Abstract
INTRODUCTION
Dementia is known as a major public health problem with serious physical and emotional consequences for patients and their caregivers and a high financial strain upon society. Most people with dementia live at home and are cared for by informal caregivers, mostly spouses and children, who play a crucial role in the supervision and therapy because of the progressive mental decline of the disease. Knowing that on the one hand there is a high cost of the disease for the society due to frequent hospitalization and permanent institutionalization in nursing homes and, on the other hand, people with dementia prefer to remain in their trusted environment as long as possible, it can be stated that informal caregivers are probably cost saving for society, but also essential in the care process [1–9]. Especially given the emerging focus on person-centered care, caregivers are indispensable to maximize well-being of people with dementia since they are best placed to identify their unique needs and desires [10].
Although caregiving for a loved one can be very satisfying, it also demands a lot from informal caregivers over long periods [5, 11]. Most research in the past decade has shown that informal caregivers have more health problems, visit healthcare professionals more frequently, suffer from isolation, and have an increased risk of depression, distress, and other illness [1, 11–15]. Because caregiver burden can reduce quality of caregiving and increases the likelihood of care-recipient placement into institutional care, there is also a substantial risk for the care-recipient [6, 17].
To prevent caregivers from getting overburdened different supportive interventions, such as psychoeducation, respite care, cognitive behavioral interventions, and occupational therapy, have been developed to improve their well-being. Although well-being remains a non-uniform concept with various definitions, in this study we defined well-being as an evaluation of a person’s quality of life, life satisfaction, or positive affect. It refers to a balance between the social, physical, and psychological resources caregivers need in order to meet their social, psychological, and physical challenges [18–21]. Although many interventions demonstrated positive impacts on different outcomes such as burden [22, 23], quality of life of the caregiver [24], and delay in hospitalization [25]; other similar studies did not find the same benefits [26–28]. Inconsistency of the current findings can be explained by the methodological issues caused by the complex nature of this type of research, e.g., [4, 29–35], such as: different duration of interventions, different intensity and characteristics of interventions, and the variety of caregivers and care-recipients characteristics. Additionally, these interventions often suffer from insufficient statistical power, short follow-up periods, and the use of various instruments to measure outcomes. To give caregivers of people with dementia the care they need, more insight is still needed in this area[1, 11, 36].
After a review of the existing reviews in this field (Supplementary Table 1), we concluded that there was a need for an updated general systematic review for different reasons. First, most of the existing reviews about supportive interventions targeting informal caregivers have focused either on a particular type of intervention: for example internet based interventions [2] or respite care [37–39]; on a particular outcome: such as anxiety [40], quality of life [41], delay in institutionalization [42], or burden [43]; or only on the impact for the caregiver [2, 45] ignoring other possible effects on the care-recipient or vice versa. Second, many reviews excluded respite care as a type of caregiver support while this should also be recognized as a supportive strategy for caregivers. Third, some reviews only included a small number of studies potentially missing important evidence of non-included but eligible interventions [29]. Finally, many reviews included both controlled and uncontrolled studies making comparison between interventions difficult [12, 45].
A general systematic review is thus desirable, only including recent studies with both an intervention arm and a control arm, taking different types of interventions targeting caregivers into account, and measuring outcome effects at least on the caregiver or/and on the care-recipient. The last review which met the abovementioned criteria was published five years ago and only included randomized controlled trials through 2008 [33]. It is to be expected that since then, there is new evidence regarding effectiveness of supportive strategies for informal caregivers.
Aim of this systematic review
The current review aims to address the following question: Are initiatives to support informal caregivers of people with dementia effective for caregivers and/or care-recipients in comparison to standard dementia care? More specifically: what impact does supporting informal caregivers of people with dementia have on the well-being of those caregivers and/or on the well-being of people with dementia?
METHODS
Search strategy
A systematic literature search was conducted in PubMed and Web of Science between 15 February 2015 and 12 March 2015. PubMed was used as a core database first identifying potential ‘MeSH Terms’ which were then included in the search strings to fully capture all relevant synonyms. Used terms in the search strings were: effectiveness*, impact*, effect*, efficacy*, support*, intervention studies (MeSH)*, intervention*, program*, programme*, therapy*, psychoeducation*, psycho-education*, occupational therapy (MeSH)*, cognitive behavioral therapy (MeSH)*, respite*, respite care (MeSH)*, day care (MeSH)*, day-care*, partial hospitalization*, intermediate care facilities (MeSH), intermediate care*, temporary admission*, short stay*, residential respite*, night care (MeSH)*, night-time care*, in-home respite*, home respite*, host family respite*, caregiver*, carer*, care giver*, spouse caregiver*, family caregiver*, informal care*, informal caregiver*, dementia (MeSH)*, Alzheimer disease (MeSH)*, Alzheimer*, Alzheimer’s disease*, and frail elderly (MeSH)*.
Study selection
The inclusion and exclusion criteria used in this systematic review are represented in Fig. 1.
Data extraction
The initial search in two databases resulted in a total of 1,033 studies. After eliminating duplicates, 881 studies remained. The remaining studies were further analyzed based on reviewing titles and abstracts resulting in 88 studies assessed for eligibility. After reading full texts of those papers, 53 studies were finally included in this systematic review.Figure 2 illustrates the data extraction process.
The first 50 abstracts assessed for eligibility were double-checked by two researchers to avoid bias. When there was disagreement about eligibility, it was discussed between the researchers. Mostly disagreement occurred when it was not clear if inclusion criteria had been met based on reviewing the abstract. In that case, the full text was read and consensus achieved.
Quality appraisal
Study quality was fully assessed by one reviewer using the 27-items checklist for the assessment of the methodological quality for both randomized and non-randomized studies of health care interventions created by Downs and Black [46]. A detailed list of the scores per item per study can be obtained upon request (Supplementary Table 2). To avoid bias, a second reviewer assessed quality of ten randomly selected papers. When disagreement occurred this was discussed between the two reviewers. If no consensus could be reached, a third reviewer was consulted. Based on the obtained consensus regarding the 10 randomly selected papers the first reviewer followed a consistent approach to assess the rest of the papers.
In line with McKechnie et al., some items from the original checklist (related to adverse events; representativeness of staff, places, and facilities; blinding study subjects; and compliance) were not applied because of not relevant to the investigated papers [44]. Finally, the item that deals with statistical power was simplified to a choice of scoring either zero or one point, in which one point means that the study had sufficient power to detect a clinical effect [47]. Given these adjustments each study could now reach a maximum score of 23. Scores from 0–7 indicate very low methodological quality, 8–13 low methodological quality, 14–18 medium methodological quality, and 19–23 high methodological quality.
Study classification
Different types of interventions can be defined but the overlap of common components in many interventions makes it difficult to create an distinct classification [15]. As a result, various classifications are used in the existing literature. In this systematic review, five main types of support for caregivers were distinguished: psychoeducational interventions, respite care, occupational therapy interventions, cognitive behavioral interventions, and miscellaneous interventions. In accordance to Pinquart & Sorensen, when a certain intervention contained components of different distinguished support types, they were classified in one of the first four categories based on their dominant component [12]. When no dominant strategy could be identified, the intervention was classified under miscellaneous interventions. In the following paragraphs, each main type of support, except miscellaneous interventions (see above) is explained.
Psychoeducational interventions are the most common used type of intervention to support caregivers of people with dementia with a strong emphasis on gaining knowledge, on developing skills to deal with disease-related problems, and on providing social support. Typically these are multicomponent containing a broad range of activities such as education, social support, counseling, problem-solving techniques, coping, and skill building activities [31, 49]. When psychoeducational interventions were solely focusing on one strategy or contained multiple activities, they were respectively further classified as single or multicomponent interventions.
Second, respite care can be defined as services providing a temporary break to caregivers and can be further distinguished into: community-based respite care (including in-home respite & host family respite) and residential respite care (including day care, temporary residential admission, and night-time care) [38, 50].
Occupational therapy interventions are originally developed to maintain and improve function, independence, and participation in activities of daily living in people with illness, injury or disability by performing meaningful activities. When occupational therapy targets caregivers the purpose is to reduce burden of the caregivers by increasing their ability to handle problematic behaviors and their sense of competence [51, 52].
Cognitive-behavioral therapy is a type of specialized mental health counseling including both behavioral and cognitive techniques. The aim ofthese structured interventions is to help caregivers be aware of automatic, dysfunctional thoughts and to see problematic situations more clearly so they can respond more effectively in the future. Cognitive-behavioral therapy interventions often partly overlap psychoeducational interventions, especially when coping-based strategies and skills training are used in psychoeducational interventions, but must nevertheless be classified as a separate type. The latter by its specialized structured character based on cognitive components with a strong emphasis on cognitive processes as a mediator for behavioral changes [53, 54].
Finally, each type of support except respite care is once more divided into individual based, group based or a combination of both depending on whether they are delivered at individual level, group level or a combination of both. Figure 3 illustrates the developed classification for this review.
RESULTS
In this chapter, the results of all 53 included studies are discussed by intervention type. Possible outcomes investigated in the included interventions for informal caregivers are related to quality of life/well-being, mental health, physical health, competence/self-efficacy, burden, attitudes, and resources.
More detailed information on the included studies is displayed in Table 1. When stated that a particular intervention had an effect, this means that there was a significant positive effect on at least one outcome measure at
Psychoeducational interventions
A total of 37 psychoeducational papers (11 of high, 23 of medium, and three of low methodological quality), of which 32 showed positive effects (86%), were identified. In general, psychoeducational interventions generated more often benefits for caregivers (83% of the interventions investigating informal caregiver outcomes) than for care-recipients (39% of the interventions investigating care-recipient outcomes). The multicomponent psychoeducational interventions seem to be most popular generating better results (90%) than the single component psychoeducational interventions (67%). In general the most frequently mentioned benefits are found in self-efficacy, depressive symptoms and burden scales of caregivers, and to a less extent in delaying nursing home placement of care-recipients.
To structure the results about psychoeducational interventions and in order to better compare those interventions, we grouped the psychoeducational interventions based on the different supportive strategies they used, such as: support group, education, training, and counseling.
Single component interventions
Four of the six included single component interventions (67% were effective) showed some small benefits. Based on the performed quality appraisal, three studies were considered to be of high methodological quality and the three others of medium methodological quality.
Individual based single component interventions
Two of the three individual delivered single component psychoeducational interventions showed small benefits for caregivers, mostly only for certain caregiver subgroups. One study showed benefits for care-recipients.
One intervention of high methodological quality consisted of a care management program to support the caregivers. This was done by a care manager
who helped caregivers create a care management manual [55]. Results indicated clear benefits for the intervention caregivers in the primary outcome adherence to guidelines and in multiple secondary resource and self-efficacy outcomes (receipt of services and resources, perceived quality of health care, confidence, mastery, social support, and receiving as much help as needed of formal caregivers). On the other hand, knowledge and quality of life of informal caregivers did not improve. Care-recipients also benefitted from the intervention. Their decline in quality of life was lower than in care-recipients of the other trial arm.
In another study of high methodological quality, informal caregivers received social and emotional support from a trained volunteer during home visits. This study could not produce any benefits in the primary outcome depression, nor in the secondary outcomes related to mental health, subjective well-being, resource use, and self-efficacy (anxiety, objective burden, resource use, QALY, life events, loneliness, affect, relationship quality, social support, and coping) [56].
Also Melis and colleagues sent a nurse to the homes of caregivers to support them. Although this intervention of medium methodological quality found no overall positive effects on the primary outcome caregiver burden and the secondary outcome time spent in caring, subgroup analysis showed benefits on burden of caregivers living with the patients [57].
Group based single component interventions
For the two group based single component psychoeducational interventions of medium methodological quality, only small benefits for caregivers were found. In one study, only a subgroup of caregivers benefitted, and in the other study, there was only some improvement found in a subscale of one of the outcomes under investigation.
The first intervention consisting of telephone-based support groups could not find benefits in the primary outcomes: depression, burden, and personal gains. However, a subgroup of older caregivers (>65 years) reported lower depression rates in the intervention group [58].
The other study that implemented group-based education showed no overall positive effects on the primary outcomes: depression and quality of life of caregivers or in the secondary outcomes: resource utilization, hours spent in caregiving, and delay of temporary or permanent nursing home placement. Only emotional role functions improved (subscale of the health survey) [26].
Combination of individual and group based single component interventions
The only study, of high methodological quality, combining individual and group-based counseling could not find any improvements. No positive effects were found in the primary outcomes: depression and anxiety or in the secondary outcomes: burden and quality of life of intervention caregivers [59].
Multicomponent interventions
Twenty-eight of the 31 multicomponent interventions (eight of high, 20 of medium, and three of low methodological quality) showed improvements (90%) on at least one of the investigated outcomes. Most frequently mentioned benefits for caregivers are a decrease in burden and depressive symptoms, and for care-recipients a delay in nursing home placement.
Individual based multicomponent interventions
Eleven studies were delivered individually and all showed improvements for the caregivers, especially in self-efficacy and burden. Five studies were evaluated as having high methodological quality, five as having medium methodological quality, and one as having low methodological quality.
Three different individual technology based interventions can be identified in this section. They all generated benefits for caregivers although results were rather mixed. A telephone-based intervention of medium methodological quality (containing emotional support, education, and teaching strategies) found reduced caregiver burden. In addition, caregivers also showed less severe reactions to memory and behavioral problems of the care-recipient. On the other hand this study could not identify a decrease in depressive symptoms of caregivers [60]. In contrast, a second study of medium methodological quality delivering a telephone-based exercise (consisting of a telephone-based exercise, a booklet, a workbook, and motivational newsletters) could not find a decrease in burden, but did find a decrease in perceived stress and exercise self-efficacy. On the other hand, efficacy in self-care and total weekly exercise did not improve, but caregivers with low exercise scores at baseline did show improved total weekly exercise [61]. A last technology based intervention of low methodological quality that provided two DVD based training sessions and communication support strategies could not demonstrate a reduction in burden either, but did have a positive impact on knowledge and positive aspects about caregiving. It also reduced the frequency of disruptive behaviors of care-recipients, but did not improve their mood [62].
Eight individually delivered interventions combined education and training. Each intervention that investigated the possible impact on caregivers indicated benefits for caregivers, especially in burden and self-efficacy outcomes. Teri et al., targeting their intervention of high methodological quality at both caregiver and recipient, through an exercise program for the recipients and training for caregivers, found positive effects on the primary outcomes physical health and depression (affective status) in recipients. Possible impact on the caregivers was not measured in this intervention [63]. An intervention of medium methodological quality, that added adaptive treatment to the education and problem-solving training, partially decreased subjective and objective burden and enhanced caregivers in terms of upset. There was also less need for ADL help and better affect. On the other hand there were no differences found for caregivers’ overall well-being and overall (objective and subjective) burden, nor in care-recipients’ behavioral problems and in their physical functioning [22]. The psychoeducational intervention program (PIP) of high methodological quality developed by Martin-Carrasco et al. also decreased burden (primary outcome), improved mental health, and did improve caregiver well-being [64]. In accordance, Gavrilova and colleagues, who also implemented a high quality multicomponent intervention containing education and training, pointed out a decrease in caregiver burden, but could not decrease distress or improve quality of life. For recipients, no differences were found in quality of life, nor in behavioral and psychological symptoms [23]. Ducharme et al. concluded that after receiving their intervention of medium methodological quality, also containing education and training, caregivers were more confident in dealing with caregiving situations. They perceived themselves better prepared for caring, more efficacious in their role, and were better able to plan for future needs. They also had a better knowledge of services and made more frequent use of coping strategies. No significant differences were found for perceived informal support and family conflicts [65]. In Ducharme et al. (medium methodological quality), the authors gave half of the intervention caregivers a booster session wherein the caregivers were asked about changes in the caregiving situation over the last 6 months. Only the preparedness to provide care differed significantly between the caregivers receiving the booster session and those who did not [66]. Finally, another intervention of high methodological quality implemented a coping based psychoeducational intervention with educational and training components. Intervention caregivers had reduced affective symptoms, i.e., depression and anxiety (primary outcome), reduced case level depression, and improved quality of life. On the other hand case level anxiety and potential abusive behavior of caregivers, and quality of life of recipients [24] did not improve. The intervention remained effective after 24 months and was also cost effective[67].
Group based multicomponent interventions
Seven of eight multicomponent studies (three of high, three of medium, and two of low methodological quality), delivered to groups, indicated positive effects. Six of them found benefits for caregivers and two for care-recipients. Although most interventions benefitted caregivers, the found effects were rather small and often only found in subscales of the measurements.
An intervention of low methodological quality containing elements of education, support, and skills training could not find overall effects on burden or satisfaction of caregivers. Only strain, disappointment (subscales of the used burden scale), and purpose (subscale of the used satisfaction scale) showed improvements [68]. The same authors also investigated possible impact of the intervention on time to nursing home placement but could not find significant delay, except for a subgroup of caregivers of adult children and daughters [69].
Berger et al. implemented a dyad intervention (medium methodological quality) by combining a support group for caregivers with music therapy for recipients. Burden and depressive symptoms of caregivers did not decrease. Neither did behavioral and psychological symptoms, and cognitive and functional status of recipients after receiving the intervention [70]. Another quasi-experimental intervention of low methodological quality also providing a support group, but this time in combination with education and information sessions, increased caregivers’ disease understanding (primary outcome) [71].
Four different studies described interventions with training and educational components to caregivers. One of them (medium methodological quality) found lower levels of depression and burden of caregivers beside less severe reactions to behavioral problems. There was, however, no decrease in the frequency of problem behaviors of recipients [72]. In contrast, an educational program of high methodological quality followed by group meetings to learn communication techniques and structured problem solving, did decrease behavioral and psychological symptoms (primary outcome) of recipients in a subgroup of female intervention caregivers. Caregiver stress (primary outcome) on the other hand was not reduced [73]. Another intervention of high methodological quality found significant improvements in disease understanding. Even more, the intervention increased caregiver ability to cope, but could not decrease depressive symptoms (primary outcome), sense of competence, and burden of caregivers. Neither could the intervention induce functional (primary outcome), cognitive, and behavioral changes in recipients [74]. The intervention of high methodological quality developed by Martin-Carrasco et al. was more recently also implemented in groups. This time, the authors could not establish the same benefits as in the individual delivered psychoeducational intervention. Burden (primary outcome), quality of life, and perceived health did not improve in caregivers. Only some benefits were identified on a subscale of anxiety and insomnia [75].
Combination of individual and group based multicomponent interventions
Twelve psychoeducational interventions, of medium methodological quality, that partially delivered their intervention in group and partially on an individual basis were included in this review. Ten of them benefitted caregivers or care-recipients. Most frequently reported benefits were in self-efficacy and depressive symptoms of caregivers, and in delay of nursing home placement of care-recipients.
Two technology-based interventions can be classified in this section. Steffen, who investigated the effectiveness of an anger management video series, containing multiple training strategies, on caregivers found significant positive effects on anger, self-efficacy in managing problem behaviors, and also on depressive symptoms in the home based viewing group [76]. On the other hand, another computer-mediated automated interactive voice system intervention—containing stress monitoring, information, counseling, a telephone support group, and a distraction call for recipients—only showed positive effects in reducing bother, depressive symptoms, and anxiety in caregivers with lower mastery. For the subgroup of wives, there was a reduction in bother [77].
Drentea and colleagues combined individual and family counseling with participation in a support group. They found higher levels of satisfaction with their support network over the first five years, higher levels of perceived emotional support, more visits, and having more network members to whom they felt close in the intervention group [78]. Another similar intervention composed of counseling and support groups found a significant reduction in caregiver reactions to behavioral problems (primary outcome), but not in patient behavioral problems [79]. They also found positive effects on depressive symptoms of caregivers (primary outcome) [80] and also pointed out that recipients of the intervention group stayed on average 557 days longer at home than those in the control group (primary outcome) [25]. Finally, they concluded that intervention caregivers had significantly better self-rated health (primary outcome) and that their number of reported illnesses decreased [81]. In contrast, other authors [82] who also combined a support group with counseling could not find to same benefits on time to placement (primary outcome). Even more, caregiver burden, and care-recipient’s ADL, IADL, cognitive functioning, and quality of life in the intervention groups didnot change.
The Savvy Caregiver program, that implemented an intervention containing educational and skills training components to support caregivers, concluded that interventions caregivers had better mastery and distress scores compared to those who did not receive the intervention [83].
Nichols et al. conducted an intervention combining education with support group provision. They found significant reduction in hours of care (primary outcome) provided by the caregiver and concluded that the intervention was also cost-effective [84].
Eloniemi-Sulkava et al. implemented an intervention containing support group meetings, at home training, and information sessions. Less recipients of the intervention group were admitted to long-term institutionalization after 1.6 years (primary outcome), but the effect did not remain after 2 years. The study also found a reduction in service use and expenditures. However, the reduction diminished when costs of the intervention were included [85].
Phung and colleagues investigated long term effects of an intervention combining counseling and education that already showed benefits on the primary outcomes depressive symptoms and quality of life of recipients at 12 months. After 36 months the earlier found benefits did not remain. The other investigated outcomes did not improve or decrease [28].
Respite care
Five respite care interventions, two of low methodological quality and three of medium methodological quality, were included. Four measured the impact of day care and one of an in-home respite care program. Although three of the four day care interventions indicated benefits for the care-recipient (75%), especially in decreasing behavioral problems, the reported benefits of day care provision for caregivers were rather small and mixed (33% were effective). The only included in-home respite care program indicated some beneficial effects for caregivers, but did not investigate the impact on the care-recipient.
An intervention of medium methodological quality providing three months of day care once or twice a week could not find an effect on caregiver burden or any benefits for the care-recipients in anxiety levels, depressive symptoms, or functional status [86]. On the other hand another intervention of low methodological quality providing day care at least twice a week did report a decrease in dementia symptoms in care-recipients and increased their well-being. Also in this intervention, there was no effect on caregiver burden or in subjective well-being [9]. These findings partially corresponds with the findings of Femia et al. (medium methodological quality) who also reported an improvement in nighttime sleep disturbances, a subscale of total behavioral problems (primary outcome) [87].
In contrast to the others, Mossello et al. did find a decrease in caregiver burden (primary outcome) besides a decrease in behavioral problems of care-recipients (primary outcome). It must however be mentioned that in this medium quality intervention the caregivers also received some additional counseling in the day care center [88].
The only in-home respite care program (low methodological quality) that was included provided two weeks of in-home help during six hours a day. Some beneficial effects for caregivers were found which can reduce their morbidity and mortality [7]. On the other hand benefits in psychological symptoms could not be found.
Occupational therapy interventions
Eight occupational therapy interventions (three of high methodological quality and five of medium methodological quality) were included. All investigated the impact on the well-being of the caregiver and on the care-recipient. All seven occupational interventions delivered to individuals showed benefits, especially in caregiver self-efficacy outcomes and in the frequency of behavioral problems of care-recipients. The only group based occupational therapy intervention could not find benefits.
Individual based occupational therapy interventions
Graff and colleagues developed a community based occupational therapy intervention of high methodological quality containing ten sessions to train both caregivers and care-recipients. Both caregivers and care-recipients benefitted from the intervention. In caregivers of the intervention group sense of competence, skills, and sense of control improved. For both caregivers and recipients mood, quality of life, and health status improved. In care-recipients daily functioning improved [51, 89].
Another research team implemented three similar occupational therapy interventions of medium methodological quality, each time in a different dose of intensity. For all three different doses, benefits were found for the dyads under investigation. During the first intervention, participants received five occupational therapy sessions. Here, only subgroups of caregivers (spouses and women) benefitted. They indicated less upset and enhanced self-efficacy. Recipients showed less decline in IADL, but not in ADL or in behavioral problems [90]. In the second intervention, they provided six sessions of occupational therapy. This time behavioral problems of the recipients did decrease. For all caregivers, reported need for help and skills improved, but no decline was found in caregiver upset. There were also no changes in affect at 6 months. Nevertheless, over a longer period affect improved [91]. In the last intervention, dyads received eight sessions. Here, the same improvements as in the previous intervention were found for behavioral problems and skills. Even more, recipients showed improved activity engagement and ability to keep busy. Caregivers showed a reduction in objective burden. On the other hand no effects were found on depressive symptoms of the dyads, on subjective burden of caregivers, and on quality of life of recipients [92]. Yet another occupational therapy intervention (high methodological quality) of the same first author as the previous studies [93], that contained nine occupational therapy sessions, showed overall positive effects on the primary caregivers’ outcomes: upset and confidence in managing target problem behavior as well as in the care-recipient primary outcome: frequency of target problematic behavior. Additionally, there was less caregiver upset with all problem behaviors, less negative communication, depression and burden decreased, and well-being improved (secondary outcomes). Finally, after the follow up phase intervention caregivers had a greater ability to keeppatients home.
Finally, Nobili and colleagues added an educational component to their occupational therapy intervention of medium methodological quality. Although the intervention could not reduce stress in caregivers, the frequency of behavioral occurrences of recipients also decreased after implementation of their intervention [94].
Group based occupational therapy interventions
Unlike the other individual occupational therapy interventions, the only occupational therapy intervention in group (medium methodological quality) did not show any benefits for the dyads [95]. None of the investigated caregiver outcomes improved (primary outcome: general health of caregivers & secondary outcomes: distress, depression, relationship with the recipient, and anxiety). Neither did the investigated care-recipient outcomes (primary outcome: quality of life & secondary outcomes: autobiographical memory, depression, anxiety, ADL, and service use). Even more, a side effect was identified. Caregivers of the intervention group had higher anxiety levels on a subscale (GHQ-28) after implementation.
Cognitive behavioral therapy interventions
Three group-based cognitive behavioral therapy interventions, each of medium methodological quality, delivered in groups were included. In each study benefits for caregivers were found especially in dysfunctional thoughts. Possible impact on the care-recipient was not measured in any of the studies.
The first group of authors who tested efficacy of a cognitive behavioral therapy intervention found positive effects on dysfunctional thoughts, frequency of leisure time, and depressive symptoms [96]. The second group of authors who added training in self-help techniques to the cognitive behavioral treatment also found a positive impact on the primary outcomes: dysfunctional thoughts and mental health of caregivers of the intervention group. The secondary outcomes quality of life and burden did not differ after receiving treatment [97]. Finally, Passoni et al. concluded that both intervention groups had less need for care or assistance compared to control group participants, but anxiety and depressive symptoms did not differ [98].
DISCUSSION
A first major strength of this paper is its general overview and update in the knowledge of supporting strategies targeting informal caregivers of people with dementia by searching for potential new insights based on recent interventions, and by balancing previous knowledge against current findings. Additionally, we explored whether the earlier identified methodological issues were still present. Finally, we attempted to develop a new classification to distinguish global and more specific tendencies in the results. As a result, we were able to state for each distinct supportive type what particular benefits are to be expected. After all, before one proceeds to invest in caregiver support, one must know what benefits to expect. Nevertheless, we have to recognize that the 53 studies describing interventions to support caregivers were multifaceted making it impossible to prevent that some interventions placed in one subdivision did not contain elements of others.
Another general limitation that should be mentioned is the lack of performing a quantitative analysis. This procedure was considered, but was not feasible due to methodological heterogeneity, i.e., variations in the outcomes and the various instruments used to measure these outcomes. Also, due to restricted resources available, the quality appraisal could not be fully evaluated by two independent authors. Nevertheless, a more pragmatic approach was used maintaining the methodological quality of the work at the highest possible level, i.e., it was opted to perform a quality appraisal of randomly selected studies by two reviewers independently, and then to compare their scores. Since the authors had very similar ratings and there was complete agreement on the level of quality the rest of the quality appraisal was done by one author following the same consistent approach.
In general most supportive interventions were to some extent effective (87% of the studies showed at least some benefits for the caregiver and/or the care-recipient). This, however, should be interpreted with caution and not be overestimated. Indeed, effects were found in 87% of the included studies, but in some cases the benefits could only be found in secondary outcomes or even in subscales of outcome measures or in subgroups of participants. Even more, not all studies clearly specified their primary and secondary outcomes. It should also be mentioned that methods and findings were rather inconsistent. This phenomenon, also previously described by others, e.g., [11, 99], can still be explained by the heterogeneous nature of supportive strategies caused by differences in content, duration, and intensity; the methodological quality of studies; the variety in outcome measures; and the different characteristics of the caregivers and the care-recipients.
Another general conclusion that can be drawn is that individual interventions are more likely to be effective. Given the fact that caregivers are a heterogeneous group with different needs, expectations, and characteristics, it is to be expected that interventions delivered at individual level can better adapt to those differences and as a result lead to better results. This conclusion corresponds with the systematic review of Selwood. This author, who investigated the effect of psychological interventions on family caregivers of people with dementia, also concluded that group interventions are less effective [45].
Methodological issues
Based on our performed quality appraisal using the Downs and Black checklist [46], we identified 14 studies of high methodological quality, 34 of medium methodological quality, and five of low methodological quality. All high quality studies were effective except for two single component psychoeducational interventions of which we know they generally are less effective than multicomponent interventions. Even more, we have established that in nine of the 12 effective high quality studies, overall effects were found in the primary outcomes. On the other hand, 29 of the 34 studies of medium methodological quality were to some extent effective, but here only 20 studies could ascribe these benefits to significant changes in primary outcomes. Finally, all five low quality interventions were to some extent effective. However, only two of those studies found clear effects in the primary outcomes targeting caregivers.
Despite several previous reviews concluded that the interventions under study are overall effective in decreasing or increasing specific outcomes that improve well-being (e.g., depression, self-efficacy), our opinion is that these overall conclusions should be made cautiously due to inconclusive results often caused by methodological issues as has been discussed in other comparable reviews, e.g., [11, 37].
As long as researchers keep using a wide range of instruments to measure effectiveness, differences in effectiveness results could rather be caused by differences in sensitivity and specificity of the used measurements rather than by the fact that a certain intervention is less or more effective in influencing a certain outcome. Hence a standardized way of measuring outcomes is preferred. The latter might be perceived as a paradox since a person-centered approach is to be preferred. Yet, a compromise can be found by applying person-centered care while using standardized yet sensitive instruments to measure outcomes, such as the Zarit Burden Interview for caregiver burden.
The current interventions in this research area are also plagued by other methodological issues. First, the active components of the implemented interventions are often not enough described in detail making it difficult to attribute significant improvements to a particular type of intervention. Second, most interventions have short follow-up periods. Given the fact that dementia is a slow degenerating process with high demands for caregivers over long periods, knowledge on long-term effects can be important as well. Yet another frequently occurring shortcoming is the absence of a well-performed power calculation to determine the appropriate sample size. This often resulted in too small sample sizes. As a result, a risk of falsely rejecting effective interventions arises because the sample was not large enough to establish effects. Simultaneously occurring with this issue is the lack of exploring the minimal needed effect size of the research outcomes.
Knowing the characteristics of participants lost to follow up and their potential significant differences with those who completed the study can provide researchers important information about the acceptability and applicability of interventions. Nevertheless, in several studies this important step was not accomplished. Also, studies often lacked to ascertain that subjects asked and prepared to participate were representative of the entire sourcepopulation.
Another more difficult methodological problem to solve is about the control group. In the included studies, the control group is mostly defined as the group not receiving the intervention or only receiving treatment as usual. Because it is often not clear what other supportive services they get besides the intervention, benefits could also be caused by those other services. Even more, subjects that are potentially most in need of support could be forgotten because most dyads are recruited via organizations already giving them some support. In the future, researchers should explore more how to find those subjects. Also including them can contribute to finding stronger evidence.
Finally, dementia can be divided into different stages. Although some interventions mention in what stages the included subjects were, it is not known what influence severity of dementia can have on certain effective interventions.
Different types of interventions
Psychoeducational interventions are the most commonly investigated type and in general an effective strategy (86% of the studies showed benefits) also showing the typical inconsistency that we already discussed. Of this broad type, the multicomponent interventions are most popular. Many multicomponent studies reported a positive impact on caregivers (83%) especially in self-efficacy, burden, and depressive symptoms. However, they had in general less effect on care-recipient outcomes (39%) except for delay of nursing home placement. As already mentioned, the single component psychoeducational interventions solely focusing on one strategy (such as education, social support) were rather not effective. This does not mean that these strategies should not be undertaken, but rather that they need to be embedded into more extensive interventions.
Although not many cognitive behavioral therapy interventions were included in this review, this supportive strategy seems to be effective in decreasing dysfunctional thoughts of caregivers. Future research of this type should, however, consider measuring the possible indirect effects on care-recipients as well.
Occupational therapy interventions also seemed to be effective in decreasing behavioral occurrences of care-recipients and self-efficacy of caregivers, especially when there were delivered individually.
Only a few studies about some types of respite care were included in this review. Four of the included studies provided day care allowing the caregiver to take a break. Unfortunately, only one community based in-home respite care intervention was included. Knowing that substantial research, especially qualitative research, was already done in this field there seems to be a lack of RCTs and quasi-experimental designs measuring the impact of certain respite care services [38]. This could be explained by the difficulty of finding a similar control group, leading researchers to choose a pre-posttest design. Although most included day care interventions indicated benefits for the care-recipients in terms of behavioral problems, caregivers outcomes were rather mixed. In accordance to the main findings of the recently updated review of Lee et al. [100] by Maayan et al. [37], this review also concludes that the mixed results and the small number of included studies in this area make it impossible to draw clear conclusions about effectiveness of respite care services. Therefore, an updated review about effectiveness of respite care services for people with dementia will be needed first. It is suggested to also include pre-posttest designs without control group in order to explore more potential effects. Second, we also recommend new intervention studies investigating different forms of respite care services most preferably using a comparable control group allowing stronger conclusionsto be made.
Conclusion
Stating which intervention type works best for all caregivers and also indirectly benefits recipients is still impossible due to the complex nature of supporting strategies and the current methodological issues. Yet we can conclude that supporting caregivers is important and effective because, when properly implemented, it has the potential to improve well-being of caregivers and their care-recipients, resulting inevitably in additional benefits forsociety.
After all, informal caregivers play an increasingly important role in the care process given the current trend toward community-based health care in all health policies to press rising health care costs and to improve quality of life. Policy makers must value and recognize their exceptional role and support them so they can sustain better.
Future research should deal with the methodological limitations of current evidence in this research field to draw more clear and unambiguous conclusions. There is also a need to further explore possible effectiveness and cost-effectiveness of respitecare.
