Systematic review of strengths and limitations of Randomized Controlled Trials for non-pharmacological interventions in mild cognitive impairment: Focus on Alzheimer's disease

Abstract

Background

Non-pharmacological interventions may improve cognition and quality of life, reduce disruptive behaviors, slow progression from Mild Cognitive Impairment (MCI) to dementia, and delay institutionalization. It is important to look at their trial designs as well as outcomes to understand the state of the evidence supporting non-pharmacological interventions in Alzheimer's disease (AD). An analysis of trial design strengths and limitations may help researchers clarify treatment effect and design future studies of non-pharmacological interventions for MCI related to AD.

Methods

A systematic review of the methodology of Randomized Controlled Trials (RCTs) targeting physical activity, cognitive interventions, and socialization among subjects with MCI in AD reported until March 2014 was undertaken. The primary outcome was CONSORT 2010 reporting quality. Secondary outcomes were qualitative assessments of specific methodology problems.

Results

23 RCT studies met criteria for this review. Eight focused on physical activity, fourteen on cognitive interventions, and one on the effects of socialization. Most studies found a benefit with the intervention compared to control. CONSORT reporting quality of physical activity interventions was higher than that of cognitive interventions. Reporting quality of recent studies was higher than older studies, particularly with respect to sample size, control characteristics, and methodology of intervention training and delivery. However, the heterogeneity of subjects identified as having MCI and variability in interventions and outcomes continued to limit generalizability.

Conclusions

The role for non-pharmacological interventions targeting MCI is promising. Future studies of RCTs for non-pharmacological interventions targeting MCI related to AD may benefit by addressing design limitations.