In total, 20 PWMCs and their CGs who lived in the community were recruited in the United States to assess (1) how their technology use shifted in response to the COVID-19 pandemic and (2) the impact that shift had on social isolation. Each PWMC-CG dyad was recruited to participate in a semistructured qualitative interview over Zoom videoconferencing. Although Zoom interviews were utilized to protect the health of older adult participants who were more likely to develop severe illness if they contracted COVID-19, this may have introduced selection bias by only including more technologically literate participants who could use Zoom [3]. Selection bias is discussed further in the Discussion section. Participants were recruited through the University of Minnesota Caregiver Registry, a list of family members of PWMCs and health professionals who have agreed to learn about research participation opportunities available on behalf of the senior author’s project team. We emailed all individuals in the registry, inviting them to participate in the study. Participants were also recruited through email advertisements in professional networks and at memory clubs and adult day programs for PWMCs.

To be eligible to participate, the PWMCs had to speak English, have no history of a serious mental illness (ie, any major psychiatric disorder), and have a diagnosis of AD/ADRD or mild-to-moderate cognitive impairment by a physician. To be eligible, CGs had to speak English, be 21 years of age and over, and self-identify as someone who assists the PWMCs because of their memory loss. Because the interviews were conducted via Zoom, 3 dyads were considered ineligible since neither member of the dyad had access to a working web camera and microphone. One dyad was ineligible because the PWMC had never received a formal diagnosis of AD/ADRD or cognitive impairment. Figure 1 describes study participant flow. In total, 40 participants were enrolled and participated in the dyadic interviews conducted by authors EA and AM. The study was approved by the University of Minnesota Institutional Review Board (STUDY00006318).

Qualitative interviews were conducted from June 25 to August 6, 2020. Before an interview began, the CGs and PWMCs had to provide verbal consent or assent. PWMCs’ capacity to consent was evaluated by administering the Mini-Cog and the University of California, San Diego (UCSD) Brief Assessment of Capacity to Consent (UBACC) [19,20]. The PWMCs provided verbal consent if they had a Mini-Cog score of 3 or higher and a UBACC score of 14.5 or higher. PWMCs provided verbal assent if they scored 2 or lower on the Mini-Cog or less than 14.5 on the UBACC. Once consent or assent was obtained, we administered a brief survey to CGs and PWMCs to determine age, sex, race, ethnicity, education, income, employment status, living arrangement, relationship to each other, and disease progression of the PWMCs.

Qualitative interviews adhered to a semistructured protocol focusing on 4 major topics: (1) technology use pre-COVID-19, (2) technology use during the COVID-19 pandemic, (3) adoption of specific technologies during the pandemic, and (4) facilitators and barriers to technology adoption. See Multimedia Appendix 1 for the semistructured interview questions. Interviews were audio-recorded, and direct observation notes were completed within 24 hours of each interview to document impressions of the participants’ location, level of comfort with Zoom technology, any nonverbal behaviors of importance, and how the PWMCs and CGs interacted.

Audio recordings were professionally transcribed and organized in NVivo 12 (QSR International). Interview transcripts and direct observation notes were thematically analyzed using Braun and Clarke’s 6 steps of thematic analysis: (1) familiarization, (2) generation of initial codes, (3) search for themes, (4) review themes, (5) define and name themes, and (6) write-up of themes analyzed [21]. An iterative process was followed to continually identify themes, linkages, and explanations, which were compared to create a codebook. Researchers then identified textual elements that emerged repeatedly (ie, codes); these codes were clustered into larger categories that were used to construct major thematic elements from the text. All authors reviewed the codebook to refine and clarify codes and themes. Next, EA and AM independently coded a subset of the interviews and revised the codebook after comparing codes and discussion with the full authorship team. EA coded the interview transcripts, and author JM reviewed the coded material and revised it, as necessary, to ensure accuracy and replicability. The iterative process of developing codes and themes ensured that saturation was achieved and the data were characterized appropriately. Peer debriefing, negative case analysis, and clear audit trails enhanced transparency and rigor in the analysis [22].

In total, 20 dyads participated in the study (Table 1). Most PWMCs were male (14/20, 70%), and most CGs were female (14/20, 70%). In addition, 16 (80%) of the 20 dyads were spouses/partners, while the other dyads were siblings or parents/adult children. Furthermore, 18 (90%) of the 20 PWMCs lived with their CGs. Half (n=10) of the PWMCs were diagnosed with AD or early-onset AD, 6 (30%) were diagnosed with mild cognitive impairment, and the remaining 4 (20%) were diagnosed with other types of memory loss.

Overall, 6 themes were identified that highlight why and how CGs and PWMCs altered their technology use during the COVID-19 pandemic. We present them in 3 groups: (1) facilitators of technology use, (2) barriers to technology use, and (3) overcoming challenges. Technological devices commonly used by participants were smartphones, smartwatches, computers, tablets, smart TVs, and assistive technology designed for PWMCs, such as pill dispensers.

The COVID-19 pandemic forced a shift in technology adoption for those who were fortunate enough to have access to it. Yet, CGs and PWMCs were not always willing to try new technology to assist them due to barriers such as each person’s level of technological literacy, the dependence on the CG for use, and limitations of the technology. For CGs and PWMCs who adopted new technology or adapted existing technology, technology was perceived as more useful during the pandemic than before, specifically in relieving boredom, maintaining social connections, and increasing CG respite.

Our research affirms that social network support, perceived value, perceived impact on the quality of life, and confidence in the ability to learn a new technology are all important influences of technology adoption [14,15]. Each of these influences played a role in technology adoption among PWMCs and CGs during the pandemic. Our research also highlights the role of technology in reducing the caregiving burden during the pandemic, along with increasing PWMC independence. GPS technology seemed to minimize CG stress by reducing the likelihood of PWMCs getting lost [23]. Similar to findings by Øderud et al [24], our research also shows that GPS technology provides CGs with respite time, while allowing PWMCs to enjoy their freedom and outdoor activities safely.

Studies suggest that assistive technology can reduce the caregiving burden yet may also pose an additional burden when technology adoption and use require too much CG help [25]. Such results are consistent with our findings that CGs are unlikely to adopt burdensome technology. In a review of 56 studies, assistive technology was perceived as removing CG stress and burden overall, although no significant change in the caregiving burden was reported in any of the 16 (29%) quantitative studies included in the review [23]. This discrepancy in CG burden results may be due to insensitivity of existing quantitative outcome measures [26]. Perhaps alternatively, no single device or app is comprehensive enough to reduce the caregiving burden, and instead, a combination of technologies is required to significantly reduce burden.

Technology can foster social connections for CGs and PWMCs by counteracting the impact of diminished social support and interactions during the pandemic. A systematic review of 25 publications concluded that information and communication technologies (ICT) can be an effective way of reducing social isolation among older adults; however, it is not suitable for all older adults [12]. Prior to the pandemic, technological interventions to reduce social isolation were understudied [27]. Not only are more ICT interventions being developed to reduce loneliness and increase social participation during the pandemic, but older adults also have a more positive view of ICT interventions now in contrast to before the pandemic [28]. Preliminary findings are showing promising results; in a qualitative study, researchers demonstrated that virtual memory cafés, where PWMCs and CGs remotely interact with other PWMCs and CGs, were able to support the social connectedness of PWMCs and CGs during the COVID-19 pandemic [29].

The findings emphasize the importance of incorporating technological barriers of PWMCs and their CGs into the design process. Although dependence on the CG to use technology cannot be eliminated entirely, it should be avoided or minimized through mindful design. Therefore, it is important to engage both the PWMC and the family CG during each stage of the design process. Additionally, it is likely not feasible for technology developers to improve a users’ technological literacy. However, testing new products by CGs and PWMCs at various stages of dementia progression could highlight difficulties for less technologically literate users. This codesign process is important for any technology used by older adults yet is most important for technology designed specifically for PWMCs.

Limitations of this research include factors that may make the study population different from the general population. Participants were recruited through email, and they had to be willing and able to participate in interviews over Zoom, which could have led to selection bias. Participants may be more likely to use Zoom and other technologies than individuals who would have only been recruited in a nontechnological manner or did not have the capability to participate in a Zoom interview. Since income, education, and race/ethnicity are major influences of technology adoption, it is important to note that this sample was highly educated, had a median income higher than the national median income, and primarily identified as non-Hispanic White [30]. Additionally, we did not include PWMCs living in long-term care, and no PWMCs were in the later stages of dementia. Due to these differences, participants in this study may have different patterns of technology use/disuse compared to other CGs and PWMCs, all of which likely limit the generalizability of the findings. The results are hypothesis generating, and future research should engage a more racially and ethnically diverse population of CGs and PWMCs and include participants with lower incomes, educational attainment, and technological literacy.

This research contributes to the literature on this population’s technology usage. Much of the existing literature focuses solely on assistive technology, while our research points out that mainstream technologies, such as smartphones or Zoom, were predominately used by this sample. Further research is needed to examine how mainstream technologies are used to support PWMCs and CGs in their everyday lives and to compare whether those who adopted more technology during the pandemic coped better with isolation than those who did not alter their technology usage.

Our research found that the COVID-19 pandemic resulted in an increase in technology use among many participants. These findings emphasize the importance of technology use among CGs and PWMCs, particularly during isolation, to provide relief from caregiving burden and afford PWMCs more independence.