Expectancy Does Not Predict 18-month Treatment Outcomes with Cognitive Training in Mild Cognitive Impairment.

Background: Computerized cognitive training (CCT) has emerged as a potential treatment option for mild cognitive impairment (MCI). It remains unclear whether CCT's effect is driven in part by expectancy of improvement.

Objectives: This study aimed to determine factors associated with therapeutic expectancy and the influence of therapeutic expectancy on treatment effects in a randomized clinical trial of CCT versus crossword puzzle training (CPT) for older adults with MCI.

Design: Randomized clinical trial of CCT vs CPT with 78-week follow-up.

Setting: Two-site study - New York State Psychiatric Institute and Duke University Medical Center.

Participants: 107 patients with MCI.

Intervention: 12 weeks of intensive training with CCT or CPT with follow-up booster training over 78 weeks.

Measurements: Patients rated their expectancies for CCT and CPT prior to randomization.

Results: Patients reported greater expectancy for CCT than CPT. Lower patient expectancy was associated with lower global cognition at baseline and older age. Expectancy did not differ by sex or race. There was no association between expectancy and measures of everyday functioning, hippocampus volume, or apolipoprotein E genotype. Expectancy was not associated with change in measures of global cognition, everyday functioning, and hippocampus volume from baseline to week 78, nor did expectancy interact with treatment condition.

Conclusions: While greater cognitive impairment and increased age was associated with low expectancy of improvement, expectancy was not associated with the likelihood of response to treatment with CPT or CCT.