Authors: Allison Lindauer, Kristin G. Cloyes, Christina Zonker, Heather Franklin, Nathan Dieckmann, Hailey Chatterton, Michelle Kinsella, Aimee R. Mooney
Categories: Article, Frontotemporal dementia, FTD, Dementia, care partners, caregivers, intervention
Source: Alzheimer's & dementia. Behavior & socioeconomics of aging
Doi: 10.1002/bsa3.70063
Authors: Allison Lindauer, Kristin G. Cloyes, Christina Zonker, Heather Franklin, Nathan Dieckmann, Hailey Chatterton, Michelle Kinsella, Aimee R. Mooney
Frontotemporal dementia is the most common dementia in people under age 60. Although FTD is often characterized by behavioral symptoms, few national interventions address management of these symptoms. Here we describe our protocol for STELLA-FTD, an intervention for FTD care partners.
STELLA-FTD is an NIH Stage 1B randomized controlled trial that tests the efficacy of the intervention and will isolate the hypothesized mechanism of action. Two groups (test and control) receive information and peer support over eight weekly sessions. The test group will receive the behavioral management intervention, the control group will not.
Prior to implementation, the advisory board (STAR Council) tested the website and participated in mock sessions. Revisions were completed and recruitment started in June 2025.
STELLA-FTD addresses the needs of an under-studied group, FTD care partners. The study will provide evidence of intervention efficacy and identify the mechanism of action.
Frontotemporal dementia (FTD), a subset of Alzheimer’s disease and related dementias (ADRD), is a term for a group of related conditions characterized by abnormal aggregation of misfolded proteins, leading to a degeneration of the frontotemporal lobes of the brain.^1^ The neuropathologic changes of FTD manifest as behavioral and language symptoms that contribute to substantial family caregiver (“care partner”) burden, often resulting in negative physical and mental health outcomes.^2^ Despite being the most common dementia in individuals under age 60, few interventions exist that help family care partners manage the FTD-related behavioral symptoms and resulting burden.^3,4^
We designed STELLA-FTD (Support via Technology: Living and Learning with Advancing FTD) to address this gap. STELLA-FTD is a psychoeducational, behavior change intervention for family care partners, provided alongside peers, and delivered via videoconferencing technology. STELLA-FTD systematically trains care partners to manage behavioral symptoms using the ABC (activator, behavior, consequences) analytic approach.^5^ Grounded in self-efficacy theory^6^ and foundational research,^5^ STELLA-FTD trains care partners to address current behavioral symptoms and prepare for future ones by providing a structure for observing and describing the context in which a behavior occurs, the behavior itself, and what follows. By incorporating this information into routine decision-making and problem-solving, care partners adapt, refine, and observe the effects of these changes on behaviors and responses over time.
The design of STELLA-FTD was informed by pilot testing. Family care partners of persons with FTD participated in the eight-week intervention and post-intervention focus groups. The quantitative and qualitative data indicate that the intervention is feasible and acceptable to care partners and potentially efficacious in reducing care partner burden. A full description of the pilot testing is described elsewhere.^7^
Dementia due to FTD affects adults at the apex of their work and family life, with the peak incidence between the ages of 45 and 65.^3^ FTD disorders are a heterogeneous group of dementias characterized by progressive impairments in behavior, language and motor function due to degeneration in brain regions responsible for executive function, social comportment and language.^1,2,8,9^ FTD includes several subtypes, with the most common being behavioral variant frontotemporal dementia (bvFTD) and primary progressive aphasia (PPA) (in this paper we use the “FTD” for all diagnoses).^1,2,9,10^ The combination of behavior, language, and physical deficits causes substantial burden for the persons caring for those engaged in daily support for people living with FTD, their family care partners.^2,11^
Care partners report burden in all types and phases of FTD. In early stages, memory is often preserved, which can contribute to diagnostic confusion and delay,^12^ accompanied by family stress and frustration.^2,10,13–15^ With disease progression, behavioral symptoms in the person with FTD, such as irritability, anxiety, agitation, apathy, and disinhibition become more prominent, predicting higher levels of care partner burden.^15,16^ Behavioral symptoms occur in all FTD types but tend to happen at different phases of the disease trajectory. For example, in bvFTD, behavioral symptoms are prominent early on, but in PPA, the behavioral symptoms are more pronounced in the later stages. Language, motor, and global cognitive deficits further contribute to care partner burden.^2,9,14,17^
FTD care partner burden contributes to multiple negative care partner outcomes, including depression, anxiety, social isolation, exhaustion, pre-death grief and impairments in physical and financial health.^10,11,18^ Families report persistent worry about symptomology and care needs, as well as feeling ill-prepared to manage their family member’s future physical, language, and social needs. Affected families report a need and desire for education and guidance to facilitate effective symptom management throughout the disease trajectory.^10,14,19,20^
Although behavioral symptoms predict significant care partner burden across all FTD subtypes,^2,15^ few programs exist to help FTD care partners assess and manage behavioral symptoms. Further, despite the World Health Organization’s (WHO) directive that rehabilitation care be a core component in FTD treatment, few interventions incorporate rehabilitation care into managing the advancing behavioral, communication, physical and social FTD-related changes.^10,21^ Families can receive help from support groups offered by national organizations, such as the Alzheimer’s Association^22^ and the Association for Frontotemporal Degeneration (AFTD).^4,23^ However, support groups do not provide formal behavior management training, and active care partner engagement in programs is hindered by access and cost,^24–27^ particularly for underrepresented and younger families. Rural families in particular report having difficulty accessing knowledgeable clinicians, close proximity to support services, and the financial cost of travel to supportive services.^28^
Telehealth-based interventions can provide access to knowledgeable professionals, effective programs, and reduce travel costs. However, leveraging online resources to increase access to behavioral support programs remains under-researched among FTD caregivers.^7,29^ While a few interventions have been tested with FTD care partners, these do not specifically address behavioral symptoms,^30^limit enrollment to a specific FTD subtype,^31^ or lack power to identify significant findings.^32–34^ To address this gap, we translated a behavioral intervention for ADRD care partners^5,35^ into an FTD-focused intervention.
The revised intervention, STELLA-FTD, is a telehealth-based behavior change intervention designed with and for family care partners of those with FTD. With funding from the AFTD, we pilot-tested the NIH Stage 1A^7^ STELLA-FTD intervention and found that it is feasible and acceptable to care partners and has the potential to reduce burden.^7^ The STELLA-FTD curriculum is informed by nursing and rehabilitation science to support care partners’ efforts to address their family members’ behavioral, physical, and communication needs. While past work in caregiving science reveals ABC interventions can significantly reduce burden,^5,36,37^ the mechanism of action is not well understood.
Since the early 2000s, researchers have employed the ABC analytic approach to address dementia-related behaviors.^5^ Building on Teri’s work,^5^ our team and others^36–40^ have found quantitative evidence that these interventions reduce care partner burden. However, to our knowledge, none of these studies have examined the critical components within the intervention that produce improvements in care partner outcomes. That is, the essential ingredient in these interventions has not been examined or explicated. The knowledge gained from this study will advance the scientific understanding of how and why ABC-type interventions, and STELLA-FTD in particular, facilitate effective management of behavioral symptoms.
The specific aims of this NIH Stage 1B^41^ clinical trial are 1) test the efficacy of STELLA-FTD in reducing care partner burden, 2) test the mechanism of action of STELLA-FTD by isolating the ABC training component (the behavior change technique) and determining its relationship to self-efficacy^42^ (the intervention target) and burden (the primary outcome); and 3) prepare STELLA-FTD for NIH Stage 3^43^ testing by standardizing training materials and fidelity assessment processes.
We hypothesize the Aim Upon completion of the STELLA-FTD 8-week intervention, care partners who receive STELLA-FTD with ABC training (the test group) will have significantly lower (better) burden scores^44^ than those in the control groupAim care partners who receive STELLA-FTD with ABC training (the test group) will have significantly higher self-efficacy scores^42^ than the control group, and that self-efficacy will partially mediate the relationship between ABC training and burden
Framed by these aims, we will demonstrate STELLA-FTD’s early efficacy in reducing care partner burden, explain STELLA-FTD’s mechanism of action, and prepare materials for future community-based research.
The triadic theoretical foundation for STELLA-FTD (Figure 1), which mirrors the hypotheses and informs the process and outcome variables, is based on three principles.
First, behavioral symptoms contribute to care partner burden,^45^ the study’s primary outcome variable. Second, care partners can be trained to use the ABC analytic approach (the behavioral change technique), which may improve care partner self-efficacy, the intervention target. Third, self-efficacy, the mediating variable, may inform the process of how ABC training and burden are related. That is, ABC training can reduce burden, but only if the care partner feels confident in applying the ABC analytic approach when experiencing distressing behaviors.^6,41,46,47^
The first principle, supported by Kales et al.’s^45^ model and informed by Silverman et al.,^2^ posits that behavioral symptoms are a result of the interplay between neurodegenerative changes, care partner well-being and disease knowledge, and the wider environment. The second principle, framed by holistic learning theory,^47^ maintains that adults learn through three knowledge explicit (instruction in ABC approach), implicit (sharing experiences with peers) and emancipatory--using new-found knowledge to liberate themselves from burden. However, knowledge of the ABC analytic approach does not mean the care partners will actually use the technique.^6^
The third principle, the capacity to plan, act and adjust behavior and actions is influenced by a person’s belief in their ability, their self-efficacy.^6,46,48,49^ Care partners who perceive themselves as competent in managing behavioral symptoms will be more successful in mitigating them and, in turn, will feel less burdened by them. Self-efficacy is grounded in knowledge, but reinforced by experiences of success, witnessing the successes of others and positive feedback from perceived experts.^6^
STELLA-FTD addresses each pillar of the triadic theoretical foundation by recognizing the role of behavioral symptoms in relation to burden, targeting self-efficacy by training care partners in a method to address the symptoms and providing opportunities for success as they employ the ABC method. STELLA-FTD occurs within a social setting of other FTD care partners, providing support and vicarious experiences by witnessing the success of others. The new knowledge and increased self-efficacy can liberate care partners from negative reactions to the behavioral symptoms and may lower their burden.^6,45,47^
STELLA-FTD was funded by the National Institute on Aging (NIA) in July 2024 (R01 AG084523). An NIA-approved data safety management plan is in place with a safety officer.^50^ The full protocol was reviewed and approved by the OHSU Institutional Review Board, IRB # 26023. STELLA-FTD is registered with ClinicalTrials.gov, NCT05338710.
STELLA-FTD is a 36-week randomized controlled, two-group, repeated measures trial will test the effect of the test intervention on care partner burden. Care partners start the study with two pre-study surveys at Weeks 1 and 4. They then engage in the 8-week intervention, followed by repeated surveys post-intervention at Weeks 13, 20 and 36. This design allows for examination of the mechanism of action of STELLA-FTD by isolating the active component (ABC training) to determine its relationship with self-efficacy and burden. Figure 2 presents the overall design and flow of study activities.
We will engage a sample of Care Partners that will be large enough to identify statistically significant findings. To our knowledge, STELLA-FTD is the first study of its kind to test an ABC-type behavioral intervention with a large group of FTD care partners. Thus, the power analysis for STELLA-FTD is hampered by the lack of larger-scale studies. We therefore calculated the power based on published ADRD behavioral research that used our primary outcome variable, reactivity on the Revised Memory and Behavior Problems Checklist (RMBPC).^5,44,46^
In Teri et al.’s^5^ RCT that tested the ABCs with family caregivers for those with Alzheimer’s disease, the effect size of difference in change in RMBPC^44^ reactivity score between the two groups was 0.38. Based on this study, we conservatively determined the sample size calculation with an effect size of difference in pre-post intervention change as 0.3. To achieve at least 80% power to detect such a difference in a study design with three repeated measurements and an Intra-Class Correlation (ICC) =0.5 to 0.6 in compound symmetry covariance structure, we need at least 117 to 128 subjects/each group, for alpha = 0.05.
We will also enroll the persons with FTD for two reasons. First, the intervention requires discussions about their behaviors, and second, in our experience, persons with dementia sometimes enter the video space during the sessions and can be recorded.^36^ To inform them of their privacy and confidentiality rights, we will seek their assent to enrolling in the study. If they cannot assent, we will ask their legally authorized representative (LAR), as dictated by OHSU regulations. The person with FTD (or their LAR) must assent to the study and stay in the study for the duration of their care partners involvement. If they leave the study, they and Care Partner will be withdrawn. In our experience,^35^ the primary reason for withdrawing of the person with dementia is due to their death. In our current study, 10% of the attrition was due to care recipient death.^35^ Thus, we will increase our sample size to account for this by enrolling 160 care partners for each group (320). We will also enroll their 320 care recipients. The care recipients will not take part in any of the intervention.
The criteria are informed by our previous work.^7,35–37^ For STELLA-FTD all participants will be adults over the age of 18 who live in the United States. Care partners must speak English and have adequate vision and hearing to participate in the telehealth-based study. Care partners must have a phone, internet access and be willing to participate in the video conference-based meetings. They need to have an email address to receive study materials and assessments. Care partners must document at least two behaviors, in the person with FTD, that are at least moderately distressing for the care partner and occurring in the last two weeks, such as care recipient pacing or voicing embarrassing comments in public.
To be included in STELLA-FTD, care recipients must have a diagnosis of any type of FTD such as bvFTD, PPA, Progressive Supranuclear Palsy, Corticobasal Syndrome, or FTD with ALS. Sampling will not be stratified by FTD subtype. We will ask care partners the diagnosis in the pre-screen survey and at enrollment. During enrollment, we will ask the care partner to attest that the care recipient has an FTD diagnosis by a qualified professional (i.e., nurse practitioner, osteopathic doctor, physician, physician associate, psychologist). We considered verifying the diagnosis with the participants’ medical providers; however, in our experience, this deterred some potential participants and caused substantial delays.^36^ We also considered limiting the sample to a single diagnosis (e.g., PPA or bvFTD), but behavioral symptoms occur in all FTD subtypes.^2^ We learned from our pilot, in which care partners had family members with bvFTD, PPA and progressive supranuclear palsy, that the care partners liked learning from each other. In post-pilot focus groups, care partners recommended that STELLA-FTD include all FTD diagnoses. Care partners need to provide care for their family member with FTD for at least four hours per week, but this care does not need to be in-person. Care partners experience burden regardless of the location of the person with dementia, thus they do not have to live with them. Care partners who attend an external support group can participate.
Recruitment for STELLA-FTD is challenging. We will need to recruit 320 care partners for persons with a rare dementia. We will recruit nationally for STELLA-FTD, engaging multiple strategies, from in-person promotion to wide-spread social media use. For example, we will promote STELLA-FTD at seminars hosted by the National Alzheimer’s Coordinating Center^51^ and on podcasts, such as “Remember Me,” which fosters community support for families living with FTD. We will recruit from our Oregon ADRC and the other ADRCs around the nation.^52^
Our pilot funder, the AFTD, will promote the study and assist us with contacting FTD support group leaders. We will also recruit from two national entities. The ALLFTD is a nationwide study that focuses on characterizing FTD conditions, collecting cognitive, clinical, and diagnostic data and sharing data with scientists.^53^ The FTD Disorders Registry includes persons with FTD and their care partners.^54^ These registries will facilitate the recruitment of care partners with varied backgrounds and ensure that our sample represents a range of care partner experiences.
Our retention approach is informed by Grill et al.^55^ who found multiple strategies result in lower rates of attrition. This includes employing a diverse team and providing a fair stipend to participants ($200). The stipend is pro-rated, so care partners who complete some study activities will still receive some funds. Quarterly newsletters provide updates on recruitment goals and presentations made by our team (e.g., posters from professional meetings). An annual holiday card with a team photo thanks participants for their engagement.^55^
Interested care partners will be directed to our website, www.stella-ftd-study, to complete the electronic pre-screen survey via Qualtrics^56^ to see if they qualify for further screening. This survey provides enough information to inform the study staff if further screening is needed. If the potential participant meets the pre-screening criteria, our research coordinator will contact them to complete a full screen via phone or video, following OHSU screening protocol. The recruitment team will track the number of care partners screened for the study, the number consented, the number of those that dropped out, the number who completed some of the intervention, and the number of those that completed the entire intervention.
If a potential care partner participant is eligible (passes the full screen), study staff will then arrange an appointment to review the Information Sheet and Authorization Form (ISAF). The coordinator will confirm comprehension by administering the Knowledge Check. If the potential participant answers all questions correctly, they are eligible to be enrolled in the study. We will save a copy of the ISAF and the Knowledge Check Form in our study’s REDCap database.^57^
The person with FTD will not partake in the STELLA-FTD intervention, but we will collect and analyze data about them from their care partner during the study period. Due to the complicated nature of FTD conditions, the care recipient may be decisionally impaired. If the person with FTD cannot agree to participate due to impairment, we will ask the care recipient’s LAR to review the care recipient-specific ISAF and agree to study participation.
After the care partners and care recipients have agreed to participate in STELLA-FTD via the ISAFs, the study staff will arrange a phone or Webex^58^ appointment to enroll care partners. At that visit, staff will collect enrollment information (e.g., demographics, contact information, confirmation of diagnosis) and provide the STELLA-FTD website orientation.
Care partners are randomized to either the test or control group using an electronic randomization process and are blinded to their group designation and details of the differences between the groups. The control group receives the basic STELLA-FTD curriculum; the test group receives the basic curriculum plus training in the ABC analytic approach.
STELLA-FTD is a technology-based intervention. We have been conducting telehealth-based care partner research since 2016,^7,20,35–37,46,59–61^ and our experience informs delivery of STELLA-FTD in a safe, acceptable manner. The intervention, assessments and all communications are delivered by electronic, videoconference, or by telephone strategies. The ISAF explains technology and privacy risks.
Each STELLA-FTD participant has access to the STELLA-FTD technology support team. This team has extensive experience in working with individuals who are technologically naïve and/or overwhelmed by their caregiving work. The team tests the participants’ access to the videoconferencing sessions. If all attempts fail to connect the care partner to the sessions, the care partner can join the sessions via phone. This is the practice in our current projects, where fewer than 4% have needed to change platforms (e.g., switch from computer to phone) during a study due to technology issues.^62^ STELLA-FTD will provide laptop devices for approximately 15% of participants that may need a computer to participate in the study.^63^
All care partners receive materials that were tested in our pilot^63^ and include group-specific handbooks, the FTD Roadmap©, and access to the STELLA-FTD interactive website with instructional resources and videos. The materials are sent via mail and/or email to all care partners after they have completed the week 4 pre-intervention assessment. Access to the STELLA-FTD website begins after pre-intervention assessments are completed. Care partners may keep their materials. Over the course of the study, we will seek feedback from care partners about the utility of the materials and revise them for a future NIH Stage 2 study.^43^
The STELLA-FTD website contains all training materials. Both test and control groups have access to written instructional resources and recorded videos; the test group has additional access to ABC planning information and documentation. Professional videos (10–20 minutes) will inform care partners about how different specialists can help address current FTD symptoms and prepare for future needs. Video Summary Sheets are available on the website. Care partners may access the STELLA-FTD website with their password as long as they are enrolled in the study. We will monitor engagement with the website to track objective data on how often participants log in and view the resources and videos.
The STELLA-FTD intervention is delivered over 8 weekly video conferencing group sessions. The intervention is provided by “Guides,” individuals with experience in FTD care and/or research. There are two Guides for each group per 8-week cohort, for both control and the test group. Guides are masked to which group they are leading.
Care partners in the control group receive the basic STELLA-FTD curriculum. The curriculum was informed by pilot testing and involves 8 modules on topics about care challenges and resources for families; curriculum topics, presenting health care professionals, and examples are listed in Table 1.
This multi-modal curriculum includes synchronous video-based instruction about FTD as well as asynchronous access to the weekly videos and written materials. During all sessions, explicit knowledge acquisition^47^ is facilitated as the Guides highlight key points of the video, encourage interaction between care partners, and discuss use of their handbooks.
The test group will receive the basic curriculum discussed above plus explicit training in the ABC analytic approach to address care recipient target behaviors. A detailed description of the ABC approach can be found in our earlier publications.^7,35^ Training in the ABC analytic approach in the test group incorporates theory-based training to promote care partner self-efficacy in managing the FTD-related behavioral symptoms in four important ways. First, care partners are trained in the ABC approach and instructed to create and actively use their ABC plans, write them down on the STELLA-FTD website, share their ideas with the group, and note their successes and modify the plans as needed. Second, care partners are advised to translate their ABC plans into action by testing them with their family members with FTD. Third, Guides use strength-based feedback to validate and reinforce care partner efforts; and fourth, peer interaction allows care partners to witness the trials and successes of other group members. These strategies provide progressive opportunities for success, which, according to Bandura, is essential to building self-efficacy (the intervention target), translating knowledge into action and ultimately, reducing care partner burden (the outcome variable).^35^
To assess enactment of the ABC analytic approach, Guides directly observe care partners in group discussion regarding their use of their ABC plans and how they applied them to their target behaviors. To verify receipt of the intervention, Guides query care partners about application of their ABC plans. We will collect the data from their ABC plans from the STELLA-FTD website to identify changes in behaviors. Care partners will record the number of ABC plans they developed each week, how many they implemented, and all the places they recorded these plans, in a weekly survey.^64^
Table 2 presents all study measures. The primary outcome variable, burden, is assessed with the 24-item RMBPC,^44^ which measures the frequency of care recipient behavioral symptoms and care partner reactions to these behaviors. The RMBPC was chosen because it aligns with our theoretical foundation that assumes burden is a result of care partner reactions to behavioral symptoms. The RMBPC has been used in Teri’s work^5^ and our telehealth studies,^35–37,61^ including the STELLA-FTD pilot, allowing for comparison across studies. We will use the traditional 24-item measure. We will also use the RMBPC FTD Supplement, which includes nine additional test questions that address FTD symptomology. In addition, both groups (test and control) will document the target behaviors they want to address, how frequently they occur, and how upsetting they are, on the STELLA-FTD website.
Self-efficacy, the mediating variable, is assessed using Fortinsky’s measurement of family care partner self-efficacy for managing dementia.^42^ Few measures address self-efficacy in the context of FTD Care, thus we identified sclaes that, while not specific to FTD, addressed the variables in our Dementia-related behaviors and community service use (e.g., home based rehabiliation services). Fortinsky’s scale queries caregivers about their certainy that they can handle behavior problems and seek community-based services. Crellin et al.’s^65^ measure is used with the Neuropsychiatric Inventory.^66^ The NPI includes behaviors found in FTD (e.g., disinhibition), and Crellin’s rating system is simple to administer. However, Crellin’s scale does not measure of community service usage.
FTD severity will be measured with the Clinical Dementia Rating Scale plus National Alzheimer’s Coordinating Center Frontotemporal Lobar Degeneration-Motor (CDR+NACC FTLD-M) assessment. This measure assesses disease severity, behavioral, language and motor symptoms. Responses about the person with FTD will be provided by the care partner.^9^
The “Orbit” survey was adapted from the weekly survey developed by the Oregon Roybal Center for CAre Support Translational Research Advantaged by Integrating Technology (ORCASTRAIT), housed at the Oregon Alzheimer’s Disease Research Center (OADRC).^67^ The survey measures care partner emotional, physical, and financial strain, medication use, and whether their care recipient moved into long term care (e.g., assisted living, memory care). The survey asks about the frequency of contact with both the care partner’s and the family member with FTD’s clinical providers for health, behavioral, or medication-related concerns. Orbit surveys also collect data on adverse events. The weekly collection of illness and injury in FTD families will provide valuable information on health risks and healthcare usage. Further, when the team is notified of a concerning adverse event, care partners are contacted, which many appreciate. The surveys take approximately 5 minutes to fill out and are emailed to the care partners weekly using Qualtrics.^56^
To assess acceptance of the intervention, care partners who agreed to be recontacted after completing the intervention will be asked to participate in optional focus groups after completing all activities, and information about the focus groups is detailed in the ISAF. Focus groups will generate data describing participants’ perceptions of intervention, including acceptability, usability, and perceived value of specific components, and identify barriers and facilitators for implementation and sustained use. Participants will also complete the Post-study System Usability Questionnaire (PSSUQ).^68^
Care partners who complete all components of the intervention, all surveys, and 80% of the Orbit surveys will receive the full stipend of 25 stipend.
All surveys will be sent to care partners using Qualtrics^56^ or REDCap^57^ platforms. The Qualtrics and REDCap platforms are HIPAA compliant and secure. We will assess completion rates and internal consistency of the online platform surveys.
This study will generate a comprehensive set of data that tests STELLA-FTD’s efficacy and its mechanism of action. In all final models, we will examine the relationships between diagnosis type, demographic, and biological characteristics (e.g., sex, race, education) and the outcome and process variables. All the results will be considered significant at α = 0.05.
For Aim 1, we will first assess the RMPBC^44^ reactivity score in the two pre-intervention assessments (completed at weeks 1 and 4) in both test and control groups. First, we will provide both statistical and graphical evaluation for the distributions of the primary outcome and transform the data to normalized values if needed (e.g., log transformation). We will descriptively explore the RMBPC^44^ reactivity score at weeks 1 and 4 for each group and the primary outcome will be reported as mean (SD). To test whether reactivity scores change from week 1 to week 4, paired t-tests will be used to compare scores within each group. Then, we will compute the change scores between week 1 and week 4 for each participant and the normality of change score will be assessed. If normality assumption is violated, we will perform a data transformation. To test the difference in change scores between weeks 1 and 4, we will conduct a two-sample t-test of the change score between the test and control groups. We expect no significant change between weeks 1 and 4 within each group and the change scores will not be significantly different between the test group and control group.
Next, to assess if the test group has significantly lower (better) burden scores than those in the control group, we will employ a mixed effects model to analyze whether the change in reactivity on the RMBPC^44^ differs across groups (test vs. control) at the conclusion of the 8-week intervention. Two fixed effects will be included in the 1) Group effect will be included as a dichotomous categorical variable (test vs. control); 2) Time effect will be included as a continuous variable. The interaction between group and time (group × time) will also be included. Random intercepts will be incorporated for subject specificity in the model to account for clustering effect within subject over time. We will perform a hypothesis test for the coefficient of the interaction term (group × time) to assess the difference in change on primary outcome in the intervention group as compared to control group, over time. We will extend the model to include diagnosis type (e.g., PPA, bvFTD), frequency of behavioral symptoms as measured on the RMBPC,^44^ socio-demographic covariates (e.g., sex, age, sex, education) and external support group use. Similar analyses will occur at Week 20 and 36.
To test the mechanism of action (Aim 2), the statistical analyses will be carried out in three steps to investigate the mediated effect of self-efficacy (Fortinsky’s measure,^42^) on the association between intervention and primary outcome, the RMBPC^44^ reactivity score.
First, we will assess self-efficacy that using mixed effect model analyses on total self-efficacy score as the outcome. A hypothesis test for the coefficient of the interaction term (group × time) will be performed to assess the difference in change on self-efficacy score in the test group as compared to control group, over time.
Second, we will construct a mixed-effect model on the RMBPC^44^ reactivity score as the outcome. Total self-efficacy score^42^ will be included as a continuous independent variable; time will be included as a continuous variable. The interaction between total self-efficacy score and time (self-efficacy score × time) will also be included. Random intercepts will be incorporated for subject specificity in the model to account for clustering effect within subjects over time. A hypothesis test for the coefficient of the interaction term (self-efficacy score × time) will be performed to assess the relationship between change on primary outcome (RMBPC reactivity score^44^ and total self-efficacy score,^42^ over time.
In the third step, we will extend the established model to include total self-efficacy^42^ score as a continuous covariate and the interaction term between total self-efficacy score and time (self-efficacy score × time) will also be included. We expect the interaction term coefficient (group × time) will be significant, and the coefficient of the interaction term between group and time (group × time) will be attenuated and may become non-significant after extending the model to add total self-efficacy score as a covariate and the interaction between total self-efficacy score and time (total self-efficacy score × time). This will indicate that the self-efficacy has a mediated effect on the association between the intervention and change in the primary outcome (RMBPC reactivity score^44^) as compared to the control group. We will also extend these models to include diagnosis, frequency of behavioral symptoms as measured on the RMBPC,^44^ socio-demographic covariates and preparedness.
To inform assessments of self-efficacy in future FTD studies, will assess construct validity of the two self-efficacy scales (Fortinsky^42^ and Crellin^65^) by analyzing the correlation between the two scales and CDR+NACC FTLD-MM. Convergent validity will be assessed by analyzing the correlation of the two self-efficacy scales.
Items from the Orbit survey^35^ will be examined as repeated weekly measures in each group.. We will first explore the data descriptively and then manage or transfer the variables from survey items if needed. Categorical variables will be displayed as frequency (%), and continuous variables will be shown as mean (SD) or median (25^th^ percentile, 75^th^ percentile). These measures will be compared between two groups each week using a two-sample t-test for continuous variables and chi-square test for categorical variables. Within each group, we also plan to evaluate the change of measure over time using line charts and compare the line pattern between two groups for continuous variables. The change in continuous variables between two time points within the group can be assessed using paired t-tests. The change in categorical variables between two time points within the group can be assessed using chi-square test.
Focus group data will be analyzed in iterative deductive and inductive cycles (e.g., deductive coding for PSSUQ subscale constructs, inductive descriptive coding of care partners’ experiences.) This will identify patterns relevant to care partners’ perceptions of intervention structure, content, usefulness, ease of use, the value of components considering their own caregiving experiences, and their suggestions for improvement and/or refinement. These data, when integrated with the PSSUQ results and interpreted in relation to engagement metrics and survey results, will inform the development of future studies
The STELLA-FTD trial employs a rigorous fidelity assessment protocol that evaluates interventionist (Guide) adherence to protocols and participant engagement with the intervention. We utilize this protocol in our current study^35^ to reduce threats to internal validity and facilitate dissemination of findings and procedures. This protocol is informed by evidence-based standards, including recommendations from the NIH Treatment Fidelity Workgroup Behavior Change Consortium^69^ and addresses five components. The first three components (design, intervention receipt, enactment) are discussed above. Training and delivery monitoring^70^ are addressed next.
Guides participate in standardized didactic and experiential training. Best practices in managing and facilitating group sessions are emphasized. Mock sessions allow Guides to practice group strategies such as management of inappropriate behavior (e.g., use of disparaging comments) and inclusion of reticent participants. Prior to leading group sessions, Guides demonstrate essential competencies in behavior identification, intervention delivery, and data documentation. All training is completed via videoconferencing and recorded to maintain consistency across Guides.
All STELLA-FTD sessions are recorded, and recordings are stored on a secure server that the fidelity assessment team can access via the OHSU secure filesharing system. Over the course of the study, the fidelity team will view a randomly identified subsample (20%) of the recordings to evaluate procedural, documentation, and theoretical fidelity. Fidelity items are scored using a binary rating of “1” (present) or “0” (absent), with clearly defined characteristics required for either rating. Inter-rater agreement on video assessments will be measured.^71^ We will hold quarterly team meetings to review fidelity compliance and give specific feedback to individual Guides as warranted after fidelity assessment.
Guides are required to review recordings of each session within 8 hours of the sessions and complete standardized protocol adherence checklists to document alignment with criteria in each fidelity domain ensure essential study components are administered and verify adherence to the protocol. Session duration is monitored to prevent potential threats of variance in dosing between trial groups. Our fidelity team will incorporate multiple strategies to enhance fidelity and prevent intervention drift, including weekly guide meetings with fidelity specialist and weekly Guide emails with protocol tips. We actively monitor for potential contamination between the test and control groups. Our study design, operationalized procedures, and ongoing and diligent fidelity monitoring will aid in early identification and management of fidelity concerns.^69^
The STELLA Team Advisory Resource (STAR) Council provides structured input to ensure study activities reflect community perspectives and are culturally and contextually appropriate. The Council is comprised of former study participants, caregivers, and community members with lived experience and an interest in dementia caregiver well-being. STAR Council members review study materials and recruitment strategies to provide user-centered feedback. Council members participated in beta testing of websites and technology tools and attended mock intervention sessions as practice participants to support staff training. They will review dissemination products (e.g., manuscripts, posters, newsletters). The STAR Council members are provided with up to $50/hour of service as funds allow. No PHI of any study participants will be shared with STAR Council members. The study team will review Council feedback to inform materials, procedures, and dissemination. Engagement quality will be assessed annually.
The STELLA-FTD intervention started recruiting participants in June 2025, with 12 care partners starting the intervention in August 2025.. The study will be active until June 2029. The study is being implemented with an adequately powered research design that will test the hypotheses. We have a fidelity assessment process in place to ensure adherence to protocol and an advisory council to ensure alignment with community values, needs, and goals. To answer our research questions, we will collect quantitative and qualitative data and employ a well-thought-out analytic plan to assess acceptability, feasibility, and efficacy of the STELLA-FTD intervention.
To our knowledge, STELLA-FTD is one of the first national studies that tests an intervention designed to reduce the frequency of FTD-related behaviors and care partners’ reactions to the behaviors.^44^ This study is unique in two important ways. First, this RCT provides an opportunity to test the mechanism of action of ABC training. The protocol aligns the intervention in both test and control group as closely as possible, allowing for isolation of the effect of ABC training on self-efficacy and burden. STELLA-FTD is powered to identify the efficacy of ABC training. Thus, if the findings point to significant improvement in behavior management in the ABC group, future behavioral interventions can confidently include this approach.
Second, the intervention was developed and informed by nursing and rehabilitation science. Nurses have recognized the essential value of person-centered care since Nightingale recognized the concept in the 1800s.^72^ STELLA-FTD manifests this concept by allowing care partners to identify the specific behavioral symptoms that are personally relevant to them, and those behaviors they find “most distressing.”. While the RMBPC provides a common list of behavioral symptoms,^44^ it does not allow for recognition of the extremely personal challenges care partners face. For example, in our earlier work, care partners reported behavioral symptoms not found on the RMBPC, such as “overfeeding the dog,” or “refusing to get off the couch.”^36^ STELLA-FTD centers the care partners and recognizes their expertise in caring for their family member.
The World Health Organization mandates inclusion of rehabilitation professionals in dementia care.^73^ Our clinical experience recognizes that families living with FTD benefit from the support of speech and language pathologists, physical therapists, and occupational therapists across the disease trajectory. Therefore, these specialists contributed to the design of the user-informed curriculum.^7 30,31,46,60,74^
In preparing the STELLA-FTD intervention we met challenges that caused delays in active implementation. From a regulatory standpoint, STELLA-FTD was delayed by notices from the NIH (NOT-OD-25–090) that required thoughtful consideration from the OHSU leadership prior to active recruitment. In addition, we faced delays in developing the STELLA-FTD website and video development.
We were able to assess the functionality of the intervention website with another pilot study that tested a fully online version of the ABC intervention for care partners of persons with any dementia (NCT06460012). We learned that the website required multiple revisions to collect the necessary data, and the vendor engaged in the pilot was unable to adequately address our concerns. We thus made the difficult decision to change vendors. This resulted in delays due to the requirement for a privacy and security review, and a formal business agreement with our institution.
Our initial plan, and the one used in our STELLA-FTD pilot, was to collaborate with our peers to develop recorded presentations for care partners to view during the weekly sessions. Early in the pre-implementation phase, however, we realized that the audio-visual quality of the videos recorded by our healthcare provider peers did not meet the professional production standard required for this 5-year national study. We thus hired a professional production team to produce higher-quality videos suitable for use in the curriculum.
Addressing these challenges resulted in delays and higher expenses. However, we recognized that we would have limited ability to change the study components once we were actively implementing the intervention. We chose to assume the financial and time-based costs prior to launching the intervention.
We have identified limitations to the study. It could be argued that including care partners for all types of FTD could affect the outcome measures or limit camaraderie among the care partners in the groups. However, prior research indicates that behavioral symptoms are noted in all FTD types, and thus, the primary outcome, burden, would not be affected by FTD type. Regarding group cohesion, our pilot study found that care partners appreciated hearing from other care partners with different experiences and perspectives.^7^ Nonetheless, the RMBPC and satisfaction surveys will be analyzed by group to identify important differences.
Recruitment for STELLA-FTD will be challenging, however, the FTD care partner community is very committed to supporting research. In the first month of recruitment, by partnering with the AFTD, we recorded over eighty care partners signing up to be contacted by our study. Nonetheless, we recognize recruitment will have marathon-like qualities that require consistent and persistent efforts to reach the target sample. We will continue to work with the AFTD, the FTD Disorders Registry, and care partner support programs, such as those with the Alzheimer’s Association. We will seek out opportunities to promote our study via media posts and on social media.
STELLA-FTD is a multi-component study that will identify the elements that improve care partner burden. The national reach, coupled with broad organizational support, will allow the recruitment of a large sample. The resulting findings will provide a foundation for wider dissemination of the intervention, thus addressing the needs of FTD families as they strive to manage this complex disease.