Harriet Demnitz‐King, Claudia Cooper, Michaela Poppe, Jessica Budgett, Sweedal Alberts, Larisa Duffy, Mariam Adeleke, J. Barber, Elisa Aguirre, Henry Brodaty, Alexandra Burton, Paul Higgs, Jonathan Huntley, Helen C Kales, Iain Lang, Natalie L. Marchant, Sarah Morgan-Trimmer, Anne Marie Minihane, Penny Rapaport, Miguel Rio, Karen Ritchie, Zuzana Walker, Kate Walters, Rachael Hunter
BACKGROUND: Previous studies have modelled long-term cost-effectiveness of dementia prevention but seldom within trials. APPLE-Tree (Active Prevention in People at risk of dementia through Lifestyle, bEhaviour change and Technology to build REsiliEnce) is a personalised, multi-domain, low-intensity dementia prevention intervention for adults experiencing subjective cognitive decline or mild cognitive impairment. Intervention receipt, relative to control, was associated with an adjusted mean difference of 0.06 (95%CI -0.001 to -0.128) in the Neuropsychological Test Battery (NTB). We evaluated cost-effectiveness from health- and social-care perspectives over 2 years. METHODS: We recruited 746 older adults from English health and community settings and randomly assigned them (1:1) to APPLE-Tree plus usual care (UC) or UC plus written dementia prevention information. Quality-adjusted life-years (QALYs) were calculated from the EQ-5D-5L. Over 24-months, intention-to-treat analyses compared QALYs (primary) and incremental cost per NTB unit (one z-score) change (secondary). RESULTS: Between October 2020 and December 2022, 374 participants were randomised to intervention and 372 to control. Mean adjusted QALYs were 1.511 (intervention) and 1.520 (control), an adjusted difference of -0.010 (95%CI -0.04 to -0.022). Mean adjusted (SD) costs were £2966 (3629) and £2551 (4439), respectively. Cost per 1-point NTB change z-score was £6809. At a £30 000 threshold, the probability of cost-effective was 12% and 96% based on QALYs and cognition, respectively. Post hoc analyses indicated higher cost-effectiveness for non-White participants (incremental costs: £374; 95%CI -590.751 to -1338.515) and socioeconomically disadvantaged participants (non-homeowners), who had higher QALYs (0.106) and lower costs (-£1830), yielding a 98% probability of cost-effectiveness at a £20 000 threshold. CONCLUSION: APPLE-Tree has a high probability of being cost-effective for a 1-point NTB gain, a difference previously associated with a three-fold reduction in 5-year dementia risk. QALY-based cost-effectiveness was not observed, but post-hoc analyses suggested it may be more effective in groups with greater needs. Findings highlight the potential of targeted, lower-intensity interventions to reduce dementia risk. Longer-term follow-up is required to determine whether cost-effectiveness is realised over time.