Abstract
Introduction
The prevalence of Alzheimer’s disease (AD) and other forms of dementia is rapidly increasing world-wide due, in part, to rising life expectancies (Prince et al., 2013). Most care for individuals with dementia is provided at home in the community by family members (Alzheimer’s Association, 2014). This is unlikely to change in light of the tremendous costs of long-term out-of-home residential care and family preferences for keeping individuals at home as long as possible. Thus, there is an urgent need for widespread dissemination of effective interventions that can delay institutionalization and improve quality of life for families living with this disease. In January 2011, President Obama signed into law the National Alzheimer’s Project Act, which mandates, in part, the development and dissemination of evidence-based interventions for persons with AD and their caregivers, and the expansion of coordinated services by state and local community-based organizations serving these individuals (U.S. Department of Health and Human Services, 2013). Since 2002, the U.S. Administration on Aging (AoA) has funded more than 39 projects to test whether interventions that have shown positive outcomes in randomized controlled trials are also feasible and effective in diverse community populations or when delivered by different community agencies and service providers (Maslow, 2012). This article describes one such recently completed AoA-supported project to translate and implement the evidence-based STAR–Community Consultants program (STAR-C) within the Oregon Department of Human Services—State Unit on Aging and Area Agencies on Aging (AAAs) serving both urban and rural areas in the state of Oregon.
STAR-C is one of a series of theoretically grounded treatment protocols for training caregivers to improve care and reduce behavioral distress for persons with dementia (e.g., Logsdon et al., 2010; McCurry, Gibbons, Logsdon, Vitiello, & Teri, 2005; Teri et al., 2003; Teri, Logsdon, Uomoto, & McCurry, 1997). The theory, clinical framework, and empirical evidence for these protocols have been reported elsewhere (Teri, Logsdon, & McCurry, 2005). Family members are taught a set of core behavioral strategies that include dementia education, effective communication, ABC (activator–behavior–consequence) problem-solving strategies, and implementation of pleasant events to reduce mood and behavioral disturbances in community-dwelling persons with dementia (Teri, McCurry, Logsdon, & Gibbons, 2005). Caregivers also learn strategies to improve their own emotional well-being and reduce their adverse reactions to challenging behaviors. STAR-C is conducted in eight weekly sessions in participants’ homes over 2 months, followed by four monthly phone calls (Logsdon, McCurry, & Teri, 2005). In the original randomized controlled trial (RCT; Teri, McCurry, et al., 2005), 95 persons with AD and their caregivers were randomly assigned to STAR-C or routine medical care (RMC) and evaluated at baseline, 2 months post-treatment, and 6-month follow-up. Consultants were master’s-level health care professionals who were practicing in the community at agencies serving older adults. Compared with RMC, caregivers assigned to STAR-C showed significant (p < .05) improvements in depression, burden, and reactivity to care-receiver behavior problems. Care-receivers experienced significant reductions in the frequency and severity of individually targeted behavior problems (p < .0001) and improved quality of life (p < .05). Both sets of results were maintained at 6-month follow-up.
In 2009, the Oregon Department of Human Services–State Unit on Aging received an AoA Alzheimer’s Disease Supportive Services Program (ADSSP) grant (#90AE0334, J.M., PI) to implement STAR-C in the state of Oregon. The goals of the grant were to evaluate the feasibility of adding STAR-C to existing Oregon AAA caregiver services, and to identify factors that would facilitate local long-term sustainability of the program. This article describes recruitment and implementation of the STAR-C Oregon Translation Study (OTS), presents caregiver and care-receiver outcomes from the 3-year project, and presents recommendations to facilitate long-term sustainability of STAR-C in Oregon and other potential sites.
Method
Human Participants
The study was approved by the Oregon Public Health Institutional Review Board (#09-31).
Recruitment
Printed brochures were developed in the first year of the grant. Existing AAA case-managed clients were reviewed to identify potential participants. AAA regional coordinators also distributed brochures at community presentations to medical providers, elder law attorneys, adult day and other dementia care programs, faith-based organizations, senior meal sites, and libraries. Lastly, information about STAR-C was provided to the Alzheimer’s Association National Helpline, as well as to local newspapers that featured articles about the program.
Study Participants
One hundred eighty-nine dyads were screened for participation in the OTS. Of these, 158 were eligible to participate, 151 chose to enter treatment, 96 returned post-treatment assessment packets, and 60 returned the 6-month follow-up packet (Table 1).
Enrollment Into the Eight-Session STAR-C OTS.
Note. STAR-C = STAR–Community Consultants program; OTS = Oregon Translation Study; AAA = Area Agencies on Aging. All group comparisons p > .05.
Completed STAR-C Program = (home visits and post-treatment assessment).
Consistent with the RCT, the target population for the OTS was community-dwelling persons with dementia and their family caregivers. Care-receivers ranged in age from 58 to 100 years, were equally divided between men and women, and were predominantly White (67%). Caregivers were predominantly spouses (66%) or adult children (31%) whose ages ranged from 40 to 102 years; 74% were female, 66% were White, and more than half had been providing care for three or more years. Fifty-nine percent of the enrolled dyads were from urban regions, and 41% were rural. As shown in Table 2, participants in the OTS were significantly older, more ethnically diverse, and caregivers were more likely to be spouses than in the original RCT.
Comparison of Baseline Characteristics of Eligible OTS Participants (Who Completed the Baseline Assessment) Versus Original RCT Data.
Note. Values are expressed as percentages unless otherwise indicated. OTS = Oregon Translation Study; RCT = randomized controlled trial (Teri, McCurry et al., 2005); QOL-AD = quality of life-Alzheimer’s Disease; SCB-SB = Screen for Caregiver Burden–Subjective Burden; CES-D = Center for Epidemiologic Studies Depression Scale.
p < .05; **p < .01; ***p < .001.
The original RCT required care-receivers to have a diagnosis of probable or possible AD from their primary physician. Because a dementia diagnosis is not routinely required for AAA services, Oregon relied on telephone screenings with caregivers to identify care-receivers who had at least 6 months duration of symptoms consistent with AD or a related dementia, and who did not have other neurological (e.g., Parkinson’s) or major psychiatric (e.g., bipolar) diagnoses. For both the RCT and the OTS, caregivers were required to live with the person with dementia or spend at least 4 hr/day with him or her, to ensure that the caregiver could reliably identify and intervene with care-receiver’s mood or behavior challenges.
Procedures
As in the original RCT, OTS assessments were collected at baseline, 2 months post-treatment, and 6-month follow-up. In both the RCT and OTS, study coordinators explained the STAR-C program and screened potential dyads over the phone. In the original RCT, all other enrollment and assessment visits were conducted in person by trained interviewers. In the OTS, all data were collected by mail, including consent, registration, initial assessment forms (completed and returned prior to the initial STAR-C home visit), and post-treatment and follow-up assessments. Assessment packets were returned by mail to the Oregon State Unit on Aging for checking and data entry.
Intervention
STAR-C treatment protocol
As in the original RCT, OTS consultants met with caregivers in their homes for eight weekly sessions, followed by four monthly phone calls. An overview of the session-by-session STAR-C protocol content is shown in Table 3 (Logsdon et al., 2005; Teri, McCurry, et al., 2005).
Outline of STAR-C Treatment Sessions.
Note. STAR-C = STAR–Community Consultants program; ABC = activator–behavior–consequence.
Community AAAs and case managers
Two Oregon AAAs serving one urban and two rural Oregon counties participated in the study (Multnomah County Aging & Disability Services [ADS]; Rogue Valley Council of Governments Senior & Disability Service [RVCOG]). The Multnomah County ADS serves a primarily urban county (including the city of Portland) with a relatively diverse racial and ethnic population, contracting with five district/senior centers spread across the county to carry out a number of services, including case management. The Multnomah County ADS identified 11 case managers from five district centers to be trained as STAR-C consultants.
Rogue Valley Council of Governments Senior & Disability Services serves a rural two-county area in southern Oregon and identified four existing AAA staff case managers and two private geriatric care managers to be trained as STAR-C consultants.
In the original RCT, master’s-level mental health professionals were contracted to serve as STAR-C consultants. In the OTS, STAR-C consultants had a range of educational backgrounds (1% high school graduate, 69% BA level college, 29% MS level college, 1% PhD). Consultants were required to have experience working with persons with dementia and family caregivers, and reported having between 4 and 29 years of experience working with aging services.
Consultant training
Potential STAR-C consultants participated in a 1½-day group training conducted by University of Washington (UW) trainers (S.M., R.L., D.L.F.). This training included presentation of information about the original RCT results, description of the core STAR-C behavioral components, review of the treatment manual, role playing, and structured practice creating behavioral problem-solving plans using a video series developed by Dr. Linda Teri and colleagues (Teri & Huda, 2004). Three STAR-C training workshops were held in Oregon between 2009 and 2013.
After attendance at these workshops, STAR-C consultants completed one pilot case for which the eight in-home sessions were audio-recorded. Recordings were either uploaded to an electronic site that could be immediately accessed by UW trainers, or transferred to CDs and mailed to UW the next day. This allowed consultants to receive corrective feedback and implement changes in a timely fashion. UW trainers reviewed the audio recordings, completed a consultant adherence checklist, and provided individualized feedback to consultants. This feedback to consultants included comments about any STAR-C material that was omitted or incorrectly presented. Feedback included suggestions to enhance presentation of the clinical material and to help caregivers brainstorm their own problem solutions with consultant guidance. Successful completion of at least one pilot case, with adequate adherence and demonstrated skill acquisition, was required for certification as a STAR-C consultant. A total of 16 individuals were certified; one additional person was trained but did not complete certification requirements.
Consultant treatment fidelity
Because one of the aims of the OTS was to examine the feasibility of training AAA case managers to administer STAR-C, after certification two randomly selected session recordings from each case were reviewed by UW trainers for fidelity to the treatment protocol. This was in contrast to the original RCT, where clinical supervisors reviewed every audio-recorded STAR-C session. OTC consultants were encouraged to contact trainers with any questions or concerns that they had about STAR-C concepts or how best to present the material to clients with particularly challenging behaviors or situations. Once certified, consultant fidelity was maintained through (a) participation in periodic group conference calls with UW investigators, (b) ongoing supervision with their individual regional coordinators to address clinical issues or logistical difficulties, and (c) attendance at subsequent training workshops, both as a way to refresh their own knowledge about the program and to enhance the group training by sharing their experiences with new consultants.
Measures
Caregiver and care-receiver outcomes
Caregiver outcomes included the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977) and the Screen for Caregiver Burden–Subjective Burden subscale (SCB-SB; Vitaliano, Russo, Young, Becker, & Maiuro, 1991). The CES-D assesses frequency of depressive symptomatology during the past 2 weeks; it has good internal consistency (Cronbach’s α = .92; Fountoulakis et al., 2007), is sensitive to depression in caregivers, and is responsive to change over time (Teri, Logsdon, Wagner, & Uomoto, 1994). Scores of 16 or higher are considered indicative of depression.
The SCB provides scores for both objective burden (number of potentially negative experiences) and subjective burden (caregivers’ reported distress in response to the experiences). The SCB was designed specifically for caregivers of persons with dementia, is reliable (α = .85-.89; Vitaliano, Young, & Russo, 1991), is sensitive to change over time, and correlates with both caregiver and care-receiver change (Vitaliano, Russo, et al., 1991).
The Quality of Life–Alzheimer’s Disease (QOL-AD; Logsdon, Gibbons, McCurry, & Teri, 2002) assessed care-receiver quality of life. Because all data were collected by mailed questionnaires, the caregiver proxy version was completed by caregivers about their care-receivers. The caregiver proxy version of the QOL-AD has good internal (α = .87) and test-retest reliability (intraclass correlation = .92), and its validity is supported by correlations with lower levels of depression, better day-to-day functioning, and higher levels of activity in care-receivers (Logsdon, Gibbons, McCurry, & Teri, 1999).
In addition to the above measures, which were part of the original STAR-C randomized trial, the OTS included two questions taken from the Tailored Caregiver Assessment and Referral (TCARE) tool (Montgomery & Kwak, 2008) about whether caregivers would consider placing the care-receiver in long-term care if his or her condition stayed the same or worsened. A modified version of the Revised Memory and Behavior Problem Checklist (RMBPC; Teri et al., 1992) was also administered, in which caregivers were asked to rate the presence or absence of 24 memory, depression, and disruptive dementia-related behaviors, and how bothered they were by behaviors that had occurred. This modified measure included the same items as the original RMBPC, which has good reliability (α = .84), validity confirmed through comparisons with other indexes of mood and behavior (Teri et al., 1992), and has been shown to be sensitive to change in other studies (Burns, Nichold, Martindale-Adams, Graney, & Lummus, 2003).
Caregiver treatment enactment
Caregivers were assigned homework at every session to be reviewed the following week. This homework was specific to the session, and was also individualized to the dyad; for example, pleasant event scheduling was assigned for the week following the fifth session, but the specific pleasant events that were scheduled varied from person to person. Caregivers demonstrated their understanding of core STAR-C concepts (realistic expectations, good communication, ABC problem solving, and pleasant events) through the development of individualized behavioral plans for their care-receiver. Plans were developed both in-session with the help of the STAR-C consultant, and between sessions using blank ABC problem-solving and pleasant event planning forms. Weekly homework was reviewed every session, and strategies for overcoming obstacles or revising behavioral plans were developed. Consultants also completed weekly progress notes rating caregiver homework compliance and understanding to monitor caregiver enactment of STAR-C concepts.
During each of the four monthly telephone calls that followed the final in-home session, consultants reviewed the STAR-C concepts, asked caregivers whether they were continuing to use them in daily care, and assisted caregivers in brainstorming new ABC and pleasant event plans.
Statistical Methods
Change scores from pre- to post-treatment were compared using paired t tests for participants with data at both time points. Longitudinal analyses included both post-treatment visits (2 and 6 months) and time, controlling for pre-treatment value of the outcome, and using generalized estimating equations with a normal link function, robust standard errors, and an exchangeable correlation structure. This regression-type model for panel data provided consistent estimates of parameters of a generalized linear model and assumed equal correlation between successive measurements. This method allowed for the inclusion of all available participant data (rather than dropping cases in which there were any missing data points).
Supplemental analyses were conducted to assess the impact of those cases lost to follow-up. Missing data were imputed at post-treatment and 6-month follow-up using the last observation carried forward, then pre- to post-treatment and longitudinal analyses were repeated. The impact of demographic and baseline characteristics on the outcomes of those who did or did not complete the post-treatment and 6-month follow-up assessment were compared.
Results
Sixty-four percent of OTS caregivers (n = 96) who provided pre-treatment questionnaires completed the STAR-C program and returned mailed post-treatment assessment packets. Caregivers had significant (p < .001) reductions in self-reported depression, subjective burden, and reactivity to care-receiver behavior problems on the modified RMBPC (Table 4). Care-receivers had significant reductions in depression (p = .02) and improvements in quality of life (p = .01). Effect sizes ranged from small to moderate. There were no significant differences in caregiver ratings of their likelihood to place the care-receiver in long-term care if his or her condition stayed the same or worsened.
STAR-C Pre- and Post-Intervention Outcomes for Participants in OTS.
Note. Post-test is 2 months after baseline. Participants with data at both time points range from 78 to 89 (one more than the degrees of freedom) depending on which completed measures were returned. Higher scores are worse outcomes for all outcomes except QOL-AD. For placement outcomes, higher scores imply greater likelihood of long-term placement. STAR-C = STAR–Community Consultants program; OTS = Oregon Translation Study; CES-D = Center for Epidemiologic Studies Depression Scale; SCB-SB = Screen for Caregiver Burden–Subjective Burden; QOL-AD = Quality of Life–Alzheimer’s Disease; RMBPC = Revised Memory and Behavior Problem Checklist.
Paired t tests, two-tailed.
RMBPC subscale items administered to rate presence or absence of each behavior problem (not frequency over past week). Caregiver reaction is summed for entire 24-item scale.
At 6 months, 40% of caregivers (n = 60) returned the follow-up assessment packets. Longitudinal analyses, adjusting for pre-treatment values, indicated that over 6 months of follow-up, significant reductions in caregiver reactivity to behavior problems (b = −1.07, z = −3.53, p < .001), and care-receiver depression (b = −0.14, z = −3.2, p = .001) were maintained.
Supplemental analyses using imputed data showed a similar pattern of results for both the pre- to post-treatment and longitudinal analyses. Demographic and outcome measure characteristics did not differ between completers and non-completers at pre-treatment except that caregivers who completed all three assessments were more likely than non-completers to have been caregiving for three or more years, χ2(2) = 6.1, p < .05.
After the final follow-up call, caregivers were asked to rate their program satisfaction. Ninety-two percent of caregivers who returned the survey (n = 79) felt the program had been very or somewhat helpful, and 80% said they would definitely or probably continue to use STAR-C skills.
Discussion
Findings from the OTS indicated that STAR-C can be conducted by AAA community partners, yields positive caregiver and care-receiver outcomes, and is viewed as helpful by both caregivers and coaches. Similar to results from the original RCT (Teri, McCurry et al., 2005), at post-treatment caregivers reported reductions in depression, subjective burden, and reactivity to care-receiver behaviors. Reductions in care-receiver depression and improvements in care-receiver quality of life were also reported. At the 6-month follow-up, reductions in caregiver reactivity to behaviors and care-receiver depression were maintained. Participating AAAs and coaches were enthusiastic about offering STAR-C to their clients, and have continued to make the program available since AoA grant funding ended in June 30, 2013.
It should be noted that the intentional design of the OTS (no randomization or comparison condition, so that all interested caregivers could participate in the program; use of mailed assessment packages rather than in-person trained interviewers; simplification of the RMBPC subscales to make it easier for caregivers to complete) precludes direct comparison of these findings with the original RCT. For example, although the pattern of results that were observed in Oregon (e.g., improvements in caregiver depression, burden, and reactivity to behavior problems) were similar to those reported in the original RCT, we are unable to compare the size of treatment effects directly because OTS data are based upon within-person comparisons over time, and the RCT was based on comparisons to a control group. Nevertheless, the OTS data support the potential for evidence-based programs like STAR-C to be delivered widely within a state system of care, and to produce positive benefits for community participants receiving them.
Lessons Learned
The Oregon STAR-C translation project has been successful in many ways. Client outcomes were similar to those obtained in the original RCT, despite having a more diverse client population, more depressed caregivers, consultants with less formal education, and less tightly controlled assessment protocols. Careful attention to treatment fidelity during the certification process has ensured that the cadre of Oregon STAR-C consultants is well trained according to standards developed by the program originators. However, the translation of any evidence-based program into the community is not without challenges. Some of the lessons learned as we engaged in this expansion will be helpful for future sites looking to implement such programs.
First, the STAR-C consultants are critical to the success and effectiveness of the program. Oregon chose to implement STAR-C in part because existing case managers could be trained to offer it, but consultants must learn to “wear a different hat” during their STAR-C sessions, focusing less on direct advice-giving and more on training STAR-C concepts and empowering caregiver participants to brainstorm their own ideas. The Oregon project intentionally did not allow consultants to provide the intervention to individuals in their own case management portfolios, but in many locations this separation of roles would be neither feasible nor desirable. Subsequent experience has shown that the “two different hats”—that of a STAR-C consultant and that of a case manager—are highly compatible and can be readily integrated in most situations.
Second, as has been found in other efforts to translate evidence-based caregiver interventions (Primetica, Menne, Bollin, Teri, & Molea, 2013), intensive community outreach is critical for successful recruitment of program participants. Initially, the Oregon agencies believed that capturing participants through existing caseloads would be sufficient. They also anticipated that large numbers of caregivers would have great interest in a free of charge, evidence-based program that was offered in their own homes. It was therefore initially a surprise to find recruitment for the program was an ongoing effort. In retrospect, both AAAs wished that they had started using paid advertisement more actively from the beginning of the grant to help develop familiarity and interest in the program. Outreach to community health care providers was also critical for generating referrals. The role of the Alzheimer’s Association Chapter was (by their choice) limited to participation in a project advisory group and provision of general information about dementia care resources. Developing more active involvement between Alzheimer’s Association Chapters and AAAs in future studies might help with outreach.
Third, data collection is valuable but must be balanced against staff burden and the other competing time and cost demands of community caregiver programs. As the Oregon AAAs do not routinely or systematically track their program outreach efforts, reasons for STAR-C program’s ineligibility, refusal, or discontinuation were not gathered during the OTS. Thus, we are unable to determine what client characteristics were associated with non-participation and how those might have differed between sites. The OTS also relied on mailed consent and assessment forms rather than hiring staff to collect in-person outcome data, and in some cases, caregivers simply did not return the mailed post-treatment and follow-up assessment packets. Once AoA funding ended, follow-up assessments had to be discontinued altogether. Although ongoing collection of program eligibility, recruitment, and outcome data would be invaluable for future program planning and evaluation, already overburdened agencies may be unable to collect such data without additional funding support.
Last, but not least, the long-term sustainability of STAR-C (indeed any program) requires diligent attention and planning. Since grant funding ended, both AAAs have continued to offer the STAR-C program but overall costs and demands on staff time are a challenge given limited budgets. AAAs raised concerns early on about the sustainability of delivering eight weekly, in-home sessions, particularly for clients who lived in more remote locations where increased travel time and expense greatly increased intervention costs. As the OTS approached the end of AoA funding, discussions were held between the State Unit on Aging, AAA directors, and UW investigators to explore alternative delivery options, including possibilities such as offering STAR-C groups, switching to a telephone-delivery model, requiring fewer in-home visits, and/or scheduling sessions in consultants’ offices rather than caregivers’ homes.
In response to the need for the AAAs to reduce the costs of service delivery, UW investigators reconfigured the original STAR-C from eight in-person sessions to four in-person sessions plus two telephone sessions (STAR-C2). This reconfigured STAR-C2 program—with in-home visits during Weeks 1, 2, 4, and 6, and telephone sessions in Weeks 3 and 5—includes the same content as the original STAR-C program. The AAAs have begun to test this program, and report that this reconfiguration has reduced concerns about staff time and program delivery costs, and has been helpful for recruitment of caregivers who had previously been reluctant to commit to so many in-person sessions requiring respite care. Although additional research is needed to evaluate the efficacy of this reconfigured version of STAR-C, caregiver and consultant feedback about the program has been very positive. Because the content of STAR-C was maintained in the reconfigured program, we believe the outcomes will be similar to the 8-week program, but at this time, systematic outcome data are not available.
In closing, translation of an evidence-based program into the community brings considerable benefits to the agencies involved in providing care, the beneficiaries of that care, and the researchers seeking to continually improve and understand effective treatment approaches. The day-to-day steps involved in such translation allow all involved to gain a unique understanding and respect for the diversity of skills needed to successfully complete such an undertaking. From the mundane (developing forms and procedures, going through IRB review, training new staff, and developing certification protocols) to the sublime (experiencing firsthand the successes of coaches and participants), there is no doubt that the effort is worth it. Based on Oregon’s work with STAR-C, there is a better process for replication that will allow other agencies to adopt the program and enable their consumers to benefit. As more states and agencies join Oregon’s ranks, we look forward to continuing to share our successes and learn from our failures. Only by engaging in this process can we hope to continue to develop and improve the care of older adults with dementia.
Footnotes
Acknowledgements
Appreciation is extended to the staff of the Northwest Research Group on Aging and the Oregon Department of Health Services for their commitment to this project and their dedication to improving the care of older adults with dementia.
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 author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data collection, analysis, and preparation of this paper was supported, in part, by grants from the Administration on Aging Alzheimer’s Disease Supportive Services Program (#90AE0334, J Mead, PI) and a contract from the Washington State Aging & Disability Services Administration (#1165-37431, MS McCurry, PI).