Convergent Validity and Classification Agreement Between Traditional Cognitive Screening Instruments and a Novel, Electronic-based, Voice-recognition Cognitive Screening Tool
Maddeline Sadoff1, Angelina Tretola1, Allison Kaup1, Jacob Hall2, Lori Alasantro1, Jay Rosen1, Vidar Vignisson1, Gregory Sahagian3
1The Neurology Center, 2Neurology Center of Southern CA, 3The Neurology Center of Southern California
Objective:
his study aimed to investigate convergent validity and classification agreement between traditional cognitive screeners and a novel, electronic-based, voice-recognition cognitive screening tool (Intraneuron Memory Exam Screening, IME-S).
Background:
Mild cognitive impairment (MCI) often goes undetected. Traditional paper/pencil cognitive screeners (e.g., MMSE, MOCA, Mini MOCA) are widely utilized, but lack sensitivity to subtle cognitive change, lack specificity for the cause of decline, and are time consuming for medical staff to administer.
Design/Methods:
73 patients (age range 37-87; 50.7% female) from a community-based neurology practice completed a traditional cognitive screener and IME-S. For IME-S, a verbal memory test and verbal fluency tests were administered on iPad using voice-recognition technology. We examined convergent validity between cognitive screener and IME-S scores using correlation analysis. To assess classification agreement, we conducted McNemar tests to compare the number of individuals classified as cognitively “normal” vs. “impaired” using the traditional screener (with standard cut-offs) versus IME-S (using 1SD or 1.5SD cut-offs).
Results:
We found statistically significant, though modest, correlations between the traditional screeners and IME-S scores (all p < .05): letter fluency (r = 0.38), animal fluency (r = 0.33), verbal learning (r = 0.29), verbal recall (r = 0.36). However, results showed significant classification disagreement between the two assessment modalities. Using a 1SD cut-off for IME-S scores, IME-S classified 89.0% of patients as impaired, while the traditional screener classified 67.1% as impaired (p < 0.001). Using a 1.5SD cut-off for IME-S scores, IME-S classified 80.8% of patients as impaired vs 67.1% with the traditional screener (p < 0.05).
Conclusions:
This study demonstrated convergent validity between traditional cognitive screeners and IME-S. Results suggest IME-S may be superior in detecting early cognitive decline compared to traditional screeners (e.g., MOCA). This study suggests IME-S may have the potential to maximize utilization and improve early detection of cognitive decline.