JMIR Publications

ABSTRACT

Background:

The role of technology in accessing resources/services for managing everyday activities is rapidly advancing. Although research suggests the digital divide between generations has decreased, most older adults remain less engaged with technology than their younger counterparts.

Objective:

This study aims to investigate the prevalence and the potential predictors of computer use in older adults over two time points.

Methods:

Data were obtained from the ASPREE (ASPirin in Reducing Events in the Elderly) Longitudinal Study of Older Persons (ALSOP). Community dwelling Australian participants aged ≥70yrs were recruited via general practices. Participants were relatively healthy at first wave. ALSOP first wave (2012-15) and second wave (2013-18) data were used. Computer use was self-reported on a five-point Likert scale; never, rarely, sometimes, often, always. Computer use prevalence was determined by a 95% confidence calculated using bootstrap resampling. Potential predictor variables included demographics, lifestyle factors and chronic conditions. Top five important predictors of computer use and its change over time were determined using Least Absolute Shrinkage and Selection Operator (LASSO) with ordinal logistic regression. In sensitivity analyses, we repeated models after imputing missing values.

Results:

At first wave, 12,896 individuals participated (90% response rate, 54.2% women, mean age 75.2yr) and 8,411 (65.2%) reported on the relevant variables and were therefore included in the analysis. At first and second waves, over half of participants reported ‘Always’ using a computer (51.7% [95%CI:50.8-52.6], 56.3% [95%CI:55.3-57.2] respectively). Around a quarter reported ‘Never’ using a computer, the second most prevalent category 26.7% [95%CI:25.9-27.5], 23.6% [95%CI:22.8-24.4]). Overall, older adults significantly increased their computer use between the waves of questionnaires (OR:1.14 [95%CI:1.11- 1.16], P-value<0.001). The top five predictors of computer use at first wave were: more years of formal education, income, writing frequency, men, and younger age. In terms of top 5 predictors of increase in computer use, gender and writing frequency were no longer predictive of more computer use, being replaced by the physical component of quality-of-life measure (derived from SF12) and the presence of diabetes. Sensitivity analyses yielded similar results.

Conclusions:

Overall older adults are increasingly using computers; however, a considerable proportion remain not engaged, or increasing their use. This research suggests that certain sociodemographic factors may predict older adult computer use, and some older adults may face additional barriers to use. Future work should investigate these barriers and investigate ways to help older adults, should they wish to overcome them.