At‐home assessment of cognitive performance: Establishing norm scores for the Cognitive Online Self‐Test Amsterdam (COST‐A)

Background Increased public awareness of cognitive decline among older adults increases the need for reliable self‐testing of cognitive performance at home. In this study, we aimed to assess the usability of unsupervised, at‐home cognitive testing using the Cognitive Online Self‐Test Amsterdam (COST‐A). We established norm scores in a reference sample and validated them in an independent sample of cognitively normal individuals. In a clinical sample of individuals diagnosed with mild cognitive impairment (MCI) or dementia, we validated the capacity of COST‐A to distinguish normal from impaired cognitive functioning. Method We recruited adults who self‐reported to be cognitively normal through the Dutch Brain Research Registry (reference sample), and invited them to complete and evaluate the COST‐A. We computed expected COST‐A scores based on age, sex and education level using multiple linear regression, to obtain demographically‐adjusted, standardized norm scores. We applied ‐1.5 SD units as a cut‐off for (ab)normal performance. Result Out of 3529 participants who started, 3088 participants (60.6±12.1 years old, 70% women, 68% highly educated) successfully completed the COST‐A (Figure 1). These participants considered COST‐A to be of adequate difficulty and duration (±17 minutes), with clear instructions. Based on the norm scores (Figure 2), and corresponding with our cut‐off definition, 242 participants (7.8%) in the reference sample scored below ‐1.5. Results from the independent validation cohort (N=2444, 58.9±12.1 years old, 42% female, 54% highly educated) corroborated this finding: 186 (7.6%) scored below the cut‐off. In the clinical sample, 75% of MCI and 82% of dementia patients scored below ‐1.5. Conclusion COST‐A is an easy‐to‐use online test of cognitive performance for individuals worried about their memory that can be used in the comfort of their own home. This is especially relevant in light of the COVID‐19 pandemic, where digital care is increasingly the norm. Moving forward, optimizing user experiences is key to overcome challenges regarding successful completion of the COST‐A at home, even in cognitively normal individuals. With a high accuracy for distinguishing normal from impaired cognition, COST‐A can serve as a screening tool to identify those individuals who require formal assessment by a healthcare professional and/or for research studies.