Artificial intelligence (AI) is showing huge promise in assisting with early diagnosis of dementia, which would yield clear benefits in optimising the quality of life of those with dementia and their carers. In this review, the authors provide an overview of AI technology, its applications and implications for clinical practice. Consideration is given to ethical issues and future challenges.

BackgroundSouth London and Maudsley NHS Foundation Trust is the largest mental health trust in the UK, serving four boroughs in South East London. In 2014, the ‘triage ward’ system was introduced in three boroughs. Similar to an acute medical admission unit, the triage ward would rapidly assess and treat all new admissions. The patients would either be discharged or admitted to a ‘locality ward’ for further treatment.ProblemThe unforeseen consequences of the ‘triage ward’ system were duplications and omissions of medical tasks on receiving wards, which affected efficiency and quality of care. This was due to a lack of formal medical handover. We aimed to improve efficiency and patient safety by formalising the junior doctor handover between triage and locality wards, ensuring every patient transferred had a documented handover in their electronic notes.MethodWe consulted our colleagues with a survey, ascertaining their views on the current system, the need for a more formalised system and what form that system should take. Using their feedback, we devised a handover template, to be completed for all patients transferred to locality wards. We then rolled the project out to the other two boroughs using the same methodology.ResultsA follow-up survey showed improvement in our baseline results and that the majority of transferred patients were formally handed over. Serious incident data showed a decrease in incident rates pre-intervention and post-intervention. The intervention was sustained a year later. The transfer of the intervention to other sites was problematic.DiscussionThe project showed the lack of handover was a concern shared by colleagues, and they considered our template a useful way of addressing this. The results suggested that the intervention was sustainable despite frequent rotations of staff. The difficulties in transferring an intervention to new sites are discussed.

Blood‐based biomarkers for identifying Alzheimer's disease (AD) pathology have been of considerable interest over the past decade and can offer a significant advantage over the currently available and validated cerebrospinal fluid and positron emission tomography biomarkers in terms of cost, invasiveness and accessibility. This review explores the role of three blood‐based biomarkers; plasma amyloid β, tau and neurofilament light chain, and their combination, in the identification of AD pathology and how they may be utilised in clinical practice.

No abstract