Evaluation of the Impact of a Coordinated Intervention Model for Complex Older Patients with Dementia

Background

Fragmented care within the healthcare system and hospital care focused on single diseases do not adequately respond to the needs of older patients with multimorbidity (two or more chronic diseases) and/or frailty [1]. This situation results in growing demand for healthcare resources and increased consumption of drugs, which have a significant impact on both costs and patient quality of life [1, 2, 3, 4]. Faced with this situation, healthcare systems must evolve by developing models that provide personalized, integrated care by multidisciplinary teams, co-ordinated across different care settings and are cost-effective. Person centred care planning should be a key element of the care planning process as it allows patients to actively participate in shared decision making (SDM) with the members of the care team in developing an individualised care plan (ICP) that considers its values and its main care goal (prolonging survival, maintaining function or prioritising symptom control) [5, 6, 7]. The paradigm shift towards a model in which care is planned based on the patient’s needs, their values and preferences can help to decrease the fragmentation of care, reduce unnecessary use of hospital and emergency services and improve the patient’s experience [3, 8].

In order to support care management for this segment of the population, Catalonia’s Department of Health implemented the Chronic Care Programme. Two patient profiles were identified: i) patients with chronic conditions and complex care needs (CCP), based on the presence of a combination of clinical complexity, and/or complex social factors, and/or complex health system factors [9] and; ii) patients with advanced chronic diseases and limited life expectancy (ACP). The NECPAL CCOMS-ICO© version 4.0 tool is used to assist in early identification of these patients [10, 11]. To respond to the needs of CCP and ACP patients, a four-stage person centred care model has been developed, consisting of: a) identification of the patient’s profile (mild frailty, moderate frailty or CCP, and advanced frailty or ACP); b) a situational diagnosis [12], which requires a multidimensional assessment and assessment of needs; c) putting together an ICP within a SDM framework involving the practitioners and the patient [13], and; d) the sharing of information between the practitioners and care settings involved in the patient’s care [9]. In the Catalan healthcare system, primary care (PC) teams record patients identified as needing CCP or having ACP and their ICPs in the shared care record (SCR), which allows this information to be available in all care settings.

Persons with dementia have more comorbidities and approximately twice as many hospital stays per year compared to older adults without dementia, resulting in significantly higher healthcare costs. In Catalonia, dementia is associated with a 44.1% increase in total healthcare costs, with an attributable additional expenditure of €1,311.7 per person per year, mainly related to hospitalisations, ED use and IC [14]. These hospitalisations are associated with poorer health outcomes such as cognitive and functional decline, delirium, falls, institutionalisation, and mortality [15, 16, 17, 18]. Beyond clinical and economic outcomes, it is important to emphasise the negative experiences of people with dementia and, in particular, their carers in the ED, emphasising communication difficulties, stress associated with the healthcare environment and the perception of potentially avoidable visits and readmissions [19]. It is therefore clear that since patients with dementia are usually older, with chronic conditions and complex care needs, require an integrated and person centred care model [20].

The objectives of this study were to: i) assess the implementation of a coordinated care model between intermediate care (IC) and PC, as determined by an increase in record-keeping by the PC teams in the SCR, the identification of patients with dementia and complex and advanced chronic conditions, and the development of ICPs for patients with dementia and complex or advanced chronic conditions, and; ii) assess the effectiveness of the model in relation to use of healthcare resources.

Method

Results

Of the 160 patients included in this study, 56.3% (n = 90) were women and the mean age was 84.9 ± 7.0 years. Overall, the patients included had moderate dependency for daily living activities, with a mean Barthel Activities of Daily Living Index score of 59.3 ± 25.8, and moderate frailty, with a mean VIG-Frail score of 0.42 ± 0.1. Patients with severe frailty accounted for 30% (n = 48) of the sample. The demographic, clinical, functional and cognitive characteristics at baseline, according to provenance and reason for admission, are described in Table 1.

On discharge, 58.8% (n = 94) were identified as having CCP, and 41.3% (n = 66) as having ACD-LLE. In 63.1% (n = 101) of cases, the goal established by consensus with patients/families was maintaining function, prioritizing symptom control in 35.6% (n = 57), and prolonging survival in 1.3% (n = 2). On discharge, the mean BI score was 47.4 ± 25.0, and the mean VIG-Frail score was 0.49 ± 0.10. Twenty six per cent (n = 40) of patients were discharged to a new care home setting.

At 3 months of discharge, 14.4% (n = 23/160) of the patients had died (Figure 1).

Figure 1

Of the 137 patients who were still alive, 77.4% (n = 106/137) were identified as having CCP or ACP in the SCR. Of these, 46.0% (n = 63/137) had already been identified pre-hospitalisation, representing a 31.4% increase in recorded identification in their medical record (p < 0.001). When analysed by the strategy applied, we found that with the first strategy, of the 56 patients still alive at 3 months, 71.4% (n = 40/56) had been identified, of whom 46.4% (n = 26/56) were already identified pre-hospitalisation, representing a 25.0% increase (p < 0.001). With the second strategy, of the 81 patients who were still alive, 81.5% (n = 66/81) were identified, and of these, 45.7% (n = 37/81) had been identified previously, representing an increase of 35.8% (p < 0.001).

At 3 months of discharge, of patients who were alive and identified (n = 106) 41.5% (n = 44/106) had an ICP recorded in the SCR. Of these, 17.9% (n = 19/106) already had a ICP pre-hospitalisation, which means there was a 23.6% increase (p < 0.001). When analysed according to the strategy applied, we found that with the first strategy 35% (n = 14/40) of patients who were identified as having CCP or ACP had an ICP in the record at 3 months, 22.5% (n = 9/40) of whom already had one pre-hospitalisation, representing an increase of 12.5% (p = 0.126). With the second strategy, there was an ICP recorded for 45.5% (n = 30/66), of whom 15.2% (n = 10/66) had a previous one, representing an increase of 30.3% (p < 0.001).

With respect to the use of healthcare resources, we found that in the 6 months pre-hospitalisation, 15% (n = 24) of patients had required admission to an IC unit, 17.5% (n = 28) to an AC hospital, and 48.8% (n = 78) made at least one visit to ED. The mean number of visits to PC services was 10.38 (SD 9.4), with a median of 8 (Q1:3-Q3:14). No statistically significant differences were observed in the use of resources in the 6 months pre-hospitalisation between patients who had died by 6 months and those who survived.

At 6 months of discharge, 25.6% (n = 41) of patients had died. Within 6 months of discharge, 15.1% (n = 18) of patients required admission to an IC unit, 15.1% (n = 18) to an AC hospital, and 34.5% (n = 41) made at least one visit to the ED. The mean number of visits to PC services was 11.03 (SD 10.4), with a median of 8 (Q1:5-Q3:15).

Overall, at 6 months post-discharge, we found a 13.4% decrease in ED use. Of the 119 patients still alive at 6 months, we compared resource use between the 6 months pre-hospitalisation and 6 months after discharge. We found a statistically significant decrease in the use of ED, 54.4% (n = 31) of the patients who had visited the ED in the 6 months prior to hospitalisation did not visit ED in the 6 months (p = 0.026) post-discharge, 24.2% (n = 15) of the patients who had not visited ED in the 6 months pre-hospitalisation, visited it in the 6 months post-discharge, with no statistically significant differences observed in the use of other resources (Table 2 and Figure 2).

Figure 2

Discussion

A model for care coordination between IC and PC, geared towards patients with dementia and complex or advanced chronic conditions and based on the proactive identification of these patients, the development of an ICP through in-depth interviews within a SDM process together with the patient/family, and the exchange of structured information between different levels of care through the discharge summary and computer alerts, has supported the increase in record-keeping in the SCR, in the identification of patients with complex or advanced chronic conditions and the drafting of their respective ICPs by the PC services in our area. In addition, the increase in the availability of information in the SCR may have contributed to greater continuity between care levels and, hence, to a decrease in the use of ED by these patients, which suggests a possible impact in terms of effectiveness.

To meet the needs of patients with CCP and ACP, patient-centred coordinated care models are needed between care levels [28]. In this study, we assess a model of care coordination between IC and PC for patients with dementia, multimorbidity and frailty who are identified as having CCP or ACP during their admission to an IC hospital, in regard to implementation of the model and its effectiveness. This study contributes to the existing literature by providing real-word clinical practice evidence of integrated care principles for people with frailty, positioning IC as a key coordination hub with PC and focusing on a particularly vulnerable population, namely people with dementia [29].

The first practical consequence of the study, which is related to implementation of the care model, is seen by the increase in the recording of patients identified as having a CCP or ACP and their ICPs by PC in the 3 months post-discharge. Although a direct causal relationship cannot be established, this improvement coincides temporally with the incorporation of two key elements in the coordination process: the newly structured discharge summary and the sending of a computer alert to the PC teams. Both elements, the structuring of clinical information and the technological support necessary for its effective exchange between levels of care, have been identified in the literature as fundamental components of integrated care [30]. Other facilitators may also have contributed, such as the existence of an already established relationship of trust and respect between the PC and IC teams [31, 32]. Nonetheless, possible hurdles that may have influenced the absence of validation or recording of information by some PC teams should also be considered. Among these, the high care burden that makes it difficult to find the time needed, high turnover of healthcare practitioners that implies less experience or knowledge in the fields of complex and/or advanced chronic conditions, and the fact that integrated work has sometimes been perceived to be a source of additional effort rather than as a more effective way of working [8, 32, 33].

The second consequence, related to the effectiveness of the implementation of the care cooperation model in the area, was seen in the use of health resources. Our study showed a lower use of ED in the six months post-discharge, with no statistically significant differences observed in the use of other healthcare resources. In this context, the absence of changes in the number of primary care visits is interpreted as appropriate and planned follow-up of patients, consistent with the objectives of the model to improve continuity of care and reduce the use of unplanned resources.

There are a wide variety of integrated care strategies with the aim of reducing the use of healthcare resources, but findings are still insufficient to draw any conclusions regarding significant benefits in the use of hospital resources [34, 35]. Still, some coordinated care models that focus on patients with complex profiles have shown significant reductions in unplanned care and hospitalisations. The Gundersen Health [36], which focuses on high-risk patients, achieved a 64% cumulative reduction in expenditures arising from ED visits and unscheduled hospitalisations over 24 months, as well as a decrease of more than 50% in the number of hospitalisations. These results reinforce the idea that well-structured interventions can improve care efficiency, especially when they are maintained over the medium term. Indeed, the authors underline that the most notable benefits of the program became apparent after at least two years of implementation, suggesting that the duration of the intervention is a determining factor in the impact it has. Much of the evaluation literature suggests that integrated care is unlikely to deliver changes to activities and costs in the short term. Hence, integrated care should be setting broader objectives about population health and patient and staff experience [37]. We need new metrics and a profound transformation to allows us to move from a system that values volume of activity and processes to an integrated care model focused on overall outcomes and on what has value for the patient [38, 39]. To achieve this objective requires not only structural reorganization, but also the development of collective capacities that favour continuous learning, interdisciplinary cooperation and effective implementation of knowledge in clinical practice. In this sense, communities of clinical practice appear as very useful complementary instruments. Their ability to facilitate the translation of evidence into concrete actions, as well as to promote continuous improvement through peer-to-peer interaction, makes them a strategic component towards achieving value-based integrated care. Integrating these approaches can not only improve clinical outcomes and the patient’s experience, but can also contribute to the sustainability of the health system in the medium to long term [40, 41].

We also want to highlight the key role that IC plays in the healthcare system. IC is an essential component of integrated care, providing interventions aimed at preventing in-hospital complications, optimising functional recovery processes and ensuring continuity of care, especially between different levels of care and care settings [35]. In our study, 46% of the patients were admitted from ED/Observation or their place of residence (STEP-UP model), which, in line with other published reports, reinforces the role of IC in the prevention of avoidable hospitalisations and in the improvement in ED performance [42]. IC provides a multidisciplinary care model that is highly specialized in geriatrics and palliative care, geared towards the needs of patients with complex or advanced chronic conditions. Its importance lies not only in the professional experience and the integrated approach of the team, but also in the duration of hospitalisation, which allows carrying out a comprehensive geriatric assessment and to plan post-discharge care.

Our coordination model is highly transferable, and its key components are potentially applicable to other settings. Its implementation would require the availability of IC or equivalent services, functional coordination mechanisms based on the structured transfer of clinical information aimed at continuity of care, as well as shared information systems, with the necessary organisational adaptations according to the local context.

Limitations, Strengths and Future Work

This study has several limitations that should be considered when interpreting the results. In the first place, the pre-post quasi-experimental study design with no control group limits the possibility of establishing causal relationships. The changes seen during the implementation of the care model and in the use of healthcare resources may have been influenced by external factors not directly related to the intervention, such as the natural clinical course. However, it should be noted that the study was conducted as part of routine clinical practice at the hospital, which justifies the choice of design. The inclusion of a control group would have meant depriving some patients of a coordinated care model that is part of standard care practice. Secondly, the results of the implementation of the care model (at 3 months) and those assessing its effectiveness (at 6 months) were analysed only for patients still alive at the time of the analyses. This may have led to a survivorship bias, as it excludes the deceased patients, who may have had different patterns of use of resources or response to the intervention. However, we analysed the prior use of resources by both the patients who were deceased at 6 months and the survivors and observed no significant differences.

It should also be considered that we have not assessed whether the absence of validation or recording of information by some in our study is due to discrepancies between IC and PC concerning the identification of CCP or ACP and the proposed ICP.

This study has several strengths. First, it presents a practical, real-life implementation of a conceptual model for an especially vulnerable population: patients with dementia and highly complex conditions. This makes the intervention an example of translational research, as it contributes to narrowing the gap between healthcare theory and practice. Secondly, it institutes an innovative strategy using a coordinated care model that combines technological and clinical factors: structuring information in the discharge summary with specific ICP proposals, sending computer alerts to the PC teams and more active use of the SCR, not only as an information repository, but as a dynamic coordination tool. The study also addresses significant gaps in knowledge around coordination between levels of care, providing preliminary results on effectiveness and presenting an organizational model with potential for scalability to other healthcare system settings. Finally, it integrates key aspects for current healthcare system transformation processes, such as digitalisation, efficient use of clinical information, and the promotion of person centred care.

In future studies, essentially for cases that are more complex, we plan to incorporate more structured methodologies, similar to case conferences, that would allow the healthcare practitioners involved in the care of the patient to agree on a single individualised plan to be shared between them, the person, and their care setting [43]. In addition, the use of qualitative approaches, such as interviews or surveys with professionals, could help explore barriers, facilitators, and potential divergences in the identification and validation of ICP across care levels. A possible avenue of future research can be to assess whether the care received in cases of decompensation is in accordance with the values, preferences and main care goal agreed upon with the patients, considering that it is the priority outcome and therefore of utmost importance in the ICPs [44]. In patients with MACA, future studies should also evaluate end-of-life care quality and family or caregiver experience as key outcomes.

Conclusion

Patients with the most complex health profiles are major consumers of healthcare resources, yet they often receive fragmented and suboptimal care, in part because of a lack of communication and coordination among healthcare professionals [36]. Attention to the needs of persons and their social circle should be one of the objectives of the healthcare response by care teams and organizations. Persons with complex or advanced chronic conditions require comprehensive, integrated responses that are appropriate according to their stage of life. The implementation of a coordinated care model between IC and PC in our area has helped improve the identification of people with complex or advanced chronic conditions, the recording of their ICPs, and reduce their use of ED. This model has relevant practical implications and may be scalable to other regions with shared information systems and intermediate care services, subject to appropriate organisational adaptations.

Patient and Public Involvement

Patients or people with lived experience were not involved in the design, conduct, reporting or dissemination of this research.

Reviewers

Jordi Amblàs-Novellas, Director of Integrated Care & Chronicity, Departament de Salut de Catalunya, Professor, Universitat de Vic – Universitat Central de Catalunya (UVic-UCC), Spain.

Genevieve Maiden, Allied Health and Integrated Care Manager, Uniting War Memorial Hospital, South Eastern Sydney Local Health District, Australia.