes care, palliative care and non-congenital brain damage care we are now applying this approach.
Suggestions for further research and practical implications
We have three suggestions for further research. Firstly, expanding this research to other countries with other (policy) contexts is to be encouraged. We think this is interesting because reducing fragmentation in care and improving integrated care is a major issue in many countries. Winning the international Karolinska /EHMA Research award 2012 encourages us to do so. In further research we encourage inviting more respondents per integrated care service. Further analyses on the difference between self assessed phases and the phases as calculated by the model is suggested. Secondly, we suggest further research on the process of integrated care development. Our study gives insight into the phases of development that can be present in practice. It is interesting to monitor and follow the development in each phase. Possible research topics include the implementation strategies taken and which partners or other circumstances are involved at what time.
Thirdly, we suggest further research into the relationship between the development phases and the delivered results in integrated care. It would be interesting to see if integrated care services in further phases of development do have better outcomes on processes, patient satisfaction, quality of life or disease-specific indicators and costs.
Our study also has a number of practical implications. During the survey study, the respondents pointed out that filling in the questionnaire was experienced as a self-evaluation exercise which gave suggestions for the further improvement of their integrated care. When sending in their data, it was notable that they asked for benchmark results. For integrated care practitioners, coordinators and managers the DMIC with its development phases could be used as a quality management tool for multiple patient groups. In quality management the use of self-assessment models is used for reflecting on current practices, for guiding improvement and for improving performance. An example of these models is the European Foundation for Quality Management model . This frequently and internationally used model also describes (groups of) elements that are important for the effective organisation of care. Empirical research shows a clear relationship between the implementation of these elements (‘enablers’) and performance, both in industry [27, 28] and in health care [29, 30]. The EFQM model is also used as evaluation and improvement tool. The DMIC could also serve as an assessment and evaluation tool to reflect on integrated care practice and may initiate discussions on how to improve and progress to further phases. The model can provide support for steering on quality and with guiding policy and improvement plans. Hence, the DMIC could be regarded as a quality model for implementation of integrated care. The National Stroke Service Network in the Netherlands has adopted the DMIC as such.
In the Netherlands the DMIC is already being used for evaluative purposes by multiple practices in dementia, stroke, youth, palliative, diabetes, non-congenital brain damage and vulnerable elderly care. To simplify filling in the questionnaire, we made a webbased tool based on the model. Other suggestions for practice are to further develop the model into an audit tool and to facilitate benchmarking for learning from comparable others as already practiced in stroke care. The National Stroke Service network has adopted the DMIC as a basis for an audit- and improvement tool for all her members. Health care insurers are also interested and are currently exploring the opportunities to use the model when purchasing integrated care. This year, in cooperation with an insurance company, the self-evaluation of 38 integrated diabetes practices is planned.