Smartphone-based measurement of cognition and physical function in older emergency department patients: A Feasibility Study

BackgroundWearables and smartphones are increasingly deployed in digital assessments of cognitive and physical function in studies of older adults. However, their use in the emergency department (ED) for patients presenting with falls has been limited. In GAPcare II, a randomized-controlled trial of an ED-based fall prevention intervention, we combined standard quantitative measures to screen for cognitive impairment and physical function limitations, with digital measures of cognition and physical function using smartphones and smartwatches. Our objective was to assess the feasibility of deploying digital assessments in the ED for older patients with a recent fall. MethodsBetween August 2021 and January 2024, community-dwelling older ([&ge;] 65 years old) ED patients presenting for a fall were screened for cognitive impairment (Six Item Screener) and physical function limitations (use of mobility equipment, modified Barthel Index). Apple ResearchKit digital assessments were administered using smartphones and smartwatches to assess cognition (Stroop, Trail-Making tests, reaction time) and physical function (gait and balance, timed walk test). Wearable devices were applied to the patients wrist for passive movement measurement. Patients were instructed and supervised by trained research staff. We assessed feasibility by determining how many patients attempted and completed each digital task, along with reasons for non-attempt and non-completion. We also assessed the association between test completion and patient characteristics in univariate and multivariable regression models. ResultsAmong 197 patients, the average age was 78.2 years (standard deviation = 7.6), and 68% were women. Twelve percent had possible cognitive impairment, and 70% had some functional dependence. Eighty-two percent attempted at least one digital task. Leading reasons for non-attempt included concerns surrounding safety and pain or discharge from the ED before the attempt, specifically for the physical tasks. Completion rates among those who attempted were moderately high (68-87%) for cognitive tests and did not vary by age, other demographic variables, or health behaviors (e.g., tobacco, alcohol use), but did vary by possible cognitive impairment (p<0.01, all cognitive tests) and physical function (Barthel index, p<0.01, reaction time test only). Reasons for non-completion of cognitive tests included injury (15%), task was "too hard" (7%), and technology issues (7%). Completion rates for physical function tasks were substantially lower (18-20%) and did not vary by demographic characteristics but did vary with standard measures of physical function (Barthel Index and use of mobility equipment, p<0.01 and p<0.05, respectively). Low completion rates for physical function tests were mainly due to safety concerns, pain, and injury. ConclusionsDigital assessment of cognitive function using publicly available smartphone-based tests is feasible in studies of older adults presenting to the ED for falls. However, patients may be reluctant to engage in physical function tests requiring mobilization immediately after an injury. Future research will investigate whether such data are predictive of clinically relevant outcomes (e.g., time to injury recovery, ED return visits) and can inform ED care (e.g., referrals to physical therapy, skilled nursing facility placement).