Integration of Health Information Technology and Promotion of Personhood in Family-Centered Dementia Care: Intervention Trial

Communication deficits are common among PLWD and are due to a variety of symptoms associated with ADRD, including various types of primary progressive aphasia, memory loss, word-retrieval anomia, and decreased attention span (Adlam et al., 2006; Ash et al., 2019; De Vries, 2013; Mesulam, 2003; Woodward, 2013). For the PLWD, these deficits can lead to several poor outcomes, such as depression, social isolation, mood and behavioral disturbances, and premature institutionalization (De Vries, 2013; Potkins et al., 2003). Communication deficits of the PLWD also pose challenges to ADRD caregiving and clinical care, which is already complex. Care partners (i.e., unpaid persons providing caregiving and support to PLWD, often family members) report feeling overwhelmed and burdened with having to make the numerous daily decisions about the PLWD’s daily care, such as selecting meals and clothing to wear (Author et al., 2019). Care partners also find proxy medical decision making for the PLWD stressful, due to communication challenges or inability to engage in shared decision-making with the PLWD (Miller, Whitlatch, & Lyons, 2016).

The communication deficits associated with ADRD may result in care partners (and others) having a conflict with the PLWD, or experiencing depression, burden, or isolation (Hendryx-Bedalov, 2000; Potkins et al., 2003). Communication deficits associated with ADRD may also pose additional challenges to clinical care. Providers report that communication problems of PLWD create barriers to collecting information needed for clinical decision making (Hinton et al., 2007). Communication problems may create barriers to assessing the PLWD’s symptoms, recognizing medical problems, or referring them to needed services, such as palliative care (De Vries, 2013; Erel, Marcus, & Dekeyser-Ganz, 2017; Johnson et al., 2009).

Rationale

Given the potential negative outcomes due to communication deficits among PLWD, exploring how to improve upon traditional AAC devices so that they can promote and enhance personhood while also further supporting the family unit providing them care is warranted. However, criticisms of AAC devices persist, including: (1) limitations to customizing their content and design to reflect the daily care preferences of the PLWD (e.g., clothing, meal choices) and; (2) their tendency to target the communication challenges of the PLWD without taking into account how communication deficits impact the larger family unit during daily caregiving activities (May, Dada, & Murray, 2019; Volkmer et al., 2019). To address these criticisms, our interdisciplinary team is developing and evaluating an electronic AAC device that goes beyond traditional AAC devices in by offering customizable content that reflects the preferences and interests of the PLWD. In addition to its applications for caregiving, it will also be designed to help the PLWD communicate their daily experiences and clinical data that can be shared with the person’s health care provider and tracked over time. Given these advancements over traditional AAC devices, we refer to our current intervention as an AAC Plus.

Previously, our team developed a web and smartphone application (app) called Care-IT, which was designed to facilitate dyadic communication between care partners of PLWD and the care recipient’s provider about clinical symptoms in real-time (Author et al., 2016). Care-IT has demonstrated its feasibility at facilitating dyadic communication (Author et al., 2016).

To promote triadic communication, the AAC Plus will be networked with this app and will address limitations to existing AAC devices for PLWD, such as: (1) it has customizable interfaces for the care partner and PLWD; (2) it is designed to promote engagement and communication between them during caregiving activities using touchscreen technology; and (3) it facilitates discussion and tracking of behavioral and other clinical information about the PLWD (e.g., sleep, appetite). This information can be shared in real-time with a health care provider for early detection, intervention, and monitoring of clinical symptoms.

Methods

Study Design

The primary aim of the planned clinical trial is to assess the effectiveness of utilizing the newly-developed AAC Plus, as an adjunct to care and caregiving for 12 months. This report describes the planned clinical trial was guided by the (SPIRIT 2013 Statement) (Chan et al., 2013a, 2013b). The trial will utilize an experimental design (parallel-group) involving 120 triads (PLWD, care partner, healthcare provider) (see Diagram 1). Triads will be randomized to one of two groups (see Study Arms below): the experimental group, which involves the “full intervention” that includes using Care-IT and the networked AAC Plus; or the “minimal intervention,” where the care partner uses the Care-IT app; however, the triad is assigned a traditional paper-version of an AAC to support communication with the PLWD. The same healthcare providers will be involved with triads assigned to the “full intervention” and “minimal intervention.” More information about the interventions is provided below.

Trial Hypotheses (H1, H2)

There are several hypotheses and research questions that will be assessed through this study. The two main hypotheses include:

1. H1: Triads (PLWD, care partner, healthcare provider) randomly assigned to the “full intervention” and use the AAC Plus will experience a greater increase in positive outcomes (e.g., improved quality of life) compared to those assigned to the “minimal intervention.”

2. H2: Triads (PLWD, care partner, healthcare provider) randomly assigned to the “full intervention” and use the AAC Plus will experience a greater decrease in negative outcomes (e.g., care partner depression, care partner burden, behavioral disturbances in ADRD care recipient) compared to those assigned to the “minimal intervention.”

The clinical trial will also explore the following research questions:

1. In what ways do usability and satisfaction of the AAC aids (both paper and IT versions) differ across sub-groups of care partners and PLWD (e.g., based on race, ethnicity, rural versus urban)?

2. To what extent does the frequency of use and sharing of information among members of the triad differ across sub-groups?

3. In what ways does the quality and usefulness of information shared between the care partner, PLWD, and the healthcare provider differ across groups?

Study Arms

Experimental: Full Intervention.

TRIADs (care partner, PLWD, and healthcare provider) randomly assigned to the experimental arm will have access to use Care-IT AND the newly developed AAC Plus app. The provider will receive information via the AAC Plus app.

Active Comparator: Minimal Intervention.

TRIADs (care partner, PLWD, and healthcare provider) randomly assigned to the minimal intervention arm will have access to Care-IT BUT only a paper version of an AAC.

Interventions

The Care-IT App

Participants in both arms of the study will have access to the Care-IT app, which is designed to support and educate caregivers of PLWD and also improve the quality of clinical information sharing between caregivers and health care providers. Care-IT is an IT that addresses the lack of care partner support and education about PLWD care that often results in depression, burden, and low caregiving self-efficacy. The CARE-IT intervention provides care partners with electronic and video-based educational materials, links to services, and self-assessments for depression and burden that offer feedback. It also allows the care partner to assess and document symptoms of the PLWD using clinical assessments, which can be shared with the provider.

The AAC Plus

The AAC Plus will only be provided to the experimental group and will rely on a touchscreen technology platform (i.e., tablet). Touchscreen technology use among PLWD has been the focus of multiple studies, with two recent systematic reviews identifying a total of 54 unique investigations on the topic (Hitch et al., 2017; Joddrell & Astell, 2016). The interface for the AAC Plus (see Image 1) offers the ability to customize the AAC to individual’s interests and preferences, rather than a set of standardized icons. The PLWD would use the AAC Plus in partnership with the care partner to create a profile of the care recipient’s preferences with pictures, and text that represents these preferences (e.g. pictures of food, clothing, and activities). For example, the care partner could create a profile with pictures of the care recipient’s actual clothing, rather than generic icons of shirts, pants, shoes, etc. The AAC Plus would have training videos built into the application to coach the dyad on how to use it during daily caregiving activities to facilitate preference choice by the PLWD. Thus the AAC Plus provides the PLWD with opportunities for engagement and interaction with the care partner, other family members, or providers.

To understand how the AAC Plus may support caregiving and promote personhood, consider the following fictitious case:

Mary is 80 years old, has multiple medical problems and was diagnosed with vascular dementia six years ago. Mary is often confused, anxious, and does not sleep well. Her family, including her live-in daughter, makes every effort to let the daily paid nursing assistants know Mary’s care preferences (e.g., types of food, clothing, grooming), interests (e.g., preferred TV shows, music, daily walks, and manicures) and abilities (e.g., able to select her clothing, able to style her own hair). Mary has verbal communication deficits, often communicating through single words or short phrases, which frequently results in family members and in-home caregivers not giving Mary a choice on what to wear, eat, or her daily activities, despite their good intentions for her daily care. Sometimes Mary exhibits care-resistant behaviors (e.g., aggression, lack of cooperation) when she seems unhappy with food or clothing suggestions. Using the AAC Plus Interface, Mary’s caregivers are able to communicate with Mary about what she wants to eat or wear by using pictures, and Mary can select them using the touchscreen, rather than having to communicate about them verbally. Therefore, Mary is more alert and engaged during mealtimes and dressing. Essentially, the AAC Plus Interface is a tool that compensates for Mary’s communication problems and provides a method for her to make her desires known.

Paper AAC

The paper version of the AAC has been specifically designed for use with PLWD (Bourgeois, 2014). Hence, the study will address whether a technology-based AAC would provide additional benefits over those that use paper.

Study Setting and Recruitment

Two memory centers are participating in the clinical trial, one located in Miami, Florida and the other in Birmingham, Alabama. In Miami, the Mind Institute is a mental health and memory clinic located on a 22-acre campus in Dade County, serving urban south Florida, which is racially and ethnically diverse. The Mind Institute was recently designated by the Florida legislature as a state memory disorder clinic and serves 500 new patients per year, who average 88 years in age and are racially/ethnically diverse (80% Caucasian, 20% African American/Caribbean American, 25% identify as Hispanic). In central Alabama, the Memory Disorder Clinic (MDC) is part of the University of Alabama at Birmingham (UAB) School of Medicine, and is located in a region with large rural and African American populations. There are three neurologists and three nurse practitioners who provide ongoing clinical services. MDC is the only interprofessional program in Alabama and conducts patient care sessions 5 days per week. Due to geographic isolation in rural areas, some PLWD and care partners travel three hours or more to the UAB clinic for services (some as far as Mississippi). Prior research has shown that memory clinics are ideal settings for successfully recruiting persons with dementia from racial and ethnic minority groups (Morrison, Winter, & Gitlin, 2016).

We will recruit participants from these two sites, which means that the 120 triads will include 120 unique dyads of care partners-PLWD (60 at each site) and their healthcare provider. Diagram 1 provides an overview of the protocol procedures. The targeted sample size for dyads takes into account a 36% attrition rate that has been consistently found in prior studies similar to this one (Czaja et al., 2018; Gitlin, et al., 2018).

Prior to the COVID-19 pandemic, PLWD typically attend medical visits at the memory clinics with their care partner. However, under current pandemic conditions almost all of the medical visits are virtual; therefore, the recruitment strategies may need modifications. We will make recruitment procedural decisions in partnership with the memory clinic sites. Both clinics serve populations that are racially and ethnically diverse. The Alabama memory clinic also serves a large population of patients and care partners who live in rural communities. The research team will continuously monitor the diversity of the sample population to determine if the recruitment strategy needs to be altered to increase the diversity of participants. For example, the research team may decide to divert all recruitment attention to caregivers and patients from underrepresented groups or to determine if new messaging strategies in recruitment would increase the sample of underrepresented groups.

Eligibility Criteria for Participants.

Figure 1 provides the eligibility and exclusion criteria for all participants (PLWD, care partner, health care provider). Both PLWD and care partners endorsing suicide ideation, (assessed by the Cornell Scale for Depression in Dementia and the Patient Health Questionnaire (PHQ-9) respectively), will not be eligible for study participation during any point of the study and will be excluded and will be referred to mental health or emergency services, as appropriate. Any participant endorsing suicidal ideation will receive a follow up call and evaluation from a licensed clinician. PLWD and care partners already receive ongoing services at the memory clinics. Therefore, we will be able to maintain communication and engagement with them throughout the course of the study.

Figure 1.

Design of the Planned Randomized Control Trial

Ethical Considerations

Research team members at the Memory Clinics have been previously trained (and will be supervised) to obtain informed consent from PLWD and each center has existing protocols for recruiting. After reviewing the informed consent document, potential participants (both PLWD and care partner) will be shown and provided with an explanation regarding the intervention with pictures from the health technology. Providing the PLWD with visual information has been shown to be effective in increasing understanding (Chang & Bourgeois, 2020). The research team member will ask the PLWD about his or her understanding of the study, these questions are based on the work of Appelbaum (2007) and Grisso and Appelbaum (1998). Specifically, we will ask 4 questions:

1. Can the subject describe their problem and the reason for being here?

2. Can the subject describe the purpose of the study and basic details?

3. Can the subject reason about their choice to participate?

4. Can the subject clearly and consistently express their choice?

If the answer to any of the questions is “no” the consent will need to be reviewed and signed by the legal authorized representative (most likely, the PLWD care partner) plus the PLWD’s assent will be obtained. In Florida several individuals are recognized to have the right to consent to the enrollment of an individual in IRB-approved medical research that do not have the capacity to consent for themselves. These individuals include a designated surrogate, court-appointed guardian, a person holding durable power of attorney, or a proxy. In most cases, the care partner is the legally authorized representative. The PLWD (or legally authorized representative) and care partner will be asked if they have any questions, given the opportunity to review the informed consent document before signing and dating. They will also be given a copy of the signed consent document. Again, it is unknown if these procedures will be conducted in-person or remotely.

Alabama does not have a state law regarding the consent to participate in research by persons with PLWD. However, UAB’s policy follows federal regulations for consenting individuals with diminished cognitive capacity, where consent may be obtained from the individual’s legally authorized representative. Per UAB IRB policy, those who may serve as a legally authorized representative for an adult incapable of consenting for him/herself for research include: 1) a legally appointed guardian; 2) a health care proxy or an individual authorized to make medical decisions in conjunction with a durable power of attorney; 3) a spouse; 4) an adult child; 5) a parent; 6) next of kin (Source: UAB Policy on Definition of “Legally Authorized Representative” for Decisionally impaired Adults – IRB POL025).

Randomization

Allocation sequencing will be done with computer-generated random numbers and will be implemented with opaque sealed envelopes. Outcome assessors will not be part of the intervention implementation. The study participants (care partners, PLWD) will be asked not to reveal their group assignments. The data core coordinator will generate the allocation sequence. The research coordinator at each site following enrollment and baseline assessments will contact the data core to obtain the assignment.

Outcome Assessments

The evaluation plan has been informed by prior studies on PLWD, caregiving, and HIT (Yen & Bakken, 2012). It will include commonly-used valid, measures (some with modification), qualitative data, and newly-created quantitative data assessments. Measures for variables of interest have been selected based on a number of qualities, including their relevance, validity, and/or the availability in Spanish language. Quality of life (QOL) is the primary outcome to be assessed for both care partners and PLWD in the study. QOL is commonly evaluated as the primary outcome in intervention studies with ADRD caregivers and PLWD (Belle et al., 2006; Cooper et al., 2012; Gitlin et al., 2010; Hoe et al., 2007; Hoe et al., 2005). QOL is often associated with aphasia, which is another reason we have selected it as a primary outcome (Hilari & Byng, 2009; Ross & Wertz, 2003). Secondary outcomes (related to QOL) include: (a) depression, burden, functional health, care self-efficacy, and relationship with PLWD for care partners; and (b) depression, functional communication, and behavioral disturbances for PLWD. We also will collect data on demographic and functional variables to control in analysis. Primary and secondary outcomes are below, with details regarding the measure and a citation. Outcome Assessors will receive ongoing training and supervision to ensure reliability. Table 1 provides a list of the outcome variables for the trial and how they will be measured.

Table 1.

Outcome variables and measures.

Outcome	Participant Group Assessed	Measurement
Primary Outcomes
Quality of Lifea	PLWD	13 Item Quality of Life In Alzheimer’s disease (QOL-AD)- Care Recipient and Caregiver (Logsdon et al., 1999)
Quality of Lifea	Care partner	13 Item Perceived Change Index- Caregiver QOL (Gitlin et al., 2006)
Secondary Outcomes
Depressiona	Care partner	Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001)
Depressionb	PLWD	The Cornell Scale for Depression in Dementia (CSDD; (Bayles & Tomoeda, 1994). Both caregiver and PLWD interviewed. Geriatric Depression Scale (GDS-15; Sheikh & Yesavage, 1986)
Positive Aspects of Caregivinga	Care partner	Positive Aspects of Caregiving Scale- Caregiver (Tarlow et al., 2004)
Caregiver Burdena	Care partner	Zarit Caregiver Burden Inventory-22 items (ZBI-22; Bédard et al., 2001; (Zarit, Reever & Bach-Peterson,1980).
Health Statusb	Care partner	12 Item Short-Form Health Survey (SF-12; Ware, Kosinski, & Keller, 1996)
Cognitive and Behavioral Disturbancesb	PLWD	Revised Memory and Behavior Problems Checklist (Roth et al., 2003; Teri et al., 1992) Caregiver interviewed.
Relationship Qualitya	Care partner, PLWD	Partner-Patient Questionnaire for Shared Activities (Reilly et al., 2006)
Functional Communicationb	PLWD	Functional Linguistic Communication Inventory (FLCI; Bayles & Tomoeda, 1994)
Usage of AAC	Care partner	Usage data collected by the AAC Plus
Satisfaction and Usefulnessb	Care partner Provider	Qualitative Data Newly-created survey items

^a.Variable will be measured at baseline, 6-months, and 12-months

^b.Variable will be measured at baseline and at 12-months

^c.Variable will be measured at 6-months and 12-months

Data will be collected at baseline (prior to randomization), at 6-months and 1 year from all participants by trained interviewers. Table 1 displays the primary measures to be used and the time point to assess outcomes. Due to the pandemic, it is unknown if these data will be collected in person or virtually. Participants will be sent text messages to remind them to use the AAC Plus and of upcoming follow-up data collection time points. If a dyad (care partner-PLWD) are no longer eligible or wishes to stop study participation, the reason for discontinuation will be recorded.

Interview data will be collected by interviewers electronically using a Qualtrics (Qualtrics.com) link. The interviewer will input data provided by the participant during the interview. This data will be automatically secured into a database using a link accessed through a laptop or mobile device. The data will be then stored temporarily using a password-protected Qualtrics file. Lastly, the data will be transferred into an analysis software program that is password secured and encrypted.

Analysis

Power and Sample Size

We have planned for a sample size that is well-powered and allows for a diverse sample. We also have considered the consistent attrition rate of 36% that has been found of similarly-designed studies with PLWD/caregiver dyads (Czaja et al., 2018; Gitlin et al., 2018). The sample size was calculated on the basis of the primary hypothesis that the care partners and persons with ADRD in the study who use the app (i.e., “full intervention”) will experience a greater increase in positive outcomes (e.g., improved quality of life) and compared to those using Care-IT and a paper tool. With a target recruitment of 120 dyads, we estimate that the final sample at the 12-month follow-up to be n=80, which will still offer enough power to detect a moderate effect size (> 0.9 power to detect a 0.45 effect size). The target effect size was determined based on the effect size reported in previous research that was similar in design and assessed the same psychosocial outcome variables with similar dyads that will be recruited for this study (Gitlin et al., 2018). A feasible sample size was also considered when determining effect size, based on each center’s experience with clinical research. In the event that recruitment becomes challenging and the sample size reduces the effect size, the research team will revisit recruitment strategies to increase the sample size.

Analysis Plan

The primary and secondary outcome measures will be assessed for change over time, differences between use by group assignments, and clinical and demographic variables. Outcomes will be analyzed at baseline, 6 months, and at 12 months. To test primary hypotheses and examine the difference of outcome variables between groups (full intervention vs. minimal intervention) at baseline will be evaluated by ANOVA/t-test for continuous outcomes and Chi-square tests for categorical outcomes. If the assumptions of tests/models are not met (i.e., normality), the nonparametric counterparts of tests will be performed, such as Kruskal-Wallis or Mann-Whitney tests. Relationships between outcomes and potential confounders (e.g., age, gender, race, ethnicity, clinical characteristics) will be examined through correlation (continuous independent variables) or point-bi-serial correlation (categorical independent variables).

To evaluate the effects of the intervention (full intervention vs. minimal intervention) on outcomes, generalized linear mixed models (GLMM) will be performed with proper specification of link function (e.g., identity link for continuous outcomes under normal distribution) and the variance-covariance structure of residuals. To account for within-subject correlation caused by repeated measurement, random effects, such as participants and slope effects will be included in the models and the variance-covariance structure of random effects will be determined by Akaike Information Criterion (AIC). Potential confounding variables will be included in the models along with the treatment as the independent variable of interest. Similarly, to evaluate the dose-response effects of the full intervention on outcomes, we will perform GLMMs within the full intervention group with the dose (categorized as low, medium, and high users) as independent variables of interest. A P-value less than .05 will be considered statistically significant. We will employ an intent-to-treat strategy, and GLMM will handle missing data by using all available information. To evaluate the impact of missing data on analysis, we will examine the patterns of missing data and perform models on datasets after multiple imputations to determine whether conclusions hold. If appropriate, alternative models will be examined to evaluate whether moderation effects exist among independent variables.

Data Safety and Monitoring

This is a multi-site project with a vulnerable population participating (PLWD), though the intervention being tested poses minimal risk. Therefore, we will form a Data Safety and Monitoring Board (DSMB). The DSMB members’ responsibilities will include: 1) reviewing protocols, informed consent procedures, and safety plans; 2) determining whether the study should be continued, modified, or terminated in light of adverse events; 3) monitoring study progress as reflected in the rate of participant recruitment and adherence to timetables; 4) reviewing baseline demographic and clinical data; 5) evaluating the study’s risk-benefit ratio as new evidence-based treatments become available; 6) consulting with the investigators regarding protocol measures that impose undue burden on study participants and clinicians or that pose possible ethical concerns or conflicts of interest; and 7) submitting of summary reports based on meetings and review of summary information. The DSMB will be provided summaries of adverse events (additionally, all adverse events are reported to the IRB within the timeframe specified by IRB regulations).

Reporting Results

The trial was funded in 2020. The study protocol was registered with ClinicalTrials.gov [trial number NCT04571502]. The ClinicalTrials.gov website will be updated periodically. Subsequent modifications of the protocol, including the formation of the DSMB, will be reported to the NIA, the institutional review board, and ClinicalTrials.gov.

Discussion

The development work for the AAC Plus is underway, and this will proceed and inform the clinical trial. We recognize the tension in developing health technology to support care partners, as technology needs to empower caregivers and make caregiving activities easier, not more burdensome. To accomplish this, we are using a design and development approach that will primarily use an iterative process as part of agile development. The prototyping processes are iterative in that the process involves giving participants mockups (Paper Prototyping) and interactive interface prototypes (Wireframing and Interactive Prototypes) to end-users (PLWD, care partners). We are using a user-centered focus in our design and development process in which users are involved as “partners” throughout the development lifecycle.

The goal of this project is to develop and test a technology that facilitates triadic communication among PLWD, their care partners, and their providers to better address care based goals and improve outcomes for PLWD and caregivers. Following the success of this project, we intend to explore implementation of Care IT and the AAC Plus through larger, randomized control trials and in other care-settings, such as skilled nursing facilities and assisted living facilities.