Background

Healthcare for people with dementia appears to be more complex and challenging due to the symptoms of dementia, associated care needs, higher risks, and more frequent complications than for older people without dementia [1,2,3,4].

International studies have found that a high percentage of people with dementia in long-term care settings show behaviors that challenge healthcare professionals, such as agitation or aggression [5, 6]. This behavior is associated with an increased burden on healthcare professionals [7] and, in the setting of long-term care, increased prescribing of psychotropic drugs for people with dementia [8,9,10]. This, in turn, leads to decreased quality of life [11] and a possible increase in adverse effects such as risk of falls, an increase in medication that may lead to a sedated status, and, in the worst case, mortality [12, 13]. Furthermore, people with dementia are more likely to be hospitalized, have longer hospital stays, develop delirium that is more often undiagnosed, and experience a decline in their capacity to perform the activities of daily living [14,15,16,17,18]. As a result, the transition process (here, discharge from hospital to home or nursing home) and postacute care needs are more complex, challenging, and are associated with poorer outcomes than for older people without dementia [19, 20].

Internationally, an increasing number of psychosocial evidence-based interventions are focusing on these challenges and aimed at improving care outcomes for people with dementia [21,22,23,24,25]. Study results show that despite the increasing number of evidence-based interventions, patients receive only 30–40% of their care in line with the current scientific evidence, and in 20–25% of patients, there is a risk of harm in care [26].

Furthermore, healthcare professionals report that they implement research findings relatively seldomly in a structured and systematic way in their care practice [27]. This implementation gap has been researched thoroughly. For example, regarding the prescription and administration of psychotropic drugs to people with dementia in long-term care to reduce behaviors that challenge healthcare professionals. Although this has been shown to increase mortality since 2005 and there is poor evidence of effectiveness in improving symptoms [12], implementation and provision of evidence-based alternatives such as psychosocial interventions [28] do not appear to be used as a first approach [9, 29]. This is partly because implementing evidence-based interventions appears to be complex for healthcare staff, and there is often a lack of knowledge about how to implement interventions in a structured way [30,31,32,33].

Implementation models, frameworks, and recommendations

To address this knowledge gap and further advance the implementation of, e.g., evidence-based interventions, various implementation models, frameworks, and recommendations for practitioners, researchers and other stakeholders exist. Among the best known are the Consolidated Framework for Implementation Research (CFIR) [34], the Expert Recommendations for Implementing Change (ERIC) [35, 36], and implementation outcomes according to Proctor, Silmere [37], which represent core concepts addressed by implementation science: facilitators and barriers to implementation [35], strategies to support implementation [32, 33], and implementation outcomes [34].

To evaluate the success of an implementation process, it is important to focus on the influencing factors for the implementation. Considering and identifying these factors can help to better select and design the implementation strategy up front [38], make appropriate adjustments during implementation, and gain a better understanding of what did or did not work and how and why after implementation has been completed. The CFIR provides a comprehensive description of these factors, which are divided into five major domains (intervention characteristics, outer setting, inner setting, characteristics of the individuals involved, and the process of implementation) [34].

The ERIC provides a comprehensive overview of 73 relevant implementation strategies that can be used individually or in combination by practitioners and researchers to implement interventions in care, for example [35, 36]. To assess whether an implementation has been successful and which implementation strategies are more effective, these strategies need to be tested and compared against predetermined implementation outcomes. Proctor, Silmere [37] have provided an overview of eight different implementation outcomes (acceptability, adoption, appropriateness, feasibility, fidelity, implementation cost, penetration, and sustainability), their level of analysis, theoretical basis, salience by implementation stage, and available measurements.

Research questions

To our knowledge, there is no comprehensive, systematized evidence on implementation strategies, implementation outcomes and factors that influence the implementation of evidence-based interventions, which address the three phenomena that arise from the challenges in dementia care described above. Therefore, we developed the following three research questions:

  • ▪ Which implementation strategies are promising for the implementation of evidence-based interventions for three preselected phenomena: (A) behavior that challenges supporting a person with dementia in long-term care, (B) delirium in acute care, and (C) postacute care needs?

  • ▪ What are the effects of these implementation strategies on implementation outcomes?

  • ▪ What are the factors that influence the implementation of evidence-based interventions?

Methods

We described our methodological approach for the scoping review in our published review protocol [39], and according to Pieper, Ge [40], we reused the text of our review protocol for the methods sections in this publication and made changes in the method section where the process differed between the planned and conducted methodological approach. For reporting our scoping review, we use the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) Checklist [41], as applicable (Supplementary Table 1). Additionally, we used the flow chart of the updated Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines [42] to report the three literature searches (A, B, and C).

Search strategies

To identify evidence-based interventions addressing the preselected phenomena (A, B, and C), two researchers (MR and TQ) conducted a narrative literature search in the MEDLINE (via PubMed), CINAHL and PsycINFO (via EBSCO) databases. We identified interventions that have been tested for feasibility and effectiveness and addressed our preselected phenomena. This led to the identification of these three key interventions: the Describe, Investigate, Create and Evaluate (DICE) approach for behavior that challenges supporting a person with dementia in long-term care [43], delirium management interventions (screening, assessment, monitoring, nonpharmacological interventions) [44], and the transitional care model (TCM) for the management of postacute care needs [45]. We used these interventions as starting points to develop our search string.

To develop a broad search string, we operationalized the interventions and their components into search terms. We also used other, broader terms for our identified interventions (e.g., person-centered care or transitional care) to avoid limiting ourselves to only those interventions identified up front. We supplemented these with search terms derived from our research questions (population, phenomena, implementation, setting). In addition, we used an initial search (MRM, JIB, CM and DP) in MEDLINE (via PubMed) and key publications to identify free search terms and indexing words. We clustered all of these search terms and indexing words according to the Population, Concept, and Context (PCC) mnemonic [46] and developed three different search strings (Supplemental Tables 2, 3, and 4). The search strings were developed by the researchers (A and B: MRM; C: CM), who have a professional background as nurses and have enhanced expertise in conducting reviews [47,48,49,50,51,52]. Furthermore, all three search strings were checked by all researchers (JIB, DP, TQ, MR) according to the Peer Review of Electronic Search Strategies (PRESS) guideline statements [53]. The search strings were first developed for MEDLINE (via PubMed) and were adapted for the other two databases (CINAHL and PsycINFO via EBSCO) according to the descriptions of RefHunter V.5.0 [54]. Search strategies for all three phenomena (A, B, and C) are reported in Supplementary Tables 2, 3, and 4. We searched MEDLINE (via PubMed), CINAHL, and PsycINFO (via EBSCO) between May and June 2021 and updated the search in June 2023. In addition, we conducted backward and forward citation tracking via reference lists and Google Scholar.

Selection of evidence sources

In the first step, the abovementioned first reviewers of each review (MRM: A and B; CM: C) imported the identified records under three separate Covidence [55] licenses, and records for each search were checked automatically in Covidence for duplicates. In the second step, the titles and abstracts of each search were screened independently by two reviewers (A and B: MRM and JIB; C: CM and DP) against the inclusion and exclusion criteria (Table 1). Discrepancies in the voting were first discussed between reviewers, and if consensus could not be reached, they were discussed and resolved by all researchers (MRM, JIB, CM, DP, TQ, MR) in regular video meetings. Third, full-text screening was conducted by the same two reviewers independently (A and B: MRM and JIB; C: CM and DP), and discrepancies in the voting were discussed and resolved in the same manner as in the title and abstract screening.

Table 1 Inclusion and exclusion criteria [39]
Full size table

Data extraction

Our data extraction form was based on the template for scoping reviews developed by the Joanna Briggs Institute [46]. We considered the following aspects: general information (primary and additional publication, country, setting), study design and methods (aim, study design, methods), participants (sites and study population), and intervention (description of the implemented intervention, target population of the intervention). Data extraction for each search was performed independently by two researchers (A and B: MRM and JIB; C: CM and DP). Deviations in the extraction were discussed first between the two researchers and, if a consensus could not be reached, with all researchers (MRM, JIB, CM, DP, TQ, MR) in regular video meetings.

Analysis of the evidence

For the analysis of implementation strategies, implementation outcomes, and factors influencing implementation reported in the identified studies, we used a deductive content analysis approach [56]. For this, we derived the categories from ERIC [35, 36, 57] to analyze the implementation strategies used in the identified studies. Because implementation outcomes were often not explicitly stated and reported in the included studies, we used the outcomes described by Proctor, Silmere [37] to identify and analyze implementation outcomes in the included studies. Additionally, we used the five domains of the Consolidated Framework for Implementation Research (CFIR) and their constructs [34] to analyze the reported influencing factors.

For the coding process of implementation strategies and outcomes, as well as influencing factors, the results of each search were independently coded by two reviewers (A and B: MRM and JIB; C: CM and DP). Afterward, the results for each coding were compared, and discrepancies were discussed in the two groups (A and B; C). Codes that could not be clearly assigned to one category were discussed with all researchers (MRM, JIB, CM, DP, TQ, MR) in a virtual meeting. After the coding process, all codings were peer checked by one of two researchers (TQ or MR) to ensure trustworthiness [58].

Presentation of results

For the presentation of our scoping review results, we mapped the implementation strategies and outcomes, as well as the influencing factors, in the form of 3 tables with tick boxes. In addition, we report further detailed information about the various identified in a descriptive way.

Results

Through our electronic database searches, we identified a total of 362 (A: behavior that challenges supporting a person with dementia in long-term care), 544 (B: delirium in acute care), and 714 records (C: postacute care needs). After removing duplicates, we screened 208 (A), 348 (B), and 616 (C) records against our inclusion and exclusion criteria. Ultimately, we included 7 [59,60,61,62,63,64,65] (A), 3 [66,67,68,69] (B), and 3 [70,71,72] (C) studies. In addition, we identified 9 [73,74,75,76,77,78,79,80] (A) and 2 [81, 82] (C) corresponding reports through our backward and forward citation tracking of the studies that were included in the review (Fig. 1).

Fig. 1
figure 1

PRISMA 2020 flow diagram [42] demonstrating the identification, screening, and eligibility assessments of records preceding scoping review inclusion

Full size image

Study characteristics

Most of the studies were from Australia (n = 6) [60, 61, 67, 69, 71, 72] or the USA (n = 6) [59, 62, 64,65,66, 70], and there was one study from the UK [63]. The study designs of the included primary studies included implementation studies (n = 4) [65, 67, 69, 72], projects (n = 3) [59, 62, 64], process evaluations (n = 2) [61, 71], pilot/feasibility studies (n = 2) [63, 66], pre/post design (n = 1) [60], and one qualitative study (n = 1) [70]. The number of participating healthcare professionals (n = 1079) was reported in ten studies [61,62,63,64,65,66,67, 69,70,71]. In addition, ten studies reported the number of participating people with dementia and/or patients and their relatives (n = 1435) [61, 63,64,65,66,67, 69,70,71,72]. Detailed information about the study characteristics (e.g., implemented interventions) of all included studies is provided in Table 2.

Table 2 Study characteristics
Full size table

Identified implementation strategies

In the included studies that reported implementation strategies, we were able to identify between 4 and 21 ERIC strategies per study (Table 3). The two clusters with the most reported implementation strategies were adapt and tailor to context (3 of 4, 75% reported on tailor strategies, promote adaptability, and use data experts) and train and educate stakeholders (8 of 11, 73% reported on conduct ongoing training, provide ongoing consultation, develop educational materials, make training dynamic, distribute educational materials, use train the trainer strategies, conduct educational meetings, and work with educational institutions) (Table 3).

Table 3 Implementation strategies across the different phenomena in dementia care
Full size table

We identified the most common implementation strategies in two other ERIC clusters. For the cluster develop stakeholder interrelationships, we identified the following implementation strategies: identify and prepare champions (n = 7) [59, 61, 64, 67,68,69, 72, 78, 79, 81], use advisory boards and workgroups (n = 7) [59, 62, 64, 67,68,69, 72, 78, 79, 81], and use an implementation advisor (n = 5) [59, 61, 67, 69, 75, 78, 79]. In the cluster use evaluative and iterative strategies, the following implementation strategies were identified: audit and provide feedback (n = 7) [59, 61, 62, 67,68,69, 72, 75, 78, 79, 81], develop a formal implementation blueprint (n = 7) [59, 62, 64, 67,68,69, 72, 78, 79, 81], and assess readiness (n = 6) [64, 66,67,68,69, 72, 78, 79, 81].

We were not able to identify 38 of the 73 ERIC implementation strategies. Most implementation strategies were not reported in these clusters: change infrastructure (7 of 8, 88% did not report on mandate change, change record system, create or change credentialing and/or licensure standards, change service sites, change accreditation or membership requirements, start a dissemination organization, or change liability laws), utilize financial strategies (7 of 9, 78% did not report on place innovation on fee for service lists/formularies, alter incentive/allowance structures, make billing easier, alter patient/consumer fees, use other payment schemes, develop disincentives, or use capitated payments), and provide interactive assistance (3 of 4, 75% did not report on provide local technical assistance, provide clinical supervision, or centralize technical assistance) (Table 3).

To gain deeper insight into the coding of the implementation strategies, we present examples in Table 4.

Table 4 Examples of codings for the most common clusters and implementation strategies
Full size table

Effectiveness of the implementation strategies and outcomes

Only one study tested the effectiveness of the applied implementation strategy [65]: the effectiveness of the EIT-4-BPSD versus education only. In this study, implementation outcomes related to adoption, fidelity, penetration, and sustainability were reported. The effects of the implementation outcome sustainability were compared between both groups (intervention and control). In both groups, a slight increase in the policies and environment in terms of promoting person-centered care was observed. No change was noted in the person-centered design of care plans in either group. Related to other implementation outcomes (adoption, fidelity, and penetration), no results were reported for either group [65].

Of the remaining 12 studies that did not evaluate the effectiveness of their implementation strategy, ten reported implementation outcomes [59,60,61,62,63, 66, 67, 69, 71, 72]. Here, the outcomes fidelity (n = 10), acceptability (n = 5), adoption (n = 4), and penetration (n = 4) were reported most frequently (Table 5).

Table 5 Reported implementation outcomes for the included phenomena in dementia care
Full size table

Identified influencing factors

We identified 28 of the 37 constructs of the CFIR in the included studies (Table 6). In the following, we describe the two most frequently mentioned constructs of each CFIR domain across the different phenomena in dementia care (a, b, and c). Due to the different structuring of the domain inner setting, the most frequent subcodes of the constructs implementation climate and readiness for implementation were also listed (Table 6).

Table 6 Influencing factors for the included phenomena in dementia care
Full size table

Intervention characteristics

The adaptability of the intervention was the most frequently reported CFIR construct within this domain. The adaptability of the intervention was described in terms of the needs of people with dementia and their relatives [61, 64, 66, 70, 71], knowledge that is needed/required [62, 78] and interests of professionals [64, 78], the user-friendliness of the intervention [66], organizational interests [62], and resources such as time [62, 69, 78] and staffing [64], as well as local sites where it would be interesting to implement the intervention [81].

Evidence strength and quality of the intervention was described as the second most common CFIR construct (Table 6) and was reported in terms of the perceived evidence strength and quality of the intervention [60, 63, 64] or related to intervention components such as the specialized staff (e.g., ANPs) and their roles, competencies, and skills [70, 71, 82]; information materials; documents [70, 71]; tools [77]; trainings [63, 77]; the environment [71]; and procedures [71].

Outer setting

We identified patient needs and resources as the most reported CFIR construct in this domain. Due to the focus on people with dementia and the importance of relatives as proxies during the care process, we additionally included aspects such as the needs and resources of families (which are not included in the original CFIR). Patient needs and resources were primarily described in relation to dementia [70] and were understood as influencing factors that impact implementation outcomes. For example, learning ability and the ability to coordinate care, the perception of the acute disease regarding severity and the implication of their symptoms [70] were described as influencing factors. In addition, intervention fidelity [70, 82], attitudes toward the intervention [70], and the ability to use the intervention and the awareness of the staff to support the use of the intervention [61], as well as patient resources (such as finances, living environment, insurance and medication coverage, access to healthcare, and the social network), were reported as influencing factors [70].

Influencing factors regarding needs and resources of the family were reported in terms of caregiver burden [70, 71, 82], skills and knowledge of the family (caregiver) related to the care [61, 70, 77, 82], and its coordination [70] as well as the knowledge about [70] and the perception of the disease (acute disease and dementia) [70]. In addition, expectations [61] and acceptance of the intervention [70], information about and participation in the intervention and its design [61, 64, 70, 77] were also described as influencing factors regarding the family.

Cosmopolitanism was described as the second most common construct in this domain. Here, the support and involvement of external networks such as the Alzheimer’s Association was described as an influencing factor on implementation [67, 82]. The fragmentation of the healthcare system and therefore the provision of care services was also reported as an influencing factor in the studies. In this context, aspects such as lack of cooperation, shared care plans and information exchange between external actors (e.g., primary care physicians, specialist clinics) were mentioned [71, 82].

Inner setting

Structural characteristics, culture, and networks and communications were identified as the most mentioned CFIR constructs in this domain. For the constructs implementation climate and readiness for implementation, the subcodes with the most frequent descriptions were learning climate and available resources.

Reported influencing factors within the structural characteristics construct were staff turnover [59, 67, 72, 77], structural changes in medical specialization [72], the physical environment [79], the work organization (e.g., shift work, double shifts and high volume of agency staff) [69, 77], and the level of awareness of cognitive impairment (dementia and delirium) [67]. The care setting itself was mentioned as a general influencing factor with an impact on the implementation [69, 77].

The construct culture was described as an influencing factor in terms of the culture of the organization in general [79] and management style [61, 64].

The construct networks and communications included exchange options such as meetings [62, 71], interdisciplinary teamwork [71, 72, 81], and time points when these options were available [71] during the implementation process as influencing factors.

Within the construct implementation climate, learning climate was the most described subcode, including influencing factors related to space for learning (for example, mentoring or supervision [61, 77]), as well as involvement [61, 77], support [72, 77, 81], and acknowledgment [61] of the staff during the implementation process, opportunities to try out new methods [70], and feeling safe [61] while using the intervention even if others (e.g., relatives or colleagues) disagree.

Reported influencing factors within the construct readiness for implementation were more often related to the subcode available resources, which includes time and workload of the staff [59,60,61,62,63, 67, 69, 72, 77,78,79, 81], staffing level [62, 71, 77], and resources for training [61]. Additionally, the physical environment, such as walking areas and activity rooms [77], and activity materials [77] and finances of the facility were mentioned [77].

Characteristics of individuals

We identified knowledge and beliefs about the intervention and other personal attributes as the most mentioned constructs for this CFIR domain.

Knowledge and beliefs about the intervention were described by the influencing factor attitude toward the intervention, for example acceptance [60, 61, 77], usefulness [60, 63, 71, 72, 77, 81], appropriateness [63, 71, 77], agreement with values [63, 72], burden [77], and extra work [77]. Moreover, the expectation of the intervention (e.g., outcomes) or its implementation (e.g., losing jobs) [64] was also described and included reports about desired or perceived outcomes for the patient and the family (e.g., well-being, quality of life, relationship, positive response) [61, 64, 77, 81], the staff (e.g., empowerment, confidence, teamwork, work satisfaction) [61, 64, 77], and the organization (e.g., reputation, public relations, requesting new entries, time, and workload) [61, 71]. Furthermore, the knowledge about the intervention and their task and roles in providing these interventions were described as additional influencing factors [61, 63, 71, 72].

Influencing factors such as motivation [64, 77], commitment [61, 69], language [61], experience [67], social skills [79], openness [62, 77], and cooperativeness [61] were identified as other personal attributes.

Process

In this CFIR construct, we identified the most frequently influencing factors related to engaging. We found influencing factors on engaging in general as well as specific influencing factors related to champions.

Engaging was reported in terms of engagement of staff in general (e.g., existing or lack of) [59, 72, 79, 81], qualities of the people engaged (e.g., interdisciplinarity and skills in dementia care) [62, 67, 77], and strategies (e.g., relocation staff members) [77].

Influencing factors related to champions were distinguished in quality (e.g., strong and passionate about dementia care, expertise in dementia care, skills and interest in the intervention) [62, 72, 77], tasks (e.g., interdisciplinary problem solving, ongoing education, brainstorming activities, staff meetings, physical presence on the ward) [59, 61, 64, 72, 77], and roles (e.g., role modeling, leading light in the implementation process) [59, 61, 77]. Moreover, the availability of a champion was reported as a general influencing factor [59, 67, 72].

Discussion

To our knowledge, this is the first study to systematically identify implementation strategies, implementation outcomes and influencing factors related to the implementation of evidence-based interventions that focus on three preselected phenomena in people with symptoms of dementia or those who have been diagnosed with dementia: (A) behavior that challenges supporting a person with dementia in long-term care, (B) delirium in acute care, and (C) postacute care needs. The strengths of our scoping review are the methodological quality and the systematic and broad scope. Consequently, we can provide a broad and theoretically guided overview of the current state of implementation research in dementia care across different healthcare settings.

In summary, we identified various multifaceted implementation strategies (between 4 and 21 per study), implementation outcomes (between 0 and 5 per study), and influencing factors (between 1 and 19 per study) across the 13 included studies [59,60,61,62,63,64,65,66,67, 69,70,71,72]. Despite the three different dementia-specific phenomena and the different healthcare settings, we did not find remarkable differences in the use of the implementation strategies, implementation outcomes, or factors influencing the implementation.

In terms of influencing factors, available resources appeared to be one of the most important factors influencing implementation, along with the adaptability of the intervention. This does not come as a surprise since acute care and nursing homes have often struggled with staffing, high staff turnover rates, funding issues, challenges with available equipment, and limited influence on changing the environment, even before the COVID-19 pandemic [85,86,87,88]. This could explain why we found hardly any differences in the reported implementation strategies and influencing factors between the different interventions and settings. Accordingly, it appears that these contextual factors tremendously influence the successful implementation of evidence-based interventions due to their general conditions and requirements for implementation under current conditions (e.g., staffing, staff workload, competencies, qualifications, turnover, finances). These current contextual factors can be understood as an implementation-hostile climate [89]. To address this challenge, the implementability of healthcare interventions seems to be a crucial point [90], and adapting the intervention to the specific care context and professionals’ workflows for higher acceptability will be key for successful implementation [91]. This highlights the importance of not developing and evaluating interventions in isolation from implementation strategies [92, 93] and/or without a process evaluation [94,95,96].

Furthermore, it seems necessary to critically discuss the added value of implementation research with a sole focus on influencing factors, even when this could lead to the identification of defining implementation strategies [38]. Here, a paradigm shift [97] from identifying and describing these influencing factors to developing concrete evidence-based implementation strategies seems necessary. Thus, for the discipline (implementation science) to move forward, it is essential to consolidate innovative study designs [98] and methods (specifically participatory research approaches [99]) to develop discrete, multifaceted, and tailored implementation strategies and to investigate/test their impact on the implementation strategy and outcome itself as well as the effect on intervention outcomes [100]. This gap in the current implementation research is confirmed by our results since we were only able to identify one study that tested the effectiveness of an implementation strategy [65]. Consequently, the effects of the implementation strategies we identified are still largely unknown, and it seems that implementation research [101] and respective process evaluations to address implementation challenges during the evaluation of an intervention [93] in dementia care have barely evolved in relation to this point.

However, there also seems to be a lack in the reporting of implementation outcomes and the use of psychometrically tested implementation outcome measurements, as well as an inconsistency in the understanding between intervention outcomes and implementation outcomes [47, 102, 103]. For example, in our included studies, implementation outcomes were often not specifically named as such and were not measured with psychometric tested assessments, or it often remained unclear to what extent the measurement of, e.g., gaining knowledge, could be either an implementation outcome or an outcome of an intervention if the focus lies on education. Therefore, it is necessary to improve reporting on implementation strategies and outcomes (in both intervention and implementation studies) to initiate the development of psychometrically tested measurements [102] and, despite the publication of Proctor, Silmere [37] in 2011, to keep in mind the tension between intervention and implementation outcomes [47].

Finally, we were able to identify dementia-specific influencing factors, in particular related to the family, their needs and resources, as a key point during the implementation of evidence-based interventions. This meant that we needed to modify the CFIR (outer setting—patient needs and resources/needs and resources of the family) for our review accordingly. Although the updated version of the CFIR was published in 2022 [104], considering family needs and resources as an influencing factor for implementation does not seem to be included. However, from our perspective, this seems to be a highly relevant factor for older people with and without dementia [105]. In addition, other dementia-specific influencing factors also appear to exist for the implementation of interventions that include this population [106]. We live in a diverse and global world, and in the health sector, embracing diversity is essential for individuals’ health [107, 108]. Here, it seems to be of interest in future (implementation) research to what extent frameworks such as the CFIR consider factors influencing diverse populations (e.g., people with dementia and/or migrants or ethnic minority groups). In summary, these aspects could lead to further and tailored development of the CFIR as well as the ERIC.

Limitations

Our scoping review has some limitations. As a first step, we derived our search terms from identified exemplary evidence-based dementia care interventions and their components (e.g., DICE) and supplemented them with other, broader terms (e.g., person-centered care). In doing so, we cannot exclude the possibility that we failed to consider very specific interventions addressing our preselected dementia phenomena. However, across the different included studies and thus the different interventions and settings, our results present a very homogeneous picture regarding influencing factors, implementation strategies, and outcomes. Second, by using the ERIC clusters, Proctor’s outcomes, and the CFIR domains and constructs, we used specific frameworks and descriptions, which makes it difficult to compare our results with others analyzed with other frameworks and descriptions. However, the ones we used are among the most established due to their high number of citations [57, 109]. Third, we need to point out that an update of the CFIR [104] and the CFIR Outcomes Addendum [110] were published after the completion of our review (2021). In particular, the update of the CFIR is characterized by a more specific and detailed classification of the different influencing factors (e.g., subdividing patient needs and resources into three different constructs and moving them into the domain of internal setting and persons). Therefore, it would be interesting to compare our results with the results of future dementia-specific studies focusing on influencing factors and using the updated CFIR. It would be interesting to analyze the extent to which the updated CFIR is in line with our understanding of influencing factors. Damschroder, Reardon [104] point out that despite the changes in the updated CFIR version, the constructs can be consistently mapped back to the original CFIR, thus allowing comparison of their conceptualization.

Finally, it should be mentioned that publication bias cannot be excluded; for example, we did not specifically and systematically search for gray literature [111].

Conclusion

Based on the ERIC, the descriptions of Proctor, Silmere [37], and the CFIR, our scoping review provides a broad but systematically conducted and structured overview of the current state of implementation research in dementia care. Furthermore, our review identifies various gaps to be addressed by further implementation research. Our results show that the factors influencing the implementation of evidence-based interventions in dementia care are highly homogeneous, regardless of the evidence-based intervention and/or healthcare setting. In addition, the influencing factors we identified most frequently (available resources and adaptability of the intervention) are factors to be expected in the context of and with an impact on the provision of dementia care. In contrast, we found almost no reports on the effects of the identified implementation strategies. Consequently, to fill this gap, it seems important to test existing implementation strategies, to address tailoring-based awareness for the known influencing factors and to advance implementation science and therefore to be able to make predictions about the effectiveness of implementation strategies. This could further promote the overall translation of evidence-based dementia care practice and sustain a high quality of care for a vulnerable population.