Clinical Experience with SATURN, a Public Domain Self-administered Cognitive Screening Test (P5-12.002)

Abstract

<h3>Objective:</h3> Validation via regular clinical use of a free, public domain, automatically scored, and self-administered cognitive screening test. <h3>Background:</h3> Emerging therapies for neurodegenerative disease are directed at patients with little cognitive impairment. How do we efficiently identify these patients? Universal screening with paper-and-pencil tests like the Mini Mental State Exam (MMSE) or the Montreal Cognitive Assessment (MoCA) is cost-prohibitive due to clinician time investment. Self-Administered Tasks Uncovering Risk of Neurodegeneration (SATURN) is a tablet-based cognitive screening test designed to bypass that cost. After a small validation study, SATURN was integrated into regular clinical use alongside paper-and-pencil tests. We hypothesized that SATURN’s psychometric properties continued to compare favorably to paper-and-pencil tests. <h3>Design/Methods:</h3> We will retrospectively analyze data from every patient (n&gt;500) visiting one dementia sub-specialist clinic over a 24-month period. To-date, we have analyzed 66 patients’ records. We documented visit diagnoses and SATURN, MoCA, and MMSE scores. Primary analyses (1) use regression to test whether SATURN is well-correlated with the other tests, and (2) compare test scores to clinical diagnosis using receiver operating characteristic (ROC) curves. <h3>Results:</h3> For this interim analysis, we analyzed patients (n=37) who met prespecified inclusion criteria and also had scores for at least two of the three aforementioned tests. Clinical diagnoses – normal (n=8), mild cognitive impairment (n=16), and dementia (n=13) – were strongly associated with (mean±sd) test scores (SATURN: 27±3, 18±5, 13±6) (MoCA: 25±3, 20±4, 15±4) (MMSE: 30±0, 25±2, 21±6) (each p&lt;0.02; one-way ANOVA). ROC curves were statistically similar for all three tests (p&gt;0.10). SATURN was well-correlated with MoCA (n=24 tested with both: r=0.87, p&lt;0.0001), and MMSE (n=20 with both: r=0.81, p&lt;0.0001). MMSE and MoCA were also correlated (n=17 with both: r=0.69, p=0.002). <h3>Conclusions:</h3> The freely-available SATURN (https://doi.org/10.5061/dryad.02v6wwpzr) is comparable to the MoCA and MMSE in regular clinical practice; a prerequisite for larger-scale efforts to efficiently screen for cognitive impairment. <b>Disclosure:</b> Dr. Florendo has nothing to disclose. Dr. Bissig has received personal compensation in the range of $100,000-$499,999 for serving as a Physician with University of California - Davis.