Exploring Family Caregivers’ Likelihood of Adopting a Novel App That Connects Care Teams of Persons Living With Dementia: A Mixed-Methods Study

Participants were eligible if they (a) were currently caring for or had previously cared for a PLWD, (b) were over the age of 18, and (c) owned a smartphone. Individuals who did not speak English were excluded because the prototype was only available in English. In addition, the perspective of healthcare providers was not included as they are addressed in a previous manuscript (Zheng et al., 2024). This previous study interviewed 22 adult day staff members (e.g., nurses, physicians) to assess the feasibility and acceptability of CareMOBI as the effectiveness of CareMOBI for care coordination relies on their participation alongside family caregivers (Zheng et al., 2024). Eligible participants were purposively recruited with the help of administrators at participating adult day centers in New York and California who were in contact with family members of PLWD. A research assistant contacted them by email and/or phone, described the study, confirmed participants’ eligibility, and obtained informed consent. All enrollees received a $50 gift card for their participation. The New York University Committee on Activities Involving Human Subjects (UCAIHS) provided Institutional Review Board approval for this study.

Data collection consisted of one-on-one semi-structured interviews and completion of the Technology Acceptance Model Questionnaire (Gagnon et al., 2012). One week before the interview, participants received a confirmation email with a link to an interactive prototype of CareMOBI and a 2-min informational video that could be accessed from a smartphone, tablet, or computer. Participants were asked to spend about 10 min navigating through the prototype to complete tasks including logging in, using filters to locate information, adding a new medication, and recording the typical day-to-day activities and health-related progress notes of a PLWD.

Online interviews were scheduled based on participants’ availability and conducted virtually via a secure online platform. UX (user experience)/UI (user interface) design professionals in user testing at the product development firm that built the CareMOBI prototype provided input to develop a semi-structured interview guide (see Supplemental Material). Participant interviews were approximately 30 min in length. Interviews were conducted by either the principal investigator (TS) who has received formal training in mixed-methods research and published extensively using qualitative methods or a junior research scientist (JB) with a Master’s in Public Health. Member checking took place informally during data collection. The interviewer would pause after major questions to summarize and confirm their interpretation of a participant’s responses. Open-ended questions allowed participants to elaborate on their reaction to the CareMOBI app and allowed the researchers to elicit information on factors influencing their perceptions of the app, aspects regarding its usability, potential barriers, and facilitators of eventual adoption. All interviews were recorded, professionally transcribed, and reviewed for accuracy. Field notes by the interviewer supplemented recorded interviews. Given the possibility of response bias, both interviewers maintained an audit trail/reflexive journal to document observations and reflections throughout the data-gathering process. This included decisions about follow-up questions asked and reactions to responses. Interviewers met regularly to review transcripts and discussed whether they may have influenced participant responses which supported an honest and reflexive process.

Qualitative data were analyzed using directed content analysis to provide a systematic coding process, and an a priori codebook was utilized as a coding scheme for all transcripts (Hsieh & Shannon, 2005). Codes were summarized within cases and then categorized within the four domains of the Technology Acceptance Model: likelihood of adoption, perceived ease of use, perceived value in clinical care, and fit within the existing workflow for initial coding categories (Gagnon et al., 2012). The research team (MFC, AZ, KBS, TS) discussed any text that could not be categorized within the codebook to determine if a new category or code needed to be defined or aligned with an existing category or code. Two coders who were PhD (MFC) and undergraduate nursing students (KBS) independently coded in Dedoose, a web-based platform for qualitative and mixed-method coding. The principal investigator (TS) analyzed a subset (20%) of transcripts to ensure reliability and rigor. The research team regularly debriefed to discuss and validate the analysis results, select illustrative quotes from a range of participants, update the codebook to reflect this iterative process, and saturation occurred when no new categories were identified (Hsieh & Shannon, 2005).

Upon finishing the interview, participants provided demographic information and completed an online adapted version of the Technology Acceptance Model Questionnaire, which was previously validated for use in healthcare settings (Gagnon et al., 2012). Responses to the 33 survey items enabled further examination of factors that could influence the eventual adoption of CareMOBI, and the anonymous nature reduced the potential for social desirability bias. Participants rated each item on a seven-point Likert scale ranging from “totally disagree” to “totally agree.” Each domain-specific question was averaged to determine scores and higher scores corresponded to higher perceived feasibility and acceptability. Descriptive statistics were used to characterize the sample, measures of central tendency and spread for continuous measures, and frequencies and percentages for dichotomous or categorical variables were performed in Qualtrics.

Qualitative and quantitative data were integrated in the third and final phase of analysis. We sought to align with the Technology Acceptance Model: perceived ease of use, perceived value in clinical care, fit within existing workflows, and end-users’ overall likelihood of adoption. Using the four themes as anchors, we developed an informational matrix (see Tables 2–5) in which qualitative data were embedded and compared to quantitative data. Using triangulation methods, we sought to understand the overall likelihood of end-users’ adoption of the app (quantitatively) and factors underpinning this across cases within each stakeholder group (qualitatively).

The total sample (n = 13) of caregivers of PLWD was majority White (53.85%) and between the ages of 50 to 59 years (30.77%) and 60 to 69 years (30.77%) (see Table 1). More than half of the respondents identified as female (61.54%), and the majority (53.85%) identified as children of PLWD for whom they were caring.

Perceived ease of use refers to the level of difficulty related to navigating and using CareMOBI. There were six questions within the Technology Acceptance Model Questionnaire presented to the caregivers that assessed perceived ease of use: overall ease, flexibility, technological skills, and general user comfortability with the mHealth app (see Table 2). The mean score of this theme was 6.23, indicating a high perceived ease of use for CareMOBI among caregivers. Within the quantitative survey, the question with the highest mean score was “I think I will find it easy to acquire the necessary skills to use this app” (M = 6.62 or strongly agree). “I think that the proposed app is a flexible technology to interact with” and “I feel comfortable with information and communication technologies” (M = 5.92 or agree) were questions with the lowest mean score.

The qualitative interviews revealed specific aspects of CareMOBI that made it easy for caregivers to manage and care for a PLWD. Aligned with the quantitative data, many of the participants reported that the design and setup of the app were user-friendly, which they found helpful in managing the care and healthcare information of PLWD. Caregivers also reported comfortability with technology, reflecting the mean score of 5.92 on the quantitative data, as they “[grew] up with computers and apps and phones . . . ” (CG-LW). Despite the positive reviews of CareMOBI’s design and ease of use, the qualitative interviews further revealed potential challenges such as no additional language options that could affect diverse caregivers’ ability to navigate and efficiently use the app to complete their caregiving duties. For example, one participant stated having a “Spanish version” of the app is warranted (CG-ESD).

The perceived value for clinical care describes the degree to which CareMOBI enhances or improves the ability of caregivers to care for and manage the health and records of PLWD. The Technology Acceptance Model Questionnaire had seven questions that pertained to this quality of CareMOBI (see Table 3). The overall mean of this theme was 6.20, indicating that most caregivers agree that CareMOBI has significant value for clinical care. The question with the highest mean score was “In my opinion, the use of the proposed app will have a positive impact” (M = 6.45 or strongly agree). The question with the lowest mean score was “I find it interesting to use the proposed mHealth app for the monitoring and management of my patients” (M = 5.85 or agree).

The qualitative data from caregiver interviews highlighted how the features of the app benefited the caregivers’ roles in caring for PLWD. Caregivers expressed that the app’s features helped enhance communication, organization, and coordination of care for PLWD. One caregiver viewed CareMOBI’s ability to keep them more organized and “on top of things” as invaluable (CG-FT). This aligns with the quantitative data where caregivers agreed that CareMOBI will make it easier to perform tasks needed to manage the care of PLWD. CareMOBI’s ability to provide healthcare information at the point of service and inform others how to care for PLWD in emergencies (e.g., advance directives) was found to be vital for caregivers. In addition, CareMOBI provides positive value for care through real-time communication of the health progress of and information on PLWD. CareMOBI’s features allow the care team to be “on the same page” (CG-ESD) regarding the health of patients and may be conducive to coordinating dementia care. For instance, one caregiver stated, “if family members had access to this information, or at least be able if you, as a caregiver, can share that with them, it’s a helpful tool so people have a snapshot of what’s happening with your loved one at any given time” (CG-FT).

The development of new mobile apps such as CareMOBI must enhance or continue the established workflow family caregivers maintain for patient care. Nine questions assessed the extent to which CareMOBI fits into the caregiver’s routine around monitoring, managing, and supporting the care of PLWD (see Table 4). The mean domain score (5.86) indicated that CareMOBI fits within most caregiver’s day-to-day workflow. The item with the highest average score was “I think it is a good idea to use the proposed mHealth app to monitor/manage my loved one” (M = 6.54 or strongly agree). The item with the lowest average score was “My loved one will welcome the fact that I use the proposed mHealth app” (M = 4.77 or neither agree nor disagree).

Qualitative data aligns with the quantitative data; however, explicit statements from caregivers describe how CareMOBI may be integrated into the current workflow. Several caregivers noted that they were aware of existing mobile apps, but presently rely on separate notes, cameras, and calendar apps to manage a PLWD care. Participants discussed that CareMOBI was a centralized application that could link other apps to conform and improve the management of caregiving duties. Participants were also unanimous in agreeing that medication management would be streamlined more easily. For example, one participant (CG-TN) stated, “As far as scanning the labels of the medication is fantastic.” Family caregivers noted that CareMOBI may prevent errors in medication administration and reduce the time spent inputting medications by hand due to scanning. This is consistent with the quantitative data that CareMOBI could efficiently manage a PLWD care, especially with medication management. Many participants also conveyed the difficulties of having minimal guidance or support as family caregivers of a PLWD, but found CareMOBI to be a modern intervention that could facilitate care coordination and reduce caregiver burden.

This domain assessed the extent to which family caregivers intend to adopt and include CareMOBI in the care management of PLWD. Nine questions assessed this domain with a mean score of 5.46, suggesting most respondents agreed that they intended to use CareMOBI once released (see Table 5). The highest average item was “I often use smartphone apps in my work or daily life” (M = 6.45 or strongly agree) and the lowest average item was “I have already used a smartphone app to manage my loved ones” (M = 3.46 or disagree).

The qualitative data revealed that caregivers believed CareMOBI would be beneficial and enhance their current workflow, but some concerns precluded the adoption of the app. One participant (CG-TT) stated, “It’s getting all the partners that are involved in caretaking to use the app . . . I thought of that as a potential challenge.” Wariness about family members and healthcare professionals’ involvement and commitment to utilizing CareMOBI existed. Furthermore, caregivers were apprehensive of the targeted population for CareMOBI due to generational differences such as age, which reflects the lowest quantitative score of not having used a mHealth app to manage the care for PLWD. Lastly, some participants emphasized the importance of confidentiality of patient health data and for information in CareMOBI to only be accessed by the care team. To increase the likelihood of adoption, caregivers suggested additional training and support for family caregivers in utilizing CareMOBI for the care management of PLWD.