Research evaluating the effectiveness of dementia interventions in low- and middle-income countries: A systematic mapping of 340 randomised controlled trials.

ABSTRACT

OBJECTIVES: More people with dementia live in low- and middle-income countries (LMICs) than in high-income countries, but best-practice care recommendations are often based on studies from high-income countries. We aimed to map the available evidence on dementia interventions in LMICs. METHODS: We systematically mapped available evidence on interventions that aimed to improve the lives of people with dementia or mild cognitive impairment (MCI) and/or their carers in LMICs (registered on PROSPERO CRD42018106206). We included randomised controlled trials (RCTs) published between 2008 and 2018. We searched 11 electronic academic and grey literature databases (MEDLINE, EMBASE, PsycINFO, CINAHL Plus, Global Health, World Health Organization Global Index Medicus, Virtual Health Library, Cochrane CENTRAL, Social Care Online, BASE, MODEM Toolkit) and examined the number and characteristics of RCTs according to intervention type. We used the Cochrane risk of bias 2.0 tool to assess the risk of bias. RESULTS: We included 340 RCTs with 29,882 (median, 68) participants, published 2008-2018. Over two-thirds of the studies were conducted in China (n = 237, 69.7%). Ten LMICs accounted for 95.9% of included RCTs. The largest category of interventions was Traditional Chinese Medicine (n = 149, 43.8%), followed by Western medicine pharmaceuticals (n = 109, 32.1%), supplements (n = 43, 12.6%), and structured therapeutic psychosocial interventions (n = 37, 10.9%). Overall risk of bias was judged to be high for 201 RCTs (59.1%), moderate for 136 (40.0%), and low for 3 (0.9%). CONCLUSIONS: Evidence-generation on interventions for people with dementia or MCI and/or their carers in LMICs is concentrated in just a few countries, with no RCTs reported in the vast majority of LMICs. The body of evidence is skewed towards selected interventions and overall subject to high risk of bias. There is a need for a more coordinated approach to robust evidence-generation for LMICs.