Premium
P4‐330: Developing a Novel Evaluation System for the Clock Drawing Test
Author(s) -
Hughes Charles E.,
Patel Nikhil S.
Publication year - 2016
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2016.07.074
Subject(s) - test (biology) , cognitive impairment , demographics , cognition , computer science , subject (documents) , psychology , medicine , psychiatry , paleontology , demography , sociology , library science , biology
Background: Prior research has shown that, if started early, both brain training exercises and drug therapies can have positive impacts on delaying the progression of Alzheimer’s disease. However, early detection is typically expensive and difficult (Rebok, 2014). The Clock Drawing Test (CDT) aimed to break that early detection barrier and has been a standard diagnostic test for cognitive impairment for the past 30 years; however, it’s interpretation can be subjective, potentially yielding inconsistent results. This study focused on creating an automatic, machine-based, and accurate scoring system for the CDT. Methods:The initial step was to develop the CDT evaluation system. To do that, web-based software was created that can, in real time, reproduce a subject’s pen-and-paper drawing and automatically calculate various metrics, such as the subject’s cognitive processing time, which cannot be captured by the current test. Next, the standard CDTwas administered to a test group of people (n1⁄451) in the age range of 45-60 who had been diagnosed with early-stage cognitive impairment through a series of complex tests and a control group (n1⁄453) with similar demographics but no diagnosed impairment. Those drawings were digitized using a Smartpen and put through the developed system. Results:On each clock, the system calculated over 40 metrics. Most important was “think time,” or the amount of time a subject spent holding their pen above the paper in between strokes. Statistically significant differences (p<.01) arose between the think times of the control (m1⁄412 seconds) and test (m1⁄473 seconds) groups. A logistic regression yielded a predictive equation that could correctly identify subjects with cognitive impairment 90% of the time. In comparison, geriatric physicians that evaluated the clocks correctly identified only 45%. Conclusions:The new CDT scoring system is both significantly more accurate at identifying thosewith early stage cognitive impairment and drastically cheaper. The test population of subjects with cognitive impairment had that impairment diagnosed through many complex tests, with their average cost for diagnosis being $9500. In contrast, this test was able to identify the impairment in 5 minutes with less than $200 of equipment. In a primary-care setting the cost would be about $15 per test.