TY - JOUR T1 - Deaths in dementia: a scoping review of prognostic variables JF - BMJ Supportive & Palliative Care JO - BMJ Support Palliat Care SP - 242 LP - 252 DO - 10.1136/bmjspcare-2020-002217 VL - 11 IS - 3 AU - Palvinder Kaur AU - Woan Shin Tan AU - Pradeep Paul George Gunapal AU - Yew Yoong Ding AU - Reuben Ong AU - Huei Yaw Wu AU - Allyn Hum Y1 - 2021/09/01 UR - http://spcare.bmj.com/content/11/3/242.abstract N2 - Objectives To identify the types of factors included in research examining mortality in patients with dementia, and to stratify the identified factors by care settings.Design We systematically searched PubMed, Embase, PsycINFO and the Cumulative Index to Nursing and Allied Health Literature (CINAHL) databases, and identified grey literature from the Networked Digital Library of Theses and Dissertations, Open Grey and Grey Literature Report. Two authors independently screened for eligibility of studies. Independent reviewers extracted relevant study information. We conducted a narrative synthesis of the data.Results We identified 8254 articles, of which 94 met the inclusion criteria. More than half (n=53) were published between 2009 and 2018 with half from Europe. Studies were conducted across hospices/nursing homes (n=25), hospital (n=23), outpatient clinics (n=21), mixed settings (n=15) and in the community (n=10). Nearly 60% adopted a prospective cohort study design with 87% performing multivariable analysis. Overall, 239 variables were identified and classified into six themes—individual factors, health status, functional ability, cognition and mental health, treatments and health system factors. Although a general set of factors were common across all studies, when stratified by care settings, variations were seen in the specific variables included.Conclusion Identifying prognostic variables relevant to the dementia population in each setting is key to facilitate appropriate care plans and to ensure timely access to palliative care options. Future research should also focus on ensuring the replicability of prognostic models and to generate a better understanding of the direct and interacting influence of the identified factors on mortality. ER -