Cognitive side effects are a common unintended outcome of electroconvulsive therapy (ECT). Routine cognitive assessment is important for monitoring patient outcomes, although it can pose challenges in busy clinical settings. Computerized cognitive testing has advantages that can facilitate routine monitoring. This study explored the construct and criterion validity of computerized cognitive testing compared with standard pen-and-paper tests for monitoring cognition in ECT patients.
The study included 24 participants with major depression who received an acute course of ECT. Cognition was assessed at pretreatment and at posttreatment with 3 computerized tests from the CogState battery (International Shopping List task, One-Card Learning, and One-Back Task) and 3 conceptually matched pen-and-paper-administered neuropsychological tests.
At pretreatment, only performance on the computer-administered test of verbal anterograde memory (International Shopping List task) was significantly correlated with the analogous pen-and-paper measure, whereas the other computerized tests were not. Of the computerized measures, only the International Shopping List task showed significant changes from pretreatment to posttreatment (P < 0.01, Cohen d > 1.0). In contrast, all the pen-and-paper-administered tests showed significant changes from pretreatment to posttreatment (P < 0.01, Cohen d range, 0.8-1.2). Pretreatment to posttreatment cognitive changes on the computerized measures were not correlated with changes on the pen-and-paper-administered tests.
Construct and criterion validity and tolerability varied between the computerized measures. The results highlighted potentially important issues related to the interpretation and utility of computerized tests in this patient population.