| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
|
|
|
|||||||
|
|||||||||
Correspondence: Address correspondence and requests for reprints to Professor Linda Teri, PhD, University of Washington, 9709 3rd Ave., N.E., Suite 507, Seattle, WA 98115-2053. E-mail: lteri{at}u.washington.edu
 |
Abstract |
|---|
 TOP Abstract MethodResultsDiscussionReferences |
|---|
Key Words: Alzheimer's disease • Anxiety • Behavior management • Caregivers • Depression
Over the past two decades, we have been working to develop a systematic approach to training community-dwelling caregivers to reduce mood and behavior disturbances in older adults with Alzheimer's disease. This approach, which has come to be called the Seattle Protocols (Teri, Logsdon, & McCurry, 2005), is grounded in gerontological and social learning theories. Gerontological theory addresses how problems experienced by older adults are influenced by the larger psychosocial context in which they occur (Lawton, 1990). Social learning theory addresses how this psychosocial context can be operationalized in order to understand how particular problems develop, maintain, and change (Bandura, 1977). Caregivers play a critical role in this formulation of treatment (Teri, Logsdon, Wagner, & Uomoto, 1994). They are responsible for organizing much of the patient's day and have frequent contact in providing daily care and social opportunities. Consequently, for individuals with dementia, improvements in their interactions with caregivers form a critical link in improving person–environment fit. Caregivers are taught to monitor problems, identify possible environmental or interpersonal triggering events, and develop more effective responses in order to improve patients' environment, maximize their abilities, and minimize their impairments. The integration of these two theories recognizes and accommodates the complexity of the environmental and interpersonal context in which problems and treatment occur. The Seattle Protocols approach is structured but tailored to address unique problems identified by individual patient–caregiver dyads. This approach been applied to a variety of dementia-related behavior problems, including the treatment of depression (Teri, Logsdon, Uomoto, & McCurry, 1997), agitation (Teri et al., 2000), physical inactivity (Teri et al., 2003), and sleep disturbances (McCurry, Gibbons, Logsdon, Vitiello, & Teri, 2005).
In the current study we investigated the extent to which health care professionals who were currently practicing in community settings serving older adults could be trained to use the Seattle Protocol in home-based counseling with family caregivers. To ensure that community consultants were, in fact, delivering training as developed, we closely monitored the extent to which consultants adhered to the treatment protocol, and we evaluated the relationship between treatment adherence and study outcomes. We hypothesized that (a) consultants would learn the protocol and be successful in implementing it, (b) caregivers receiving this intervention would show significant improvement on measures of depression, burden, and stress, and (c) care recipients would show improved mood, decreased behavioral disturbance, and improved quality of life.
|
Method |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
Participant Dyads
We recruited 95 family caregivers and care recipients with Alzheimer's disease for the study. Caregivers were a spouse or adult relative caring for a person with dementia in the home. Caregivers' ages ranged from 22 to 91 years; 69% were female, 86% were White, and 55% were spouses.
Care recipients had a diagnosis of probable or possible Alzheimer's disease (McKhann et al., 1984) from their primary physician. Their ages ranged from 53 to 93 years; they were predominantly female (66%), White (85%), and had had dementia for an average of 5.0 years. Their mean Mini-Mental State Examination (MMSE) score was 14.0 (SD = 7.0), which placed them in the moderate range of cognitive impairment. All care recipients were required to have three or more agitated or depressed behavior problems reported by their caregivers, occurring three or more times during the past week, at study entry (M = 6.0 problems, range = 3–11).
Procedures
We randomly assigned caregivers and care recipients to one of two treatment arms: STAR-caregivers (STAR-C) or routine medical care. We had assessments conducted at baseline, after 2 months (posttreatment), and at 6-month follow-up by interviewers blind to treatment assignment (Figure 1).
|
After consultants familiarized themselves with the manual, supervisors held a second training session to answer questions, discuss treatment logistics, and review the forms that consultants would complete during treatment. Supervisors also gave the consultants additional reading and videotape materials describing the Seattle Protocols approach (Teri, 1990; Teri, Logsdon, & McCurry, 2002; Teri & Schmidt, 1993) as well as general strategies for managing dementia-related behavior problems (Alzheimer's Association, 1990; Robinson, Spencer, & White, 1996).
Each consultant was assigned a pilot case, which had to be satisfactorily completed prior to working with enrolled participants. We arranged for all treatment sessions to be audiotaped. For this pilot case, consultants met weekly with a clinical supervisor who reviewed the tapes, answered questions, and provided feedback regarding consultants' adherence to the treatment manual.
Supervisors trained five consultants to performance criterion. (The sixth consultant began a pilot case, but, after the participants dropped out because of scheduling conflicts, the consultant chose to discontinue her involvement with the study.) Throughout the study, consultants turned in audiotapes and paperwork on a weekly basis. Clinical supervisors systematically reviewed treatment documentation and gave feedback to consultants as needed to ensure that caregiver training was being administered in a consistent fashion. We assigned consultants a varying numbers of cases, depending on scheduling availability given their other clinical commitments.
Treatment Protocol: STAR-Caregivers
Consultants met with caregivers in the caregivers' homes for 8 weekly sessions, followed by four monthly phone calls. Consultants conducted training privately with caregivers to ensure that they could talk freely without frequent interruptions or fear of upsetting the persons for whom they were providing care.
The first three treatment sessions focused on teaching caregivers the rationale and use of the A-B-C problem-solving approach to behavior change (Teri, 1990, 1994). Using examples from the caregiver's weekly diary, the caregiver and consultant brainstormed strategies for modifying antecedents or consequences of problem behaviors, and developed written behavior-management plans for the following weeks. Subsequent sessions focused on improving caregiver communication, increasing pleasant events as a means to improve care recipients' mood, and developing strategies to enhance caregiver support. In the four monthly follow-up calls, consultants helped caregivers develop behavior-management, communication, pleasant-event, and caregiver-support strategies for any new problems that arose. (A complete copy of the STAR-C treatment manual is available from L. Teri; more session-to-session details about the intervention are available in Logsdon, McCurry, & Teri, 2005).
We designed the STAR-C treatment to give consultants the freedom to use clinical judgment and modify the order of session topics in response to a caregiver's needs. In a few cases, it was necessary to extend treatment by one or two sessions to make up weeks that were missed as a result of illness or other unforeseen circumstances. All modifications were discussed with clinical supervisors, and all of the key content of the treatment protocol was addressed at some point during the intervention period. In this way, treatment reflected real-world practices in which clinicians have the flexibility to make on-the-spot decisions, without sacrificing standardization of the intervention.
Routine Medical Care
Caregivers and care recipients in the control condition received routine medical care, including acute medical or crisis intervention as typically provided by their own health care providers. This included nonspecific advice and support routinely provided by nurses and primary physicians or community support services. Control caregivers did not receive specific behavior-management training.
Measures
Treatment Integrity
One of our primary aims of this study was to assess our ability to train community consultants to administer the STAR-C training program. Several procedures helped evaluate consultant adherence to the treatment protocol: (a) Consultants completed a checklist, which rated the degree to which 12 protocol content domains were covered, including education about dementia symptoms and realistic expectations, understanding of the ABCs of behavior change, monitoring of assigned homework (readings, behavior diary, and implementation of problem-solving strategies), effective communication, assigning pleasant events, caregiver issues, and maintaining gains. (b) Consultants completed forms reporting caregiver compliance with homework and training recommendations. (c) Consultants rated whether caregivers had completed their assigned homework during the previous week (homework not assigned, not attempted, attempted but not completed, or completed).
In addition, clinical supervisors coded audiotaped treatment sessions by using a six-item Therapist Adherence Checklist form, which rated adherence to behavioral principles in the treatment sessions. These ratings provided a consistent format for supervisors to give feedback to consultants, and they provided a standardized way of rating consultant adherence to behavioral principles when delivering the STAR-C intervention. (This was important because not all consultants were behaviorally trained prior to participating in this study.) Supervisors rated consultants according to their review of homework; facilitation of behavioral problem solving; focus on current, observable events; avoidance of directive advice giving; and responsiveness to caregiver's more urgent issues. Supervisors also rated overall session quality. Supervisors reviewed consultant adherence and caregiver-compliance forms when reviewing the audiotapes, so they were able to verify that consultants were accurately reporting session events.
Finally, consultants measured caregiver exposure to the STAR-C training (the treatment dose) by recording the number and duration of sessions completed by each caregiver.
Target Problems
In the first session, consultants asked caregivers in an open-ended way to identify three problematic care recipient behaviors on which they would like to work. Once caregivers identified three behaviors, consultants helped them to clarify the behaviors of concern so they would lend themselves to ongoing observation and monitoring. For example, if the caregiver said the care recipient didn't sleep at night, consultants would determine if the problem of concern was that the care recipient wouldn't go to bed, that he or she would get up repeatedly during the night and wander, or some other aspect of sleep disruption such as nightmares or excessive daytime napping. Next, caregivers rated each problem on a 5-point Likert scale (0–4) indicating how frequently the problems had occurred in the past week (no occurrence to daily or more often), how severe the problems were when they occurred (trivial to severe), and how disturbing the behaviors were to the caregiver (not at all to extremely disturbing). Table 1 provides a sample of the behaviors chosen by caregivers. These were the target behaviors that caregivers monitored by using a daily behavior diary and that consultants used as the focus for behavior-management training. At the final visit (Session 8), consultants asked caregivers to again rate the frequency, severity, and level of disturbance they experienced from the same target behaviors that had been identified in the first session.
|
We evaluated caregiver stress and burden by using the Perceived Stress Scale (Cohen, Kamarck, & Mermelstein, 1983) and the Screen for Caregiver Burden (Vitaliano, Russo, Young, Becker, & Maiuro, 1991). The Perceived Stress Scale was designed to measure nonspecific, appraised caregiver stress during the past month. The Screen for Caregiver Burden is a 25-item questionnaire designed specifically for caregivers of persons with Alzheimer's disease. It provides scores for both objective burden (number of potentially negative experiences) and subjective burden (caregivers' reported distress in response to the experiences).
We used the Short Sense of Competence Questionnaire (SSCQ; Vernooij-Dassen et al., 1999) to evaluate the caregiver's feelings of competence to care for a demented person. The SSCQ has seven items measured on a 5-point Likert scale; published validity and reliability data are available (range =.76–.88; see Vernooij-Dassen et al.).
Care Recipient Outcomes
In the Seattle Protocols approach, caregivers are taught to view problems experienced by care recipients as modifiable behaviors that can be changed by using observation, goal setting, problem solving, and reinforcement. Care recipient outcomes provided a standardized measure of changes in overall behavioral disturbance and quality of life from the intervention. We rated overall behavioral disturbance in care recipients by using a combination of clinician ratings and caregiver proxy reports. The Neuropsychiatric Inventory (NPI; Cummings, 1997; Cummings et al., 1994) is an interviewer-administered instrument that evaluates 12 common dementia-related behaviors. Each behavior is rated for its severity and frequency during the past month, and a total NPI score is computed. Caregivers also rated the frequency of behavior problems in care recipients during the past week by using the Revised Memory and Behavior Problem Checklist (RMBPC; Teri et al., 1992). The RMBPC contains 24 items covering a range of memory, depression, and disruptive behavior problems. Both the NPI and RMBPC scales include ratings of caregiver reactions to patient behavior problems and have been used in clinical trials with persons with dementia (Teri et al., 2002).
The Quality of Life in Alzheimer's Disease measure (QOL-AD; Logsdon, Gibbons, McCurry, & Teri, 2002) was used to assess care recipients' quality of life. The QOL-AD is a 13-item measure designed for individuals with memory loss that focuses on quality-of-life domains identified as important for memory-impaired older adults. In this study, the QOL-AD was completed by care recipients, proxy caregivers (about their family members), and caregivers about themselves.
Descriptive Data
At screening, we obtained demographic information on participants' age, gender, ethnic group, education, relationship, age of dementia onset, and dementia duration. At each visit, we evaluated the cognitive status of care recipients by using the MMSE (Folstein, Folstein, & McHugh, 1975).
Adverse Reactions
Caregivers completed health status change forms at the posttest and follow-up visits. The forms indicated if the caregiver or care recipient had experienced any illness, injury, hospitalization, or institutionalization since the last assessment, and, if so, the dates of onset and the final resolution of the problem. Forms were reviewed weekly by a data safety-monitoring committee to determine if any changes met criteria for a serious adverse event (including serious and unanticipated adverse effects and deaths, or problems involving the conduct of the study or individual participation).
Statistical Methods
We conducted between-group comparisons of baseline covariates by using Fisher exact tests, t tests, or nonparametric Wilcoxon tests. We used Cox proportional hazards survival analyses to determine whether baseline characteristics significantly predicted participant attrition. We used paired t tests to measure changes in target behaviors identified by the STAR-C caregivers.
Our primary analyses were based on intent to treat (ITT), using all randomized participants, regardless of adherence to the intervention. We carried forward previous values for participants missing the posttest or 6-month data (Unnebrink & Windeler, 2001). Secondarily, we repeated analyses without imputation for missing posttests to compare against ITT results. We compared change scores from pretest to posttest for the STAR-C group with the usual-care control group by using t tests, or a Wilcoxon rank sum test if normality assumptions did not hold.
In our longitudinal analyses we used both posttreatment visits (2 and 6 months) and time, controlling for the baseline value of the outcome, using generalized estimating equations with a normal link function and robust standard errors (Liang & Zeger, 1986). Generalized estimating equations offer advantages over other statistical procedures for longitudinal analyses such as an analysis of covariance, because it allows the inclusion of all available participant data (rather than dropping cases in which there are any missing data points). We computed mean group differences with 95% confidence intervals. We assessed Time x Group interactions with the same model structure, and we included them when they were significant. We evaluated potential confounders (age, MMSE score, depression, and ethnicity) by entering their values as covariates.
|
Results |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
|
Supervisor Ratings of Consultant Adherence
Consultants received high ratings on the Therapist Adherence Checklist. On the six checklist items, moderate or extensive ratings ranged from 90% for behavior-change planning to 100% for trainer responsiveness to the caregiver (Mdn = 96% for all six items). These data indicate that community consultants were able to learn the behavioral intervention and deliver the content of the STAR-C protocol as intended.
Treatment Dose
Eighty-three percent of caregivers attended eight or more treatment sessions (M = 7.6 sessions; range = 1–10 sessions). The average duration for the home visit was 61.4 min (SD = 11.58; range = 25–105 min).
Ability of Consultants to Help Caregivers Target Problem Behaviors
Eighty percent of STAR-C caregivers identified three target problem behaviors in the first session (range = 1–4 problems; see Table 1). We defined improvement on any target behavior as a reduction in at least one rating category (frequency, severity, or caregiver reactivity) after 8 weeks of treatment. We observed significant (p <.0001) decreases in the average frequency, severity, and caregiver-reactivity scores when we compared caregiver ratings at the final session with those in Session 1 (Table 3). All STAR-C caregivers reported improvement (either in frequency, severity, or reaction) for at least one of their originally identified target behaviors. Ninety-five percent of caregivers reported some improvement on the first problem they identified (30% improved in all three frequency, severity, and reactivity ratings for the first problem); 85% of caregivers reported some improvement on their second problem (26% reported improvements in all three rating categories); and 86% reported some improvement on their third identified problem (28% reported improvements in all three rating categories). When we considered the target behavior problems overall, 62% of the problems had improvements in caregiver-reactivity scores, 57% had reductions in frequency of problem occurrence, and 52% were reported to have reductions in problem severity.
|
|
Caregiver Outcomes and Treatment Integrity
Among treatment completers, the number of sessions attended was significantly correlated with decreased caregiver-reaction ratings on the RMBPC (Spearman r = –.34; p <.05). Caregiver compliance with treatment recommendations was associated with higher caregiver ratings of competence on the SSCQ (r =.54; p <.01), and decreased depression on the HDRS (r = –.31; p <.05). Ratings on the Therapist Adherence Checklist were not related to any outcome variable.
Longitudinal Outcomes
Over 6 months of follow-up, significant (p <.05) differences in the desired direction remained between caregivers in the STAR-C group and the control group on caregiver depression [CES-D difference, M = –2.3, 95% CI (–0.6, 0.0)], subjective burden [–4.2 (–7.6, 0)], and reactivity to dementia-related behaviors [–3.2 (–6.1, –0.2); see Table 3]. Two additional caregiver outcomes were significant at 6 months: decreases in depression on the HDRS [–1.2 (–2.4, –0.0)] and reductions in self-reported sleep problems [–1.1 (–2.2, –0.1)]. STAR-C caregivers also reported fewer memory-related problem behaviors [–0.3 (–0.5, –0.0)]. There was a trend for active treatment caregivers to report higher quality of life for the care recipients; when we controlled for ethnicity differences between the two groups, this outcome was statistically significant [1.4 (0.1, 2.7)]. When we repeated analyses without imputation for missing values, longitudinal results were the same.
Rates and Reasons for Dropouts
Of the 95 caregivers and care recipients who began the study, 83 (87%) completed posttest assessment, and 66 (70%) completed the 6-month assessment. There were no significant differences at study exit between the treatment groups in rates of attrition, reasons for attrition, or significant predictors of attrition (Figure 1).
Adverse Reactions
No unexpected or serious adverse events were attributed to the STAR-C intervention, and there was no difference between active and control conditions in adverse symptoms.
|
Discussion |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
This is, of course, not the first study to demonstrate that caregivers can improve their own level of depression and burden as well as decrease the problems experienced by those for whom they provide care. Interventions to help caregivers enhance care for their loved ones with Alzheimer's disease have made considerable gains in recent years, including the multisite Resources for Enhancing Alzheimer's Caregiver Health research program, known as REACH, which tested the effectiveness of multiple different interventions in ethnically diverse caregiver populations (Burgio et al., 2003; Gitlin et al., 2003; Schulz et al., 2003). The REACH protocols are particularly noteworthy for their emphasis on monitoring treatment fidelity across participating sites (Burgio et al., 2001; Schulz et al.). Unfortunately, the gains obtained in these studies have been obtained in research-rich environments. To our knowledge, there have been few attempts to date to demonstrate that evidence-based caregiver interventions can be effectively implemented by nonresearchers in the average clinical setting. Most studies have relied on highly skilled research interventionists based in university geriatric and dementia clinics (c.f., Mittelman, Ferris, Shulman, Steinberg, & Levin, 1996). The STAR-C study is the first study that we know of to train community clinicians practicing in a variety of health care settings to deliver a standardized dementia-management intervention that is applicable to a variety of problems common among persons with Alzheimer's disease. Such studies are important if researchers are to see their interventions disseminated and used by health care professionals in real-world clinical settings.
This investigation found that consultant adherence to the STAR-C treatment protocol was very high, despite the fact that consultants varied widely in their previous exposure to, and understanding of, research procedures and behavioral theory. STAR-C was designed to allow consultants considerable flexibility and clinical judgment with regard to the timing and precise nature of their interventions with caregivers, as long as they followed basic behavioral principles and covered the core protocol components during the 8-week treatment period. Although all care recipients were required to have a minimal number of depression or anxiety symptoms prior to entering the study, STAR-C was also designed to be responsive to idiosyncratic challenges that caregivers faced in caring for their family members. Thus, STAR-C training and supervision built on consultants' and caregivers' existing skills, perspectives, and expertise, with the goal of increasing generalizability of the intervention.
Not all treatment outcomes were significantly improved by the STAR-C intervention. For example, STAR-C caregiver depression levels on the interviewer-based HDRS were significantly lower than those of controls at 6-month follow up, but not posttest. However, self-reported depression on the CES-D was significantly lower for STAR-C caregivers at both time points. Furthermore, an examination of the data shows that depression as assessed by both measures steadily increased in the untreated routine medical care caregivers over the 6-month study period. Thus, it appears that STAR-C positively affected caregiver depression, although the evidence of that impact was not immediately fully apparent.
In reviewing our findings, we find it important to consider this study's limitations. The level of supervision provided to consultants was beyond what would normally be provided in a typical geriatric clinic. However, because our goal was to evaluate whether a heterogeneous group of consultants could learn and deliver the STAR-C intervention, we felt it was essential to provide ongoing supervision to monitor adherence to the prescribed treatment protocol. In addition, ratings of consultant adherence were not done by independent raters, because rating instruments served a dual function of providing a standardized basis for ongoing feedback to the consultants. Future studies are needed to examine whether consultant training can be sustained without this level of ongoing supervision. Additional research is also needed to determine whether the effects obtained here can be replicated or improved in additional community settings.
Study findings are based on a relatively small sample size. We do not know whether differences in background characteristics between control and experimental participants would have emerged as more important or how study outcomes would have been affected if we had enrolled more participants or had a greater number of consultants, or if consultants had had different educational backgrounds. Follow-up data from the STAR-C study also are only available up to 6 months. Future research with larger numbers of caregivers and consultants, over longer periods of follow-up, is needed to replicate study findings. Although our findings indicate that community clinicians can be trained to implement behavioral interventions with caregivers in a systematic way, additional studies also are needed to determine what level of ongoing supervision or episodic retraining is necessary to ensure that consultants maintain treatment integrity over longer periods of time and changes in care recipient stage of disease.
In summary, this study demonstrates that community-based consultants can successfully implement a behavioral intervention with family caregivers of persons with Alzheimer's disease. As a first step into the real clinical world, STAR-C appears to be a practical and reasonable beginning approach to an evidenced-based method of caregiver training.
|
Footnotes |
|---|
1 School of Nursing, University of Washington, Seattle.
Decision Editor: Linda S. Noelker, PhD
Received for publication February 24, 2005. Accepted for publication July 13, 2005.
|
References |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Devor and M. Renvall An Educational Intervention to Support Caregivers of Elders With Dementia American Journal of Alzheimer's Disease and Other Dementias, June 1, 2008; 23(3): 233 - 241. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Bookman and M. Harrington Family Caregivers: A Shadow Workforce in the Geriatric Health Care System? Journal of Health Politics Policy and Law, January 1, 2007; 32(6): 1005 - 1041. [Abstract] [PDF]
|
|
|
|
 |
|
 M. S. Mittelman, W. E. Haley, O. J. Clay, and D. L. Roth Improving caregiver well-being delays nursing home placement of patients with Alzheimer disease. Neurology, November 14, 2006; 67(9): 1592 - 1599. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Ayalon, A. M. Gum, L. Feliciano, and P. A. Arean Effectiveness of nonpharmacological interventions for the management of neuropsychiatric symptoms in patients with dementia: a systematic review. Arch Intern Med, November 13, 2006; 166(20): 2182 - 2188. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| All GSA journals | Journals of Gerontology Series A: Biological Sciences and Medical Sciences | Journals of Gerontology Series B: Psychological Sciences and Social Sciences |
