Predicting Facilitators’ Behaviors During Alzheimer’s Family Support Group Meetings

Abstract

A social ecological model was used to predict facilitators’ behaviors observed during Alzheimer’s Association sponsored family support group meetings (N = 66). Information about group leaders was obtained via individual telephone interview prior to the observation of their support group by trained raters. Family dementia caregivers (N = 296) provided basic demographic and caregiving-related information at the end of the observed meeting. Caregiver- and group-level variables predicted ratings of observed leader support during support group meetings. The addition of leader characteristics such as years of experience or professional roles did not significantly contribute to the model. This study lends credence to ecological models emphasizing the influence of social context. Recommendations are made for facilitator training and management of Alzheimer's Association sponsored support group meetings.

Support groups continue to be a popular resource for caregivers of individuals diagnosed with Alzheimer’s disease or another dementing illness, providing caregivers with the opportunity to discuss common problems, receive information, and develop new coping skills. ^{1 –4} Even with the growing interest in telephone- and Internet-based support groups, in-person groups constitute the most common group format for dementia family caregivers. The Alzheimer’s Association in the United States, for example, reported a 2010 estimate of 6663 support groups offered by their area chapters, attended by 39 551 participants. ⁵ These types of programs are seen as a source of accessible and inexpensive assistance for families, providing emotional support and reducing isolation. ^6,7 Perceived benefits extend from an increase in the breadth and depth of social contacts, exploration of new coping strategies, and increased self-efficacy for handling stressful caregiving events. Support groups are viewed as an important resource to foster personal growth and allow members to develop and try out new caregiving skills. ⁷ The acquisition of social support is particularly important for caregivers because of its association with improved psychological and physical health outcomes. ⁸

Researchers, however, face significant challenges in conducting program evaluations of open, community-based dementia family support groups. Quasi-experimental designs that accommodate nonrandom assignment are typically unavailable due to the lack of a comparison condition. There is a tremendous amount of heterogeneity among caregivers in terms of key characteristics and participation rates, with caregivers entering and leaving groups at various times across their caregiving career. Members enter (and leave) groups in ways that could be related to public information campaigns or new community services that promote support group attendance (while leading to improvements on their own). All these threats to internal validity prevent us from directly attributing improvements over time in mental health or caregiving outcomes to the support group experience. Even with the challenges of demonstrating their effectiveness, however, we know that dementia family support groups are grounded in the experimentally established benefits of social support ⁸ and the larger literature on mutual-help and self-help groups. ⁶ Perhaps as important, support groups are highly valued by those family caregivers who choose to participate. ^5,7

Because of the above difficulties in program evaluation, the majority of the existing literature on dementia support group facilitation is theoretical or experiential in nature. Authors offer important practical advice, emphasizing the skills to (1) manage caregivers’ need for emotional support, (2) provide accurate information related to caregiving and AD, and (3) handle difficult interpersonal interactions during group meetings. ^7,9 The direct provision of support by group facilitators may be most critical for new members and also helpful during the times of transition when caregiver participants are less able to respond to the needs of other members. For these reasons, we focus our attention in this article on supportive behaviors of group facilitators. Supportive group leadership requires facilitators to balance the needs of individual members with the needs of the group, all while being perceived as helpful and emotionally connected to caregivers. Medvene and Teal ¹⁰ found that the group identity (ie, affiliation with a larger network of support groups), rather than facilitator-level variables, was most important in identifying which set of needs is given priority during group meetings. This suggests that the context in which the facilitator works may be more important in determining the goals for group meetings and facilitator behavior than the personal characteristics or perspective of the facilitator. Such research is important, promising greater understanding of support group processes and improved facilitator training procedures to maximize effectiveness within the leadership role. Attention to how contextual factors impact facilitator behaviors during dementia family support groups may provide a key to these efforts.

The relationship between individuals and their environments has historically been conceptualized as a linear and unidirectional process in which individuals intentionally create and shape physical and social settings to cope with specific stressors. Much of the literature on support group processes and benefits has come from this perspective, with a focus on the individual group member and her or his experiences. In contrast, an ecological viewpoint emphasizes more complex and bidirectional relationships, focusing instead on the interdependence between people and their physical and sociocultural environments. ^11,12 Support groups fit especially well within an ecological perspective and can be conceptualized as an “alternative social environment” that shapes health and health behaviors. ¹³ Support groups occupy a specific niche within the larger community as well as within mental health services and are largely defined by their focal problems and interactions with the larger community. ⁶ Such groups emphasize the reciprocal nature of social support, the importance of social climate, and the influence of other environmental variables on individual well-being and coping efforts. The conceptual and empirical efforts of Moos ^11,14 as well as Stokols and colleagues ^12,13 posit that the support group events and processes occur within a complicated and interrelated network of personal, group, community, and environmental variables. Thus, attention to multiple variables and levels of analysis are necessary to fully understand supportive behaviors by facilitators during dementia family groups meetings.

Utilizing an ecological perspective in dementia support group research can be helpful because conditions external to the content of group discussions often affect the outcome in powerful ways. Discussion of the role of contextual factors in the intervention participation and response ^11,14 suggests that the individual characteristics of group members, group composition, and availability of other sources of assistance can affect an individuals’ decision to enter the group as well as other features of the group experience. In this process study, we intentionally focused on 1 key aspect of group facilitation: the facilitator’s provision of support to group members. We hypothesized that caregiver-, group-, and facilitator-level variables would predict observed supportive behaviors by group facilitators during dementia family support group meetings.

In our review of the small literature specific to this topic, however, we gained few hints as to which variables hold the most promise as predictors of supportive behaviors by facilitators. We directed our attention first to variables most commonly assessed in a range of intervention research with dementia family caregivers. ^15,16 We combined this with some consideration of the practical concerns about support group management most often raised during our decade-long involvement with a local chapter of the Alzheimer’s Association. This led us to include characteristics of family caregivers (ie, length of time in caregiving role and relationship with care recipient), groups (ie, size, average history of attendance, presence of co-facilitator, and relative meeting emphasis on dementia-specific information), and facilitators (ie, length of experience as group facilitator, professional vs peer role, and personal experience as a dementia family caregiver). We hypothesized that these variables would predict the amount of supportive behaviors enacted by support group leaders in observed meetings and that facilitator-level variables would significantly improve the model after caregiver and group-level variables were accounted for.

Method

Participants

Facilitators or co-facilitators of the 73 active monthly support groups sponsored by a specific chapter of the Alzheimer’s Association within the United States were invited to participate in this study; these groups spanned 36 counties across 2 Midwestern states. Several groups (n = 3) were dropped from the analyses because of zero attendance on the designated date of the observation, and several facilitators (n = 4) were unable to be scheduled for the premeeting interview. Thus, facilitator interviews and group observation data were available for 66 (90%) of these 73 groups. In addition, 296 group members completed a brief anonymous survey at the end of the observed group meeting, with an average of 5.18 (standard deviation [SD] = 4.12) attendees per meeting.

Facilitator Characteristics

Of the 66 facilitators who were both interviewed and observed, the vast majority were Caucasian women (n = 63; 95%) who ranged in age from 23 to 81 years (M = 43.6, SD =12.9). The observed facilitators were generally well educated (M = 15.9, SD = 2.3 years), with 80% (n = 53) reporting a college degree. Many were employed as social workers (n = 32; 48%), but facilitators generally indicated that leading a support group was a voluntary activity not considered a part of their job (n = 43; 65%). On average, facilitators had long-standing relationships with the Alzheimer’s Association, with significant time as a facilitator of their current group (M = 3.6 years; SD = 3.5). A majority reported that their group had a designated co-facilitator (n = 38; 58%), but only a small proportion (12%; n = 8) described that their co-facilitator is actively leading the group more than 50% of the time. A small proportion (23%; n = 15) of the facilitators reported that they had been the primary caregiver for a family member with Alzheimer’s disease.

Group Member Characteristics

Table 1 displays the characteristics of support group members and their care recipients. Of the 296 family members completing the survey at the conclusion of the observed group meetings, most described themselves as the primary caregiver for a family member with a dementing illness (n = 226; 76%). The overwhelming majority (n = 276; 94%) indicated that a physician had diagnosed their family member with Alzheimer’s disease or another dementing illness; in the remaining cases, a formal diagnosis had not yet been made. The average length of time since diagnosis was 3.9 years (SD = 3.4) with a range of 0 to 19 years. The average amount of time that they had attended caregiver support groups was 1.6 years (SD = 2.1). Interestingly, almost half of the living care-recipients resided in an assisted living or skilled nursing facility (n = 132; 49%); a small proportion of care-recipients had died but their caregivers (n = 24; 8%) continued to attend a support group that was not specific to bereaved individuals. Group participants reported strong average levels of satisfaction with their support group experiences overall (M = 4.4, SD = 0.4) and for the specific observed meeting (M = 4.4, SD = 0.5).

Table 1.

Characteristics of Support Group Participants (N = 296)

Variables	M (SD)	N (%)
Age (range 23-95 years)	65.28 (13.90)
Time since first meeting (range 0-11 years)	1.56 (2.11)
Gender
Male		73 (25)
Female		223 (75)
Ethnicity
European American		255 (86)
African American		27 (9)
Other		14 (5)
Relationship to care recipient
Spouse		145 (49)
Adult child		118 (40)
Other		33 (11)
Primary caregiver status		226 (76)
Physician confirmed dementia diagnosis		276 (93)
Years since diagnosis (range 0-19 years)	3.89 (3.40)
Living status of care recipient
Alive		272 (92)
Deceased		24 (8)
Stage of dementia for living recipients (n = 272)
Early		43 (16)
Middle		142 (52)
Late		68 (25)
Do not know		19 (7)
Residential status of living recipients (n = 272)
Living with cg respondent		67 (25)
Living in a community with other caregiver		32 (12)
Living in community alone		9 (3)
Assisted living or skilled nursing		132 (49)
Missing response		32 (12)
General satisfaction with group (range 1-5)	4.4 (0.4)
Satisfaction with specific meeting (range 1-5)	4.4 (0.5)

Abbreviation: SD, standard deviation.

Measures

Facilitator interview

The structured telephone interview of the facilitators was designed for this study and included questions about demographics, professional roles, and personal characteristics (eg, personal experience as a dementia family caregiver). Facilitators were also asked a series of questions related to their facilitator role, including the length of time as group facilitator, status as a peer or health-care professional, whether their group also had a designated co-facilitator, and perceived levels of responsibilities for facilitators or co-facilitators. All facilitators trained as mental health professionals (eg, social workers and counselors) were coded as “professional” for the analyses, regardless of whether their group involvement was considered part of their formal work responsibilities.

Group Environment Scale (GES)—Third Edition

The GES ¹⁷ describes the social atmosphere present in groups and can be used by the independent observers to provide ratings on 10 different facets of group experience. Due to our specific research interests and the limitations of the group observation procedures, the Leader Support subscale of the GES was selected for use in this study. This subscale is described in the manual as “the amount of help, concern, and friendship the leader shows for the members.” ¹⁷ For the purpose of this study, “facilitator” was substituted for “leader” in the specific wording of items. Each of the 9 items in the subscale is rated as either true or false, and the number of items indicative of leader support are totaled (eg, some items are reverse worded). For example, items include “The facilitator helps new members get acquainted with the group,” “The facilitator spends very little time encouraging members,” “The facilitator takes a personal interest in the members,” and so on. Internal consistency and test–retest reliability for this subscale are adequate, with an α = .74 and 1-month test–retest reliability ¹⁷ of .73.

In the present study, the observed GES Leader Support subscale scores were normally distributed and did not require transformation prior to analyses. Only very limited interrater reliability data for the present research are available, due to the constraints placed on the study procedures by the chapter’s support group oversight committee (ie, prohibitions against audio/video recording of meetings and allowing 2 observers to be present at a maximum of 10 meetings). Two trained observers attended the same 10 meetings and independently completed Leader Support ratings. Only 5 of these meetings were attended by 3 or more caregivers and thus formed the basis of our inter-rater reliability estimates using interclass correlations (ICC = 45, reflecting a proportion of the variance shared in the acceptable range).

Frequency of dementia-specific information

Project-related discussions with the support group oversight committee identified a perceived distinction between groups that had a clear emphasis on education or information and those with a greater emphasis on emotional support or encouragement. However, this information versus emotional support dimension of these groups had never been formally assessed in any way. For the purposes of this study, an idiographic verbal behavior tracking form was created to record the frequency of Alzheimer’s- or other dementia-specific information verbalized by the facilitator or co-facilitators. Throughout the meeting, any time that the designated facilitator or a co-facilitator provided the group or a specific group member with specific factual information about Alzheimer’s disease or another dementing illness (eg, diagnosis, progression, treatment, risk factors, etc), the rater made a simple hash mark on the frequency form. These frequencies were tallied and used in the analyses to reflect the relative emphasis in group meetings on the provision of dementia-specific information.

Chapter visitor form

The Alzheimer’s Association chapter had previously developed an annual site visit form for quality improvement purposes. We included use of these forms in the present study as a way to document treatment fidelity. During each observed support group meeting, the rater completed this checklist reflecting facilitator attention to organization and logistics (facilitator arrives on time, sets up room, reminds participants of rules/guidelines such as confidentiality, and establishes ending time), references to the Alzheimer’s Association (printed materials are displayed, members are referred to chapter services, welcome packet given to new members, and other verbal references to Alzheimer’s Association), and facilitation of group interactions (kept discussion focused on relevant issues, etc). Data generated from this form were used for the descriptive purposes to describe program fidelity but were not included in the analyses.

Group member form

Caregivers participating in the meeting were asked to complete a short anonymous questionnaire at the conclusion of the observed meeting. The questionnaire included basic demographic information (ie, age, gender, and race or ethnicity) as well as their relationship to the care recipient, status as primary caregiver, and perceived onset of caregiving responsibilities. Caregivers estimated the month and year of their first support group meeting and reported on the current status of the care recipient (ie, place of residence, stage of dementia, and living status). Respondents were asked to rate their level of satisfaction with their support group experiences in general and for the specific observed meeting (1 = quite dissatisfied and 5 = very satisfied).

Procedures

Our research staff was provided with the names, phone numbers, and addresses of all the individuals listed as a support-group facilitator or co-facilitator of any of the 73 groups sponsored by this specific chapter of the Alzheimer’s Association. These facilitators or co-facilitators (n = 120) were mailed a description of the study and an invitation to participate in a study “to better understand the experiences of caregiver support group facilitators and their needs.” Facilitators were called and, after providing verbal informed consent, interviewed over the telephone using a structured interview protocol lasting approximately 20 minutes. At the end of the phone interview, the observation of the next planned group meeting was arranged. Our goal was to interview the facilitators and visit their respective group within the same 30-day period. This process resulted in 102 (85%) completed interviews by facilitators or co-facilitators.

Project staff members were trained using the GES ¹⁷ manual, a written manual created for the study, and group training and supervision sessions. The project manual included examples of statements and behaviors that would, and would not, be indicative of dementia-specific information (ie, the idiographic ratings made during the live group observations). In the group training sessions, additional examples were provided and discussed by the raters, with raters making independent judgments before discussing as a group. Weekly supervision sessions were held for all raters, with the goal of preventing rater drift during the time of the study.

The designated observer arrived 15 minutes before the group was scheduled to begin and approached the facilitator or facilitators. If 2 co-facilitators were present at the meeting, the facilitator who had completed a telephone interview was chosen as the target for observation and the rater noted the co-facilitator’s presence (coded as 0 for no and 1 for yes) on the rating form. In meetings that included active co-facilitators who had each completed the telephone interview, the one to be observed was selected at random before the meeting began. The behavior of only 1 facilitator was rated per group, and each group was observed once. Before the meeting began, facilitators reviewed and signed the informed consent form for the project.

The observer was briefly introduced to the group by the facilitator at the same time as the member introductions. Observers generally remained silent during the meetings but were trained to respond appropriately if directly questioned or otherwise addressed by a group member. During the meetings, observers tracked the frequency of dementia-specific information and facilitator behaviors listed on the Chapter Visitor’s Form. At the conclusion of the meeting, the observer distributed and collected the survey for family caregivers attending the support group. After the meeting had ended, project staff met with the facilitator and asked how well the observed meeting resembled their “typical” support group meetings. Facilitators were also given an opportunity to provide the rater with additional information about the meeting that might not have been directly observable (ie, lower attendance than expected, addition of several new members, etc). The leader support subscale of the GES ¹⁷ was completed immediately after the support group concluded, without the observers reviewing the surveys from group members.

Results

Observations made by our trained raters using the Chapter Visitor Form suggest that a strong majority of support group meetings were in compliance with the standards set by the Alzheimer’s Association chapter. As shown in Table 2, compliance rates ranged from 69% (facilitator reminded participants of group rules and guidelines) to 92% (facilitator arrived on time and set up the room). These high compliance rates are an indication of program fidelity and justify our focus on supportive facilitator behaviors.

Table 2.

Indicators of Fidelity to Alzheimer’s Association Chapter Standards (N = 66)

Facilitator behaviors	N (%)
Arrived on time and set up room	61 (92)
Reminded caregivers of group rules	46 (69)
Clearly established ending time	58 (88)
Displayed Alzheimer’s Association printed materials	52 (79)
Referred caregivers to other chapter services	49 (74)
Made verbal reference to Alzheimer’s Association	56 (85)
Kept discussion focused on relevant issues	59 (89)

We expected that the caregiver- and group-level variables would predict observed levels of supportive behaviors by group facilitators, and that facilitator-level variables would improve the fit of this model. A hierarchical multiple regression was performed with the observed GES Leader Support subscale score ¹⁷ used as the dependent variable. Averaged caregiver- and group-level variables were entered in the first block of the multiple regression; included in this block were the average length of time as caregiver, caregivers’ average length of time since their first meeting, number of caregivers present, and proportion of adult children at the observed meeting. The presence of a co-facilitator and the frequency tally of dementia-specific information were also included in this block. Characteristics of the facilitator (ie, length of time as group facilitator, role as a peer vs health care professional, and personal experience as a dementia family caregiver) were added in the second block to determine whether these would significantly improve the overall model.

As shown in Table 3, the overall model predicting observed Leader Support was statistically significant (F _9,57 = 4.96, P < .05, R² = .48). As a group, the variables included in the first block significantly predicted Leader Support (F _6,57 = 6.09, R² = .42, P < .05). Examination of the statistically significant regression weights for the individual variables in this first step revealed several associations. Higher ratings of leader support occurred in groups without a co-facilitator present and with members reporting a more recent onset of caregiving. Leaders who were observed as providing a greater frequency of dementia-specific information were also rated as higher in support at the conclusion of the meeting. Interestingly, facilitator support was also rated as higher in the groups where members had on average been attending longer.

Table 3.

Caregiver-, Group-, and Individual-Level Predictors of Observed Facilitator Support (N = 66)

Block	β	R 2	R2 change	Significance of change (P)
Block 1
Average time in caregiving role	−.28 a	.42	.42	<.01
Average time since first support group	.28 a	.42	.42	<.01
Number of caregivers at meeting	−.06	.42	.42	<.01
Adult children (%)	−.06	.42	.42	<.01
Presence of co-facilitator	−.29 a	.42	.42	<.01
Frequency of dementia-specific information	.48 a
		.42	.42	<.01
Block 2
Years of group facilitation	.18	.48	.06	.19
Peer versus health care professional	−.09	.48	.06	.19
Experience as family caregiver	.13	.48	.06	.19

^a P ≤ .05

However, contrary to our predictions, the addition of facilitator characteristics in the second block did not significantly improve the model (R² change = .06, P = .19). Examination of the regression weights revealed nonsignificant associations between GES Leader Support scores and length of time as a group facilitator, role of peer leader versus status as health care professional, and personal experience as a dementia family caregiver.

Discussion

This group process study was conducted to determine whether a specific set of caregiver, group, and facilitator variables would predict facilitators’ supportive behaviors during dementia family support group meetings. Independent variables were chosen to reflect the characteristics of the facilitators as well as the context in which they operated. This attention to the context of the facilitators’ behaviors is congruent with the recommendations by Moos, ^11,14 for designing research within an ecological paradigm. Variables from different sources were chosen to develop a more contextual understanding of facilitators’ behaviors. Our results indicate that a combination of group- and caregiver-related variables appear to influence supportive behaviors of group facilitators, accounting for almost half of the variance of observed support. This is a remarkably strong prediction model for psychosocial research. Facilitator variables did not, however, have the expected impact on their in-group behaviors. This suggests that the characteristics of the caregivers and the group may be more important in determining the perceived supportiveness of a group facilitator than the characteristics of the facilitator herself.

One byproduct of this study is the descriptive snapshot of the family caregivers attending these groups. Indeed, we are unaware of any previous studies documenting the characteristics of family dementia caregivers attending a large number of groups across a chapter’s entire multicounty territory. In examining these data, we encountered more variability in dementia caregiving than we had anticipated. Only 25% of group members were living with the diagnosed individual and close to 50% of living care recipients were in residential care. These figures point to the difficulty in selecting specific outcome variables that remain appropriate across caregiving situations and to the challenges that the facilitators face in managing these groups. The heterogeneity among group members places extra demands on the skills of facilitators, who must be (1) knowledgeable about Alzheimer’s across the stages of the disease, (2) able to help caregivers see what they share in common with other group members, and (3) responsive to a variety of individual and group needs. Although a host of interrelated skills are involved in group facilitation, we are satisfied with the choice of facilitator supportive behaviors as our focus.

Group-level analyses are ambitious and create the challenge of obtaining data with sufficient power. Variables at the levels of caregivers, groups, and facilitators were chosen to capture the major sources that might influence this aspect of support group processes. Our strategy was consistent with the ecological perspective proposed in this study; however, it also decreased the power of the analyses and limited our ability to examine additional variables. Observational procedures posed other limitations for this study. The ratings of leader support and the observed number of informational statements by leaders share method variance and are correlated, despite clear differences in content (eg, only 1 item of the leader support scale refers to providing information “The facilitator explains things to the group.”). Restrictions placed by the chapter’s support group committee (eg, disallowing audio or video recording as well as limiting number of raters and visits per group) reflected organizational anxiety about program evaluation and made it more difficult for us to interpret the data. This study would have been strengthened by the opportunity to record group meetings, leading to broader and stronger coding procedures, and an increased ability to document inter-rater reliability.

An important area of caregiver heterogeneity lacking in our sample was ethnicity. Facilitators and dementia family caregivers were overwhelmingly European Americans, preventing us from exploring ethnicity in our analyses. Although the ethnic profile of caregiver participants (85% European American and 9% African American) mirrors the demography of this chapter’s territory, it nevertheless raises a number of questions regarding how race and ethnicity of group members and leaders interact. The psychotherapy literature on therapist/client matching suggests that ethnic similarity between facilitators and group members would be most important in the earliest phases of group entry and become less important over time. More research attention to the needs of ethnically diverse family caregivers could help Alzheimer’s Association chapters in their programming decisions.

Practice Implications for Support Group Programs

Through the work of Ballard ⁷ and Greif, ⁹ among others, we have the benefit of excellent reviews of practical group facilitation skills; these will be useful for readers involved in training support group leaders. The research literature also includes outcome studies of empirically supported psychoeducational groups for dementia family caregivers ¹⁵ ; these time-limited interventions are quite different from the open support groups most commonly available in the community. The existing studies on support processes provide guidance on issues that are both important and practical. ^1,4 For example, research does not confirm the need for high degrees of similarity as the basis for effective peer support. Sabir and colleagues ³ found that relationship status (eg, adult-child vs spousal caregivers) is less important in establishing supportive peer contact than the length of time that participants have been in the caregiver role. In our study, the majority of family dementia caregivers described themselves as being very satisfied with their support group experiences. Special attention should be paid to the newest group members, who are vulnerable to dropping out before becoming comfortable or aware of concrete benefits to participation. Our data suggest that facilitators tend to provide more direct support in groups where the members have been attending a longer time. This is understandable in terms of common social exchange patterns but is not the best approach to group facilitation. In-service trainings of facilitators might focus on tools for assessing the needs of new group members ¹⁶ and strategies for being more active in providing support to newer group members. For example, group facilitators may benefit from reviewing the 9 items of the GES Leader Support subscale prior to meetings, as a reminder of specific behaviors to increase. Our data also suggest that the provision of concrete dementia-specific information is important. In addition to having strong group facilitation skills and generally supportive behaviors, support group facilitators are an important source of information about dementia and services. Thus, facilitators need to hold themselves accountable in providing information and referral. As the literature is now examining the potential contributions of telephone ^18,19 and internet-based ²⁰ support groups, attention to group format will be important. Future research studies will improve our knowledge of how groups assist individuals with their difficulties, how facilitators operate within the group context, and the individual characteristics of people who decide to pursue this form of assistance. ^21,22 With this information, it is hoped that Alzheimer’s Association support group programs will continue to aid individual family caregivers, as well as help facilitators provide thoughtful and empirically validated assistance to group participants.