Routine cognitive screenings for older adults during primary care visits can facilitate timely referrals to neuropsychologists, help establish a cognitive baseline while older adults are still healthy, monitor cognitive trajectories, and ultimately improve early detection of cognitive dementia. Unfortunately, time constraints present major barriers to cognitive screening in primary care, and less than half of cognitive impairment cases are identified in these settings. A remote cognitive screening smartphone app has the potential to save in-person visit time and replace inadequate or nonexistent screening practices in primary care, particularly for the majority of older adults who are cognitively healthy. To achieve this goal, MyCog Mobile was developed as a cognitive screening app which patients could complete remotely on their own smartphones prior to a primary care visit. The goal of this presentation is to introduce the MyCog Mobile system and describe results from a pilot study of its feasibility, acceptability, and initial psychometric properties. 

A sample of adults 65 years and older self-administered MyCog Mobile remotely on their own smartphones twice 13 to 18 days apart. After their first administration, participants also completed the Simplified System Usability Scale (S-SUS) and provided feedback on their experience. We used the Spearman-Brown split-half formula (rsb) to calculate internal consistency for each of the measures except for MyFaces, for which we found the empirical reliability of the theta scores (ρxx). We examined correlations with education level and test-retest reliability through intraclass correlations (ICCs). 

All participants in the sample (N=51; Mean age = 74.20 years, SD = 6.25) successfully completed each measure and there was no evidence of floor or ceiling effects. Regarding educational attainment, 33.33% completed high school, 19.61% completed some college, 23.53% completed a bachelor’s degree, and 23.53% completed an advanced degree. Internal consistency was adequate or better and test-retest reliability was moderate for all measures, respectively (Table 1). Each test demonstrated moderate correlations with educational attainment except MyPictures. Participants’ ratings on the S-SUS indicated good usability of MyCog Mobile overall (M=73.78, SD = 20.11).  

Table 1. Reliability of the MyCog Mobile Measures 

Note: All correlations significant at p<.001 unless otherwise noted. NS = non-significant (p>.05). 

These preliminary findings suggest that MyCog Mobile can be reliably self-administered by older adults on their own smartphones in a remote setting. Test-retest reliability was moderate across measures, suggesting MyCog Mobile may not be appropriate for frequent re-administration. Three measures demonstrated moderate correlations with education as expected, and education will be considered in future scoring models. Usability ratings indicated a satisfying user experience overall, and free response feedback resulted in actionable design recommendations. We will discuss plans for MyCog Mobile’s construct validation in a clinical sample as well as changes to the app to improve the experience for older adults.