Abstract
Dementia is characterized by progressive deterioration in cognitive ability as well as declining capacity for independent living and has become a health- and social-care priority worldwide (World Health Organization, 2012). Future scientific achievements are likely to produce new diagnostic criteria for earlier detection and treatment, thus postponing disease progression and resulting in even more people diagnosed and living longer with the disease who will require more care than the current estimates predict.
Most dementia is of the Alzheimer’s type (AD), with damage to brain cells and connections among cells producing changes in cognitive functioning that result in challenging behavioral symptoms. Difficult care situations may arise because of functional decline and behavioral symptoms, which can occur at any disease stage, with multiple behaviors sometimes appearing simultaneously (Lyketsos et al., 2011). Community-dwelling persons with dementia (PWD) have estimated rates of behavioral symptoms ranging from 61% to 88% (Lyketsos et al., 2000), and these symptoms are almost universally observed in dementia regardless of etiology (Lyketsos et al., 2011). Examples of difficult-to-manage dementia symptoms include combative, aggressive, withdrawn, and distressed behaviors (Gitlin, Kales, & Lyketsos, 2012). Safety is also a concern when dealing with unexpected, “acting out” or catastrophic behaviors that may be of an aggressive or violent nature or the PWD can wander away from home when unsupervised. Other safety issues reported by caregivers include, for example, the PWD leaving the stove on, allowing pots to boil down and catch fire, and access to alcohol, medications, and guns or knives that may be used inappropriately or harmfully. Information about how to manage behavioral symptoms is needed to prevent these potentially hazardous situations that may endanger the PWD and others in the environment (Alzheimer’s Association [AA], 2015).
Systematic reviews of caregiver interventions are beginning to identify characteristics of effective dementia-caregiving interventions such as being psychotherapeutic, multicomponent, tailored to the specific needs of the caregiver and of adequate duration (Mittelman, Haley, Clay, & Roth, 2006; Van’t Leven et al., 2013). Because caregiving situations are complex, multicomponent interventions seem more effective than interventions targeting only one aspect of caregiver functioning (Zarit & Fema, 2008). Another key element of success was tailoring intervention strategies to identified needs of caregivers and PWD. Fixed, standardized interventions are less effective than those tailored to caregiver-specific needs (Gitlin, Hodgson, & Choi, 2016). Very few studies have addressed long-term positive outcomes of caregiver interventions over time; thus, there is limited evidence that positive outcomes can be sustained over time (Mittelman et al., 2006). In addition, Thompson et al. (2007) concluded that there is no evidence that information and support interventions for caregivers of persons with dementia are uniformly effective for all PWD. Therefore, the purpose of the current study was to evaluate whether an individualized, two-component, community-based psychoeducational intervention decreased the frequency of behavioral symptoms in PWD as well as caregiver response to these symptoms over an 18-month period of time. Three research questions guided the study:
Literature Review
A considerable body of literature identifies that caring for a PWD at home is demanding and may produce negative social, economic, and emotional outcomes for caregivers (Schoenmakers, Buntinx, & Delepeleire, 2010). Challenging behaviors affect up to 90% of all PWD over the course of disease, and symptoms are associated with distress among both PWD and their caregivers. Apathy, depression, irritability, agitation, and anxiety are the most prevalent behavioral symptoms (Cerejeira, Lagarto, & Mukaetova-Ladinska, 2012). Even infrequent incidents of distressing behavior may result in nursing home admission. Gaugler et al. (2010) reported caregivers of PWD who developed dangerous behavior (an incident) after baseline were more likely to experience increases in caregiver burden with resultant admission of the PWD to a nursing home. However, persistent occurrence of less potent behaviors (particularly memory problems) was the strongest predictor of more time to nursing home admission. Thus, considering patterns such as incident versus persistent behaviors and potency (danger) of specific challenging behaviors may help facilitate more timely and targeted delivery of clinical interventions for dementia symptoms.
A variety of behavioral symptoms may occur in the context of dementia. Typically, these symptoms evolve over the course of the illness. Multiple symptoms may occur simultaneously; anxiety, depression, mood lability, apathy, sleep disturbances, agitation, psychosis, combativeness, wandering, inappropriate vocalizations, and inappropriate sexual behaviors are just some of the possible behavioral issues. Generally, nonpharmacological interventions are preferable. However, when medication is required, certain approaches may be useful. Clearly defining and measuring the behavior to be treated are helpful. Deciding on the most significant symptoms to address, rather than attempting to treat all symptoms, helps limit polypharmacy. There is no specific Food and Drug Administration (FDA)–approved medication for behavioral symptoms of dementia, making all such treatments “off label.” A variety of treatments may be used according to the psychiatric or behavioral symptom being treated. Common medications include antidepressants (especially Selective Serotonin Reuptake Inhibitors [SSRIs] and Serotonin and Norepinephrine Reuptake Inhibitors [SNRIs]) and mood stabilizers such as valproate, benzodiazepines, and antipsychotics. All of the medications are controversial for this use, and data on their effectiveness are limited. Moreover, all may be related to adverse events such as sedation, falling, or increased mortality, prompting the FDA to place “black box” warnings for all antipsychotics in dementia. However, these drugs are still utilized for lack of better alternatives (Casey, 2015).
Behavioral symptoms provide challenges for the caregiver and other family members in understanding principles of management during the disease process. Professional ability to provide a quality plan of care for the PWD requires knowledge of the family caregiver’s situation and needs. Due to the complexity of the disease process, heterogeneity of the PWD and caregiver dyad, and too few behavioral management interventions translated into practice, it is difficult to establish best practices and successful interventions to improve quality of care for PWD in home settings (Zabalegui et al., 2014).
Successful interventions have used various strategies to enhance the capacity of caregivers to succeed in the caregiver role, including acquisition of knowledge and skills related to the care tasks they perform and development of more objective attitudes and problem-solving skills. For example, a psychoeducational behavioral management program delivered to family caregivers in the home setting using principles of the Progressively Lowered Threshold Model (PLST) had a positive impact on both the frequency and response to behavioral symptoms among spousal caregiver/recipient dyads (Gerdner, Buckwalter, & Reed, 2002). Furthermore, a combined education and individualized counseling program with family meetings was successful in delaying institutionalization for a year and a half more than caregivers receiving usual care (Mittelman et al., 1995; Mittelman, Ferris, Shulman, Steinberg, & Levin, 1996; Mittelman et al., 2006). Another more recent adult child caregiver adaptation of the NYU Caregiver Intervention occurred in Minnesota with demonstration of similar benefits including delayed residential treatment of 228 days (Gaugler, Reese, & Mittelman, 2013) and resulting cost savings (Long, Moriarty, Mittelman, & Foldes, 2014).
Many researchers have found that education and support interventions for PWD–caregiver dyads reduced the frequency and severity of behavioral symptoms (Selwood, Johnston, Katona, Lyketsos, & Livingston, 2007). The increasing burden of AD in the health care system places a high priority on identifying what interventions work best for which PWD–caregiver dyads. A systematic international review by Gallagher-Thompson et al. (2012) identified 159 nonpharmacological best practices for family caregivers of people with dementia that are ready to be implemented worldwide and incorporated into national dementia policies, including individual/family counseling programs that are individualized and provide relatively long-term support (1-2 years), psychoeducational programs that use a structured format to teach specific coping skills such as “Coping With Caregiving” (Au et al., 2010), and multicomponent programs that include education, respite, and support (e.g., REACH II program; Gallagher-Thompson et al., 2012). Yet another systematic review of nonpharmacological interventions for behavioral symptoms in dementia (Brodaty & Arasaratnam, 2012) found 23 high-quality, randomly controlled trials of interventions with comparable results to those of pharmacological (antipsychotic) interventions. Significant treatment effects were found for reductions in behavioral symptoms (effect size = 0.34, p = .01) and caregiver negative responses (effect size = 0.15, p = .006; Brodaty & Arasaratnam, 2012). Unlike antipsychotic medications, most caregiver interventions have no negative side effects for caregivers or PWD. In an editorial commenting on the Brodaty and Arasaratnam (2012) review of behavioral interventions, Gitlin (2012) noted that the effect sizes are clinically meaningful because improvements compared more than favorably with pharmacological trials.
Two-Component Evidence-Based Intervention
Development of the first component of this psychoeducational intervention, teaching behavioral management to family caregivers, was based on content derived from the PLST model and prior national and international research using the model (Gerdner et al., 2002; Huang, Shyu, Chen, Chen, & Lin, 2003; Smith, Gerdner, Hall, & Buckwalter, 2004) over the past 29 years (Figure 1). The PLST model (Hall & Buckwalter, 1987) proposes that because of declining cognitive and functional abilities, PWD benefit from modified environmental demands and decreased internal and external stressors. Reducing stress caused by these demands can enhance functional adaptive behavior. Interventions based on the PLST model have reduced caregiver burden, impact, and depression (Buckwalter, Gerdner, Hall, et al., 1999; Buckwalter, Gerdner, Kohout, et al., 1999) and have enhanced caregiver satisfaction, reducing behavioral symptoms (Huang et al., 2003) over time. Logsdon, McCurry, and Teri (2007), two of whom developed the Revised Memory and Behavior Problem Checklist (RMBPC) used in this study, identified the PLST model as one of two effective psychological interventions for the treatment of behavioral symptoms in dementia. Development of the PLST approach was grounded in the stress and coping model (Lazarus, 1966), suggesting that caregiver well-being is affected not only by primary stressors originating from the illness or care of the PWD but also from environmental stressors. Individualized plans of care based on the PLST model provide for environmental modifications because of declining cognitive/functional abilities of the PWD. Family conflict is an example of an environmental stressor, which illustrates the need for the second component in the intervention tested here.
Conceptual framework for the two-component intervention.
The second component of the intervention uses an underlying stress process model (Pearlin, Mullan, Semple, & Skaff, 1990) and is based on the hypothesis that improving family support and decreasing family conflict will improve the ability of the caregiver to withstand difficulties of caregiving and defer the need for institutionalization of the PWD (Mittelman et al., 1995). A secondary benefit of both intervention components is that through ongoing education and support, the caregiver’s appraisal of the behavioral symptoms exhibited by the PWD will be minimized, thereby decreasing their impact and stress on caregivers and potentially resulting in decreased frequency of behavioral symptoms (see Figure 1 for diagram of conceptual model).
In the study reported here, ongoing assessment and support with care planning occurred every 6 months over an 18-month period. The follow-up and offer to be available as needed resulted in sustained support not only for the primary caregiver but also for other family members. Thus, the second component was implemented to include family education during at least one 3- to 4-hr family session focused on education about management of upsetting behavioral symptoms and encouragement to use community resources.
Procedure
The aim of the two-component community intervention was to provide support and education to caregivers of PWD. Determination of location for the intervention was based on caregiver choice and convenience. Caregivers were provided individualized community education after a baseline assessment of their care situation. The intervention was designed to individualize services according to each caregiver’s needs assessed during each interview. Education was provided by the trained Research Assistant (RA), who reviewed content contained in a resource binder that included a standardized module describing the PLST approach with content on types of dementia, stages, symptoms, behavioral characteristics of each stage, communication, seeking help, and using community resources. Each caregiver was given a binder inclusive of this module to keep and use for further reference during each 6-month visit. Content was updated every 6 months via the same module as disease stage progressed and caregiving problems changed. Care planning was prioritized based on the behavioral symptoms identified as causing the most distress to caregivers. This approach of targeting specific behavioral symptoms for intervention management was also endorsed by Gitlin, Winter, Dennis, Hodgson, and Hauck (2010).
Skills management thus was focused on specific PWD behaviors identified by the caregiver as most upsetting. Based on PLST content, caregivers were taught to alternate stressful periods of stimulation and activity with periods of low stimulation and rest to keep the person with dementia below his or her own stress threshold, above which challenging behaviors tend to emerge.
A second component of the intervention included a family session. The primary caregiver was responsible for inviting all significant family, friends, and neighbors to join in educational sessions. Two 1.5-hr educational sessions were offered, but most families chose to meet for one long 3- to 4-hr session depending on the needs of the family. Rationale for the family focus was to build caregiver support and enhance communication among family members. Without family-focused education, family members may inadvertently give inappropriate help or advice to the primary caregiver. Caregivers and family members were informed about how to access and use local resources available to help caregivers and PWD. For example, the closest adult day care center and support group were identified and participants were encouraged to access and use these services.
Delivery of the intervention occurred in two phases with education and a family session planned within 1 month or as soon as possible after the baseline interview. Ongoing maintenance and support continued after the initial intervention session, with results of each assessment used to develop the care plan for the next 6-month period. To assure fidelity that all caregivers received all components of the intervention, process sheets were randomly checked in each caregiver record. Random checks included review of the date that each component of the intervention was received; thus, assurance was provided that each caregiver received the same components including same assessments, training content, and timeline of data collection.
In summary, reinforcement occurred using an individualized care plan developed every 6 months for each caregiving situation based on behavioral responses identified by caregivers as most upsetting. Follow-up interviews continued every 6 months for 18 months to reassess behavioral challenges, develop new care plans, and reinforce education according to newly identified problems.
Eligibility
All participants were caregivers enrolled in a School of Nursing (SON) dementia caregiver program of intervention research. A caregiver was defined as a person who spent more than 4 hr per week providing unpaid assistance to the PWD. The primary caregiver self-identified as the person spending the most time caregiving among the care network. Caregivers were eligible for this study if the PWD had a medical diagnosis of any type of irreversible dementia, resided in a home setting outside a nursing home facility, and lived within a geographical travel distance not to exceed 1 hr from the study site. Referrals of caregivers for the research program came from a wide network of health providers, physicians, and neurology and family practice clinics, support groups, and by word of mouth. Thus, most referrals came from medical practice sites where caregivers were usually at moderate stages of dementia before initiating their own enrollment into the SON Caregivers Program of Intervention Research. Eligibility did not exclude PWD or caregivers receiving psychopharmacological treatment. The intervention was viewed as a supplement to medical treatment received in the clinics. All caregivers enrolled expecting to receive an intervention. Because random assignment and a control group were not possible, this study allowed a “real world” implementation of the intervention whereby more understanding is gained about the implementation process outside a controlled research environment with methodology similar to that described by Menne et al. (2017). Following approval from the institutional review board, data were collected by trained research assistants who were nurses or social workers employed by the SON. The current study evaluated whether the two-component intervention based on PLST content and a family meeting encouraging use of community services improved (a) frequency of behavioral symptoms in PWD and (b) caregiver response to behavioral symptoms. We enrolled 127 dyads of caregiver/recipients with a diagnosis of irreversible dementia and exposed caregivers to the individualized intervention. Each outcome was measured at four time points (baseline, 6, 12, and 18 months) allowing us to test whether the intervention improved outcomes over time.
Instruments
The following instruments were used to measure behavioral symptoms in PWD and reactions in caregivers.
RMBPC
PWD behavioral symptoms and caregiver reactions were studied by asking caregivers to respond to the RMBPC developed by Teri et al. (1992). The RMBPC is a 24-item caregiver report measure of the frequency of and reaction to behavioral problems exhibited by the PWD. Three subscale scores and one total score were provided for frequency of behavior problems related to memory loss, depression, and disruption, with corresponding measures reported for caregiver reaction to these three types of problems.
The RMBPC Likert-type scale allowed caregivers to rate the frequency of behavioral symptoms exhibited by the PWD as 0 = never occurs, 1 = occurs frequently, but not in the past week, 3 = occurred 3 to 6 times in the last week, and 4 = occurs daily or more often. Caregiver reactions to PWD behavioral symptoms were scored by the degree to which the behaviors “bothered or upset” the caregiver: 0 = not at all, 1 = a little, 2 = moderately, 3 = very much, and 4 = extremely.
Overall scale reliability for the RMBPC was reported by Teri et al. (1992), with total alpha scores of .84 for frequency of behavior problems and .90 for corresponding caregiver reaction. Subscale alpha scores ranged from .67 to .89.
Scores on the RMBPC were calculated following the advocated scoring methods for this tool (Teri et al., 1992), for each caregiver at four time points. The frequency and reaction total scales and subscales had high internal consistency and reliability. For frequency of behaviors, the Cronbach’s alphas were .82, .71, .75, and .83 for the Total, Memory Problems, Disruption, and Depression Behavior scales, respectively. For the reaction to the behaviors, the Cronbach’s alphas were .87, .79, .76, and .86 for the Total, Memory Problems, Disruption, and Depression scales, respectively.
Geriatric Depression Scale (GDS)
Scores on the GDS were used to measure caregiver depression. The GDS is a 30-item self-report measure that takes about 10 min to complete in a yes/no format. Scores were calculated for caregivers at baseline to determine impact of their level of depression on response to PWD behavioral symptoms.
Statistical Analysis
All data were analyzed using SAS Version 9.4 (Cary, North Carolina). Descriptive statistics were calculated to describe the PWD and the caregiver, with means and standard deviations for all continuous variables and frequencies and percentages for all categorical variables. Cronbach’s alphas were used to determine the internal consistency and reliability among the frequency and reaction scales and subscales (Memory, Disruption, and Depression). To determine differences among the demographic characteristics and the reaction scales and subscales, independent two-sample t tests and ANOVA were used.
To evaluate the impact of the intervention on the reaction to behavior outcomes, linear mixed models were developed. Caregivers who had at least one follow-up data collection period were included in the analysis. A linear mixed model was used because of the repeated measures for each participant and its ability to handle missing data over time. In other words, individuals are not removed from the model if partial information is contained (i.e., time points). This reduces the bias in the statistical analysis that may be present if we only analyzed cases with complete data. In each model, time was considered a fixed effect. Subgroup analyses were conducted to examine differences among the demographic characteristics (gender, employment status, relationship to caregiver, financial strain, education level, and depression status). Interactions with time were considered for all models and each model controlled for year of entry into study. Post hoc tests were conducted using Tukey’s Method. All p values <.05 were regarded as statistically significant.
Results
Sample Characteristics
A total of 127 unpaid caregivers and PWD were included in the study. Among the PWD, the mean age was 77.9 years (SD = 8.2). The majority of the PWD were female (54.4%), and approximately 62.0% were in middle to late stages of their illness. Among the caregivers, the mean age was 66.5 years (SD = 12.4), and the majority of the caregivers were female (69.3%), were White (92.9%), were spouses (81.9%) to the PWD, had a college degree or higher (49%), and were unemployed (73.6%). Only 17.5% of the caregivers reported experiencing financial strain from caregiving (Table 1).
Baseline Demographic Characteristics of Patient and Caregivers (N = 127).
Other consists of sibling, offspring, and others.
Participant attrition was quite high over the 18-month data-collection period probably due to 62% of caregivers being enrolled while in moderate/severe stage of disease progression. Attrition resulted primarily from PWD institutionalization (n = 38) or death (n = 20). Other reasons for caregiver attrition included unable to contact (n = 12), no longer interested (n = 9), or other reasons (n = 11). Some additional caregivers were lost between enrollment and baseline (n = 15), possibly due to a lack of readiness to actively engage in an educational, skills-building intervention. Gitlin and Rose (2014) discuss importance of a concept of “readiness,” which is necessary before caregivers are able to actively engage in behavioral management interventions. Outcomes in behavioral management interventions thus depend upon readiness of caregivers to implement and continue to use new care strategies over time. Findings indicated that by treatment conclusion, more than 25% of the caregivers were not activated enough to use behavioral management strategies (Gitlin & Rose, 2016). Activation was needed to affect behavioral management outcomes but not other caregiver outcomes such as burden and well-being.
No significant differences were found in the demographic characteristics and scales and subscales between those who were lost to follow-up and those who remained in the study. These differences were checked overall (e.g., between those who remained in the study and those who were lost to follow-up at any point during the study) and at each time point (e.g., those who remained in the study at 6-month follow-up vs. those who dropped out by the 6-month follow-up, etc.) to assess potential bias (see Table 3 for numbers of caregivers remaining at each data-collection period).
The mean overall frequency rating of behavior scores among the patients was 39.5 (SD = 13.9). The three subscales were Memory Problems, Depression, and Disruption, and their means were 21.0 (SD = 5.2), 10.3 (SD = 7.1), and 8.6 (SD = 6.0), respectively. For those PWD who exhibited the selected behaviors, their caregivers were asked to rate their reaction to the behaviors. The overall reaction score to the behaviors was 28.7 (SD = 16.6), and the mean subscales were 8.9 (SD = 5.1) for Memory Problems, 9.3 (SD = 6.7) for Disruption, and 11.0 (SD = 8.2) for Depression.
Table 2 presents individual behaviors, the PWD’s mean frequency ratings, and the caregiver’s mean reaction to those behaviors. Although the highest rates of occurrence were among memory problems, the strongest reactions were prompted by depression and disruption behaviors.
Mean PWD Frequency and Caregiver’s Reaction Scores at Baseline (N = 127).
Note. PWD = persons with dementia.
Includes those individuals where the action happened at least once.
Mean includes behaviors that had never occurred.
Mean computed only for those where the behaviors occurred at least once.
Frequency of PWD’s behaviors
A total of 85 caregivers (66.9%) had at least one follow-up measurement and were included in the linear mixed models. There were no significant decreases over time in overall frequency of behaviors (p = .19). The mean frequency scores of behaviors at baseline, 6 months, 12 months, and 18 months follow-up were 40.1 (SE = 1.5), 37.3 (1.5), 37.2 (1.8), and 37.9 (1.8), respectively. However, when observing just baseline and 6-month follow-up data, there was a significant decrease in frequency of behaviors (p = .03). There were no significant decreases over time for frequency of memory problems (p = .35), disruption behaviors (p = .38), and depression behaviors (p = .54) among PWDs, as the frequencies at 6 months remained stable over the 18-month follow-up.
Caregiver’s reaction to PWD behaviors
Over time, the reaction to the challenging behaviors significantly decreased (p = .009) suggesting a potential effect of the intervention. The mean reaction to the behaviors decreased from 29.5 (SE = 1.8) at baseline to 23.5 (SE = 2.1) at the 18-month follow-up (Table 3). A post hoc analysis showed significant differences in mean reaction to behaviors between baseline and 12 months (p = .02) and baseline and 18 months (p = .03). There were also significant decreases in the reaction to memory problems (p = .002) and disruption over time (p = .04; Table 3). Figure 2 shows the mean scores for the overall reaction to behaviors and the three subscales over the four time points.
Linear Mixed Model Results for Each Reaction to PWD Behavior Scales and Subscales With Adjusted Means by Time (N = 85).
Note. Each model controlled for time of entry into study. PWD = persons with dementia.
Plots of the mean reactions of the caregivers to the patient’s behaviors over the four time periods.
Table 4 presents the mean reaction to behavior scores (total score and subscales) at baseline by caregiver demographic characteristics. Caregivers with a financial strain had significantly higher mean reaction scores to overall behaviors compared with those with no financial strain (M = 35.9 vs. M = 27.2, p = .03). For the subscales, spouses had significantly higher reaction scores to memory behaviors compared with others (M = 9.4 vs. M = 7.0, p = .045), and those with a financial strain had significantly higher reaction scores to depression behaviors compared with those with no strain (M = 14.4 vs. M = 10.3, p = .03).
Mean Caregiver’s Reaction to Behavior Scores at Baseline by Caregiver Demographics (N = 127).
To examine differences in caregiver’s reaction to PWD’s behavior over time among subgroups (e.g., gender, race, education), separate linear mixed models were developed with time as a fixed effect, year of entry into the study, and a time by subgroup interaction. In the mixed models, there was a significant interaction between relationship of caregiver to PWD and time with regard to overall reaction scores (p = .007). Figure 3 shows the mean scores over time for spouses compared with others. From baseline to 6-month follow-up, both spouses and others had decreases in reaction scores, but the adult children/other caregivers observed increases in reaction scores, while spouses’ reactions continued to decline. There was also a significant interaction between relationship and time with regard to the Disruption subscale (p = .004). Figure 3 presents the mean reaction to disruption behaviors over time for the two groups. Figure 3 also shows significant differences between gender and reaction to memory behaviors over time (p = .01). The reaction to memory behaviors remained constant among female participants throughout the follow-up period, whereas males reported slight decreases in their reaction to memory behaviors.
Plots of the mean reactions of the caregivers to select patient’s behaviors over the four time periods by caregiver characteristics.
Discussion
The individualized, two-component intervention targeted the most upsetting PWD behavioral symptoms for management strategies to decrease symptoms and included a family session to educate participants about the disease process, improve communication, and enhance use of community resources. Overall, the intervention significantly decreased the caregivers’ reaction to PWD behavioral symptoms over the 18-month study period, thus improving their responses to behavioral symptoms. Although improved responses to behavioral symptoms were sustained over time, a significant decrease in frequency of behavioral symptoms was seen only from baseline to 6 months. After 6 months, frequency of behavioral symptoms remained stable.
Behaviors That Produced the Most Negative Response
At baseline, prior to introduction of the intervention, the most negative caregiver reactions occurred for depressive behavioral symptoms. Although the most frequently occurring behavioral symptoms were memory symptoms, the highest upset reactions occurred for individual dangerous behaviors such as “threats to hurt self.” This suggests that particular behaviors, albeit even with low rates of occurrence, are associated with stronger reactions from caregivers. For example, the mean rate of occurrence for “Threats to hurt self” was 0.2 (SD = 0.56), but the mean reaction to the behavior was 2.6 (SD = 1.41). Future behavioral interventions must include more specific management skills for PWD with depressive behaviors, including how to recognize and manage potentially dangerous behaviors. Because caregivers dealing with dangerous behaviors are at highest risk of placement of PWD in a nursing home (Gaugler et al., 2010), these caregivers may benefit from additional services (such as personal counseling) in addition to receiving nonpharmacological interventions.
Caregivers Who Benefited the Most
Caregivers under financial strain had higher reaction scores, especially for depressive behaviors. Caregivers under financial strain may be so stressed about money issues that they lose the capacity to cope with PWD depressive behaviors.
Prolonged stress was a primary predictor of nursing home placement, and financial strain was identified as a main stressor (Spillman & Long, 2009). Readiness to use a nonpharmacological intervention improved as financial status improved. Because distress from financial strain may prevent readiness to use an intervention, providers must address caregiver financial status as part of a comprehensive assessment and assist caregivers with financial needs by making referrals for financial counseling or legal services (Gitlin & Rose, 2014).
Spouses had the highest reaction scores to memory behaviors. Gender differences also occurred in reaction to memory behaviors with a slight reduction in reaction to memory problems occurring for men while reaction scores for women remained constant. Similarly, an earlier community-based PLST investigation (Gerdner et al., 2002) found that male caregivers reported fewer problematic behaviors and were also less bothered by behaviors than female caregivers. In addition, in the current study, at baseline, older caregivers (primarily spouses) reported the highest number of behavioral problems, but it was the youngest cohort of caregivers (primarily adult children) who reported the most distress in reaction to the behaviors.
Findings from the study presented here support an earlier controlled trial (n = 132 experimental participants; n = 105 comparison group participants) using a similar PLST-based intervention to examine the impact of a longitudinal, multi-site, community-based intervention on frequency and caregiver response to behavioral problems in PWD (Gerdner et al., 2002). In that study, the PLST-based intervention had a positive effect on both frequency of and reaction to PWD behaviors, although effects varied by relationship. More specifically, an intervention effect was found on frequency of memory/behavioral responses for non-spouses only. Similarly, early research using the PLST model in a nursing home setting also found no effect for the intervention on the cognitive or functional status of residents; however, staff caregivers reported positive outcomes, and residents with dementia experienced fewer catastrophic behaviors (Swanson, Maas, & Buckwalter, 1993).
Importantly, the intervention benefited all caregivers, that is, both spouse and non-spouse, in terms of their response to memory/behavioral problems. Other studies investigating psychoeducational family group interventions (Ostwald, Hepburn, Caron, Burns, & Mantell, 1999) reported no impact on the frequency of behaviors as reported by caregivers but did note that negative responses of caregivers to PWD disruptive behaviors were significantly reduced.
Limitations
The major limitation of the current study is that a randomized control group was not possible, which thus limits the conclusions that might be made about the effectiveness of this intervention. The setting in which this study took place (SON Caregivers Program of Research) requires that all caregivers receive the intervention. Discussions in the literature do offer support for intervention research without a control group because most randomized controlled trials (RCTs) do not lead to translation at the community level. Wethington and Burgio (2015) argued for giving attention to findings from other research designs rather than use of only RCTs when designing translational trials. In particular, attention must be given to those studies that test interventions in more diverse settings. Most communities do not have the needed resources to feasibly adapt the RCT intervention for use. For example, cost effectiveness is another issue of little concern in RCTs but becomes a priority regarding feasibility in real-world settings.
Generalizability of findings to a much larger population of caregivers is another criterion of validity. The national profile of caregivers in the United States (AA, 2015) describes caregivers as follows: 65% women; 21% aged 65 or older; 44% having some college or a college degree; 64% currently employed, a student, or homemaker. The majority were married (71%), Caucasian (66%), and took care of parents (55%). Thus, another limitation may be that our results lack external validity as our sample was better educated (49% vs. 44%), Caucasian (92.9% vs. 66%), spouses (81.9 vs. 45%), and generally financially solvent with only 17% reporting caregiving to be a financial hardship. In addition, caregivers who enroll in a SON caregiver program of intervention research may be quite different from caregivers who just try to cope on their own.
As noted earlier, rates of attrition in this study were high. As previously discussed, the second component (family meeting) of our intervention came from the work of Mittelman who had similar attrition problems. Mittelman and Bartels (2014) discussed implications of their high attrition rate (54.8%) before completion of all six sessions. Their higher attrition rate found in translational community interventions contrasts to high completion rates in Mittelman’s initial RCT (98.3%). Reasons for higher attrition in real-life settings include differences in caregivers who agree to participate in a RCT versus those who participate in real-life interventions, the latter may be more similar to caregivers seeking social services.
A last limitation is that multicomponent interventions result in an inability to tell which component achieved the desired outcomes. Understanding which component contributes to which outcome will be useful for future tailoring of the intervention to specific caregivers and families (Van’t Leven et al., 2013).
Implications for Future Research
Currently available pharmacological interventions for behavioral and psychological symptoms of dementia suffer from limited efficacy and a high potential for adverse reactions (Casey, Antimisiaris, & O’Brien, 2010). There is an urgent need to identify effective nonpharmacological measures, which may assist caregivers in managing depressive and dangerous behaviors (Casey, 2015). If best-practice dementia guidelines included information about nonpharmacological interventions reviewed by Brodaty and Arasaratnam (2012), physicians and other providers might be guided to initially recommend psychosocial interventions as a first line of treatment over pharmacological treatments (Vasse et al., 2012). Thereby, referrals would be made before initiating pharmacological treatments, and caregivers would be enrolled in supportive, educational interventions much earlier in the disease process. In this way, more knowledge might be gained to answer questions concerning when to start and stop interventions, dosing, and which components are most effective for improving behavioral symptoms and caregiver reactions (Brodaty & Arasaratnam, 2012). Because 62% of caregivers in our study reported caring for someone in the moderate to late disease stage, future studies must enroll caregivers and PWD at earlier stages of the disease process to decrease attrition due to nursing home admission or death. Caregivers of PWD approaching later, more severe disease stages are probably receiving the intervention too late.
Clearly, future replications of this work will need to introduce random assignment and a control group to better delineate the effects of this specific intervention versus treatment as usual, or even the natural history of the disease. In this way, it may be possible to determine whether the significantly decreased caregiver response to challenging symptoms observed in this study was due to the intervention or a result of caregiver adjustment and accommodation to symptoms as they became more seasoned in the caregiving role. Future studies might exclude PWD and caregivers using certain psychopharmacological treatments to differentiate results of the PLST intervention versus antipsychotics.
In the race to find pharmacological “cures” to prevent and delay onset of dementia, results of nonpharmacological studies such as the one reported here deserve more attention, so that caregivers and PWD will be more strongly supported with education about management of dangerous behaviors and how to cope with depressive responses. Similarly, health policy changes must address financially strained PWD and their caregivers, in an effort to improve their quality of life and potentially delay premature nursing home admission.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.