This study used data collected in the “Active aging – resilience and external support as modifiers of the disablement outcome (AGNES)” cohort study conducted at the University of Jyväskylä, Finland (Rantanen et al., 2018). The data were collected in a structured home interview and with a postal questionnaire between October 2017 and December 2018, and with a postal questionnaire in April–June 2020. The 2020 postal questionnaire was an addition to the baseline, sent to the baseline participants during the state of emergency due to COVID-19, with the aim to investigate the consequences of the pandemic and social distancing recommendations among older adults. The AGNES cohort study baseline data (Rantanen et al., 2018) and the 2020 postal questionnaire (Rantanen et al., 2021) have been reported in detail elsewhere. In brief, the AGNES cohort study participants comprise a population-based sample of men and women aged 75, 80, and 85 at baseline, residing independently in the city of Jyväskylä in Central Finland. Those living in the study area, able to communicate, and willing to participate were invited to take part in the AGNES study. The initial baseline sample comprised 1021 persons. The 2020 postal questionnaire was sent to the 985 baseline participants who had not withdrawn their consent, of whom 809 responded.

The Ethical Committee of the Central Finland Health Care District approved the AGNES study on 23 August 2017 and the 2020 questionnaire on 13 May 2020. The Declaration of Helsinki has been followed throughout the AGNES study. All participants signed an informed consent form before participating in each phase of the study.

Digital device use was measured with the postal questionnaire in 2020 (Eronen et al., 2021; Rantanen et al., 2021). The participants were asked to report whether they had difficulty in using digital devices, such as computers, tablets, or smart phones, with response options (1) no difficulty, (2) with minor difficulty, (3) with major difficulty, (4) unable without help, and (5) I don’t use any digital devices. For the analyses, these were grouped as fluent users (1), learners (2–4), and non-users (5).

Predisposing, need and enabling variables were assessed at the study baseline with home interviews and postal questionnaires. Predisposing variables included age group, gender, being partnered (married/cohabiting vs. divorced, widow, unmarried) education, occupation, and self-perceived financial status. Age and gender were drawn from the Finnish National Population Register. Information on marital status, as well as education, occupation, and self-perceived financial status were obtained via the study questionnaire. Education level was assessed by asking participants to report their highest education attainment, which was categorized as low (primary school or less), medium (middle school, folk high school, vocational school, or secondary school), or high (high school diploma or university degree). The participants were asked to report their longest-held occupation during working age. Occupations were categorized as manual, lower non-manual, and upper non-manual. Participants were asked to rate their perceived financial status as very good, good, satisfactory, or poor. For the analyses, the categories satisfactory and poor were combined.

The need factors included poor self-rated health, presence of depressive symptoms, loneliness, and number of chronic conditions. Self-rated health was assessed with a question on current general health with the response options very good, good, fair, poor, very poor. For the analyses, the categories very good/good and poor/very poor were combined. Depressive symptoms were assessed with the Center for Epidemiological Studies Depression Scale (CES-D, range 0–60) (Radloff 1977). The CES-D score was dichotomized into absence (0–16) and presence (> 16) of depressive symptoms (Sawyer Radloff and Teri 1986). Loneliness was asked with a question “How often do you feel lonely,” with response options very rarely or never, rarely, often, and almost always. For the analyses this was dichotomized as never/very rarely and at least sometimes (rarely, often, and almost always). Physician-diagnosed chronic conditions were self-reported during the home interview. Participants were asked to report whether they had any of the conditions listed under ten categories of chronic conditions and if so, to specify these. The ten categories were respiratory conditions, cardiac conditions, vascular conditions, cerebrovascular condition or brain injury, musculoskeletal condition, visual or auditory impairment, diabetes mellitus, malignant cancer, neurological conditions, and depression. In addition, participants were asked an open-ended question about any other physician-diagnosed chronic conditions they had, and their responses were later categorized by a nurse. For the present analyses, the number of chronic conditions each participant had was calculated by summing all the illnesses and conditions reported by the participants (range 0–12).

The enabling variables used in this study were social support and health literacy. Social support was assessed with three questions on the frequency of meeting children or other relatives, friends, and other acquaintances. The response options for these were (1) daily, (2) weekly, (3) monthly, (4) a few times a year, (5) only seldom, and (6) I do not have children/friends/acquaintances. Persons who met their children, friends, or acquaintances at least weekly were categorized as having social support. Health literacy was measured with the Finnish translation of the short form of the European Health Literacy Survey, HLS-EU-Q16 (Sørensen et al. 2013). Following the guidelines provided by the developers of the HLS-EU-Q, we calculated a health literacy score only for participants who had answered at least 80% of the items, i.e., given answers to at least 13 items in the questionnaire (Sørensen et al. 2013). The health literacy score was computed by recoding the response options very easy and easy as 1 and difficult and very difficult as 0. Summing the responses yielded a health literacy score ranging from 0 to 16, which in accordance with the guidelines issued by the developers of the instrument, was categorized as inadequate (0–8), problematic (9–12), and sufficient (13–16) health literacy, as recommended by the developers of the instrument (Sørensen et al. 2013).

The analyses were adjusted for baseline cognitive capacity, as assessed with the Mini-Mental State Examination (MMSE, range 0–30) (Folstein et al. 1975).

Participant characteristics by category of digital device use are presented as percentages and means with standard deviations (SD). Associations of factors with digital device use were analyzed with multivariate logistic regression analyses. We report odds ratios (OR) with 95% confidence intervals (CI). We fitted four models: model 1 included age group, gender, education, occupation, and self-perceived financial status (predisposing variables); model 2 included poor self-rated health, loneliness, presence of depressive symptoms, and number of chronic conditions (need factors); and model 3 included social support, marital status, and health literacy (enabling variables). Model 4 included all statistically significant variables from models 1, 2, and 3 along with cognitive capacity. All analyses were performed with IBM SPSS version 28 and statistical significance was set at 0.05.