111
Validation of BrainCheck’s Digit Symbol Substitution Task (DSST) in a mixed clinical sample
Sruthi Gade, The University of Texas at Austin, Austin, United States
Dariella Fernandez, The University of Texas at Austin, Austin, United States
Illyana Runnels, Dell Medical School at UT Austin, Austin, United States
Bertha Ortiz, Dell Medical School at UT Austin, Austin, United States
Bonnie Scott, Dell Medical School at UT Austin, Austin, United States
Jared Benge, Dell Medical School at UT Austin, Austin, United States
BrainCheck is a cognitive health platform that includes several well-known cognitive tests that have been digitized for online use, including a digit symbol substitution task (DSST). Initial validity studies of the DSST are promising, but further validation in clinical samples is needed before widespread use for clinical decision making is appropriate.
Participants were 66 adults (50% women) referred for neuropsychological evaluation of cognitive concerns. Average age of participants was 67.7 years (SD = 11.4), and average education level was 16.6 years (SD = 2.3). Racial/ethnic composition of the sample was 84% non-Hispanic White, 12% Hispanic White, 2% Native Hawaiian or Pacific Islander, and 2% American Indian or Alaskan Native. All participants completed a battery of traditional paper and pencil neuropsychological tests after which participants were invited to take part in this research study. Participants provided written informed consent and then completed the BrainCheck DSST on an iPad. The DSST instructs examinees to tap on numbers that are paired with unique symbols as quickly as possible in 60 seconds. Forty-seven percent of participants were diagnosed as cognitively normal (CN), 38% with mild neurocognitive disorder/mild cognitive impairment, and 15% with major neurocognitive disorder/dementia. Due to the small sample size of individuals with mild dementia, all participants diagnosed with a neurocognitive disorder were collapsed into a single group, cognitively impaired (CI), for analysis.
The CI group was significantly older (70.5 vs. 64.7 years; t(64) = -2.1, p = 0.04) and was more likely to be male (66% vs. 34%; X2(1) = 7.4, p = 0.007) than the CN group. After accounting for age and gender, the DSST raw score of the CI group was significantly lower than the CN group (11.3 ± 5.8 vs. 18.5 ± 5.2; F(1) = 18.6, p < 0.001). DSST raw scores correlated significantly (p £ 0.001) in the expected direction with raw scores of the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) Coding (r = 0.71), Trail Making Test Parts A and B (r = 0.67 and 0.70, respectively) and Stroop Color and Word subtests (Word: r = 0.44; Color: r = 0.57; Color-Word: r = 0.53). DSST raw scores did not correlate significantly with the Geriatric Depression Scale – Short Form, Multilingual Naming Test, or RBANS Picture Naming, Figure Copy, or Line Orientation subtests.
BrainCheck’s DSST exhibited good construct validity, including convergent and divergent validity, in a mixed clinical sample of patients referred for neuropsychological evaluation. Advantages of the DSST over similar measures are a slightly shorter administration time (60 seconds) and ability to administer remotely.
Keyword 1: computerized neuropsychological testing
Keyword 2: psychometrics
Keyword 3: technology