BackgroundPeople with dementia and their families need support in different forms, but currently services are often fragmented with variable quality of care. Case management offers a way of co-ordinating services along the care pathway and therefore could provide individualised support; however, evidence of the effectiveness of case management for dementia is inconclusive.ObjectiveTo adapt the intervention used in a promising case management project in the USA and test its feasibility and acceptability in English general practice.DesignIn work package 1, a design group of varied professionals, with a carer and staff from the voluntary sector, met six times over a year to identify the skills and personal characteristics required for case management; protocols from the US study were adapted for use in the UK. The feasibility of recruiting general practices and patient–carer dyads and of delivering case management were tested in a pilot study (work package 2). An embedded qualitative study explored stakeholder views on study procedures and case management.SettingFour general practices, two in the north-east of England (Newcastle) one in London and one in Norfolk, took part in a feasibility pilot study of case management.ParticipantsCommunity-dwelling people with dementia and their carers who were not already being case managed by other services.InterventionA social worker shared by the two practices in the north-east and practice nurses in the other two practices were trained to deliver case management. We aimed to recruit 11 people with dementia from each practice who were not already being case managed.Main outcome measuresNumbers of people with dementia and their carers recruited, numbers and content of contacts, needs identified and perceptions of case management among stakeholders.ResultsRecruitment of practices and patients was slow and none of the practices achieved its recruitment target. It took more than 6 months to recruit a total of 28 people with dementia. Practice Quality and Outcome Framework registers for dementia contained only 60% of the expected number of people, most living in care homes. All stakeholders were positive about the potential of case management; however, only one of the four practices achieved a level of case management activity that might have influenced patient and carer outcomes. Case managers’ activity levels were not related solely to time available for case management. Delivery of case management was hindered by limited clarity about the role, poor integration with existing services and a lack of embeddedness within primary care. There were discrepancies between case manager and researcher judgements about need, and evidence of a high threshold for acting on unmet need. The practice nurses experienced difficulties in ring-fencing case management time.ConclusionsThe model of case management developed and evaluated in this feasibility study is unlikely to be sustainable in general practice under current conditions and in our view it would not be appropriate to attempt a definitive trial of this model. This study could inform the development of a case management role with a greater likelihood of impact. Different approaches to recruiting and training case managers, and identifying people with dementia who might benefit from case management, are needed, as is exploration of the scale of need for this type of working.Trial registrationCurrent Controlled Trials ISRCTN74015152.FundingThis project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full inHealth Technology Assessment; Vol. 18, No. 52. See the NIHR Journals Library website for further project information.