[P1–601]: DESIGN OF CONTROLS IN TRIALS OF COMPUTERISED COGNITIVE TRAINING IS INEFFECTUAL: A META‐ANALYSIS IN HEALTHY OLDER ADULTS

• DOI: 10.1016/j.jalz.2017.06.617
• Publication Date: 2017-10-20T08:31:03Z
• Authors: Harry Hallock; Anna Radowiecka; Kathryn M. Broadh...; Isabella Hoi Kei ...; Michael Valenzuela; Amit Lampit
• Keywords: Blinding; Cognitive Training; Cognitive Decline

Rapid development and evaluation of interventions to modify ageing-related cognitive decline is key for global efforts prevent dementia. Computerised training (CCT) a popular safe intervention enrichment in older adults. CCT has been illustrated be efficacious improving overall cognition variety populations including healthy adults, MCI PD. However, as effective masking the therapeutic effect challenging blinding participants implausible, it difficult quantify contribution non-specific effects such expectancy bias (‘placebo effect’) or importance trial design choices. Using large set randomised controlled trials (RCTs) we therefore aimed examine determinants outcomes potential moderators within control arm trials. Medline, Embase, PsychINFO were systematically searched RCTs investigating (≥4 hours) on performance neuropsychological tests adults without neuropsychiatric disorders diagnosed impairment. Fifty-seven encompassing 63 groups 2,712 found. A series subgroup analyses revealed no statistically significant differences between: active (k=30, g=0.18, 95% CI 0.12 0.24, I2=38.23) passive (k=33, g=0.12, 0.08 0.16, I2=0); blinded (k=32, g=0.15, 0.09 0.20, I2=32.84) non-blinded assessors (k=31, 95%, g=0.16, 0.10 0.21, adherence intention-to-treat analysis (k=30 g=0.14, 0.18, I2=0) non-adherence 0.07 I2=37.57). We also found difference between various approaches 30 studies, as: type, location frequency sessions; participants. meta-regression relationship size controls their corresponding (β=0.22, p<0.01, R2=0.86). Contrary common practice, supposed ‘gold standards’ empirical science appear ineffectual. Given costs associated with implementing designs, shift head-to-head recommended better inform clinical community translation.