Dementia caregiving is associated with a variety of negative health consequences. Dementia caregivers consistently report higher levels of depression and anxiety and exhibit high rates of mood disorders (Schulz & Matire, 2004). There is also evidence that providing caring for an individual with dementia can lead to increased risk for physical health problems and death (Hooker et al., 2002; Vitaliano et al., 2003). Although caregiver demographic characteristics (e.g., age, gender, race), severity of dementia, and behavior problems contribute to emotional burden and distress (Davis, & Tremont, 2006; Hooker et al, 2002; Torti et al., 2004), caregiver appraisal of problems, depression and guilt, social support, quality of life, coping skills, and feelings of self-efficacy are important predictors of caregiver burden (Connell, Janevic, & Gallant., 2001; Gilliam & Steffen, 2006; Markowitz et al., 2003). Recent evidence also shows that satisfying pre-illness relationships between caregivers and persons with dementia and good family functioning are associated with lower reactivity to memory and behavior problems and less burden compared to individuals with unsatisfying prior relationships and poor family functioning (Steadman, Tremont, & Davis, 2007; Tremont, Davis, & Bishop, 2006). Despite the negative consequences for most dementia caregivers, some individuals experience postive effects of providing care (Cohen et al., 2002). In addition, length of caregiving may result in stability or decreases in burden and may increase time to nursing home placement, suggesting adaptation to the caregiving role (Gaugler et al., 2005; Towsend et al., 1989). Overall, factors accounting for emotional distress in dementia caregivers seem to be partially explained by objective caregiver and patient characteristics, although there is evidence that modifiable subjective factors play a major role. Given these findings, psychosocial interventions focused on changing perceptions, improving coping skills, and encouraging resource use are particularly important.

Meta-analyses of psychosocial interventions for dementia caregivers demonstrate small to moderate statistically significant effects on a broad range of mental and physical health outcomes, with particularly strong effects on service utilization, delayed institutionalization, and psychiatric symptomatology (Brodaty, Green, & Kochera, 2003; Knight, Lutzky, & Macosky-Urban, 1993; Schulz et al., 2002; Sorensen, Pinquart, & Duberstein, 2002). Comprehensive caregiver interventions involving regular, individualized contact with a counselor/social worker delivered over an extended period have resulted in reduced caregiver distress compared to control conditions (Brodaty & Gresham, 1989; Mittelman et al., 1993, 1995, 1996, 2004). Well-controlled, large-scaled psychosocial interventions consistently show caregiver satisfaction and subjective benefit from the intervention, improved knowledge about dementia caregiving and the disease course, reduced depression, anxiety, anger, and hostility, and delayed institutionalization. However, reduction in perceived burden has been less consistently identified (Acton & Kang, 2001). It appears that multi-component interventions may result in the best outcomes for perceived caregiver burden and distress. Additionally, individually-tailored interventions and interventions of longer duration (6 months or more) appear more efficacious than fixed, briefer interventions (Belle et al., 2006; Brodaty et al., 2003; Gitlin et al., 2003; Schulz et al., 2002; Sorensen et al., 2002;). Therefore, there is good evidence that psychosocial interventions can improve dementia caregiver well-being. The next step in developing interventions for dementia caregivers is to improve cost-effectiveness, ease of implementation, and accessibility.

Telephone interventions are generally highly acceptable to patients and have effects on a broad range of mental and physical health outcomes and medical resource utilization. For example, telephone interventions have been successful in all of the following areas: (1) reducing mortality, hospitalizations, and office visits in the primary care setting (Wasson et al., 1992); (2) improving compliance with medical procedures (Cargill, 1992); (3) increasing knowledge about medical problems (McCullough, Day, Herlihy, & Rapp, 1989); (4) minimizing stress and anxiety and depressive symptoms in medical patients (Evans & Jaureguy, 1982); and (5) bringing about health-related lifestyle changes including smoking cessation (Ockene, 1991). A frequent concern raised in the literature is whether telephone-based interventions work because of nonspecific factors such as increased attention from telephone contacts or are a direct result of the specific therapeutic techniques. Austin et al. (1996) examined this question by assigning patients with systemic lupus erythematosis to either telephone counseling or telephone symptoms monitoring. Counseling was shown to be superior to symptoms monitoring in improving functional status, arguing against the nonspecific effects hypothesis. There are few studies of telephone-based interventions for dementia caregivers. An uncontrolled telephone counseling program for dementia caregivers showed improvement in stress, burden, and anxiety, though was methodologically limited (Strawn & Hester, 1998). Telephone voice mail and computerized interventions for caregivers have had mixed results, with some positive findings in subsets of caregivers (Mahoney, Tarlow, & Jones, 2003), possibly related to the amount of time caregivers spend accessing these interventions.

Because of concerns about cost and accessibility to dementia caregivers, we developed an intervention, Family Intervention: Telephone Tracking – Dementia (FITT-D). The intervention is a modified version of a program designed to help caregivers of stroke victims that has demonstrated preliminary efficacy (Miller et al., 1998). It is delivered entirely over the telephone. Theoretical underpinnings of the intervention are based on psychosocial transition (Tyhurst, 1958), transactional stress and coping (Lazarus & Folkman, 1984), and a systems view of family functioning (i.e., McMaster Model of Family Functioning; Epstein et al., 1978). Taken together, the three underlying theories of the FITT-D are geared toward enhancing coping within the caregiver through active problem-solving and facilitating positive changes within the family system. According to these models, dementia caregiving is comprised of potentially stress-inducing transitions that prompt an appraisal process by the caregiver to identify whether resources (particularly those within a family) can be mobilized to cope with the changes. Within these overlapping models, caregiver burden is viewed as a situation in which demands (both perceived and objective) exceed the coping and resources (both perceived and objective). As burden persists, it becomes a stressor in its own right, leading to feelings of helplessness, hopelessness, and depression. Overall, the FITT-D directly targets caregiver appraisal and coping processes to directly reduce stress, improve caregiver mood and quality of life, and positively impact the family system.

The current study reports preliminary findings of the FITT-D. It was hypothesized that caregivers receiving the FITT-D would show reductions in burden, depression, and reactions to memory and behavior problems compared to caregivers receiving standard care.

Conceptually, FITT-D is based on three theoretical contructs (i.e., stress and coping theory, psychosocial transition, and the McMaster Model of Family Functioning). We can only hypothesize about the potential mechanisms of the FITT-D that brought about improvement in the caregivers, but the intervention emphasized active problem-solving, coping, and changes within the family system. Furthermore, although the intervention is manualized and structured, it maintains considerable flexibility to allow for individualized application of interventions as deemed necessary and appropriate. The length of the intervention may also be important. Most caregiver interventions are 6 months or less. Although the telephone contacts were generally short (less than 30 minutes), FITT-D was delivered over one year. This provided an opportunity for the therapist to help caregivers apply problem-solving skills in a variety of situations and in the face of cognitive and behavioral decline in the care recipients. Multiple intervention techniques were used based on the problem and situation. Caregivers were encouraged to identify the positive aspects of caregiving throughout the intervention and to identify their effective problem solving by the end of the intervention. Overall, we believe that the multifaceted, individualized approach to dementia caregiving was responsible for bringing about improvement in burden and reaction to memory and behavior problems. Future studies of the FITT-D can examine the most effective ingredients and appropriate treatment duration.

Surprisingly, FITT-D was not associated with any statistically significant change in depression scores. Comparing our sample to other dementia caregiver studies, we had lower rates of caregivers reporting depressive symptom levels above the clinical cut-off at baseline. For example, Mittelman et al. (2004) reported 42% of caregivers had GDS-30 item scores of 11 or greater whereas in our sample, only 24% had scores in this range. In addition, their mean GDS scores at baseline were considerably higher than in our sample. It is possible that our sample is not representative of community-based dementia caregivers due to some bias in our recruitment strategies. We found evidence that caregivers who dropped from the study (either because of care recipient death or for other reasons) had higher baseline depression scores than those who completed the study, which may account for less depression in our caregivers than is typically seen. Caregivers assigned to the control condition who did not complete the study showed higher baseline depression scores than control caregivers completing the study, which may have disguised any antidepressant effect of FITT-D. However, our caregivers were significantly burdened, and caregivers receiving FITT-D reported significant reductions in burden. Therefore, it is unclear how caregivers with moderate to severe depressive symptoms would respond to FITT-D.

We found that more caregivers in treatment placed care recipients in supervised living settings by the end of treatment than caregivers in standard care. The difference did not reach statistical significance and did not appear to have any significant impact upon results of the intervention. In many intervention studies, goals of treatment are to prevent or postpone nursing home placement (e.g., Mittelman et al., 1993). Inappropriate or premature placement would clearly reflect a negative outcome for an individual with dementia, although there are many situations in which nursing home placement is timely and appropriate. In the current study, we chose to continue treating caregivers after placements. We believed that the FITT-D was adaptable enough to be effective with these caregivers. In addition, there is evidence that caregiver burden and distress levels do not change after institutionalization of the care recipient (Schulz et al., 2004). Our group has developed and preliminarily testing a specific intervention to deal with this important transition.

Importantly, this study demonstrated that master's level individuals can be effectively trained to deliver the FITT-D intervention. We trained two individuals who had no prior experience with dementia or caregiving in general. Therefore, we expect that the intervention could be easily learned and delivered by other healthcare professionals, such as nurses and social workers. In addition, because the FITT-D is manualized, it has the potential to be implemented in the community on a large scale. For example, the intervention methods could be adapted and integrated into care consultation provided by staff of local Alzheimer's Association chapters. Similarly, state elderly affairs agencies could use the intervention approach to assist rural or isolated caregivers.

The current study has some limitations, and we experienced a relatively high attrition rate, mostly due to care recipient death. The sample size was small, possibly limiting our statistical power. Despite this, we were able to show statistical significant effects on burden and reaction to behavior problems. Future studies of FITT-D will need a larger sample to explore the relationship between treatment, burden, and depression. Our small sample limited our ability to identify variables that predict best response to treatment, such as dementia type, relationship type, ethnic identity, gender, and living arrangement. An additional limitation of a small sample size was that we could not address the relationship between the treatment process and outcome. The study did not involve a long-term follow-up of the intervention, which would address whether the intervention effects persist past the end of treatment. The intervention included only in-home caregivers providing substantial care. Therefore, results have somewhat limited generalizability. It is anticipated that future studies of this intervention can investigate its effects in a larger more diverse sample to address these limitations.