Joining the dots: Day to day challenges for practitioners in delivering integrated dementia care

Despite the increasing policy focus on integrated dementia care in the UK, little is known about the opportunities and challenges encountered by practitioners charged with im-plementing these policies on the ground. We undertook an extensive, mixed-methods analysis of how a contemporary multidisciplinary dementia pathway in the UK was experienced and negotiated by service providers. Our pragmatic mixed methods design incorporated three types of research interaction with practitioners: (a) Semi-structured interviews ( n = 31) and focus group discussions ( n = 4), (b) Practitioner ‘shadowing’ observations ( n = 19), and (c) Service attendance and performance metrics reviews ( n = 8). Through an abductive analysis of practitioner narratives and practice observations, we evidenced how inter-practitioner prejudices, restrictive and competitive commissioning frameworks, barriers to effective data sharing and other resource constraints, all challenged integrative dementia care and led to unintended consequences such as practice overlap and failure to identify and respond to people's needs. In order to more successfully realise integrated dementia pathways, we propose innovative commissioning frameworks which purposefully seek to diffuse power imbalances, encourage inter-pro-vider respect and understanding, and determine clear lines of responsibility.

Integrated Care Systems, NHS England have set ambitious demands and targets in their efforts to implement sustainable, intersectoral partnerships and systems which aim to transform healthcare practice and place England at the forefront of the Western world in integrated provision (Forder et al., 2018;Ham, 2018;NHS England, 2019).
While current integration strategies are centrally guided, frameworks for delivery are non-prescriptive and encourage geographical regions to draw on the needs of local people, communities, as well as pre-existing care provision and commissioning structures (e.g., NHS England, 2019). As a result, there is no prototype of local service provision, rather an eclectic mix of different forms of integration happening in different ways, over varied time phases, and to mixed effectiveness (Forder et al., 2018;Charles et al., 2018;National Audit Office (henceforth, NAO), 2017). Moreover, since integrated services and systems are still in their infancy, to date there have been limited efforts to evaluate and evidence the effectiveness and parity of these different models and, where such evaluations have occurred, only marginal improvements in some piloted areas have been identified (Charles et al., 2018;Forder et al., 2018;NAO, 2017).
A further challenge in our understanding of contemporary integrated care partnerships and systems is that our endeavours to evaluate effectiveness are focused on implementation experiences at the leadership, strategic and structural levels (such as, senior NHS, local government & third sector management; e.g., Alzheimer's Society, 2018a, 2018b; Charles et al., 2018;Ham, 2018). As such, we lack nuanced understanding of the extent and complexity of barriers practitioners face at the ground level in their efforts to successfully work and communicate in integrated ways. We propose that to fully understand such challenges, it is essential we broaden our evaluative lens to encompass perspectives and experiences at the inter-practitioner level and not just at the inter-agency and leadership level.

| Integrated pathways in dementia care
An area where important integrated care provision remains underwhelming in terms of development and evaluation is in the field of dementia. The historical and forecasted escalation in dementia support and treatment costs in the UK and worldwide (Knapp et al., 2014;Luengo-Fernandez, Leal, & Gray, 2010;Wittenberg, Hu, Barraza-Araiza, & Rehill, 2019) has prioritised the need for evidence-based, multi-disciplinary and equitable dementia care pathways (Bragato & Jacobs, 2003;NHS England, 2017;Samsi & Manthorpe, 2014). While not anchored in a singular definition (Samsi & Manthorpe, 2014), dementia care pathways aim to uphold best integrated health and social care practice principles and are thought to be pivotal in directing and coordinating appropriate intervention and rehabilitative services for people, families and carers affected by dementia (Banerjee, 2010;Brooker, 2006;Department of Health, 2016;NHS England, 2017). In reality, contempo-

| The current study
While significant investment into health and social care provision is planned (Department of Health & Department for Communities and Local Government, 2017; Department of Health, 2016), we remain unclear about the efficacy of current community-based services, and the types and breadth of challenges we are faced within providing or accessing integrated disciplinary models of care. Dementia care practice is one particular example of integrated care provision which requires closer scrutiny if it is to fulfil its potential to fully support people with dementia and their families in the community, and to help people remain outside of acute services (Banerjee & Owen, 2009;Forder et al., 2018;NAO, 2017;Torjesen, 2020;Wittenberg et al., 2019). In the absence of good evaluative data, integrated care is at risk of being premised upon idealised principles (and rhetoric), which in turn can detrimentally impact on health and well-being outcomes and treatment costs for people affected by dementia (Banerjee, 2010;Banerjee & Owen, 2009;NAO, 2017). This paper offers insights from a commissioned independent review of a dementia care pathway in the South East of England. The review aimed to understand and evidence the finer workings of the dementia provision with a view to help future local service configurations. To do so, the review adopted a pragmatic mixed-methods approach (Morgan, 2007) to explore how the current multidisciplinary dementia pathway was experienced and negotiated by service What is known providers and recipients. A pragmatic mixed-methods approach 'organically' develops understanding and evidence through flexibly, and reactively, utilising a methods framework in response to the research context as it evolves (e.g., Feilzer, 2010;Morgan, 2007), rather than rigidly applying a fixed method design (e.g., Creswell & Plano Clark, 2007). The mixed methods framework in question, included: interviews and focus group discussions; practitioner 'shadowing' observations, and service attendance and performance metrics, which were used to explore the barriers to effective integrated and collaborative practice for older people mental health services, as well as other statutory and third sector dementia service providers. This paper presents our understanding of the challenges in operationalising an integrated pathway from the service provider perspective. A forthcoming publication will present service user experiences.

| Design
The current analysis was part of a process evaluation (Robson, 2000) commissioned by an Integrated Commissioning Unit in South East England. The research team consisted of the authors and research assistants (n = 9). A 'pragmatic' mixed-methods framework (Feilzer, 2010;Morgan, 2007) was utilised to capture engagement and practice experiences of people with dementia, their carers and pathway practitioners.
As such multiple research methods were incorporated 'organically' and 'in situ' (i.e., a methods framework was utilised reactively in response to the research context as it developed, rather than prescriptively applied; Morgan, 2007). The methods framework included: semi-structured and structured interviews, focus groups, practitioner 'shadowing' and a quantitative review of service attendance and performance metrics.
Adopting and adapting these methods 'organically' meant that researchers aimed to validate and challenge findings through evidence emerging from the multiple sources of data throughout the study (Morgan, 2007), e.g., shadowing observations and service attendance and performance metrics could be used to inform interview questions for both people affected by dementia and service providers, as well as challenge practitioners' answers and other initial findings (see Sources of data & Procedures for fuller description). Due to resource and time restraints our analyses did not address 'end of life' or nursing and residential care experiences.

The NHS Health Research Authority and the University of East
London ethically approved the study. All interviews and focus groups were conducted in a quiet and confidential setting, either within participants' places of work or at the premises of the host University for the study.  Table 1).
1. Interviews and focus groups concentrated on service aims and practitioners' perspectives on their roles and experiences when working with people with dementia, family carers and other service providers. All pathway service provider managers were interviewed at least once. Other interviews and focus groups were with practicing professionals, which were either prearranged with the managers, or initiated after shadowing observations in the field. Questions were guided by previous shadowing observations and performance metrics. Where possible, or necessary, practitioners were interviewed on more than one occasion (Table 1). For example, service managers were initially interviewed and asked to outline the aims of their services, how they were implemented and their remits in relation to other services. These scoping interviews were used to inform team practitioner interview questions and shadowing observations and, on some occasions, re-interviewing of managers e.g., if practitioner role clarification was needed.
2. Shadowing observations were pre-agreed with service provider managers and their team members. These involved practitioners being observed in their day-to-day roles by one or more researchers. Where possible the same practitioners were observed by two different researchers on separate occasions. Shadowing observations were inductive and interpretative (Quinlan, 2008) and focused on the quality of practitioner interactions with people with dementia and carers, as well as on inter-and intra-service relationships with colleagues. Data generated from these sessions included informal discussions with practitioners and service users which were written up with observation notes after session interactions were completed. The accuracy and quality of observations were verified between observers and the participants being shadowed;either less formally at the time of observation or more formally elsewhere (e.g., subsequent interviews or focus groups that followed the shadowing; further detail below).
3. Attendance and performance metrics of core services included recorded outcome measures captured by providers for quarterly and annual commissioner reports and internal audits, e.g., service attendance and satisfaction metrics.

| Analyses
Interview transcriptions, shadowing observation notes, as well as performance indicators were collated. A variant of Thematic Analysis (Braun & Clarke, 2006) was used to systematically identify emergent patterns and themes within and between multiple data sources.
Analysis was abductive and therefore integrated in the data gathering (Morgan, 2007), where emergent knowledge and theory were interrogated 'through action' throughout the course of the study, e.g., emergent findings from shadowing observations informed subsequent interview questions and analyses for service providers, and, likewise, earlier service manager interview narratives (e.g., outlining roles and aims of service) were used to inform shadowing observations. The main analysis of data was conducted by MC, but there were regular research team meetings throughout the study where themes and research directions were discussed, challenged and negotiated between co-researchers. Consistent with the data gathering approach, final analysis was abductive; shifting between inductive and deductive interpretative approaches to develop and triangulate emergent themes (Jick, 1979). Our findings were finalised after sharing our analysis with local dementia pathway commissioners and practitioners as well as wider stakeholder audiences (e.g., Chase, 2019).

| FINDING S
Drawing on multiple data sources in the analyses (interviews, focus groups, shadowing and attendance and performance metrics), the following themes concerning barriers to effective integrated and collaborative practice were identified (Table 2).
Throughout the review, people and carers affected by dementia spoke highly of the expertise, compassion and commitment of dementia practitioners involved in their lives. However, our analyses raised important questions about the efficacy of provision across services; particularly with respect to integrative and collaborative working. Barriers to optimal practice identified were Attitudinal and Practical & Policy driven (Table 2). These were observed to lead to: service overlap (e.g., where two or more services were identified to be doing the same job); recognised and unnecessary demands on people with dementia and their carers to repeat their needs or re-tell their stories to different professionals; information or intervention overload (at a time when people with dementia and carers may not be ready, or equipped to assimilate it); and missed practice opportunities (e.g., times when practice was duplicating a service already provided, rather than helping people meet other important needs).

TA B L E 2 Barriers to effective integrated and collaborative practice
For statutory and third sector dementia service providers and less than helpful to requests for important information which reablement workers did not have direct access to (e.g., patients' discharge dates or next of kin's contact details; see e.g., Figure 1).
Although hospital practitioners' workloads appeared to be an im-

| A 'contractual competitiveness'
Attitudinal barriers were evident in other ways. Contracted for only one or two years, locally commissioned third sector 'pilot' providers were concerned about their longer term employment and were acutely aware of how 'competing' service providers were a threat to contract extensions. While less apparent in interview data at the manager level (e.g., managers maintained their staff had an understanding of roles of other pilot services, and were willing and keen to communicate and work effectively with them), shadowing observations and interviews with practitioners 'on the ground' evidenced limited knowledge of, and willingness to promote, other third sector providers ( Figure 2).
Attitudinal barriers and resistance to working and signposting effectively between third sector providers was further evidenced through interview data at a managerial level:

| Ambiguities in interpretation of roles and financial responsibilities
For statutory health and social care practitioners (e.g., social workers and community psychiatric nurses), lack of clarity of roles Shadowing observations further identified that existing local data and patient protection protocols meant that commissioned third sector rehabilitation and advisory support services were unable to directly access health service databases and contact details for new existing people diagnosed with dementia. They were therefore reliant on referrals from general hospital admissions, GP and memory outpatient clinics appointments, or self-referrals.
Consequently, commissioned third sector services failed to run at capacity, with a large amount of their practitioner time spent networking at community events or setting up 'stalls' in locations likely to elicit self-referrals e.g., targeted GP surgeries and memory clinic reception areas.

F I G U R E 2 Shadowing field notes: contract competitiveness
Visited, with K (third sector reablement practitioner), LB (male living with dementia) and SC (LB's wife and carer) in the community. LB was four weeks out of hospital and it was planned to be K's last visit. K asked each individually how they were doing and appeared to have a good rapport with both of them. In preparation for their discharge, K was enthusiastically disseminating information in the form of leaflets and verbal summaries of services that were available to them both. This appeared to be too much information, too quickly, and neither LB nor SC seemed to welcome or understand the barrage of information and activity. Interestingly, the information provided no mention of the dementia advisors or memory cafés as possible resources until I mentioned them at the end; these fellow pilot services would have been more local and appropriately suited to monitor the wellbeing of LB and SC following their discharge from the reablement team.

(Third sector, Reablement Practitioners, shadowing notes; MC)
In a follow up interview, K was asked about why she had not signposted the Dementia  (Banerjee, 2010).

| Study limitations
A pragmatic and abductive data-driven mixed-methods approach (Morgan, 2007) was used to develop a deeper understanding of how a dementia services pathway works and was experienced in one locality in the UK. A particular strength of the chosen methodology (over mono or more 'fixed' approaches) was that it allowed, within a team or researchers, reactive 'triangulation' of data generated from multiple methods (e.g., shadowing observations could be used to validate interview data). In this way, we were able to delve beneath the 'surface layer' of how people and practices worked; and are confident in the internal validity of findings and the level of practice understanding which emerged.
While we acknowledge that our analysis represents only one geographical area in the UK (e.g., NHS England, 2019), and data col-

| Implications
With current local dementia pathway services struggling to accommodate demand at a time of scant financial support, effective integrated health and social care working is essential to accommodate need, alleviate strains on services and to encourage 'preventative' over crisis-response modes of practice (Alzheimer's Society, 2018a; Ham & Walsh, 2013;NAO, 2017NAO, , 2020 Committee, 2016), it is essential that the risk averse data sharing culture that has evolved in health and social care is replaced with one which authentically places the needs of people affected by dementia at the forefront of their care (Forder et al., 2018;NAO, 2017). A final recommendation relates to situating the voices of people affected by dementia at the centre of decisions for how best to configure and commission dementia care pathways and keep people out of hospital (Alzheimer's Society, 2018b), an issue which is picked up and discussed further in the subsequent paper by the same authors.

ACK N OWLED G EM ENTS
Our sincere appreciation goes to all the people living with dementia and their carers who were willing to share their stories with us, and