Improving quality of life remains a significant challenge in long-term care (LTC) and dementia care environments. Montessori-based interventions (MBIs) were developed to support life quality in dementia by emphasizing participation in meaningful roles and/or activities (Wilson et al., Reference Wilson, Camp, Judge, Bye, Fox and Bowden1997). Currently available research suggests that, among other positive effects, MBIs have the potential to increase active engagement, improve mood, and reduce agitation (Sheppard, McArthur, & Hitzig, Reference Sheppard, McArthur and Hitzig2016). Nevertheless, increasing the use of MBIs in LTC may require additional resources.
Researchers have begun to examine whether volunteerism has the potential to enhance psychosocial care in resource-strained LTC environments. van der Ploeg, Mbakile, Genovesi, and O’Connor (Reference van der Ploeg, Mbakile, Genovesi and O’Connor2012) confirmed that it was feasible to involve volunteers in group activities and individual visits with LTC residents with dementia, and Söderhamn, Landmark, Aasgaard, Eide, and Söderhamn (Reference Söderhamn, Landmark, Aasgaard, Eide and Söderhamn2012) found that volunteers experienced their work in psychosocial dementia care positively. There has also been limited study of volunteer or non-expert delivery of MBIs. For example, van der Ploeg, Walker, and O’Connor (Reference van der Ploeg, Walker and O’Connor2014) examined the feasibility of volunteer involvement in MBIs for LTC residents who had dementia and agitated but non-aggressive behaviour. Overall, volunteers found the experience very rewarding, but some felt discouraged on occasion, when their visits did not alleviate agitation or if they were unable to engage the resident in the MBI. Schneider and Camp (Reference Schneider and Camp2003) trained family members to implement an MBI with LTC residents who had dementia, and found improved life satisfaction, mastery, reduced burden, and better visiting experiences. Finally, Hunter, Thorpe, Hounjet, and Hadjistavropoulos (Reference Hunter, Thorpe, Hounjet and Hadjistavropoulos2020) examined the acceptability and feasibility of a volunteer-delivered MBI from the perspective of LTC staff. Staff reported that the MBI provided essential psychosocial support to residents; however, they believed that residents with late-stage dementia benefited less.
When non-experts, including volunteers, are responsible for intervention delivery, intervention integrity, or fidelity, is a particular concern. Intervention fidelity is the degree to which programs or interventions are implemented as intended (Carroll et al., Reference Carroll, Patterson, Wood, Booth, Rick and Balain2007). Developing an understanding of the determinants of intervention fidelity provides a foundation for the study of intervention effectiveness (Carroll et al., Reference Carroll, Patterson, Wood, Booth, Rick and Balain2007).
The Conceptual Framework for Intervention Fidelity
To encourage more comprehensive assessments of intervention fidelity, Carroll et al. (Reference Carroll, Patterson, Wood, Booth, Rick and Balain2007) proposed a Conceptual Framework for Intervention Fidelity (CFIF). The CFIF emphasizes adherence (intervention content, frequency, duration, and coverage) as the key measure of fidelity. Other elements of the framework are considered moderators of adherence (see Figure 1). These include (1) intervention complexity, (2) facilitation strategies to optimize and standardize implementation, (3) delivery quality, and (4) participant responsiveness. Complexity refers to the structure of the intervention (simpler intervention models, higher levels of structure, and specific requirements are generally associated with higher adherence; Carroll et al., Reference Carroll, Patterson, Wood, Booth, Rick and Balain2007). Facilitation strategies are resources (e.g., manuals, training, or monitoring and feedback) to promote adherence. These vary by intervention, and the need for them depends partly on intervention complexity. Delivery quality refers to how closely delivery approximates a theoretical ideal; for example, if intervention relies on reading a manual, and there is evidence that many of those who delivered the intervention did not read the manual, this suggests poor quality. Finally, participant responsiveness refers to the extent to which participants were engaged in or satisfied with the intervention (a result usually distinguishable from the anticipated intervention outcome). A poor response can be seen as an impediment to high-fidelity implementations.
Figure 1: Implementation fidelity evaluation. Adapted from Carroll et al. (Reference Carroll, Patterson, Wood, Booth, Rick and Balain2007)
Methods
Objective
The objective of this study was to evaluate the fidelity of a volunteer-led MBI, including any associated strengths and limitations, using the CFIF.
Context
The study took place within a special dementia unit (SDU) in a non-profit LTC facility in Western Canada. The LTC is home to 129 residents, and the SDU is home to 49 residents with dementia. Although the SDU provides regularly scheduled group recreational activities, prior to the study, volunteer presence on the SDU was limited.
The MBI
The research team created an array of more than 30 activity kits corresponding to the interests and abilities of participating residents, as reported in care records or by staff and family members. Activities were based on a manual (Elliot, Reference Elliot2012) and activity guides for dementia care (e.g., Camp, Reference Camp2006), and spanned four domains emphasized in MBIs: practical, sensorial, cognitive, and sociocultural. Activity kits were stored in transparent plastic containers, numbered, and placed in a conveniently accessible storage area. Each contained brief instructions on a large-font laminated page, including suggestions to make the activity more or less challenging. After pre-testing, a chart identifying activities that best matched each resident’s interests and abilities was posted in the storage area. A profile outlining each resident’s activity interests and current functioning (i.e., sensory, motor, cognitive) was also made available.
Participants
A convenience sample of 18 community volunteers was recruited by the partnering LTC home to participate in a trial of the MBI (Devers & Frankel, Reference Devers and Frankel2000). After expressing interest, volunteers were provided with detailed information about the nature of the requested commitment and given the opportunity to consent or refuse participation. Those who consented were assigned to one of two volunteer experiences, only one of which is a focus of the current study. Specifically, nine volunteers received MBI training immediately, and participated in the MBI evaluation that we report on here. Nine additional volunteers were assigned to more traditional roles (e.g., assisting residents at mealtimes; social visits), and participated in a separate evaluation. Three of the nine participants in this study (33%) were men, and six (67%) were women. Their ages ranged from 22 to 55 (
$ \overline{x} $= 35, SD = 13). On average, volunteers had 15 (SD = 2) years of education. Only one (11%) had prior volunteer experience in LTC, and none had prior training or experience working with residents with dementia, or working in an SDU. One was studying to be a teacher and was generally familiar with the Montessori approach to education.
Volunteer Training
Volunteers received a total of approximately 5 hours of training across three sessions. During the first session, volunteers learned about dementia and received an overview of the MBI. During the second, they learned how to conduct a Montessori visit, acquired experience with several activities, and reviewed a profile of the interests and abilities of the resident with whom they had been matched. During the third session, they learned how to safely enter and leave the SDU and how to access resources for the MBI. At this time, they also met the residents for the first time.
Procedure
Following institutional review board ethical clearance, volunteers were asked to complete a minimum of 10 scheduled visits and 10 additional visits at their own convenience. Scheduled visits were scheduled to take place between the hours of 1300 and 2000 p.m. If either member of the dyad was unavailable, volunteers were asked to visit at their own convenience between 1:00 and 8:00 p.m. A research assistant closely observed the participating resident during each scheduled visit to assess resident outcomes (these are reported separately). This assistant also documented whether the volunteer and/or the resident attended, and administered a questionnaire to the volunteer following the visit. A second research assistant interviewed each volunteer after the first five visits, the next five visits, and after 2 months. The first two interviews lasted approximately 15–30 minutes, and the last, approximately 30–45 minutes. Interviews took place in a private meeting room at the LTC facility.
Measures
Interviews
During the first two interviews, volunteers were asked to describe the range of activities that they had invited residents to participate in, how they believed the visits had gone, how comfortable they had been, and whether they had sufficient support and resources. During the final interview, participants were asked about changes to their visiting routine since the first 10 visits (which were pre-scheduled), general experiences initiating MBIs, and beliefs about the value of MBIs to LTC residents. They were also asked about whether the program met expectations, what they learned or gained, suggestions for improvement, and (again) if they believed that they had adequate support and resources throughout the program.
Visiting Quality Questionnaire (VQQ)
Immediately after each of the first 10 visits, volunteers completed the 18-item VQQ created for this study to assess the perceived quality of a visit. The nine-item Volunteer Response subscale is a self-assessment of volunteer visiting experiences such as pleasure, interest, and comfort. The nine-item Resident Response subscale represents the volunteer’s assessment of similar experiences on the part of the resident. Items are rated on a seven-point Likert type scale (higher ratings implying more positive responses). Both subscales of the VQQ evidenced high internal consistency (both α = 0.97) and high test–retest reliability at a 4-day interval (r = 0.96 and r = 0.97, respectively).
Analysis
We analyzed interview, questionnaire, and observational data using the CFIF (Carroll et al., Reference Carroll, Patterson, Wood, Booth, Rick and Balain2007), which assesses intervention adherence and associated moderators.
Adherence
Adherence to the MBI was assessed as: (1) the proportion of volunteers who gave specific evidence, during interviews, of using an MBI approach as opposed to relying on a general conversational visit; (2) the number of total possible visits received by residents; (3) the rate of volunteer attrition; and (4) the number of total possible visits made by volunteers.
Moderators of adherence
Moderators of adherence, including delivery quality, facilitation strategies, and participant responsiveness, were analyzed using qualitative content analysis, facilitated by Dedoose, version 8.0.42. One moderator, participant responsiveness, was additionally assessed using VQQ subscale means and standard deviations (SD) as indicators of participant responsiveness. Another CFIF moderator, intervention complexity, was excluded from the analysis because we found that all related interview data overlapped the category facilitation strategies.
Qualitative content analysis involves the categorization of textual units (Graneheim & Lundman, Reference Graneheim and Lundman2004). We employed a deductive approach to content analysis, which involves classifying units of text from interview transcripts into an already-existing framework, often drawn from theory (Sandelowski, Reference Sandelowski2000; Vaismoradi, Turunen, & Bondas, Reference Vaismoradi, Turunen and Bondas2013). In this case, the framework was an adaptation of the CFIF (see first two columns of Table 1). In order to provide a more meaningful description of the interview data associated with two categories, facilitation strategies and participant responsiveness, we inductively analyzed these data to generate subthemes. The first author completed this analysis, and the second author completed an audit of the trustworthiness of results. Specifically, the auditor reviewed the coding of all interviews against the coding framework, and generated a Likert agreement rating (on a scale of 1–5) of the extent to which each code was grounded in the data.
Table 1: Overview of research findings using an adapted conceptual framework for intervention fidelity
Note. These results are based on an adaptation of the Conceptual Framework for Intervention Fidelity (CFIF) (Carroll et al., Reference Carroll, Patterson, Wood, Booth, Rick and Balain2007). The leftmost column includes elements of the CFIF. The center column includes the fidelity indicators evaluated in this study. The rightmost column summarizes results.
Results
Eight volunteers completed an initial interview after their first week volunteering, a second after their second week of volunteering, and a third at the end of the study period. This represented a total of 24 interviews, and a participation rate of 100 per cent, excluding from consideration one volunteer who resigned before the first interview. An average audit rating of 4.63 supported the trustworthiness of the qualitative analysis of interview data. In addition, we found that the CFIF moderator categories delivery quality, facilitation strategies, and participant responsiveness were well saturated; that is, interview content substantially addressed each of these themes. Results are further described, in the two major categories of fidelity according to the CFIF: adherence and moderators of adherence. See Table 1 for an overview of results.
Intervention Adherence
Coverage
Overall, residents attended 78 of 80 (98.5%) scheduled visits: one resident missed two visits because of fatigue.
Frequency
Of the nine volunteers who originally agreed to participate in the evaluation of the MBI, eight (89%) fulfilled their commitment to visiting throughout the study period; one (11%) withdrew after four visits, reportedly because of an overcommitted schedule. Considering only the eight remaining volunteers, 66 of 80 planned visits were completed (82.5%). The most commonly reported reason for volunteer absence was an unexpected event (e.g., car trouble).
Content
During the interviews, all volunteers gave evidence of attempting to use the MBI. Four volunteers consistently used the activity kits designed for the MBI, and described generally good responses to these visits. Two volunteers appropriately integrated personal interests into the visits to a significant degree, by modifying existing activities or suggesting new ones; for example, a musician created activities that involved keeping time to music by shaking a maraca. The remaining four volunteers gave evidence of having some difficulty adhering to the pre-designed activities consistently. In most cases, activities that supported important roles were substituted, consistent with a Montessori approach; for example, a volunteer who worked with a retired salesperson favoured a conversational approach. In three of the four cases, volunteers were matched to residents with late-stage dementia. Following some experimentation with pre-planned activities, these volunteers began to favour everyday activities that the resident seemed to enjoy (e.g., walking, spending time outdoors, or playing piano).
Moderators of Intervention Adherence
Facilitation strategies
A number of facilitators were described during the interviews. Those most commonly mentioned included initial training, an initial meeting with the resident, access to a profile of the resident’s abilities and interests, a diverse array of activities to choose from, a workable visiting time, and support from staff. It was common for participants to mention these strategies spontaneously, and to name a number of these together, suggesting both that volunteers were aware of program resources and that they interpreted these as supports. See Table 2 for representative quotations.
Table 2: Facilitative strategies
Note. In parentheses following each code is the number of interview excerpts to which the code was applied, followed by the number of interviews these excerpts occurred within, and lastly, the number of interviewees who raised the topic. This provides an indication of theme saturation.
Additional recommendations
Volunteers were asked for recommendations during the final interview, and these included: having an increased range of activities, providing more time to get to know the resident before using the MBI, providing more support when a resident is not engaging, providing more training on non-verbal communication, and providing more training on problem-solving (e.g., responding to critical incidents or other unusual occurrences). Nevertheless, other than the first theme, these did not have good saturation, suggesting that they were more important to some volunteers or dyads than to others (further addressed in the following text). See Table 3 for representative quotations.
Table 3: Recommendations
Note. In parentheses following each code is the number of interview excerpts to which the code was applied, followed by the number of interviews these excerpts occurred within, and lastly, the number of interviewees who raised the topic. This provides an indication of theme saturation.
Regarding volunteers’ perceived need for an improved range of activities, most acknowledged that they could access several activities, but that once they discovered those that best engaged their visiting partners, they wished for a larger array of similar activities. The interaction between volunteer and resident interests was also important to volunteers’ activity preferences. For example, some volunteers and residents were not comfortable using high-sensory activities (e.g., colourful or textured manipulatives), perhaps perceiving these as too childish, whereas other volunteers and residents responded positively to these.
Only two volunteers expressed a need for more support with resident engagement, but did so consistently across three interviews. These volunteers were working with residents with late-stage dementia, for whom direct engagement in the activities was limited, despite passive interest or occasional attempts to participate.
Delivery Quality
A number of key recommendations were communicated to volunteers during training, and interview data were analyzed for evidence that volunteers were following these recommendations, which included: use of Montessori activities (e.g., as opposed to strictly conversational visits), modeling the activity first, accommodating disability, making every visit a success, and supervising the activity closely. See Table 4 for representative quotations.
Table 4: Delivery quality
Note. In parentheses following each code is the number of interview excerpts to which the code was applied, followed by the number of interviews these excerpts occurred within, and lastly, the number of interviewees who raised the topic. This provides an indication of theme saturation.
Use of Montessori activities
All interviewees described the use of one or more pre-prepared MBI activities during volunteer visits; nevertheless, several interviews also contained descriptions of portions of visits that did not include the use of MBI activities. Further analysis suggested that because some volunteers were working with residents who had good conversational skills, they had incorporated conversation into their visits; for example, one volunteer led an activity first, and stayed to talk afterward. Nevertheless, some volunteers included conversation in their visits even when the resident showed little evidence of understanding.
Overall, interview content suggested that all volunteers used prepared MBI activities for some visits. However, there was evidence that three volunteers reduced their use of prepared activities over time, preferring shared participation in music or walks outside. These volunteers were all working with residents with late-stage dementia, who responded best to sensorial activities. Another volunteer ultimately chose conversational visits over an MBI approach; in this case, the resident was a former salesman who enjoyed conversation. Although an additional volunteer claimed not to be using prepared activities, further exploration during the interview clarified that he had appropriately modified activities to suit his and the resident’s interests.
Modeling the activity first
Most participants discussed modeling the activity before asking the resident to join in. Interview data suggested that this was an ideal way to engage a resident’s interest in an activity, particularly when residents had lower verbal communication ability. Often, the resident decided to join the activity; however, sometimes, the resident’s ability to mirror the activity proved to be limited. For example, after one volunteer modeled a practical/sensorial activity that involved transferring coffee beans from one container to another, the resident attempted to eat the coffee beans.
Accommodating disability
Examples provided during the interviews suggested that most volunteers adjusted their approach to visiting after learning about residents’ needs, interests, and abilities. Although some of this information was available to volunteers at the outset of the program, volunteers also described relying on direct experience to identify adjustments that facilitated visits with a particular resident (e.g., using a louder voice, offering assistance only after fatigue set in). Some volunteers noted that residents’ responses varied from day to day, or in response to fluctuations in cognitive functioning and fatigue. Interview data suggested that volunteers adjusted well to these situations. Some mentioned returning at another time, whereas others gave examples of spontaneously modifying activities or introducing an alternate activity.
Making every visit a success
As volunteers described their use of the Montessori activities, it was clear that they regularly modified activities to ensure that the resident could successfully complete them. For example, one volunteer emphasized stacking dominoes instead of matching patterns. Another volunteer mentioned that it was helpful to keep visits brief in order to ensure that the resident did not fatigue, and therefore disengage. In general, there was good evidence that volunteers adapted activities so that residents felt engaged and efficacious.
Supervising the activity
Evidence that volunteers were closely supervising activities was implicit in volunteers’ descriptions of the visits. For example, volunteers described their roles and the residents’ roles in the activities in detail, and most offered detailed observations about residents’ reactions to a range of activities. Occasionally, volunteers gave examples of having circumvented possible or actual risk because of close supervision. In particular, it was common for residents with late-stage dementia to put small colourful objects in their mouths, a possibility that had been addressed during volunteer training.
Participant Responsiveness
Residents’ responses to visits
Scores on the Resident Response subscale of the VQQ ranged from 4 to 7 (
$ \overline{x} $= 5.46; SD = 0.88), suggesting that volunteers perceived residents as moderately satisfied with the visits, on average. Volunteers described residents’ responses to visits in terms of acceptance, engagement, and recognizing the volunteer. All themes were well saturated. See Table 5 for representative quotations.
Table 5: Resident response to visits
Note. In parentheses following each code is the number of interview excerpts to which the code was applied, followed by the number of interviews these excerpts occurred within, and lastly, the number of interviewees who raised the topic. This provides an indication of theme saturation.
Overall, acceptance of volunteer visits seemed high. Refusals seldom occurred, and interview data suggested that some residents were immediately accepting of volunteer visits, whereas other residents’ acceptance seemed to increase over time. During initial visits, volunteers perceived residents as somewhat hesitant to participate in activities. During later visits, volunteers perceived residents as more engaged and expressive. Some residents began to call the volunteers friends.
The level and quality of engagement in the MBI varied from one resident to another, and sometimes varied across visits. Some volunteers described residents engaging in these activities with ease. Others noted that residents preferred to watch with interest, or were engaged one day but not the next, or tired or lost interest after briefly engaging. Volunteers working with residents who had late-stage dementia were more likely to describe inconsistent or low engagement. Overall, residents’ responses were described as ranging from passive interest and observation to active participation. As time passed, most volunteers discovered activities that engaged the residents more often than not, but variations in engagement continued to occur.
Nearly all volunteers spontaneously noted that the resident was recognizing the volunteer. Some noted that the resident seemed to “light up” or smile when they arrived, or that the resident now introduced them as a friend. In one case, a resident learned the volunteer’s name. Although none of the volunteers spoke directly to the significance of being recognized, these statements tended to be associated with descriptions of residents’ increased receptivity and engagement.
Volunteers’ responses to visits
Scores on the Volunteer Response subscale of the VQQ ranged from 5 to 7 (
$ \overline{x} $ = 6.17; SD = 0.75), suggesting that volunteers were very satisfied with the visits. Most volunteers described feeling comfortable, connected, and rewarded. A few spoke to experiences such as feeling drained, sad, or discouraged (minor themes). See Table 6 for representative quotations.
Table 6: Volunteer responses to visits
Note. In parentheses following each code is the number of interview excerpts to which the code was applied, followed by the number of interviews these excerpts occurred within, and lastly, the number of interviewees who raised the topic. This provides an indication of theme saturation.
Volunteers who expressed that they were comfortable related this to their training, noting that they felt prepared for the circumstances that they might encounter, and that even though they were still learning, training gave them an adequate foundation to build on. In addition, one volunteer related her level of comfort to having been in LTC before, and a few volunteers noted that residents’ positive responses quickly put them at ease. One female volunteer became uncomfortable when a male resident expressed sexual interest. She was reassigned to work with another resident, and described her overall experience of the program as positive and comfortable.
A sense of feeling rewarded by volunteering was apparent in expressions of positive feelings (e.g., compassion) towards the specific residents regularly visited, or gratitude and pride about being involved in volunteer work. Several volunteers mentioned that they looked forward to visits and felt happy after visiting. Some stated that they were making a difference in residents’ quality of life; for example, noting the rarity of residents’ opportunities to engage with the wider community, or commenting on specific outcomes such as making the resident smile. A few volunteers mentioned their own personal learning or growth as rewarding; for example, learning a new card game, learning about the resident’s history, or becoming more flexible.
Volunteers also mentioned feeling connected to the residents as they continued to visit. Many described the ways that their relationship was growing, or described the value of this new relationship. For example, some volunteers expressed pleasure that residents with dementia recognized them on return visits, seemed more interested in visiting with them over time, or had shared information about their lives. One volunteer described feeling as though the resident was an uncle, and another described feeling as though the resident was a new friend.
Although volunteer responses were positive overall, some volunteers had mixed feelings. Two volunteers mentioned feeling discouraged when a resident was uninterested in a particular activity, or when responses to the visits varied. Both were working with residents who had late-stage dementia, and neither had prior experience in dementia care. The same two volunteers mentioned that they felt drained after visiting. Two other volunteers mentioned feeling sad for residents in between visits or when leaving their visits. In both cases, the volunteers empathized with residents’ apparent loneliness. Nevertheless, this sadness was not prohibitive, as both felt that their visits made a difference in the moment.
Discussion
We used the CFIF to assess the intervention fidelity of a volunteer-delivered MBI for people with dementia living in one LTC home. We also explored potential moderators of fidelity, including the quality of volunteers’ efforts, program set-up, and the responses of participants. Overall, we found that resident and volunteer adherence to scheduled MBI visits was excellent, and most volunteers consistently employed Montessori activities or made adaptations consistent with Montessori principles. With respect to potential moderators of intervention fidelity, we found that adherence was supported by a highly structured program, by the perceived rewards associated with visiting, and by volunteers’ attention to training recommendations. Some limitations were also noted. For example, we found that many volunteers emphasized conversation during their visits, even when LTC residents had significant aphasia, and that volunteers who worked with residents with late stage dementia tended to favour everyday activities over the Montessori activities used in this intervention. In addition, some volunteers, particularly those working with people who had late-stage dementia, did not feel as efficacious as others.
Some concern has previously been expressed about involving volunteers with people who have dementia. For example, Robinson and Clemens (Reference Robinson and Clemons1999) reported that volunteers involved in providing home-based respite care to people with dementia were initially quite apprehensive. In the current study, which primarily involved new volunteers with little to no experience with dementia, volunteers’ commitment to the MBI program was excellent. The majority of visits were fulfilled, and there was evidence that most volunteers used the planned activities. These findings align with other emerging evidence that it is realistic to engage volunteers in dementia care (Seitz, Knuff, Prorok, LeClair, & Gill, Reference Seitz, Knuff, Prorok, Le Clair and Gill2016; Söderhamn et al., Reference Söderhamn, Landmark, Aasgaard, Eide and Söderhamn2012; van der Ploeg et al., Reference van der Ploeg, Mbakile, Genovesi and O’Connor2012), and with evidence that volunteers can implement Montessori resources and principles in their visits (Camp, Reference Camp2010; Schneider & Camp, Reference Schneider and Camp2003; van der Ploeg et al., Reference van der Ploeg, Walker and O’Connor2014). A potential reason for volunteers’ strong commitment to the MBI program, supported by volunteers’ comments, is that it was highly structured. According to the CFIF, a well-structured intervention with good program facilitation strategies is likely to enhance fidelity. A systematic review of volunteers’ experiences working with patients who have dementia in acute care settings suggests that a high-structure approach is common (at least in clinical research), and is appreciated by volunteers (Hall, Brooke, Pendlebury, and Jackson, Reference Hall, Brooke, Pendlebury and Jackson2019). In LTC, the results of one observational study of casual visits between volunteers and residents suggested that a superimposed structure might also contribute to a more equitable distribution of volunteer resources, avoiding the possibility that only a few residents would benefit from volunteerism (Damianakis, Wagner, Bernstein, & Marziali, Reference Damianakis, Wagner, Bernstein and Marziali2007).
In some qualitative reports of volunteers’ experiences with LTC residents who have dementia, volunteers have mentioned difficult experiences. For example, Söderhamn et al. (Reference Söderhamn, Landmark, Aasgaard, Eide and Söderhamn2012) report on the experiences of a volunteer who found it difficult to contemplate her volunteer experience without sadness, and ultimately withdrew from the role. Similarly, van der Ploeg et al. (Reference van der Ploeg, Walker and O’Connor2014) described a sense of futility expressed by a small proportion of volunteers, and Guerra, Demain, Figueiredo, and De Sousa (Reference Guerra, Demain, Figueiredo and De Sousa2012) reported that some volunteers were afraid that their interventions were inadequate. In this study too, although most volunteers reported positive experiences, a few volunteers reported feeling discouraged. These volunteers were all working with residents with late-stage dementia. The similarity of this result with results of prior descriptive studies might suggest that it can be anticipated that a small proportion of new volunteers are unlikely to adapt well to working with people who have dementia. However, in this study, the result was more consistently observed when the resident had late-stage dementia, suggesting that the interaction between volunteer and resident characteristics should also be considered. Volunteers who lack familiarity with the progression of dementia might need more information about what signs indicate a meaningful positive response to a visit in late-stage dementia. Consistent with results from this study, other research on volunteer-led support for people with dementia has suggested that having access to ongoing, scheduled support should be considered to increase volunteer skill and efficacy, or to aid retention (Hall et al., Reference Hall, Brooke, Pendlebury and Jackson2019; van der Ploeg et al., Reference van der Ploeg, Mbakile, Genovesi and O’Connor2012). According to the CFIF, the responses of program participants ultimately positively or negatively impact program fidelity; therefore, it is important that these concerns be addressed in future volunteer-led MBIs.
Although several studies of MBIs have been conducted over the last three decades, very few of these have involved volunteers, and no other studies of fidelity are available for comparison. Given the need for stronger non-pharmacological psychosocial interventions in LTC, it is helpful to understand the full potential of volunteer workforce contributions (Seitz et al., Reference Seitz, Knuff, Prorok, Le Clair and Gill2016). This study was implemented in close to real-world conditions, with inexperienced volunteers, brief training, consultative support, and a fairly small commitment of time. On the whole, we found that MBI volunteers adhered to the key points communicated during their training, including staying with residents to ensure psychological and physical safety; choosing shared activities over a conversational approach; initiating an activity by modeling it first; being attentive to the needs, interests, and abilities of residents; and adjusting activities to ensure success (cf. Elliot, Reference Elliot2012). Nevertheless, a few limitations in quality were observed.
One observed limitation in volunteers’ approach to the MBI was a reliance on conversation during MBI activities. Although this emphasis was sometimes appropriate to resident’s roles and abilities, at other times, it was less appropriate. In addition, a gradual movement away from pre-designed MBI activities took place among three resident–volunteer dyads, all including residents with more severe cognitive impairment. In these cases, the MBI was gradually replaced with music, outdoor activities, and conversation. This might suggest an interaction between participant characteristics and program fidelity. The CFIF suggests that fidelity is strongest when program participants (in this case, both residents and volunteers) respond well to the intervention. Volunteers’ movement away from the kinds of pre-designed activities recommended in MBI guides (Camp, Reference Camp2006; Elliot, Reference Elliot2012) might suggest that MBIs, as communicated in these guides, are mismatched to the needs of people with late-stage dementia; alternatively, it might suggest that some volunteers are uncomfortable providing MBIs to people with late-stage dementia. A useful next step would be to study the fidelity of volunteer-led interventions tailored specifically for late stage dementia (e.g., Hunter et al., Reference Hunter, Kaasalainen, Froggatt, Ploeg, Dolovich and Simard2017; Simard & Volicer, Reference Simard and Volicer2010;).
Volunteer activities thrive in an atmosphere of mutual gain, including rewards such as finding meaning, feeling efficacious, and having a sense of community (Dwyer, Bono, Snyder, Nov, & Berson, Reference Dwyer, Bono, Snyder, Nov and Berson2013; Greenslade & White, Reference Greenslade and White2005; Okun & Michel, Reference Okun and Michel2006). Both questionnaire and interview results from this study suggested that all volunteers were experiencing at least some of these rewards. This is consistent with results from other studies in LTC (Damianakis et al., Reference Damianakis, Wagner, Bernstein and Marziali2007; Söderhamn et al., Reference Söderhamn, Landmark, Aasgaard, Eide and Söderhamn2012; van der Ploeg et al., Reference van der Ploeg, Mbakile, Genovesi and O’Connor2012), and, given that the CFIF suggests that the positive experiences of program participants are among the most important moderators of intervention adherence, volunteer-led psychosocial interventions would seem to have strong potential for success.
A few recommendations are warranted based on study results. First, on the basis of volunteers’ comments about program supports, we recommend that volunteer-led MBIs incorporate the following elements: initial training, a meeting with the resident, information about the resident’s abilities and interests, a diverse array of activities, and access to consultative support. Second, we recommend ensuring that resident and volunteer availability is well-matched. Third, we recommend that volunteers working with residents who have late-stage dementia have additional supports, potentially including: information about late-stage dementia, training in non-verbal communication, intervention skill modeling, and co-delivery of the intervention at the outset, as well as periodically scheduled check-ins. These recommendations are generally consistent with those made in prior studies (e.g., Holmberg, Reference Holmberg1997; Robinson & Clemons, Reference Robinson and Clemons1999; van der Ploeg et al., Reference van der Ploeg, Walker and O’Connor2014).
Research Strengths and Limitations
The use of the CFIF strengthened this analysis by fostering attention to a full range of intervention fidelity considerations. As a result, we were able to describe some issues that need further attention in clinical research on MBIs, including questions about the design and use of MBIs for people with late-stage dementia, and possible enhancements to training for volunteers supporting residents with late-stage dementia. Nevertheless, the fact that our analytic framework was not chosen a priori precluded a more robust and objective assessment of adherence to the intervention, which might have included directly observing volunteers’ use of Montessori activities and principles during visits with residents. In addition, one advantage of our sampling strategy, which relied on the participation of a small group of inexperienced volunteers, was that it allowed us to explore intervention fidelity in real-world conditions. However, studies with larger sample sizes will help to assess the generalizability of our observations, and controlled trials of volunteer-led MBIs will be necessary to examine program outcomes. Finally, we note that the VQQ was created for this study to address an absence of similar alternatives for the LTC sector. Although it had good inter-item consistency and test–retest reliability, it would benefit from additional validity testing.
Conclusions
This study examined fidelity to a Montessori-based intervention among volunteers who were new to dementia care. We found that most volunteers fulfilled their commitments. In addition, most implemented the key principles emphasized in their training, including active supervision, introducing activities by modeling, and modifying activities in response to residents’ interests and abilities. Moreover, volunteers suggested that both they and residents responded positively to the MBI. Although volunteers generally felt well supported by training and other available resources, those working with residents with late-stage dementia requested more support. Overall, this study suggests that with a small investment of time in training and ongoing support, volunteers who are new to dementia care settings can be engaged to provide a quality psychosocial intervention to residents with dementia. Yet, these results also illustrate the importance of ensuring access to support throughout the course of volunteer involvement, and of further examining the suitability of MBIs for residents with late-stage dementia.
