Much is already known about principles that should guide care, supports, and services for persons living with dementia, as well as core components of care that should be provided throughout the course of the condition (see Chapter 2). Unfortunately, many persons living with dementia lack access to or do not receive these core components; nonetheless, further study is not needed to conclude that they should be provided to all. It should be noted, moreover, that at the individual or family level, persons living with dementia, care partners, and caregivers may want to experiment with such pleasurable activities as listening to music that can be tailored to their personal interests and carry little potential harm to see what works for them, knowing this may change as the condition progresses. This report, however, focuses on what is known about the effectiveness of specific care interventions, services, and supports to serve as the basis for decision making about their broad dissemination and implementation and to inform the relative prioritization of interventions that could be helpful but will require resources that are limited.
This chapter begins with a review of the evidence supporting the two types of dementia care interventions for which the Agency for Healthcare Research and Quality (AHRQ) systematic review found sufficient evidence to support conclusions about effectiveness: collaborative care models and a multicomponent intervention for family caregivers (REACH [Resources for Enhancing Alzheimer’s Caregiver Health] II and its adaptations). It then examines gaps in and opportunities for improving and expanding the evidence base on other dementia care interventions.
INTERVENTIONS READY FOR IMPLEMENTATION IN REAL-WORLD SETTINGS WITH MONITORING, EVALUATION, QUALITY IMPROVEMENT, AND INFORMATION SHARING
This section first details the committee’s approach to assessing the evidence on readiness for broad dissemination and implementation of the above two types of interventions, including application of the GRADE (Grading of Recommendations Assessment, Development and Evaluation) Evidence to Decision (EtD) framework. It then reviews the evidence, drawn, as explained below, not only from the AHRQ systematic review’s findings on effectiveness but also from supplemental sources addressing not only effectiveness but also the criteria of the EtD framework.
Approach to Assessing the Evidence on Readiness for Broad Dissemination and Implementation
Ideally, decisions to broadly disseminate and implement interventions for persons living with dementia, care partners, and caregivers would be informed primarily by evidence from multiple large, rigorous randomized controlled trials (RCTs) that tested an intervention in all relevant settings where it was to be provided; included participants from all representative populations; and tested key factors related to successful implementation, such as integration into existing workflows and contextual factors related to the settings in which the intervention was designed to be delivered. However, the AHRQ systematic review found limited such evidence, finding sufficient evidence to draw conclusions about effectiveness for only the two types of interventions noted above—collaborative care models and multicomponent interventions for family caregivers (REACH II and adaptations)—each found to be supported by low-strength evidence of benefit on specific outcomes for persons living with dementia or care partners and caregivers (Butler et al., 2020) (see Box 5-1).
Given the limitations of the evidence base, the AHRQ systematic review was unable to draw conclusions regarding all other interventions examined. According to the authors, “Ultimately, we uncovered very little evidence to support interventions and programs for active, widespread dissemination because evidence was insufficient to draw conclusions about the effects of the vast majority of interventions studied” (Butler et al., 2020, p. ES-3). This does not necessarily mean that those interventions are not helpful for persons living with dementia, care partners, or caregivers. As the authors note,
Rather, it means that current available evidence cannot yet provide clear answers about which interventions offer consistent benefits. Therefore, the uncertainty of the evidence is too high for us to draw conclusions, at present. Furthermore, when the evidence overall does not find a difference between groups, uncertainty is even higher about whether the lack of difference is truly because the interventions being compared did not differ in effect, or because the studies were designed to detect differences rather than no difference. (Butler et al., 2020, p. 107)
It is important to emphasize that the AHRQ systematic review was designed specifically to inform the question of which interventions, if any, are ready for broad dissemination and implementation, and the review
authors made decisions through this lens that inform the interpretation of the review findings and conclusions. The AHRQ systematic review excluded studies judged to be in Stages 0–II of the National Institutes of Health Stage Model for Behavioral Intervention Development (small-sample or pilot studies) and those judged to have high risk of bias (Butler et al., 2020).1 Stages 0–II describe early-stage research that has not yet included testing of interventions in real-world settings. Excluding studies that have small sample sizes or high risk of bias is standard in systematic reviews. However, exclusion of the heterogeneous category of pilots is important to interpreting the systematic review findings. While many pilots are small and preliminary, the review also excluded some studies that are described by the study authors as pilots although they used relatively large sample sizes (i.e., hundreds of participants), and in some cases were conducted in the community with relatively long follow-up times. An example of such a study is an assessment of the Maximizing Independence at Home (MIND) program in a pilot RCT involving 303 community-dwelling individuals that included two arms, with outcomes being measured at 18 months (Samus et al., 2014). Had the AHRQ review targeted specific interventions in more depth and included research-setting efficacy studies without applying the lens of readiness for dissemination and implementation, the analysis and conclusions might have been different. Because these pilot studies were excluded from the systematic review, it is unknown what portion of them could potentially be informative for determining efficacy and what portion would be excluded because of sample size or quality concerns. Considerations related to the trajectory of development in this field and approaches for developing the evidence base needed to support implementation in the real world are explored later in this chapter and in Chapter 6.
The limitations described above make it challenging to answer the core question that motivated this study of which dementia care interventions, if any, are ready for broad dissemination and implementation. To provide the most complete view of the evidence available to inform decision making in real-world settings, the committee supplemented the AHRQ review findings by applying the GRADE EtD framework and considering supplemental evidence, as described below.
1 Of the 627 unique studies eligible for analysis, 409 were excluded because they had small sample sizes or were pilots, and an additional 218 were assessed as having high risk of bias. Recognizing the challenges of conducting research in this area, the AHRQ systematic review authors set the sample-size criterion generously: studies were excluded only if they had fewer than 10 participants per arm. Similarly, the review authors characterize their approach to assessing risk of bias as “generous, relative to how risk of bias is assessed in more targeted systematic review topics.” For example, studies were assessed as having high risk of bias due to attrition only if attrition was greater than 40 percent (Butler et al., 2020, pp. 20, 108).
The findings of the AHRQ systematic review with regard to intervention effectiveness were used to identify the above two types of interventions as potentially ready for broad dissemination and implementation. In addition to effectiveness, however, many factors need to be considered in determining whether broad implementation of an intervention is appropriate. To inform the development of its recommendations on the two types of interventions identified in the systematic review as supported by low-strength evidence of benefit (see Figure 5-1), the committee applied the GRADE EtD framework. As described in Chapter 4, this framework can be used to consider evidence on factors in addition to effectiveness in making and using clinical recommendations, coverage decisions, and health system and public health recommendations and decisions (Alonso-Coello et al., 2016; Moberg et al., 2018). Factors taken into account in the assessment can include the priority of the problem, how substantial the benefits and harms are, the certainty of the evidence, the value of the outcomes to stakeholders, how the intervention or option in question compares with others, resource requirements, cost effectiveness versus comparable options, the impact on health equity, the acceptability of the intervention to stakeholders, and the feasibility of implementation.
For the EtD assessment, the committee culled available evidence on each criterion from studies that met the AHRQ review criteria for inclusion. The committee also examined evidence in other available published studies on these interventions. In particular, REACH II has been adapted for different populations and implemented in a variety of settings. The AHRQ review authors rated some of these implementation studies as having high risk of bias and excluded others because they used methodologies that failed to meet their inclusion criteria, such as studies with a single pre–posttest. Nevertheless, these studies provide such information as feasibility, equity, and resources required that is important for making decisions regarding implementation in the real world. The committee identified these additional studies by reference mining the studies that met the AHRQ review inclusion criteria, reviewing studies mentioned in the AHRQ review that did not meet the inclusion criteria, conducting PubMed and hand searches, and reviewing the Best Practice Caregiving database and Center for Medicare & Medicaid Innovation (CMMI) evaluations.
Together, these studies provide a rich source of evidence regarding effectiveness beyond the outcomes on which the AHRQ review was able to draw conclusions. Thus, without negating the AHRQ review’s conclusions regarding the low strength of evidence or the uncertainties that prevented the review from reaching conclusions on many outcomes, the committee deemed that compiling this comprehensive set of available evidence on effectiveness could be informative for those considering implementing these two types of interventions in a range of settings. Accordingly, the sections below include discussion of trends in intervention benefits across a range of outcomes beyond those for which the AHRQ review was able to draw conclusions. These findings include primary outcomes from individual studies for which the AHRQ review found insufficient evidence to draw a conclusion (often because of inconsistencies in this outcome across studies), primary outcomes not included in the AHRQ review,2 secondary outcomes mentioned in the AHRQ review, and results from implementation studies that did not meet the AHRQ inclusion criteria. It is important to note that these additional data on effectiveness are provided for descriptive purposes; the committee’s conclusions and recommendations rest on the AHRQ review findings and the EtD assessment, as well as on the
2 “For the KQs, we assessed the effects of outcomes using clinically important differences if well-established, but for many outcomes this was not the case. Because of the very wide range of outcomes of interest across the panel of potential interventions, we did not list specific priority outcomes beyond those noted in Table 1.1. For any individual study, we examined no more than five to seven outcomes per PLWD or caregiver population, prioritizing person-centered outcomes, (e.g., quality of life, function, and harms), over intermediate outcomes (e.g., laboratory test values, subscales of outcome measurement tools). Our rationale for this decision is that excessive reporting of outcomes generally happens with the latter type of outcome” (Butler et al., 2020, p. 18).
committee’s analysis of the most effective path forward for improving the evidence base for these interventions, as described later in this chapter.
Collaborative Care Models
The AHRQ systematic review identified 13 unique studies of collaborative care models from 32 publications (Butler et al., 2020). Of these 13 studies, 7 were determined to have low or medium risk of bias and proceeded to analysis. These studies report outcomes for persons living with dementia and care partners/caregivers, but the AHRQ review found sufficient evidence to support conclusions only on three outcomes for persons living with dementia. The review found insufficient evidence to support conclusions on the other outcomes for persons living with dementia evaluated in collaborative care interventions, as well as on benefits to care partners/caregivers. Box 5-1, presented earlier, provides the AHRQ summary of findings.
The AHRQ systematic review describes collaborative care models as care delivery interventions that use multidisciplinary teams integrating both medical and psychosocial approaches to the care of persons living with dementia (Butler et al., 2020). Collaborative care models may include individuals with multiple comorbidities, as is the case for both the ACCESS and Care Ecosystem models (Boustani et al., 2019; Possin et al., 2019; Vickrey et al., 2006). This section provides an overview of the AHRQ review’s analysis of evidence on collaborative care models, examines limitations of the evidence and the analysis, and applies the GRADE EtD framework to inform discussion of the readiness of these models for broad dissemination and implementation.3Box 5-2 provides a brief description of those collaborative care models that were included in the AHRQ analytic set and demonstrated benefit on at least one outcome.
Summary of AHRQ Findings on the Effectiveness of Collaborative Care Models
As described in Box 5-2, the seven studies of collaborative care models included in the AHRQ systematic review’s analytic set examined six different model interventions. The three outcomes for persons living with dementia for which the AHRQ review found sufficient evidence to draw conclusions were (1) quality of life, (2) quality indicators, and (3) emergency room visits. Box 5-1, presented earlier, provides the AHRQ descrip-
3 For the purposes of this assessment, benefit as reported here is based on statistical significance, consistent with the AHRQ systematic review approach. This was extracted from the AHRQ systematic review for those studies included in its analytic set, and from the results reported in the original study for those studies not included in the analytic set.
tion of this intervention category and a summary of the review findings. Findings for persons living with dementia and for care partners/caregivers are discussed in turn below.
Findings for persons living with dementia
The AHRQ review evaluated the effect of collaborative care interventions on seven outcomes for persons living with dementia: quality of life, neuropsychiatric symptoms, function, depression, quality indicators, emergency room visits, and nursing home placement (Butler et al., 2020). At least one study reported evidence of benefit for quality of life, neuropsychiatric symptoms, quality indicators, emergency room visits, and nursing home placement. The AHRQ review found no benefit for function or depression in persons living with dementia.
The AHRQ review’s analytic set for collaborative care included four studies that evaluated quality of life in persons living with dementia. Of these, two observed benefit: a study of the ACCESS model (Vickrey et al., 2006) and a study of the Care Ecosystem model (Possin et al., 2019). The study of the ACCESS model reported benefit at 18 months follow-up and that of the Care Ecosystem model at 12 months. Two other studies evaluated in the AHRQ review found no benefit for quality of life in persons living with dementia (Chodosh et al., 2015; Thyrian et al., 2017). The conclusion of the AHRQ review was that there was low-strength evidence that collaborative care models improved quality of life in persons living with dementia (Butler et al., 2020). Despite the mixed findings for this outcome, the weighting of larger pragmatic trials in the analysis enabled the AHRQ review to reach this conclusion about low-strength evidence.
One study was included in the AHRQ review’s analytic set for the effects of collaborative care on neuropsychiatric symptoms. For the Indiana University/Purdue University collaborative care model, benefit was observed for neuropsychiatric symptoms in persons living with dementia at 12 months follow-up (Callahan et al., 2006). However, this was a relatively small, explanatory study, and the AHRQ review found this evidence insufficient to support a conclusion regarding the effectiveness of collaborative care models for reducing neuropsychiatric symptoms in persons living with dementia.
Two studies assessing quality indicators were included in the AHRQ review’s analytic set on collaborative care. Both of these studies evaluated the ACCESS model, and both reported benefit for quality indicators as measured by adherence to dementia care guidelines at an average of 22.5 months (Vickrey et al., 2006) and either 6 months or 12 months follow-up (Chodosh et al., 2015). Given these findings, the AHRQ review concluded that there was low-strength evidence that collaborative care models improved quality indicators.
The AHRQ review’s analytic set for collaborative care models included one study that evaluated emergency room visits. This study assessed the Care Ecosystem model, and observed benefit for decreasing emergency room visits for persons living with dementia during 12 months of followup (Possin et al., 2019). While this was the only collaborative care study evaluating emergency room visits, AHRQ’s analysis was weighted for larger
pragmatic trials, leading to the AHRQ review’s conclusion that there was low-strength evidence that collaborative care models decreased emergency room visits for persons living with dementia.
The analytic set for collaborative care included three studies assessing nursing home placement for persons living with dementia. The study of the collaborative care model of the Central Union for the Welfare of the Aged in Helsinki reported benefit for nursing home placement at 1.6 years, but not at 2 years (Eloniemi-Sulkava et al., 2009). The other two studies included in the AHRQ review’s analysis found no benefit for nursing home placement (Callahan et al., 2006; Thyrian et al., 2017). The AHRQ review determined that there was insufficient evidence to draw a conclusion on the benefit of collaborative care for nursing home placement.
Two studies included in the AHRQ analytic set assessed the effects of collaborative care models on the function of persons living with dementia. Neither of these two studies reported a benefit for function (Callahan et al., 2006; Thyrian et al., 2017).
The AHRQ review included one study evaluating depression among persons living with dementia in the analytic set for collaborative care. This study found no benefit of the collaborative care model for depression (Callahan et al., 2006).
Findings for care partners and caregivers
The AHRQ review’s analytic set for collaborative care models included studies evaluating five outcomes for care partners and caregivers: quality of life, strain, depression, self-efficacy, and quality measures (Butler et al., 2020). At least one study reported benefit for strain, depression, and quality measures. The AHRQ review found no benefit for quality of life or self-efficacy in care partners and caregivers.
Of the studies in the AHRQ review analytic set that evaluated the effect of collaborative care on care partner and caregiver strain, benefit was observed for both the Care Ecosystem model (Possin et al., 2019) and the Dementia Care Management model (Thyrian et al., 2017) at 12 months follow-up. The other studies included in the analytic set found no benefit for care partner and caregiver strain (Bass et al., 2013; Chodosh et al., 2015; Vickrey et al., 2006). The AHRQ review determined that this inconsistent evidence was insufficient to draw a conclusion on the effects of collaborative care models on care partner and caregiver strain.
The analytic set for collaborative care included three studies evaluating care partner and caregiver depression. Among these three studies, the Care Ecosystem model was the only one for which benefit was observed, having improved care partner and caregiver depression at 12 months follow-up. The other two studies found no benefit for this outcome (Bass et al., 2013; Callahan et al., 2006). Given this evidence, the AHRQ review found that there was insufficient evidence to draw a conclusion on the effect of collaborative care on care partner and caregiver depression.
One study included in the analytic set for collaborative care assessed quality measures. This study of Partners in Dementia Care observed benefit for certain indicators at 6 months follow-up, including unmet needs of care partners and caregivers and use of care partner and caregiver support services, but no benefit for the number of unpaid helpers that assisted with the care of the person living with dementia (Bass et al., 2013).4 The AHRQ review determined that the evidence was insufficient to draw a conclusion regarding the effects of collaborative care models on quality measures.
The AHRQ review included one study evaluating care partner and caregiver quality of life in the analytic set for collaborative care. This study did not demonstrate benefit for quality of life in care partners and caregivers (Vickrey et al., 2006).
The collaborative care analytic set included one study that measured care partner and caregiver self-efficacy, but it found no benefit for this outcome (Possin et al., 2019).
In addition to the outcomes assessed in the AHRQ review, at 18 months, caregiver participants in ACCESS were more likely to report confidence in caregiving and caregiving mastery, better social support, and receipt of adequate help with patients’ problem behaviors and such services as respite care or home health aide services (Vickrey et al., 2006).
Findings for related interventions
Several interventions that share features with collaborative care models, including several identified in the AHRQ systematic review in categories other than collaborative care, have been implemented in various settings. These related interventions include programs that restructured the delivery of care through care management teams from a centralized hub (Amjad et al., 2018) or within a specific health care system (Jennings et al., 2020; LaMantia et al., 2015), as well as case management programs designed to help coordinate care and services for persons living with dementia and their care partners/caregivers (Chien and Lee, 2008). These interventions either did not meet the AHRQ systematic review’s inclusion criteria or were found to have insufficient evidence for the review to draw a conclusion about their effectiveness.
Limitations of the evidence
The AHRQ review aggregated results across different interventions in the collaborative care category to draw conclusions regarding particular outcomes. This approach was not designed to distinguish effectiveness among interventions within a category, a limitation
4 A translation of Partners in Dementia Care implemented in two sites in Ohio, evaluated with a quasi-experimental, pre–post design, also showed benefit for a decrease in embarrassment about memory problems and unmet needs for persons living with dementia (Bass et al., 2019). The study also found benefit for care partners and caregivers for decreased lack of caregiving confidence and unmet needs.
exacerbated by the lack of common outcome measures (discussed further in Chapter 6). As an illustrative example, studies of four different collaborative care models assessed caregiver burden, with two showing benefit and two showing no effect. Because of this inconsistency across studies, the AHRQ review judged this evidence insufficient to support a conclusion regarding the effect of collaborative care models on caregiver burden. In contrast, the effect on number of emergency room visits was assessed for only one model (Care Ecosystem), and a benefit was reported. In this case, the AHRQ review was able to draw a conclusion based on a single study because it was a large pragmatic study. However, the frequency of inconsistent findings across outcomes that were studied for more than one collaborative care model raises questions about whether similar inconsistencies would have been found had more than one study assessed the emergency room visit outcome. To reflect the AHRQ analytic approach and the complex set of evidence analyzed, and to avoid implying greater certainty about specific models than is warranted by the evidence and the approach to its review, the committee followed the AHRQ analytic approach in examining collaborative care models as an intervention category.
Additional Evidence on Effectiveness
Several collaborative care interventions for which no studies met the AHRQ review inclusion criteria were assessed in CMMI evaluations. The two interventions that had sufficient data enabling CMMI to perform a rigorous analysis were the Aging Brain Care Medical Home program and the University of California, Los Angeles (UCLA), Alzheimer’s and Dementia Care program (NORC at the University of Chicago, 2016). Analyzing 3 years of claims data, the CMMI evaluation found no benefit for utilization or cost outcomes for the Aging Brain Care Medical Home participants living with dementia. However, a qualitative analysis of participant focus groups and interviews reported that participants and their care partners and caregivers had improved quality of life and quality of care, and decreased stress, although this finding was not reported by condition, and more than half of the program participants were not living with dementia.
The CMMI evaluation of the UCLA Alzheimer’s and Dementia Care program reported a decrease in hospitalizations for ambulatory care–sensitive conditions and in 30-day readmissions for persons living with dementia (NORC at the University of Chicago, 2016). The study of the intervention also reported a decreased risk of persons living with dementia being admitted to long-term care facilities. The cost of care for persons living with dementia who were enrolled in the program was less than that for the comparison group, with a cost savings of $605 per patient per quarter (90% confidence interval: $120, $1,090). A qualitative analysis of participant
focus groups reported that persons living with dementia and care partners and caregivers reported an improvement in quality of life and quality of care.
GRADE Evidence to Decision Framework
This section describes available evidence for the criteria of the GRADE EtD framework for the four collaborative care interventions for which positive results were reported for at least one outcome and that were implemented within the United States: ACCESS, Care Ecosystem, the Indiana University/Purdue University model, and Partners in Dementia Care. Two of the collaborative care models for which effectiveness evidence is described above—Central Union for the Welfare of the Aged in Helsinki and the Dementia Care Management program in Germany—were implemented in contexts very different from that of the United States, so that data on such criteria as feasibility and cost would not be applicable.
Collaborative care models have been implemented in racially and ethnically diverse populations spanning various geographic areas. It is important to note, however, that these models have been studied primarily in Black, Hispanic or Latino, and white populations, with less recruitment of individuals of other racial and ethnic groups, such as Asian Americans and American Indians.
The ACCESS trials were designed to be delivered in English or Spanish (Chodosh et al., 2015; Vickrey et al., 2006). Whereas the initial ACCESS intervention drew on a largely white population (Vickrey et al., 2006), another iteration of ACCESS was carried out in a low-income Latino community, and the majority of participants were Hispanic or Latino (Chodosh et al., 2015). The ACCESS program has been implemented in persons living with various types of dementia, including Alzheimer’s and vascular dementias, and along the spectrum of severity, with the original trial enrolling mainly individuals with mild and moderate dementia (Vickrey et al., 2006) and an adaptation comprising primarily persons living with moderate and severe dementia symptoms (Chodosh et al., 2015). The ACCESS program also has been implemented in populations of persons living with dementia and care partners/caregivers with less than a high school education (Chodosh et al., 2015; Vickrey et al., 2006), and one study reported that the ACCESS program decreased disparities in quality of care among care partners/caregivers with lower educational attainment (Brown et al., 2013). However, the requirement of phone access to participate in the intervention could pose a barrier to equal access, as some caregivers concerned about limited phone minutes would not respond to calls from program staff (Chodosh et al., 2015).
The Care Ecosystem program was delivered in English, Spanish, or Cantonese (Possin et al., 2019) to participants that identified as Asian
American, Black, Hispanic or Latino, white, or other/mixed race. Participants resided in urban and rural settings across California, Nebraska, and Iowa. About half of the persons living with dementia participating in the intervention had mild symptoms, and persons with moderate and severe dementia each constituted about one-quarter of the participants. The Care Ecosystem model was implemented successfully in a population that included individuals with less than a high school education and with low annual income. While the reliance on phone and Internet access for delivery of the intervention may disproportionately exclude older, rural-dwelling, individuals with lower educational attainment and low-income individuals—groups with lower Internet use (Pew Research Center, 2019a)—the Care Ecosystem model leverages telemedicine to connect individuals in resource-poor areas with specialized dementia care (Possin et al., 2019).
The Indiana University/Purdue University collaborative care model was delivered in an urban Indianapolis health care system that serves low-income individuals, as well as the Indianapolis Veterans Affairs Medical Center (Callahan et al., 2006). Of the persons living with dementia who participated in the intervention, half identified as Black, two-thirds were Medicaid recipients, most had multiple chronic comorbid conditions, and the average years of education attained was approximately 9. The intervention was delivered exclusively in English, and excluded individuals without phone access and persons living with dementia who did not have a consenting caregiver or care partner.
Partners in Dementia Care is an intervention available to veterans living with dementia who receive their primary care from the U.S. Department of Veterans Affairs (VA) and its care partners and caregivers (Bass et al., 2013, 2014, 2015, 2019; Judge et al., 2011; Shrestha et al., 2011). The intervention has been delivered in geographically distinct areas, including Massachusetts, Ohio, and Texas. Between 16 and 21 percent of veterans in the studies of this intervention identified as racial or ethnic minorities. Participants represented a wide range of dementia severity, as well as diversity in levels of educational attainment. Nearly all of the veterans living with dementia were male, and nearly all of the care partners and caregivers were female.
The ACCESS intervention has demonstrated some level of acceptability for various stakeholder groups. More than 75 percent of participants remained enrolled in the intervention 6 months after enrollment, although the intervention staff initially experienced difficulties in recruiting dyads to participate (Chodosh et al., 2015). Additionally, the coordination between health care systems and community organizations reduced duplication of effort for dementia care professionals, which may help reduce costs (Vickrey et al., 2006).
Participants in the Care Ecosystem intervention reported high levels of satisfaction with the program. Seventy-eight percent of caregivers said they were satisfied or very satisfied with the services provided, and 97 percent of caregivers indicated that they would recommend the program to another caregiver (Possin et al., 2019).
At the end of the 12-month intervention, 83 percent of the care partners and caregivers that participated in the Indiana University/Purdue University model rated the primary care received by the individual living with dementia in whose care they were involved as very good or excellent (Callahan et al., 2006).
In a survey of the Partners in Dementia Care care coordinators, respondents described the acceptance and use of the program by persons living with dementia and their families as a moderate challenge to implementation (Judge et al., 2011). Respondents also found physician participation to be a minor difficulty.
Collaborative care models tend to leverage existing health care and community resources, a feature that may make implementing such models across diverse settings more feasible.
The authors of the ACCESS intervention emphasize that the ability of the program to link patients and caregivers with existing community resources facilitates the adaptation of ACCESS to other settings (Vickrey et al., 2006). The in-home arm of one of the ACCESS studies encountered difficulties with fully implementing the home contacts, with the average number of in-person contacts received being just one rather than the planned six (Chodosh et al., 2015). In addition, staff involved in that trial offered suggestions for addressing issues they encountered with recruitment, especially among underserved populations, which may be useful to those considering the implementation of this intervention. These suggestions included hiring intervention staff from the target community, who would help build trust and anticipate challenges, and leveraging community resources, such as religious institutions, for outreach.
The Care Ecosystem intervention has several features that allow for greater scalability. The intervention does not require face-to-face visits, enabling dementia care to be provided to those in rural or resource-poor settings (Possin et al., 2019). In addition, the use of intervention staff as first points of contact allowed dementia specialists to attend to work that required their expertise. The intervention also can be delivered from a centralized hub to participants across large geographic areas and different health systems or within a single health care system.
The Indiana University/Purdue University model is centered in primary care practices, which takes advantage of the fact that primary care settings are where most older adults receive medical care (Callahan et al., 2006).
However, the authors note that the financial costs and practice design requirements of implementing this intervention may be impractical for many primary care practices.
Partners in Dementia Care leverages the skill sets and resources of health care in VA medical centers and of community organizations in local chapters of the Alzheimer’s Association or Area Agency on Aging (Bass et al., 2015, 2019; Judge et al., 2011; Shrestha et al., 2011). While the intervention has been tested only in veterans receiving care from the VA, the VA operates the largest health care system in the United States, and the intervention has been translated to various VA medical centers and community partnerships (Bass et al., 2019). The study authors describe these partnerships as successful, noting that neither organization was overly burdened and that intervention activities were equally distributed between the VA medical centers and community partners (Shrestha et al., 2011).
The feasibility of implementing collaborative care models has also been demonstrated in additional interventions for dementia and for other diseases. A systematic review (Dham et al., 2017) and a meta-analysis (Archer et al., 2012) found that such models were effective interventions for adults with various psychiatric conditions. A narrative review of emerging collaborative care interventions for dementia also describes a general finding of feasibility and sustainability for the interventions across settings, as well as responsiveness to the needs of health care systems (Heintz et al., 2020). These authors urge ongoing research and the use of implementation science to advance and improve collaborative care models. At the same time, the optimism surrounding the feasibility of these models may be tempered by the fact that existing health care models are resistant to change, and that broad implementation of the models will require the commitment of intervention staff, health care administrators, payers, and regulatory bodies.
In a cost analysis of the ACCESS intervention, the start-up cost is estimated as $70,256 and ongoing intervention costs at $118 per patient per month (Duru et al., 2009). In another study, the estimated cost of an ACCESS adaptation is $358 per patient per month for the in-home intervention and $216 for the telephone-based intervention, with intervention costs taking salaries, mileage, and organizational overhead into account (Chodosh et al., 2015). It is not clear whether the differences in intervention costs between these two ACCESS studies is due to differences in how the calculations were performed. In the Duru and colleagues (2009) study, average monthly health care costs, including service costs, out-of-pocket expenditures, costs of family caregiving, and spending for end-of-life care, are estimated as $6,479 for intervention participants and $6,381 for the usual care control. However, health care costs were higher for the control
group than for the intervention group when costs of end-of-life care were excluded. In the Chodosh and colleagues (2015) study, the costs related to health service utilization over a 1-year period are calculated as $5,595 for the in-home intervention and $7,761 for the telephone-based intervention. The authors do not discuss potential reasons for the difference in costs between the in-home and telephone-based implementations.
The cost of the Care Ecosystem intervention averaged from $202 to $762 per person living with dementia per month, depending on the location and stage of the intervention (Rosa et al., 2019). The calculation of intervention costs includes personnel, supplies, equipment, training, and facilities. Overall, intervention costs decreased as caseloads for intervention staff increased as a result of the efficiencies of scaling the program to cover more participants. A CMMI evaluation of Care Ecosystem observed slightly lower (5 percent) Medicare expenditures for program participants compared with control participants, although this difference may not be a result of the program (Gilman et al., 2020).
There is no formal cost analysis for the Indiana University/Purdue University collaborative care model (Callahan et al., 2006). However, the study authors estimate the cost of the care manager to be $1,000 per patient annually, with a caseload of 75 patients. The authors suggest that with the cost savings from a reduction in neuropsychiatric symptoms, a cost analysis may find the intervention to be cost-beneficial.
A cost analysis of the Partners in Dementia Care intervention estimates the cost of the intervention to range from $780 to $960 per dyad per year (Morgan et al., 2015). Costs of the intervention include, among others, staff salaries, equipment, supplies, and training. Total health care costs per intervention participant, including such expenditures as inpatient and outpatient care and pharmacy costs, were $1,006 (standard deviation $9,607) higher for the intervention group than for a control group when adjusted for baseline characteristics. Total health care costs for the intervention and control groups demonstrated a great degree of variability and were highly skewed.
Conclusion Regarding Collaborative Care Models
As a whole, the evidence supporting collaborative care models is encouraging. These interventions have been studied in multiple and diverse populations and with individuals along the spectrum of disease severity. While these interventions have been disseminated and used in relatively limited ways to date, some evidence related to acceptability, feasibility, and resources is available to inform implementation.
CONCLUSION: Collaborative care models—which use multidisciplinary teams integrating both medical and psychosocial
approaches to the care of persons living with dementia—have demonstrated some effectiveness under clinical trial conditions and are already being implemented in care settings with promising results. These interventions are ready for the next stage of field testing to support their widespread adaptation to and adoption in the variety of settings where people seek dementia care. Those efforts will enhance understanding and information dissemination with respect to key factors in addition to effectiveness that are important for deciding whether and how to implement an intervention, such as determining the workforce and space needed; testing payment models and integration into workflow; and ensuring adaptations for different populations (e.g., racial/ethnic groups) and settings (e.g., rural areas).
A Multicomponent Intervention for Family Caregivers: REACH II and Its Adaptations
The AHRQ systematic review identified 22 unique studies of multicomponent interventions for care partners and caregivers, 7 of which were rated as having a low or medium risk of bias and were included in the analysis (Butler et al., 2020). These 7 studies examined three different multicomponent interventions. The AHRQ review found sufficient evidence to draw conclusions about one of these multicomponent interventions (REACH II), but not about the other two. Box 5-1, presented earlier, provides the AHRQ summary of findings.
The three multicomponent interventions analyzed by the AHRQ review comprised different components; the common element of interventions in this category is simply having multiple components. Because the question at hand is what specific interventions are supported by sufficient evidence to be considered ready for broad dissemination and implementation, the committee limited further exploration of this category to the one intervention found to be supported by low-strength evidence of effectiveness in the AHRQ review.
This section provides an overview of available evidence on REACH II and its adaptations, examines limitations of the evidence, and applies the GRADE EtD framework to inform a discussion of the readiness of this intervention for broad dissemination and implementation. Box 5-3 describes the components of REACH II and related adaptations for different populations.
Summary of AHRQ Findings on the Effectiveness of REACH II
The AHRQ systematic review found low-strength evidence that REACH II improved care partner/caregiver depression at 6 months (Butler et al., 2020). Other outcomes examined included care partner/caregiver health, stress, and strain, but the review was unable to draw conclusions about these outcomes because the definition, measurement, and reporting of outcomes varied so widely. The AHRQ REACH II findings were based on three RCTs. The original trial studied the intervention in a population that was one-third Black or African American, one-third white or Caucasian, and one-third Hispanic or Latino (Belle et al., 2006). That study reported that the intervention was effective for decreasing the prevalence of caregiver depression. While the prevalence of caregiver depression decreased across racial and ethnic groups, only white caregivers exhibited a benefit in the decrease in depression when the results for the three groups were disaggregated. A German adaptation observed a decrease in caregiver depression (Berwig et al., 2017), a finding based on the psychological component of health-related quality of life as measured by the SF-12, which has been used as a tool to detect depressive disorders (Vilagut et al., 2013). And a study with Hispanic caregivers that compared the community-based REACH OUT adaptation with another intervention, the New York University Caregiver Intervention, found no difference between those two interventions and no change in caregiver depression from baseline (Luchsinger et al., 2018).
In addition to the low-strength evidence on reduction in caregiver depression, the AHRQ review identified a reduction in caregiver strain
associated with the REACH II adaptation DE-REACH (Berwig et al., 2017). However, this reduction was limited to Black caregivers in one of the studies (Belle et al., 2006). REACH OUT improved caregiver strain compared with baseline but not with the comparator intervention (Luchsinger et al., 2018). Ultimately, the available evidence was insufficient to draw a conclusion on the effectiveness of REACH II in reducing caregiver strain (Butler et al., 2020).
REACH II has been studied in and adapted for diverse populations to a greater extent than is usual in the field, as discussed in greater depth below. In addition, the AHRQ review concluded that there was “more development along the NIH [National Institutes of Health] Stage Model in this set than in most other intervention categories. This literature set demonstrates growth over time toward the development of both pragmatic trials as well as dissemination/implementation research” (Butler et al., 2020, p. 78).
Additional Evidence on Effectiveness
In additional studies of REACH II and its adaptations, study authors have observed a range of beneficial outcomes. With the exception of caregiver depression, the AHRQ review found insufficient evidence to support conclusions about these outcomes. The additional studies described in this section were not included in the AHRQ review analytic set, in many cases because they used an ineligible study design, such as a single pre–posttest. Their findings are described briefly here to illustrate the trends in benefits observed with this intervention.
The results of these studies suggest benefits for reductions in caregiver strain or stress (Burgio et al., 2009; Cheung et al., 2014; Cho et al., 2019; Czaja et al., 2013, 2018; Lykens et al., 2014; Nichols et al., 2011; Stevens et al., 2012), caregiver depression (Burgio et al., 2009; Cheung et al., 2014; Cho et al., 2019; Czaja et al., 2018; Lykens et al., 2014; Nichols et al., 2011), challenging behaviors of the persons living with dementia (Berwig et al., 2017; Burgio et al., 2009; Cheung et al., 2014; Cho et al., 2019; Nichols et al., 2011; Stevens et al., 2012), caregiver frustration or bother (Cheung et al., 2014; Czaja et al., 2018; Nichols et al., 2011), and physical symptoms of psychiatric conditions (Berwig et al., 2017). Studies of REACH II adaptations and implementations also reported improvements in self-reported social support (Burgio et al., 2009; Cho et al., 2019; Czaja et al., 2013, 2018), self-reported caregiver health (Burgio et al., 2009), caregiver reactions to challenging behaviors (Berwig et al., 2017; Czaja et al., 2018), positive aspects of caregiving (Burgio et al., 2009; Cheung et al., 2014; Czaja et al., 2013), and safety of persons living with dementia (Stevens et al., 2012).
Application of the GRADE Evidence to Decision Framework
This section reviews available evidence for the GRADE EtD domains related to REACH II and its adaptations.
REACH II and its adaptations have been carried out in racially and ethnically diverse study populations within the United States, including Asian American, Black, Hispanic or Latino, and white caregivers (Belle et al., 2006; Burgio et al., 2009; Cho et al., 2019; Czaja et al., 2013, 2018; Lykens et al., 2014), and the role of demographic characteristics in moderating the effectiveness of REACH II has been described in detail (Lee et al., 2010). However, much of the research on REACH II has been focused on Black, Hispanic or Latino, and white participants, with other racial and ethnic groups, such as Asian Americans and American Indians, representing small proportions of the populations studied. Of note, REACH II, REACH-TX, REACH II (North Texas), and REACH II via videophone were designed to be delivered in English or Spanish (Belle et al., 2006; Cho et al., 2019; Czaja et al., 2013; Lykens et al., 2014). Several of the REACH II iterations have been implemented successfully in low-income communities or with low-income participants (Belle et al., 2006; Cheung et al., 2014; Czaja et al., 2018). Many REACH II interventions require the use of a cellphone interface, which could potentially preclude the involvement of some caregivers, especially those aged 65 and older, who are less likely than those under age 65 to own a cellphone (Pew Research Center, 2019b). On the other hand, this feature could be helpful for adapting to the current environment in which many activities are being carried out remotely because of the COVID-19 pandemic.
An analysis of the results of REACH II according to racial and ethnic groups observed that the intervention improved quality of life for white and Hispanic or Latino family caregivers, but for Black caregivers, quality of life improved only for those who were caring for a spouse (Belle et al., 2006). The benefit observed for Hispanic or Latino caregivers may be explained in part by the intervention’s linguistic and cultural adaptation to this community, which has typically had less access to community services and resources. In the REACH OUT adaptation, white and urban-dwelling caregivers reported a greater reduction in strain relative to Black and rural-dwelling caregivers, respectively (Burgio et al., 2009). Additionally, Black caregivers experienced a larger improvement in positive aspects of caregiving compared with their white counterparts. The REACH II via videophone adaptation was reported to decrease strain for Hispanic but not Black caregivers, although the authors posit that this finding may be due to lower levels of strain among Black participants at baseline (Czaja et al., 2013). In REACH-TX, both young and Black caregivers were more
likely to be lost to follow-up compared with other groups, and the authors suggest that strategies for addressing this disparity should be explored (Cho et al., 2019).
The REACH II intervention and its adaptations appear to have broad acceptability for participants, intervention staff, and the systems in which the intervention is implemented. Nine hospital units and six clinic care teams were targeted by intervention staff to implement the REACH-TX intervention, and all agreed to participate (Stevens et al., 2012). Only 44 percent of enrolled caregivers completed follow-up. Of these, 82 percent said the services offered were helpful, and 93 percent were satisfied with the quality of those services. Moreover, all participating caregivers reported satisfaction with the information provided and the phone contacts from intervention staff. In a satisfaction survey of the 77 percent of caregivers who completed 6-months’ follow-up for Community REACH, 96 percent said that they had benefited from the intervention (Czaja et al., 2018); 93 percent reported that it had made their life easier, and 61 percent agreed that it had improved the life of the person living with dementia. In a REACH II adaptation using videophones, 82 percent of caregivers reported that the intervention was helpful, and 85 percent reported that the support groups were valuable (Czaja et al., 2013). In the REACH OUT adaptation, of the 87 percent of enrolled caregivers that provided responses, 99 percent and 98 percent were satisfied with the type and quality of the intervention, respectively (Burgio et al., 2009). Adherence to the intervention tended to be high, with 95 percent of caregivers receiving all the treatment components during at least one session of the intervention. Similarly, 81 percent of caregivers enrolled in DE-REACH completed at least 10 of 12 sessions (Berwig et al., 2017). Of the 83 percent of caregivers enrolled in REACH-HK who completed follow-up, 92 percent said they would recommend the intervention to other caregivers (Cheung et al., 2014), and all participants in the Community REACH adaptation who completed followup said they would recommend it to others (Czaja et al., 2018).
Intervention staff also have reported high levels of satisfaction with REACH II interventions. Case managers for REACH OUT all agreed that they perceived the intervention to be helpful to participants (Burgio et al., 2009). However, many of these same case managers described feeling constrained by time, primarily because the session duration and number of sessions were insufficient to enable them to fully understand and address caregiver concerns, and other aspects of the intervention were time consuming.
A number of adaptations of the REACH II intervention have been implemented in community settings, in the VA system, and in various
locations around the United States and globally (see Box 5-3 presented earlier). This range of settings provides evidence of the feasibility of adapting the REACH II intervention to suit different settings and cultures and to fit within existing community organizations and health care systems. Of particular note, the VA offers REACH VA as a routine program through its Program of General Caregiver Support Services (VA, 2020). REACH II and its adaptations have been administered by individuals from diverse professions, including nursing, social work, and counseling, in real-world care settings (Benjamin Rose Institute on Aging and FCA, 2020; Nichols et al., 2011). While the quantity and duration of in-home and telephone sessions vary among the different REACH II implementations, the authors of the REACH-TX adaptation note that the dose and intensity of the intervention can be gauged and modified through an initial risk assessment to evaluate personal and environmental challenges and needs for caregivers (Cho et al., 2019). They also explain that ongoing communications between the community organization and an evaluation team improved implementation and helped sustain a partnership with the intervention’s funder. The authors state further that sustaining the intervention was possible through monetary and institutional support from the health care system, continuing education for health care providers, and incorporation of the intervention training into orientation for new nurses (Stevens et al., 2012).
Several challenges to implementing REACH II interventions were also uncovered across the various adaptations. In DE-REACH, just 70 percent of the basic modules described in the intervention manual could be executed (Berwig et al., 2017). Local funding for the community organization implementing REACH II (North Texas) was insufficient to support a comprehensive evaluation of the program (Lykens et al., 2014). A survey of organizations that have implemented REACH II or its adaptations found that the most frequent barriers to successful implementation were lack of internal organization resources, insufficient understanding of the program, and issues with participant enrollment and completion (Benjamin Rose Institute on Aging and FCA, 2020).
The authors of various REACH II adaptations have offered suggestions for future research that would help propel REACH II toward broad implementation and adoption. According to Cheung and colleagues (2014), an important step toward broad implementation would be pragmatic clinical trials. Similarly, Luchsinger and colleagues (2018) recommend modifying trial designs in future research on REACH II, conducting long-term studies that last more than 6 months, and including diverse sociodemographic groups in studies that are appropriately powered. Authors of two separate adaptations advocate for research on how modifying the intensity, dose, and duration of the intervention may impact its effectiveness (Berwig et al., 2017; Czaja et al., 2018).
The total cost of the original REACH II intervention was $1,214 per caregiver, which included costs of staff training, staff labor for intervention delivery, caregiver time, travel expenses and travel time, and intervention materials (Nichols et al., 2008). The cost for a caregiver to gain an additional hour of time spent on noncaregiving activities was calculated as $4.96 per day for each caregiver enrolled in the program. Twelve months following the intervention, the persons living with dementia whose caregivers had been enrolled in REACH VA exhibited a cost savings (including drug costs) to the VA system of 25 percent compared with the control group and with 12 months prior to the intervention (Nichols et al., 2017). With implementation of REACH VA, the estimated average annual savings to the VA was predicted to be $4,338 per participant (Nichols et al., 2017).
Conclusion Regarding REACH II and Its Adaptations
As a whole, the evidence supporting REACH II and its adaptations is encouraging. Of particular note, REACH II has been studied in and adapted for diverse populations to a greater extent than is usual in the field. A moderate amount of evidence related to intervention acceptability, feasibility, and resource requirements is available.
CONCLUSION: REACH II and its adaptations—interventions that provide support for family care partners/caregivers through a combination of strategies that include problem solving, skills training, stress management, support groups, provision of information and education, and role playing—have demonstrated some effectiveness under clinical trial conditions and are already being implemented in a variety of community settings with promising results. These interventions are ready for the next stage of field testing to support their widespread adaptation to and adoption in a variety of settings where people seek dementia care. Those efforts will enhance understanding and information dissemination with respect to key factors in addition to effectiveness that are important for deciding whether and how to implement an intervention, such as determining the workforce and space needed; testing payment models and integration into workflow; and ensuring adaptations for different populations (e.g., racial/ethnic groups) and settings (e.g., rural areas).
Together, collaborative care models and REACH interventions are practical instantiations of many, but not all, core components of care, services, and supports that are important for persons living with dementia, care partners, and caregivers, as outlined in Chapter 2. These core components are detection and diagnosis; assessment of symptoms to inform planning and deliver care; information and education; medical management; support in activities of daily living; support for care partners and caregivers; communication and collaboration; coordination of medical care, long-term services and supports, and community-based services and supports; supportive and safe environment; and advance care planning and end-of-life care. These interventions also respond to priorities identified by persons living with dementia and their care partners and caregivers at the committee’s public meetings, including the need for education, practical guidance, skills, and support, as well as challenges related to navigating a patchwork of often uncoordinated care systems, providers, and services.
The state of the evidence base for these two intervention types as assessed by the AHRQ review complicates making recommendations for a path forward. The AHRQ finding of low-strength evidence of effectiveness suggests limited confidence in the effectiveness of these interventions and indicates that additional evidence is likely to change the estimate of effect. Nevertheless, the committee recommends a path forward based on the following argument. First, given the inherent challenges of studying this topic—including the complexity of dementia care interventions, the diversity of populations affected, and the importance of contextual effects, as described in Chapter 3—the fact that these two interventions produced low-strength evidence of effectiveness is important. Second, there is a notable trend in benefits across multiple outcomes beyond those for which the AHRQ review was able to draw a conclusion, and the consistency of evidence of benefit across sources of evidence is encouraging. Third, there is a moderate amount of evidence to inform implementation as assessed against the EtD criteria. Particularly important, while more evidence is needed regarding the full range of populations that could benefit from these interventions, they have already been studied in diverse populations, although additional evidence is needed to expand understanding of their use in all populations.
Taken together, these considerations led the committee to conclude that the evidence is sufficient to justify implementation of these two types of interventions in a broad spectrum of community settings, with evaluation conducted to continue expanding the evidence base for future implementation. The committee believes that this approach to expanding the evidence base is likely to bring greater gains and better inform real-world implementation relative to focusing on additional large RCTs aimed at generating
moderate- or high-strength evidence in a future systematic review before any further dissemination can be supported. These concepts are discussed in detail in the next chapter.
It is important to stress that these recommendations should not be taken to imply that these are the only two types of interventions that should be pursued. As discussed next, additional research on a full range of interventions should be undertaken to continue to innovate and develop better ways of meeting the urgent needs of persons living with dementia, care partners, and caregivers.
For the majority of dementia care interventions included in the AHRQ systematic review, the evidence was insufficient to draw conclusions regarding their effect on outcomes for persons living with dementia and/or their
care partners and caregivers. However, a finding of insufficient evidence does not mean that an intervention is ineffective or that it should not be implemented. Rather, such a finding simply reflects the high uncertainty resulting from the limitations of the evidence base and the approach used in the AHRQ systematic review to synthesize and assess the strength of the existing evidence to support conclusions on readiness for broad dissemination and implementation. As discussed in Chapter 4, different stakeholders may use different criteria to inform decisions on the implementation of dementia care interventions, and the AHRQ systematic review acknowledges that even low-strength evidence is a difficult bar to reach given the complexity of dementia care interventions and the settings and systems in which they are implemented (as discussed in Chapter 3). If the magnitude of the effect of an intervention is small, or moderate but only for a specific subpopulation, that effect will be more difficult to detect. Thus, it is possible, and in some cases likely, that a dementia care intervention with insufficient evidence to support a conclusion on effectiveness may be beneficial for some populations in certain circumstances. To guide research investments going forward and to extract the maximum value from the large body of interventions for which the AHRQ review found the evidence to be insufficient, the committee sought to identify gaps in and opportunities to improve and expand the evidence base for dementia care interventions.
This section first details the committee’s approach to assessing the state of the evidence for dementia care interventions other than the two interventions discussed in the previous section—collaborative care models and REACH II and its adaptations—and identifying gaps in and opportunities for expanding and improving that evidence base. The sections that follow detail the findings of this assessment first for interventions targeting the community, policy, and societal levels and then for those targeting the individual level.
Approach to Assessing the State of the Evidence for Other Interventions and Identifying Gaps and Opportunities
Consistent with the study charge, the committee’s approach to assessing the state of the evidence for interventions other than collaborative care models and REACH II and adaptations relied heavily on the findings from the AHRQ systematic review. However, the committee also considered additional sources of evidence, including expert and stakeholder input and such resources as Best Practice Caregiving, a database resulting from a joint project of the Benjamin Rose Institute on Aging, the Family Caregiver Alliance, and the Gerontological Society of America. Best Practice Caregiving provides information derived from real-world implementation of interventions (Benjamin Rose Institute on Aging and FCA, 2020) and was helpful
in identifying interventions (or adaptations thereof) evaluated in the AHRQ systematic review that had been implemented in practice settings, although the committee did not systematically evaluate the evidence of effectiveness captured in the database.
The committee also mapped the interventions in the AHRQ systematic review against the framework for dementia care interventions presented in Chapter 3. This mapping exercise made it possible to assess the balance among interventions targeting the individual, community, policy, and societal levels, all of which are important to meeting the needs of persons living with dementia, care partners, and caregivers.
In addition, the committee sought to understand the degree to which the AHRQ review’s findings of insufficient evidence resulted from a lack of evidence or from other limitations of the evidence base that prevented drawing conclusions about readiness for broad dissemination and implementation. Where evidence was lacking, the committee leveraged stakeholder input from its information-gathering process to identify opportunities for future research that would expand the evidence for interventions that have been identified by persons living with dementia, care partners, and caregivers as important to their health and well-being. (Opportunities to expand the evidence base for collaborative care models and multicomponent interventions such as REACH II that provide support to caregivers in a variety of ways are discussed in the preceding section.) For those categories of interventions (e.g., exercise, psychosocial interventions) for which the AHRQ review’s analytic set5 includes a multitude of RCTs and for which there was some signal of benefit and little or no evidence of harm, the committee identified important gaps that posed barriers to the synthesis and interpretation of the evidence. Signal of benefit was determined based on the observation of benefit for a given outcome in multiple independent RCTs evaluating the same (or a similar) intervention, even if the overall body of evidence was mixed for that outcome (i.e., one or more RCTs found no benefit for that outcome).6 Although a signal of benefit may be insufficient to recommend interventions for broad dissemination and implementation—the focus of the AHRQ systematic review—this approach enabled the committee to
5 As discussed previously, the AHRQ review excluded studies from the analytic set if they were judged to be pilot studies, had small sample sizes, or were rated as having high risk of bias.
6 For the purposes of this assessment, benefit was defined based on statistical significance, consistent with the AHRQ systematic review. However, the committee recognizes that the ability to achieve statistical significance depends on the sample size and that failure to detect a statistically significant effect does not necessarily mean that an intervention failed to provide benefit for some people or in some circumstances. Trends in the primary data that show improvement over time (or with increased intervention intensity) may support conclusions regarding benefit even in the absence of statistical significance.
highlight opportunities to address gaps and advance evidence-based practice for dementia care.
Gaps and Opportunities for Interventions Targeting the Community, Policy, and Societal Levels
In addition to individual-level interventions, discussed in the next section, community-, policy-, and societal-level interventions have the potential to improve the health and well-being of persons living with dementia and their care partners and caregivers by changing the systems and settings in which they receive care, services, and supports (e.g., by targeting the organization, financing, and delivery processes). However, much of the focus on promising interventions in the field has been on those targeting individual persons living with dementia, care partners, and caregivers. The AHRQ systematic review and a recent Lancet Commission report (Livingston et al., 2020) used different approaches to evaluate the evidence, but both focused heavily on individual-level interventions, highlighting an evidence gap related to community-, policy-, and societal-level strategies.7
In addition to the collaborative care models described earlier in this chapter, other community-level interventions evaluated in the AHRQ systematic review included case management, implementation of care protocols (descriptions of procedures, processes, and tools for providing care in an organization or care delivery system), and care staff education and training. With the exception of collaborative care models, however, the AHRQ analytic set included few if any studies for most of these interventions. Moreover, all of these interventions target systems for the delivery of care, services, and supports. Completely absent from the AHRQ systematic review are policy- and societal-level interventions, such as dementia-friendly community initiatives and social insurance policies that would provide coverage for home- and community-based long-term care.
The paucity of evidence identified for interventions beyond the individual level in the AHRQ systematic review may be due in part to the challenges involved in studying these interventions. They often are not well suited to evaluation using the kinds of study designs that are likely to meet the evidence criteria used by AHRQ for its systematic reviews. Notably, the AHRQ review included no nonrandomized studies, but an RCT of
7 Of note, the AHRQ systematic review refers to respite care and social support as programs delivered at the community and societal levels. While the availability of such programs may depend on community resources and policies, their implementation is at the individual level, and the committee therefore classifies them as individual-level interventions for the purposes of this report.
an intervention that increases paid leave for caregivers, for example, is unlikely to be feasible. Rather, evaluation methods such as those used for policy demonstration projects may be more appropriate for assessing the effectiveness of such policy interventions. As discussed further in Chapter 6, the committee urges that investment in future research on dementia care interventions include a focus on how better to study these kinds of interventions with rigor. Also vital is that evidence from studies not designed as RCTs be incorporated into future efforts to synthesize the evidence on dementia care interventions, even if there is a greater risk of bias related to the nonrandomized design. For example, one evidence synthesis methodology designed specifically to consider evidence on community-level interventions targeting population-level outcomes is that used by the Guide to Community Preventive Services (The Community Guide). Because RCTs are often difficult or inappropriate to conduct for public health interventions, The Community Guide does not privilege evidence from RCTs, but considers the suitability of the study design and the quality of execution for each quantitative study included in the body of evidence (Briss et al., 2000).
Additional challenges stem from the way interventions are defined and the consequences for search strategies used in the AHRQ systematic review. Policies and community-level programs and organizational structures may not be recognized as interventions per se. For example, the AHRQ review did not include studies on dementia villages—residential settings designed and operated around the care and support needs of persons living with dementia. At the committee’s public workshop in April 2020, Mary Butler of the Minnesota Evidence-based Practice Center indicated that those studies had been excluded because they were considered to be evaluations of the effectiveness of care delivery settings rather than intervention studies.8 The AHRQ systematic review notes that some community services and supports approaches, such as referral services and awareness-raising outreach, may have been missed because of the challenges of designing effective search strategies for such interventions in the context of a review with such broad scope. Going forward, adopting a broader definition of what constitutes a dementia care intervention may ensure that resources are invested in evaluating community-, policy-, and societal-level interventions and that such evaluations are included in future efforts to take stock of the state of the evidence.
CONCLUSION: The evidence base for dementia care interventions appears to be biased toward those targeting the individual
8 Presented by Mary Butler of the Minnesota Evidence-based Practice Center at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
level. The gap in the evidence for interventions targeting the community, policy, and societal levels may result from the way interventions are defined and the challenges of studying these kinds of interventions with rigor. Expanding the evidence base for such interventions will require investment in research that uses appropriate study designs, engages key stakeholders, and characterizes critical features of implementation. To support conclusions on the effectiveness of such interventions, synthesis methods will need to enable the evaluation of evidence on complex interventions derived from nonrandomized studies.
Gaps and Opportunities for Interventions Targeting the Individual Level
Although the evidence base for interventions targeting the individual level is larger as a whole relative to that for community-, policy-, and societal-level interventions, the AHRQ review determined that for all but collaborative care models and REACH II and its adaptations, the evidence was insufficient to draw conclusions regarding their effect on outcomes for persons living with dementia and/or their care partners and caregivers. The committee identified several gaps related to the quality and heterogeneity of the evidence for other individual-level interventions that need to be addressed to better support decision makers seeking guidance on which interventions are ready for broad dissemination and implementation. These gaps are described in the sections below, with interventions included in the AHRQ systematic review used to illustrate the issues involved and the opportunities to expand knowledge about what works, for whom, and in what conditions.
Gaps in High-Quality Evidence
For the majority of the individual-level interventions evaluated in the AHRQ systematic review, including some identified by persons living with dementia, care partners, and caregivers as important for their health and well-being, few if any studies met the criteria for inclusion in the analytic set, indicating a paucity of high-quality evidence to support conclusions regarding intervention effectiveness. As discussed previously, in many cases, studies captured through the AHRQ literature search were classified as small-scale and/or pilot studies or were assessed as having high risk of bias. While it is possible that inclusion of small trials and pilot studies could have bolstered the number of studies contributing evidence on effectiveness for some interventions, the systematic review team deemed them unsuitable for supporting conclusions on readiness for broad dissemination and implementation.
As discussed further in Chapter 6, future research can address these gaps by facilitating the evaluation of interventions through larger and longer-duration studies in real-world settings and using methodological approaches that decrease the potential for bias to reduce certainty in the findings of the study. As noted in the AHRQ systematic review, changes in research funding requirements in the past 5 years are already driving improvements in the methodological rigor of studies of dementia care interventions (Butler et al., 2020). Ongoing and future studies characterized by more stringent data monitoring and reporting, therefore, are likely to give rise to an evidence base that supports stronger conclusions regarding intervention effectiveness. The sections below highlight opportunities to expand and improve the evidence base for the following categories of interventions with the potential to make meaningful differences in the lives of persons living with dementia, care partners, and caregivers:
- late-stage care interventions,
- respite care,
- social support, and
- training and support for direct care workers.
Late-stage care interventions
The importance of models of care that meet the needs of persons living with dementia and care partners/caregivers across the full continuum of dementia stages, including early- and late-stage care, has been recognized (Gitlin and Maslow, 2018). Interventions relevant to early-stage care (e.g., educational interventions) are discussed in other sections below; this section focuses specifically on those late-stage care interventions included in the AHRQ systematic review.
Late-stage care interventions encompass care, services, and supports designed to anticipate and meet the unique care needs of persons in the late stages of dementia and their caregivers. Late-stage care interventions evaluated in the AHRQ systematic review include decision aids and supportive interventions for decision making about feeding options, advance care planning, and palliative care (Butler et al., 2020). Decision aids, a set of evidence-based tools, can be used to guide caregivers in decision making regarding care for persons living with advanced dementia. For example, such decision aids may provide information about feeding options, including feeding tubes and assisted oral feeding, for persons living with dementia who are experiencing problems related to eating, such as difficulty swallowing (dysphagia). Decision aids may also be used to facilitate advance care planning, which can help reduce uncertainties about the wishes and goals of persons living with dementia as the disease progresses and may increase the incorporation of palliative care content into care plans (Livingston et al., 2020). The overarching aim of palliative care services is to reduce bother-
some symptoms, distress, and hospitalization burden while increasing the comfort of persons living with dementia and their caregivers (Butler et al., 2020).
Participants in the 2017 National Research Summit on Care, Services, and Supports for Persons with Dementia and Their Caregivers emphasized the need for future such activities to focus specifically on care and services for late-stage dementia and end of life (Gitlin and Maslow, 2018). Included in one of the recommendations resulting from the 2017 summit was the need to identify effective approaches for helping persons living with dementia participate in their health care decisions, including person-centered advance care planning and end-of-life decisions. This recommendation is consistent with the tenets of supported decision making, which focuses on enabling people to make decisions about their own life and to be involved in decisions that affect their care (Donnelly, 2019).
Despite the importance ascribed to these issues, the AHRQ systematic review found little in the way of high-quality evidence to guide effective late-stage dementia care practices and advance supported decision making. The AHRQ review included one cluster RCT for an advance care planning intervention (a video for medical decision makers of persons living with dementia), but, compared with usual care, no benefit was observed for the outcome of burdensome treatments for persons living with dementia, such as hospital transfers or feeding tube insertions, or for the caregiver outcomes related to “do not hospitalize” directives, goals-of-care discussions, and decision makers’ preference for comfort care (Mitchell et al., 2018). Another cluster RCT compared a print decision aid for feeding options with usual care and reported a benefit for persons living with dementia, as measured by the number of persons receiving a specialized dysphagia diet after 3 months (Hanson et al., 2011). Benefits were also reported for caregivers, as measured by a reduction in decisional conflict compared with the control group and an increase in the frequency of feeding discussions held with medical care providers. No studies on palliative care were included in the AHRQ analytic set, primarily because of high risk of bias.
Respite care interventions provide a means for care partners/caregivers to have temporary breaks from caregiving that may range from a few hours in a day to one or more full days (Butler et al., 2020). These interventions may include services through which in-home care is provided for persons living with dementia, adult day programs, and institutional respite services.
The AHRQ systematic review process identified three unique studies on respite care interventions (Lawton et al., 1989; Vandepitte et al., 2019; Zarit et al., 1998), but none were included in the analytic set because of high risk of bias. The paucity of evidence on respite care captured in the
systematic review may have resulted in part from the challenges of evaluating such services through traditional RCTs (Zarit et al., 2017), and may be addressed through the application of methodologically robust nonrandomized study designs. Although the AHRQ systematic review found little evidence to support its effectiveness, respite care has been identified by caregivers as important for their well-being (Jennings et al., 2017).9 At the committee’s public meetings, persons living with dementia emphasized the need to address access issues related to personal expense for respite care (in-home and adult day programs),10,11 suggesting the potential for policy interventions to improve the real-world effectiveness of respite care.
Social support interventions, which may be delivered through in-person meetings, over the phone, or via web-based platforms (e.g., chat groups), are designed to provide care partners/caregivers with social interaction in addition to information and resources (Butler et al., 2020). Social support interventions, including peer support groups, were identified by caregivers as beneficial in their personal experience and deserving of priority attention in future research.12,13 At the committee’s April 2020 public workshop, Douglas Pace of the Alzheimer’s Association discussed the results from 3,000 listening sessions with persons living with dementia and care partners/caregivers from across the organization’s chapter network around the country, and noted that social support groups and the education they can provide were consistently identified as very important.14 However, the AHRQ systematic review found little evidence to support conclusions on the effectiveness of these interventions. The analytic set included only two studies of these interventions: one RCT compared in-person, peer-led mutual support groups for caregivers with usual care (Wang et al., 2012), and another evaluated an automated phone support system for caregivers (Mahoney et al., 2001, 2003). Caregiver outcomes evaluated in the two studies differed, precluding the aggregation of results across studies. Beneficial effects of the in-person support groups were reported for caregiver distress and quality of life (Wang et al., 2012).
9 Presented by Janet Michel at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
10 Presented by Janet Michel at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
11 Presented by Brian Van Buren at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
12 Presented by Brian Van Buren at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
13 Presented by Maria Martinez Israelite at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
14 Presented by Douglas Pace of the Alzheimer’s Association at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
Training and support for direct care workers
The AHRQ systematic review evaluated interventions targeting the well-being of direct care workers separately from those targeting the well-being of unpaid care partners and caregivers (Butler et al., 2020). Interventions for direct care workers were focused on reducing stress and burnout, for example, through peer support and training in stress management and relaxation techniques. The results from the systematic review indicate that the evidence base for such interventions is very preliminary, with only three small pilot studies being captured in the literature search (Barbosa et al., 2015; Davison et al., 2007; Visser et al., 2008), none of which was included in the analytic set.
CONCLUSION: While there have been important advances in knowledge regarding ways to better provide care, support, and services for persons living with dementia and their care partners and caregivers, significant gaps remain in the evidence base for many interventions evaluated in the AHRQ systematic review, including interventions identified by persons living with dementia, care partners, and caregivers as important to their health and well-being. To address these gaps, future research investments will need to ensure that studies are appropriately designed and conducted with methodological rigor, and progress beyond pilot and efficacy studies to include the evaluation of interventions in real-world settings.
CONCLUSION: Direct care workers often play important roles in meeting the needs of persons living with dementia throughout the different stages of the disease. Despite the potential for such care providers to experience work-related stress and burnout, however, the vast majority of interventions targeting caregiver well-being were aimed at unpaid care partners and caregivers. The benefits of training and support for direct care workers are understudied and represent a notable gap that warrants emphasis in future research.
Gaps Related to Heterogeneity and Complexity
For some interventions included in the AHRQ systematic review, the challenge related to a finding of insufficient evidence was not a lack of quality studies as much as the difficulty of synthesizing the evidence and drawing conclusions about effectiveness given the substantial heterogeneity in the sampled populations, intervention implementation (e.g., components, settings, other contextual factors), and measurement and reporting of outcomes. As a result of these challenges, discussed further in the paragraphs that follow, mixed results across studies are difficult to interpret. Consequently, despite the availability of evidence from multiple RCTs, little is
known regarding for whom these interventions work and how they need to be implemented.
As noted in the AHRQ systematic review, the diverse etiology and progressive nature of dementia add complexity to the evaluation of care interventions. Given that the need for care, services, and supports and the settings in which they are delivered change over the course of the disease, interventions may be more or less effective for individuals at different stages of disease progression. Yet, participants in a given study may have dementia ranging from mild to severe, introducing the potential for variable effects that may mask the benefit of the intervention in a subgroup. Additional variability is introduced when synthesizing evidence across studies that enrolled participants in different stages of dementia. Moreover, results from some studies that conducted subgroup analyses suggest that the effects of interventions can differ across subpopulations with different types of dementia (Alzheimer’s versus other dementias) (Toots et al., 2016).
It is also important to note the potential for variation in the experiences and circumstances of care partners and caregivers who may be targeted by interventions or involved in the delivery of interventions to those for whom they are providing care and/or support. For example, the needs and capabilities of first-time care partners or caregivers may differ from those of individuals who have previously served in these roles. In some cases, multiple care partners or caregivers may be sharing the responsibility for providing care for a person living with dementia, and in other cases, people with mild cognitive impairment or early-stage dementia may themselves be care partners or caregivers for someone experiencing a more advanced stage of the disease. Such variation in care partner and caregiver circumstances may have implications for the perceived value of interventions, the fidelity of implementation, and intervention effectiveness that need to be better understood.
Beyond the challenges related to heterogeneity in the circumstances (types and stages of dementia, caregiving situations) of those individuals enrolled in studies, the dearth of data for specific demographic subpopulations further hinders drawing conclusions about the real-world effectiveness of dementia care interventions. Important subpopulations to consider in the context of research on dementia care interventions include major racial/ethnic groups, LGBTQ populations, people with significant comorbidities (e.g., hearing loss or vision impairment), people of low socioeconomic status, and those who reside in low-resource areas (e.g., rural and tribal populations). In addition to variation in the efficacy of interventions across these groups, access to interventions and feasibility of implementation may vary as well. Thus, even for those interventions showing promise in clinical trials, the applicability of the evidence to the full range of populations experiencing dementia is unclear.
Given these challenges related to population heterogeneity, improved reporting of participant demographics and subgroup analyses may help
ensure that future research is better able to support conclusions on which interventions are effective for whom and under what circumstances. However, there are also opportunities for future research to help elucidate how interventions may be tailored to better meet the needs of specific subpopulations with respect not only to cultural appropriateness but also to such factors as degree of dementia-related impairment. For example, care interventions designed primarily for an older population may need to be adapted for individuals with early-onset dementia. Similarly, interventions designed for implementation in the context of long-term care facilities may need to be adapted for community-dwelling individuals with early-stage dementia. Recruitment of the diverse populations an intervention purports to serve, along with improved reporting and subgroup analyses, can further strengthen the evidence used to make decisions regarding implementation.
For complex interventions such those evaluated in the AHRQ systematic review, grouping interventions for the purpose of synthesizing the evidence often requires some trade-off between ensuring a body of evidence of adequate size and introducing heterogeneity that can make interpretation of the findings difficult. The AHRQ review acknowledges this challenge, especially in the absence of a field-accepted taxonomy for classifying dementia care interventions (Butler et al., 2020). Given the scope of the review, the categories of interventions were necessarily broad, with, in some cases, substantial variability in the components of interventions and/or how they were implemented. Such variability can contribute to a lack of consistency in results across studies of interventions within a category, making it difficult to draw conclusions about intervention effectiveness. Added to these challenges was insufficient detail in studies’ descriptions of interventions, which hinders determining the comparability of studies, assessing fidelity, and interpreting differences in findings across studies.
In contrast to studies of interventions aimed at the prevention or treatment of dementia, which generally focus on a limited number of clinical endpoints, care intervention studies cover a wide range of outcomes for both persons living with dementia and care partners/caregivers enrolled in the studies. These outcomes include (but are not limited to) depression; agitation; anxiety; daily function; neuropsychological symptoms; and even highly specific outcomes, such as hyperphagic (excessive eating) behavior. Adding to this complexity is the use of different scales to measure the same outcome, potentially resulting in discrepant results even within studies. For some interventions, moreover, process outcomes (such as discussions with clinical providers) may be reported in addition to outcomes related to health and well-being. This variability in reported outcomes hampers assessment of the consistency of findings across studies of similar interventions and precluded quantitative pooling of data and meta-analysis for most interventions included in the AHRQ review. Harmonization of outcomes may help
address these challenges, but efforts to define core sets of outcomes will need to include a focus on those used to evaluate benefits and harms, with particular emphasis on endpoints important to persons living with dementia, care partners, and caregivers. As noted in the AHRQ systematic review, few studies have reported on harms and other unintended consequences or on some outcomes (personhood, identity, well-being) valued by persons living with dementia, care partners, and caregivers (Butler et al., 2020). Evaluation of such outcomes is difficult using quantitative metrics and may necessitate further investment in qualitative and mixed-method research designs.
The sections below highlight types of interventions that illustrate the challenges to evaluating dementia care interventions posed by the heterogeneity issues reviewed above, as well as the value of future research and evidence synthesis approaches aimed at elucidating the populations/settings for which interventions are effective and how they should be implemented to achieve outcomes that are important to persons living with dementia, care partners, and caregivers. The following categories of interventions are discussed:
- psychosocial interventions, and
- cognitive interventions.
Exercise and physical activity are important contributors to healthy aging, with benefits related to both physical and cognitive function (Livingston et al., 2020; NASEM, 2017). In discussions with the committee, persons living with dementia and care partners/caregivers indicated that exercise is an important part of staying active15 and that physical activity helps with mental health and coping.16
Exercise was the intervention category with the second largest body of studies meeting the criteria for inclusion in the AHRQ systematic review; the analytic set included 10 studies, 3 of which were cluster RCTs (Butler et al., 2020). However, the implementation of exercise interventions and the outcomes evaluated varied across studies. Exercise interventions included aerobic, strength, or balance training, alone or in combination and with variable levels of intensity. Settings and formats (i.e., individual versus group) also varied. Most study participants had mild to moderate dementia (a mix of Alzheimer’s, vascular, and mixed dementias).
15 Presented by Janet Michel at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
16 Presented by Maria Martinez Israelite at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
Despite this heterogeneity, benefit was observed for several outcomes for persons living with dementia in more than one RCT, providing a signal of effectiveness. Of eight studies that evaluated the effect of exercise on daily function (Bossers et al., 2016; Chen et al., 2019; Hoffmann et al., 2016; Huang et al., 2019; Lamb et al., 2018; Pitkälä et al., 2013; Telenius et al., 2015; Toots et al., 2016), two reported beneficial effects of group exercise (Bossers et al., 2016; Pitkälä et al., 2013). Another study observed a positive effect of a hand exercise program on autonomous eating (Chen et al., 2019). Two studies reported on balance outcomes, both finding a benefit at one of two measured time points (Telenius et al., 2015; Toots et al., 2016). The effect of a high-intensity exercise on balance was greater for adults with non-Alzheimer’s dementia than for those with Alzheimer’s dementia (Toots et al., 2016). Of four exercise studies that measured the intervention effect on neuropsychiatric symptoms (Hoffmann et al., 2016; Huang et al., 2019; Lamb et al., 2018; Telenius et al., 2015), benefit was associated with group exercise in one study (Hoffmann et al., 2016) and with Tai Chi in another (Huang et al., 2019). In some cases, exercise was included as part of a multicomponent intervention. For example, group exercise is a core component of the Reducing Disabilities in Alzheimer’s Disease (RDAD) multicomponent intervention, which has been shown to benefit persons living with dementia in the areas of health, depression, and function (days of restricted activity) (Teri et al., 2003). Given the broad benefits of physical activity and the lack of any clear link to serious adverse events (Butler et al., 2020), future research may be most impactful if focused on elucidating the optimal type, duration, format, and intensity of exercise interventions and the expected benefits at different stages of disease progression.
Music interventions are generally intended to be calming or pleasurable activities and may target cognitive and sensory stimulation. At the committee’s public workshop, one caregiver described how enjoyment of music was an important aspect of quality of life for her and her husband, for whom she is providing care.17 Another speaker noted that, despite the mixed research results, music has been helpful in practice for some people experiencing such feelings as loneliness and helplessness.18
The analytic set for the AHRQ systematic review included five RCTs of music interventions, but implementation and reported outcomes varied across studies, so that for most outcomes, assessment of strength of evi-
17 Presented by Janet Michel at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
18 Presented by Douglas Pace of the Alzheimer’s Association at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
dence was based on a single RCT (Butler et al., 2020). The music interventions studied, which were often delivered in group settings, ranged from playing musical instruments to listening to recorded songs or singing and in some cases also involved body movement. The effect of music interventions on agitation in persons living with dementia was evaluated in three RCTs (Cheung et al., 2018; Lin et al., 2011; Sung et al., 2012), one of which reported benefit (Lin et al., 2011). Benefits were also observed for quality of life (Särkämö et al., 2013), mood (Särkämö et al., 2013), anxiety (Sung et al., 2012), and depression (Chu et al., 2014) in the single studies in which those outcomes were measured. In addition to the observed benefits for persons living with dementia, one study reported an improvement in family caregivers’ burden associated with participation in a singing group (Särkämö et al., 2013). While these results provide some signal of benefit, future research may increase understanding of the comparative effectiveness of different types of music interventions for different populations and, given the variation in individual preferences, the importance of tailoring such interventions.
The psychosocial intervention category within the AHRQ systematic review encompasses a diverse set of psychotherapeutic and psychoeducational interventions (Butler et al., 2020). The review classified psychosocial interventions based on whether they were targeted at addressing behavioral and psychological symptoms of dementia in persons living with dementia (e.g., through cognitive-behavioral training) or at improving the overall well-being of persons living with dementia or their care partners and caregivers (e.g., through skills training or counseling). Psychosocial interventions are generally delivered by highly trained health or social service professionals in one-on-one or group settings. In some cases, sessions involve dyads of persons living with dementia and care partners/caregivers. Although psychotherapeutic and psychoeducational interventions have distinct definitions, in practice they often share intervention components. As noted in the AHRQ systematic review, this makes it difficult to define subgroups of psychosocial therapies more narrowly for analysis at a more granular level. The resulting heterogeneity in interventions within this category poses a challenge for aggregating data and drawing conclusions about more specific interventions (or intervention components) that are effective.
The category of psychosocial therapies for care partner/caregiver well-being represented the largest body of included studies among the interventions evaluated in the AHRQ systematic review (29 studies) (Butler et al., 2020). Notably, however, the AHRQ analytic set included no studies of psychosocial therapies directly targeting persons living with dementia (for outcomes related to either behavioral and psychological symptoms of dementia [BPSD] or well-being). Benefits of psychosocial therapies for care
partner/caregiver well-being were reported for multiple outcomes for both the caregivers to whom the interventions were targeted and the persons receiving care. While the results were mixed for all outcomes, this is not surprising given the variability in the interventions included in the category. For caregivers, psychosocial therapies were associated with benefits related to depression (in 8 of 12 studies measuring short-term outcomes and in 2 of 8 studies measuring long-term outcomes), burden (in 5 of 13 studies), bother/distress (in 4 of 9 studies), quality of life (in 6 of 14 studies), and caregiving confidence (in 3 of 6 studies). Approximately half of the studies included in the analytic set also reported on outcomes for persons living with dementia who were receiving care from those caregivers to whom the intervention was targeted. Benefits for persons living with dementia were observed for measures related to function (in 2 of 3 studies measuring short-term outcomes and in 1 of 5 studies measuring long-term outcomes), depression (in 2 of 5 studies), neuropsychiatric symptoms (in 3 of 6 studies measuring short-term outcomes but none of the 3 studies measuring long-term outcomes), quality of life (in 2 of 8 studies), institutionalization (in 1 of 4 studies), and health care usage (in 1 of 5 studies). It should also be noted that the majority of components in REACH II and its adaptations, described earlier in this chapter, map to the psychosocial intervention category.
During discussions with the committee, psychosocial therapies (e.g., counseling, education on both the disease and skills for caregivers) were identified as valued and beneficial interventions by both persons living with dementia19,20,21 and care partners/caregivers.22,23 In addition, the myriad psychotherapeutic and educational/skills-building interventions in the Best Practice Caregiving database (Benjamin Rose Institute on Aging and FCA, 2020) suggest that stakeholders operating in community-level practice-based settings see value in implementing these types of interventions.24
19 Presented by Brian Van Buren at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
20 Presented by Cynthia Huling Hummel at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
21 Presented by John Richard (JR) Pagan at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
22 Presented by Janet Michel at the Care Interventions for Individuals with Dementia and Their Caregivers workshop on April 15, 2020.
23 Presented by Maria Martinez Israelite at the Care Interventions for Individuals with Dementia and Their Caregivers public meeting on May 29, 2020.
24 Of note, whether educational and skills-building interventions included in the Best Practice Caregiving database would have been classified as psychosocial or multicomponent interventions using the AHRQ systematic review taxonomy is often not clear because of the limitations of the intervention descriptions and the lack of consensus on taxonomies for dementia care interventions.
Monitoring and reporting of experiences with implementing different psychosocial interventions in these real-world settings, along with pragmatic research studies, may provide opportunities to elucidate the critical components of psychosocial interventions for persons living with dementia and care partners/caregivers. The evaluation of such interventions, however, would benefit from a clearer taxonomy.
The AHRQ systematic review evaluated several types of interventions targeting the cognitive function of persons living with dementia and/or their ability to carry out daily activities (Butler et al., 2020). These interventions included cognitive stimulation therapy, cognitive training, cognitive rehabilitation, and reminiscence therapy. While individual RCTs of such cognitive interventions suggest the potential for benefits ranging from improved quality of life (Orrell et al., 2014; Spector et al., 2003), to reduced depression (Li et al., 2019; Wang, 2007), to improvements in eating behavior (Hsu et al., 2017; Kao et al., 2016) in persons living with dementia, assessment of effectiveness was complicated by the lack of clear differentiation among the different intervention types. For example, cognitive training interventions aim to improve cognitive function (e.g., memory, reasoning, speed of processing) through repetitive or progressive drill-like exercises, while cognitive rehabilitation interventions similarly focus on cognitive abilities (such as memory and executive function) and may target the recovery or restoration of daily functions. Even within individual studies, terms for these two types of cognitive interventions are often used interchangeably. More consistent terminology and consensus on intervention taxonomies would aid in the comparison of interventions across studies and synthesis of the evidence to support conclusions on intervention effectiveness.
CONCLUSION: For some intervention categories, evidence is insufficient to support conclusions on readiness for broad dissemination and implementation, despite a multitude of RCTs providing some signal of benefit. As a result of heterogeneity in study populations, intervention implementation, and measured outcomes, little is known regarding which interventions are likely to be effective for persons living with dementia, care partners, and caregivers experiencing different stages of disease progression and how they should optimally be implemented.
CONCLUSION: Evidence is lacking with respect to the effectiveness of dementia care interventions in diverse populations, such as specific racial/ethnic groups, LGBTQ populations, people with significant comorbidities or of low socioeconomic status, and those
from low-resource areas (e.g., rural and tribal populations). Consequently, the applicability of the existing evidence base to the full range of persons living with dementia and their care partners and caregivers is not supported, even for those interventions showing promise in clinical trials.
CONCLUSION: The evidence needed to inform decisions about policy and the implementation of specific interventions broadly at the organizational and community levels—including informing the relative prioritization of many interventions that could be helpful but will require resources—is limited. Some challenges in the existing evidence base are due to the inherent complexity of the area of study, including the multifaceted nature of dementia, its heterogeneity across populations and settings, and its progression across different stages. However, the AHRQ systematic review also brings to the forefront some addressable limitations in the existing research base, such as a lack of diversity in study populations, underpowered and limited-duration studies, heterogeneity of outcome measures that precludes aggregation of results, lack of reporting on contextual factors that facilitate or impede intervention effectiveness, and research that is divorced from practical implementation needs.
Alonso-Coello, P., H. J. Schünemann, J. Moberg, R. Brignardello-Petersen, E. A. Akl, M. Davoli, S. Treweek, R. A. Mustafa, G. Rada, S. Rosenbaum, A. Morelli, G. H. Guyatt, and A. D. Oxman. 2016. GRADE Evidence to Decision (EtD) frameworks: A systematic and transparent approach to making well informed healthcare choices. 1: Introduction. BMJ 353:i2016.
Amjad, H., S. K. Wong, D. L. Roth, J. Huang, A. Willink, B. S. Black, D. Johnston, P. V. Rabins, L. N. Gitlin, C. G. Lyketsos, and Q. M. Samus. 2018. Health services utilization in older adults with dementia receiving care coordination: The MIND at home trial. Health Services Research 53(1):556–579.
Archer, J., P. Bower, S. Gilbody, K. Lovell, D. Richards, L. Gask, C. Dickens, and P. Coventry. 2012. Collaborative care for depression and anxiety problems. Cochrane Database of Systematic Reviews (Online) 10.
Barbosa, A., M. Nolan, L. Sousa, A. Marques, and D. Figueiredo. 2015. Effects of a psychoeducational intervention for direct care workers caring for people with dementia: Results from a 6-month follow-up study. American Journal of Alzheimer’s Disease & Other Dementias 31(2):144–155.
Bass, D. M., K. S. Judge, A. L. Snow, N. L. Wilson, R. Morgan, W. J. Looman, C. A. McCarthy, K. Maslow, J. A. Moye, R. Randazzo, M. Garcia-Maldonado, R. Elbein, G. Odenheimer, and M. E. Kunik. 2013. Caregiver outcomes of partners in dementia care: Effect of a care coordination program for veterans with dementia and their family members and friends. Journal of the American Geriatrics Society 61(8):1377–1386.
Bass, D. M., K. S. Judge, A. Snow, N. L. Wilson, R. O. Morgan, K. Maslow, R. Randazzo, J. A. Moye, G. L. Odenheimer, E. Archambault, R. Elbein, P. Pirraglia, T. A. Teasdale, C. A. McCarthy, W. J. Looman, and M. E. Kunik. 2014. A controlled trial of partners in dementia care: Veteran outcomes after six and twelve months. Alzheimer’s Research & Therapy 6(1):9.
Bass, D. M., K. S. Judge, K. Maslow, N. L. Wilson, R. O. Morgan, C. A. McCarthy, W. J. Looman, A. L. Snow, and M. E. Kunik. 2015. Impact of the care coordination program “partners in dementia care” on veterans’ hospital admissions and emergency department visits. Alzheimer’s & Dementia: Translational Research & Clinical Interventions 1(1):13–22.
Bass, D. M., T. Hornick, M. Kunik, K. S. Judge, B. Primetica, K. Kearney, J. Rentsch, C. McCarthy, and J. Grim. 2019. Findings from a real-world translation study of the evidence-based “partners in dementia care.” Innovation in Aging 3(3):igz031.
Belle, S. H., L. Burgio, R. Burns, D. Coon, S. J. Czaja, D. Gallagher-Thompson, L. N. Gitlin, J. Klinger, K. M. Koepke, C. C. Lee, J. Martindale-Adams, L. Nichols, R. Schulz, S. Stahl, A. Stevens, L. Winter, and S. Zhang. 2006. Enhancing the quality of life of dementia caregivers from different ethnic or racial groups: A randomized, controlled trial. Annals of Internal Medicine 145(10):727–738.
Benjamin Rose Institute on Aging and FCA (Family Caregiver Alliance). 2020. Best practice caregiving. https://bpc.caregiver.org/#searchPrograms (accessed August 12, 2020).
Berwig, M., S. Heinrich, J. Spahlholz, N. Hallensleben, E. Brähler, and H.-J. Gertz. 2017. Individualized support for informal caregivers of people with dementia—Effectiveness of the German adaptation of REACH II. BMC Geriatrics 17(1):286.
Bossers, W. J., L. H. van der Woude, F. Boersma, T. Hortobágyi, E. J. Scherder, and M. J. van Heuvelen. 2016. Comparison of effect of two exercise programs on activities of daily living in individuals with dementia: A 9-week randomized, controlled trial. Journal of the American Geriatrics Society 64(6):1258–1266.
Boustani, M., C. A. Alder, C. A. Solid, and D. Reuben. 2019. An alternative payment model to support widespread use of collaborative dementia care models. Health Affairs 38(1):54–59.
Briss, P. A., S. Zaza, M. Pappaioanou, J. Fielding, L. Wright-De Agüero, B. I. Truman, D. P. Hopkins, P. D. Mullen, R. S. Thompson, S. H. Woolf, V. G. Carande-Kulis, L. Anderson, A. R. Hinman, D. V. McQueen, S. M. Teutsch, and J. R. Harris. 2000. Developing an evidence-based guide to community preventive services—Methods. The task force on community preventive services. American Journal of Preventive Medicine 18(1 Suppl):35–43.
Brown, A. F., S. D. Vassar, K. I. Connor, and B. G. Vickrey. 2013. Collaborative care management reduces disparities in dementia care quality for caregivers with less education. Journal of the American Geriatrics Society 61(2):243–251.
Burgio, L. D., I. B. Collins, B. Schmid, T. Wharton, D. McCallum, and J. Decoster. 2009. Translating the REACH caregiver intervention for use by area agency on aging personnel: The REACH OUT program. The Gerontologist 49(1):103–116.
Butler, M., J. E. Gaugler, K. M. C. Talley, H. I. Abdi, P. J. Desai, S. Duval, M. L. Forte, V. A. Nelson, W. Ng, J. M. Ouellette, E. Ratner, J. Saha, T. Shippee, B. L. Wagner, T. J. Wilt, and L. Yeshi. 2020. Care interventions for people living with dementia and their caregivers. Comparative Effectiveness Review No. 231. Rockville, MD: Agency for Healthcare Research and Quality. doi: 10.23970/AHRQEPCCER231.
Callahan, C. M., M. A. Boustani, F. W. Unverzagt, M. G. Austrom, T. M. Damush, A. J. Perkins, B. A. Fultz, S. L. Hui, S. R. Counsell, and H. C. Hendrie. 2006. Effectiveness of collaborative care for older adults with Alzheimer disease in primary care: A randomized controlled trial. Journal of the American Medical Association 295(18):2148–2157.
Callahan, C. M., M. A. Boustani, M. Weiner, R. A. Beck, L. R. Livin, J. J. Kellams, D. R. Willis, and H. C. Hendrie. 2011. Implementing dementia care models in primary care settings: The aging brain care medical home. Aging & Mental Health 15(1):5–12.
Chen, L. L., H. Li, X. H. Chen, S. Jin, Q. H. Chen, M. R. Chen, and N. Li. 2019. Effects of hand exercise on eating action in patients with Alzheimer’s disease. American Journal of Alzheimer’s Disease and Other Dementias 34(1):57–62.
Cheung, K. S., B. H. Lau, P. W. Wong, A. Y. Leung, V. W. Lou, G. M. Chan, and R. Schulz. 2014. Multicomponent intervention on enhancing dementia caregiver well-being and reducing behavioral problems among Hong Kong Chinese: A translational study based on REACH II. International Journal of Geriatric Psychiatry 30(5):460–469.
Cheung, D. S. K., C. K. Y. Lai, F. K. Y. Wong, and M. C. P. Leung. 2018. The effects of the music-with-movement intervention on the cognitive functions of people with moderate dementia: A randomized controlled trial. Aging & Mental Health 22(3):306–315.
Chien, W. T., and Y. M. Lee. 2008. A disease management program for families of persons in Hong Kong with dementia. Psychiatric Services 59(4):433–436.
Cho, J., S. Luk-Jones, D. R. Smith, and A. B. Stevens. 2019. Evaluation of REACH-TX: A community-based approach to the REACH II intervention. Innovation in Aging 3(3):igz022.
Chodosh, J., B. A. Colaiaco, K. I. Connor, D. W. Cope, H. Liu, D. A. Ganz, M. J. Richman, D. L. Cherry, J. M. Blank, P. Carbone Rdel, S. M. Wolf, and B. G. Vickrey. 2015. Dementia care management in an underserved community: The comparative effectiveness of two different approaches. Journal of Aging Health 27(5):864–893.
Chu, H., C.-Y. Yang, Y. Lin, K.-L. Ou, T.-Y. Lee, A. P. O’Brien, and K.-R. Chou. 2014. The impact of group music therapy on depression and cognition in elderly persons with dementia: A randomized controlled study. Biological Research for Nursing 16(2):209–217.
Czaja, S. J., D. Loewenstein, R. Schulz, S. N. Nair, and D. Perdomo. 2013. A videophone psychosocial intervention for dementia caregivers. The American Journal of Geriatric Psychiatry 21(11):1071–1081.
Czaja, S. J., C. C. Lee, D. Perdomo, D. Loewenstein, M. Bravo, J. H. Moxley, and R. Schulz. 2018. Community REACH: An implementation of an evidence-based caregiver program. Gerontologist 58(2):e130–e137.
Davison, T. E., M. P. McCabe, S. Visser, C. Hudgson, G. Buchanan, and K. George. 2007. Controlled trial of dementia training with a peer support group for aged care staff. International Journal of Geriatric Psychiatry 22(9):868–873.
Dham, P., S. Colman, K. Saperson, C. McAiney, L. Lourenco, N. Kates, and T. K. Rajji. 2017. Collaborative care for psychiatric disorders in older adults: A systematic review. Canadian Journal of Psychiatry 62(11):761–771.
Donnelly, M. 2019. Deciding in dementia: The possibilities and limits of supported decision-making. International Journal of Law and Psychiatry 66:101466.
Duru, O. K., S. L. Ettner, S. D. Vassar, J. Chodosh, and B. G. Vickrey. 2009. Cost evaluation of a coordinated care management intervention for dementia. American Journal of Managed Care 15(8):521–528.
Eloniemi-Sulkava, U., M. Saarenheimo, M.-L. Laakkonen, M. Pietilä, N. Savikko, H. Kautiainen, R. S. Tilvis, and K. H. Pitkälä. 2009. Family care as collaboration: Effectiveness of a multicomponent support program for elderly couples with dementia. Randomized controlled intervention study. Journal of the American Geriatrics Society 57(12):2200–2208.
Gilman, B., D. Whicher, R. Brown, N. McCall, S. Dale, L. Felland, R. Schmitz, J. Schurrer, L. Vollmer Forrow, M. Flick, J. Dezron, and R. Lakhani. 2020. Evaluation of the health care innovation awards, round 2: Final report. Cambridge, MA: Mathematica.
Gitlin, L. N., and K. Maslow. 2018. National research summit on care, services, and supports for persons with dementia and their caregivers: Report to the National Advisory Council on Alzheimer’s Research, Care, and Services. https://aspe.hhs.gov/system/files/pdf/259156/FinalReport.pdf (accessed September 1, 2020).
Hanson, L. C., T. S. Carey, A. J. Caprio, T. J. Lee, M. Ersek, J. Garrett, A. Jackman, R. Gilliam, K. Wessell, and S. L. Mitchell. 2011. Improving decision-making for feeding options in advanced dementia: A randomized, controlled trial. Journal of the American Geriatrics Society 59(11):2009–2016.
Heintz, H., P. Monette, G. Epstein-Lubow, L. Smith, S. Rowlett, and B. P. Forester. 2020. Emerging collaborative care models for dementia care in the primary care setting: A narrative review. American Journal of Geriatric Psychiatry 28(3):320–330.
Hoffmann, K., N. A. Sobol, K. S. Frederiksen, N. Beyer, A. Vogel, K. Vestergaard, H. Brændgaard, H. Gottrup, A. Lolk, L. Wermuth, S. Jacobsen, L. P. Laugesen, R. G. Gergelyffy, P. Høgh, E. Bjerregaard, B. B. Andersen, V. Siersma, P. Johannsen, C. W. Cotman, G. Waldemar, and S. G. Hasselbalch. 2016. Moderate-to-high intensity physical exercise in patients with Alzheimer’s disease: A randomized controlled trial. Journal of Alzheimer’s Disease 50(2):443–453.
Hsu, C.-N., L.-C. Lin, and S.-C. Wu. 2017. The effects of spaced retrieval training in improving hyperphagia of people living with dementia in residential settings. Journal of Clinical Nursing 26(19–20):3224–3231.
Huang, N., W. Li, X. Rong, M. Champ, L. Wei, M. Li, H. Mu, Y. Hu, Z. Ma, and J. Lyu. 2019. Effects of a modified Tai Chi program on older people with mild dementia: A randomized controlled trial. Journal of Alzheimer’s Disease 72:947–956.
Jennings, L. A., A. Palimaru, M. G. Corona, X. E. Cagigas, K. D. Ramirez, T. Zhao, R. D. Hays, N. S. Wenger, and D. B. Reuben. 2017. Patient and caregiver goals for dementia care. Quality of Life Research 26(3):685–693.
Jennings, L. A., S. Hollands, E. Keeler, N. S. Wenger, and D. B. Reuben. 2020. The effects of dementia care co-management on acute care, hospice, and long-term care utilization. Journal of the American Geriatrics Society 68(11):2500–2507.
Judge, K. S., D. M. Bass, A. L. Snow, N. L. Wilson, R. Morgan, W. J. Looman, C. McCarthy, and M. E. Kunik. 2011. Partners in dementia care: A care coordination intervention for individuals with dementia and their family caregivers. Gerontologist 51(2):261–272.
Kao, C.-C., L.-C. Lin, S.-C. Wu, K.-N. Lin, and C.-K. Liu. 2016. Effectiveness of different memory training programs on improving hyperphagic behaviors of residents with dementia: A longitudinal single-blind study. Clinical Interventions in Aging 11:707–720.
LaMantia, M. A., C. A. Alder, C. M. Callahan, S. Gao, D. D. French, M. G. Austrom, K. Boustany, L. Livin, B. Bynagari, and M. A. Boustani. 2015. The aging brain care medical home: Preliminary data. Journal of the American Geriatrics Society 63(6):1209–1213.
Lamb, S. E., B. Sheehan, N. Atherton, V. Nichols, H. Collins, D. Mistry, S. Dosanjh, A. M. Slowther, I. Khan, S. Petrou, and R. Lall. 2018. Dementia and physical activity (DAPA) trial of moderate to high intensity exercise training for people with dementia: Randomised controlled trial. BMJ 361:k1675.
Lawton, M. P., E. M. Brody, A. Saperstein, and M. Grimes. 1989. Respite services for caregivers: Research findings for service planning. Home Health Care Services Quarterly 10(1–2):5–32.
Lee, C. C., S. J. Czaja, and R. Schulz. 2010. The moderating influence of demographic characteristics, social support, and religious coping on the effectiveness of a multicomponent psychosocial caregiver intervention in three racial ethnic groups. The Journal of Gerontology, Series B, Psychological Sciences and Social Sciences 65b(2):185–194.
Li, M., J.-h. Lyu, Y. Zhang, M.-l. Gao, R. Li, P.-x. Mao, W.-j. Li, and X. Ma. 2019. Efficacy of group reminiscence therapy on cognition, depression, neuropsychiatric symptoms, and activities of daily living for patients with Alzheimer disease. Journal of Geriatric Psychiatry and Neurology 33(5):272–281.
Lin, Y., H. Chu, C.-Y. Yang, C.-H. Chen, S.-G. Chen, H.-J. Chang, C.-J. Hsieh, and K.-R. Chou. 2011. Effectiveness of group music intervention against agitated behavior in elderly persons with dementia. International Journal of Geriatric Psychiatry 26(7):670–678.
Livingston, G., J. Huntley, A. Sommerlad, D. Ames, C. Ballard, S. Banerjee, C. Brayne, A. Burns, J. Cohen-Mansfield, C. Cooper, S. G. Costafreda, A. Dias, N. Fox, L. N. Gitlin, R. Howard, H. C. Kales, M. Kivimäki, E. B. Larson, A. Ogunniyi, V. Orgeta, K. Ritchie, K. Rockwood, E. L. Sampson, Q. Samus, L. S. Schneider, G. Selbæk, L. Teri, and N. Mukadam. 2020. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. The Lancet 396(10248):413–446.
Luchsinger, J. A., L. Burgio, M. Mittelman, I. Dunner, J. A. Levine, C. Hoyos, D. Tipiani, Y. Henriquez, J. Kong, S. Silver, M. Ramirez, and J. A. Teresi. 2018. Comparative effectiveness of 2 interventions for Hispanic caregivers of persons with dementia. Journal of the American Geriatrics Society 66(9):1708–1715.
Lykens, K., N. Moayad, S. Biswas, C. Reyes-Ortiz, and K. P. Singh. 2014. Impact of a community based implementation of REACH II program for caregivers of Alzheimer’s patients. PLoS One 9(2):e89290.
Mahoney, D. M., B. Tarlow, R. N. Jones, S. Tennstedt, and L. Kasten. 2001. Factors affecting the use of a telephone-based intervention for caregivers of people with Alzheimer’s disease. Journal of Telemedicine and Telecare 7(3):139–148.
Mahoney, D. F., B. J. Tarlow, and R. N. Jones. 2003. Effects of an automated telephone support system on caregiver burden and anxiety: Findings from the REACH for TLC intervention study. Gerontologist 43(4):556–567.
Mitchell, S. L., M. L. Shaffer, S. Cohen, L. C. Hanson, D. Habtemariam, and A. E. Volandes. 2018. An advance care planning video decision support tool for nursing home residents with advanced dementia: A cluster randomized clinical trial. JAMA Internal Medicine 178(7):961–969.
Moberg, J., A. D. Oxman, S. Rosenbaum, H. J. Schünemann, G. Guyatt, S. Flottorp, C. Glenton, S. Lewin, A. Morelli, G. Rada, P. Alonso-Coello, E. Akl, M. Gulmezoglu, R. A. Mustafa, J. Singh, E. von Elm, J. Vogel, and J. Watine. 2018. The GRADE Evidence to Decision (EtD) framework for health system and public health decisions. Health Research Policy and Systems 16(1):45.
Morgan, R. O., D. M. Bass, K. S. Judge, C. F. Liu, N. Wilson, A. L. Snow, P. Pirraglia, M. Garcia-Maldonado, P. Raia, N. N. Fouladi, and M. E. Kunik. 2015. A break-even analysis for dementia care collaboration: Partners in dementia care. Journal of Genernal Internal Medicine 30(6):804–809.
NASEM (National Academies of Sciences, Engineering, and Medicine). 2017. Preventing cognitive decline and dementia: A way forward. Washington, DC: The National Academies Press.
NASEM. 2020. Evidence-based practice for public health emergency preparedness and response. Washington, DC: The National Academies Press.
Nichols, L. O., C. Chang, A. Lummus, R. Burns, J. Martindale-Adams, M. J. Graney, D. W. Coon, S. Czaja, and Resources for Enhancing Alzheimer’s Caregivers Health II Investigators. 2008. The cost-effectiveness of a behavior intervention with caregivers of patients with Alzheimer’s disease. Journal of the American Geriatrics Society 56(3):413–420.
Nichols, L. O., J. Martindale-Adams, R. Burns, M. J. Graney, and J. Zuber. 2011. Translation of a dementia caregiver support program in a health care system—REACH VA. Archives of Internal Medicine 171(4):353–359.
Nichols, L. O., J. Martindale-Adams, C. W. Zhu, E. K. Kaplan, J. K. Zuber, and T. M. Waters. 2017. Impact of the REACH II and REACH VA dementia caregiver interventions on healthcare costs. Journal of the American Geriatrics Society 65(5):931–936.
NORC at the University of Chicago. 2016. Third annual report: HCIA disease-specific evaluation. Bethesda, MD: NORC at the University of Chicago.
Orrell, M., E. Aguirre, A. Spector, Z. Hoare, R. T. Woods, A. Streater, H. Donovan, J. Hoe, M. Knapp, C. Whitaker, and I. Russell. 2014. Maintenance cognitive stimulation therapy for dementia: Single-blind, multicentre, pragmatic randomised controlled trial. British Journal of Psychiatry 204(6):454–461.
Pew Research Center. 2019a. Internet/broadband fact sheet. https://www.pewresearch.org/internet/fact-sheet/internet-broadband (accessed September 4, 2020).
Pew Research Center. 2019b. Mobile fact sheet. https://www.pewresearch.org/internet/fact-sheet/mobile/#mobile-phone-ownership-over-time (accessed September 3, 2020).
Pitkälä, K. H., M. M. Pöysti, M.-L. Laakkonen, R. S. Tilvis, N. Savikko, H. Kautiainen, and T. E. Strandberg. 2013. Effects of the Finnish Alzheimer disease exercise trial (FINALEX): A randomized controlled trial. JAMA Internal Medicine 173(10):894–901.
Possin, K. L., J. J. Merrilees, S. Dulaney, S. J. Bonasera, W. Chiong, K. Lee, S. M. Hooper, I. E. Allen, T. Braley, A. Bernstein, T. D. Rosa, K. Harrison, H. Begert-Hellings, J. Kornak, J. G. Kahn, G. Naasan, S. Lanata, A. M. Clark, A. Chodos, R. Gearhart, C. Ritchie, and B. L. Miller. 2019. Effect of collaborative dementia care via telephone and Internet on quality of life, caregiver well-being, and health care use: The care ecosystem randomized clinical trial. JAMA Internal Medicine 179(12):1658–1667.
Rosa, T. D., K. L. Possin, A. Bernstein, J. Merrilees, S. Dulaney, J. Matuoka, K. P. Lee, W. Chiong, S. J. Bonasera, K. L. Harrison, and J. G. Kahn. 2019. Variations in costs of a collaborative care model for dementia. Journal of the American Geriatrics Society 67(12):2628–2633.
Samus, Q. M., D. Johnston, B. S. Black, E. Hess, C. Lyman, A. Vavilikolanu, J. Pollutra, J. M. Leoutsakos, L. N. Gitlin, P. V. Rabins, and C. G. Lyketsos. 2014. A multidimensional home-based care coordination intervention for elders with memory disorders: The Maximizing Independence at Home (MIND) pilot randomized trial. American Journal of Geriatric Psychiatry 22(4):398–414.
Särkämö, T., M. Tervaniemi, S. Laitinen, A. Numminen, M. Kurki, J. K. Johnson, and P. Rantanen. 2013. Cognitive, emotional, and social benefits of regular musical activities in early dementia: Randomized controlled study. Gerontologist 54(4):634–650.
Shrestha, S., K. S. Judge, N. L. Wilson, J. A. Moye, A. L. Snow, and M. E. Kunik. 2011. Utilization of legal and financial services of partners in dementia care study. American Journal of Alzheimer’s Disease and Other Dementias 26(2):115–120.
Spector, A., L. Thorgrimsen, B. Woods, L. Royan, S. Davies, M. Butterworth, and M. Orrell. 2003. Efficacy of an evidence-based cognitive stimulation therapy programme for people with dementia: Randomised controlled trial. The British Journal of Psychiatry 183:248–254.
Stevens, A. B., E. R. Smith, L. R. Trickett, and R. McGhee. 2012. Implementing an evidence-based caregiver intervention within an integrated healthcare system. Translational Behavioral Medicine 2(2):218–227.
Sung, H.-c., W.-l. Lee, T.-l. Li, and R. Watson. 2012. A group music intervention using percussion instruments with familiar music to reduce anxiety and agitation of institutionalized older adults with dementia. International Journal of Geriatric Psychiatry 27(6):621–627.
Telenius, E. W., K. Engedal, and A. Bergland. 2015. Long-term effects of a 12 weeks high-intensity functional exercise program on physical function and mental health in nursing home residents with dementia: A single blinded randomized controlled trial. BMC Geriatrics 15(1):158.
Teri, L., L. E. Gibbons, S. M. McCurry, R. G. Logsdon, D. M. Buchner, W. E. Barlow, W. A. Kukull, A. Z. LaCroix, W. McCormick, and E. B. Larson. 2003. Exercise plus behavioral management in patients with Alzheimer disease: A randomized controlled trial. Journal of the American Medical Association 290(15):2015–2022.
Thyrian, J. R., J. Hertel, D. Wucherer, T. Eichler, B. Michalowsky, A. Dreier-Wolfgramm, I. Zwingmann, I. Kilimann, S. Teipel, and W. Hoffmann. 2017. Effectiveness and safety of dementia care management in primary care: A randomized clinical trial. JAMA Psychiatry 74(10):996–1004.
Toots, A., H. Littbrand, N. Lindelöf, R. Wiklund, H. Holmberg, P. Nordström, L. Lundin-Olsson, Y. Gustafson, and E. Rosendahl. 2016. Effects of a high-intensity functional exercise program on dependence in activities of daily living and balance in older adults with dementia. Journal of the American Geriatrics Society 64(1):55–64.
VA (U.S. Department of Veterans Affairs). 2020. Program of general caregiver support services (PGCSS). https://www.caregiver.va.gov/Care_Caregivers.asp (accessed September 29, 2020).
Vandepitte, S., K. Putman, N. Van Den Noortgate, S. Verhaeghe, and L. Annemans. 2019. Effectiveness of an in-home respite care program to support informal dementia caregivers: A comparative study. International Journal of Geriatric Psychiatry 34(10):1534–1544.
Vickrey, B. G., B. S. Mittman, K. I. Connor, M. L. Pearson, R. D. Della Penna, T. G. Ganiats, R. W. Demonte, Jr., J. Chodosh, X. Cui, S. Vassar, N. Duan, and M. Lee. 2006. The effect of a disease management intervention on quality and outcomes of dementia care: A randomized, controlled trial. Annals of Internal Medicine 145(10):713–726.
Vilagut, G., C. G. Forero, A. Pinto-Meza, J. M. Haro, R. de Graaf, R. Bruffaerts, V. Kovess, G. de Girolamo, H. Matschinger, M. Ferrer, and J. Alonso. 2013. The mental component of the short-form 12 health survey (SF-12) as a measure of depressive disorders in the general population: Results with three alternative scoring methods. Value in Health 16(4):564–573.
Visser, S. M., M. P. McCabe, C. Hudgson, G. Buchanan, T. E. Davison, and K. George. 2008. Managing behavioural symptoms of dementia: Effectiveness of staff education and peer support. Aging & Mental Health 12(1):47–55.
Wang, J. J. 2007. Group reminiscence therapy for cognitive and affective function of demented elderly in Taiwan. International Journal of Geriatric Psychiatry 22(12):1235–1240.
Wang, L. Q., W. T. Chien, and I. Y. Lee. 2012. An experimental study on the effectiveness of a mutual support group for family caregivers of a relative with dementia in mainland China. Contemporary Nurse 40(2):210–224.
Zarit, S. H., M. A. Stephens, A. Townsend, and R. Greene. 1998. Stress reduction for family caregivers: Effects of adult day care use. Journal of Gerontology, Series B, Psychological Sciences and Social Sciences 53(5):S267–S277.
Zarit, S. H., L. R. Bangerter, Y. Liu, and M. J. Rovine. 2017. Exploring the benefits of respite services to family caregivers: Methodological issues and current findings. Aging and Mental Health 21(3):224–231.