The Feasibility of At-Home iPad Cognitive Testing For Use in Clinical Trials.

BACKGROUND Technological advances now make it feasible to administer cognitive assessments at-home on mobile and touch-screen devices such as an iPad or tablet computer. Validation of these techniques is necessary to assess their utility in clinical trials. OBJECTIVES We used a Computerized Cognitive Composite for Preclinical Alzheimer's Disease (C3-PAD) developed for iPad 1) to determine the feasibility of performing the C3-PAD at home by older individuals without the presence of a trained psychometrician; 2) to explore the reliability of in-clinic compared to at-home C3-PAD performance and 3) to examine the comparability of C3-PAD performance to standardized neuropsychological tests. DESIGN SETTING PARTICIPANTS Forty-nine cognitively normal older individuals (mean age, 71.467.7 years; 20% non-Caucasian) were recruited from research centers at the Massachusetts General Hospital and Brigham and Women's Hospital. Participants made two in-clinic visits one-week apart and took five 30-minute alternate versions of the C3-PAD at-home measuring episodic memory, reaction time and working memory. MEASUREMENTS A reliability analysis explored equivalence of the six alternate C3-PAD test versions. A feasibility assessment calculated the percentage of individuals who completed all at-home tests correctly, in contrast to incomplete assessments. Correlational analyses examined the association between C3-PAD-clinic compared to C3-PAD-home assessments and between C3-PAD performance and standardized paper and pencil tests. RESULTS Excellent reliability was observed among the 6 C3-PAD alternate versions (Cronbach alpha coefficient=0.93). A total of 28 of 49 participants completed all at-home sessions correctly and 48 of 49 completed four out of five correctly. There were no significant differences in participant age, sex or education between complete and incomplete at-home assessments. A single in-clinic C3-PAD assessment and the at-home C3-PAD assessments were highly associated with each other (r2=0.508, p<0.0001), suggesting that at-home tests provide reliable data as in-clinic assessments. There was also a moderate association between the at-home C3-PAD assessments and the in-clinic standardized paper and pencil tests covering similar cognitive domains (r2= 0.168, p< 0.003). CONCLUSIONS Reliable and valid cognitive data can be obtained from the C3-PAD assessments in the home environment. With initial in-clinic training, a high percentage of older individuals completed at-home assessments correctly. At-home cognitive testing shows promise for inclusion into clinical trial designs.