BACKGROUND: NIDUS-Family is a psychosocial and behavioural intervention comprising six to eight sessions, delivered by non-clinical facilitators, and tailored to goals set by dyads of people with dementia and their unpaid or family carers. The intervention has been shown to be effective for attainment of personalised client goals. The current study aimed to determine whether the intervention is cost-effective. METHODS: In this cost-utility and cost-effectiveness analysis within a two-armed, single-masked, multisite, superiority randomised controlled trial, we recruited 302 people with dementia living in their own homes and their family carers from National Health Service community settings and social and print media across England. Participants were randomly allocated (2:1) to the NIDUS-Family intervention group or control (goal setting and routine care) group. Randomisation was blocked and site-stratified, with allocation by a remote web-based system. We calculated the probability that NIDUS-Family is cost-effective for a client with dementia based on quality-adjusted life-years from a health and social care perspective and from a societal perspective (additionally including family carer time and out-of-pocket costs), at £20 000-30 000 decision thresholds for additional quality-adjusted life-years compared with usual care over 12 months. Analyses were done using the intention-to-treat population. This study is registered with ISRCTN, ISRCTN11425138, and is completed. FINDINGS: Between April 30, 2020, and May 9, 2022, we randomly allocated 204 participants (109 [53%] women and 95 [47%] men) to the intervention group and 98 (60 [61%] women and 38 [39%] men) to the control group. 218 (72%) participants at 6 months and 178 (59%) at 12 months provided cost data. At both a £20 000 and £30 000 decision threshold, there was an 89% probability that NIDUS-Family was cost-effective compared with usual care from a health and social care perspective, and an 87% probability from a societal perspective. Intervention participants accrued on average £8934 (37%) less in costs than control participants (95% CI -£59 460 to £41 592). INTERPRETATION: NIDUS-Family is the first personalised care and support intervention to show both cost-effectiveness from the perspective of the quality of life of people with dementia as well as clinical effectiveness and should therefore be part of routine dementia care. FUNDING: Alzheimer's Society.