Why do older adults seek emergency care? The impact of contextual factors, care, health, and social relations

Background: Emergency department (ED) visits are becoming more prevalent globally. EDs provide care for acute health conditions, but some of these visits are driven by needs unmet by primary health care and social care for older adults, indicating ineffective social care and healthcare systems. The ED is often an inappropriate setting for older adults because of the lack of interdisciplinary teams with clinical competence in the care of an increasingly ageing population and because of poor continuity of care which entails the risk of adverse health outcomes. The Andersen model of health services use proposes contextual and individual factors to understand health care utilisation better. However, there are knowledge gaps in research on ED care in relation to contextual factors, home help receipt, and aspects related to inadequate informal care. Moreover, selection bias often limits previous research on ED care. Aim: The overall aim was to study factors associated with ED care use in older adults. Design: Prospective cohort study Study sample: All adults ≥65 years with ED visits in 2014 living in two Swedish regions (Dalarna, N=16 688 and Stockholm, N=101 017) participated in study I. The study population in study II was all community-living older-old adults (≥80 years) who were registered residents of Dalarna on 31 December 2014, excluding those who moved into residential care facilities during 2015 (N=16 543). In study III, the participants were adults ≥60 years who participated in the Swedish National Study on Aging and Care-Kungsholmen (SNAC-K). The data were pooled from three waves (W) of SNAC: W1= 2001-2003, W3=2007-2009, and W5=2013-2015. Persons living in residential care facilities were excluded from study III providing an analytical sample of N=3 066 at W1, N=1 885 at W3, and N=1 208 at W5. In study IV, adults ≥66 years who participated in the SNAC-Blekinge study (W3: 2007-2009) and who provided information on the exposure variable were included (N=673). Data sources: The four studies of this thesis were based on national and regional registers and survey data. The registers were the Longitudinal Integration Database for Health Insurance and Labour Market, the Social Services Register, the Swedish Prescribed Drug Register, the National Patient Register, and the health care databases of Region Blekinge, Dalarna, and Stockholm. Municipal-level data were accessed from Kolada, a publicly accessible, comprehensive national database. Survey data were based on the Swedish National study on Aging and Care in Blekinge and Kungsholmen. Dependent variables: The dependent variables were ED visits, at least one ED revisit within 30 days of an initial ED visit, and frequent ED use. Independent variables: The independent variables included contextual factors (the proportion of adults aged ≥80 years in the total population, annual social care expenditures per person aged ≥80 years, home help quality, median days in residential care, and distance to the ED), individual-level predisposing factors (age, gender, and education), individual-level enabling resources (living arrangements, social connections, social support, and informal care), individual-level need factors (subjective and objective health status), and health care and social care use (primary health care visits, specialist care visits, hospital admissions, ED visits in the previous year, disposition at initial ED visit (admission to inpatient care/discharged home), residential care receipt, and home help receipt). Data analyses: Logistic regression models were used to analyse the associations between independent variables and dichotomous dependent variables (ED visits, ED revisits, frequent ED use). Cox regression models were computed to determine the association between independent variables and time to the first ED visit. Associations between independent variables and the number of ED visits were assessed using generalised estimating equations with negative binomial regressions. In studies III and IV, all analyses were stratified by age group (Study III: younger-old, 80 years). Results: Analysis of contextual factors showed that the proportion of adults aged ≥80 years in the total population and shorter distance to the ED were associated with ED visits in older-old adults (Study II). There were mixed findings on age, gender, and education level for individual-level predisposing factors. Regarding individual-level enabling resources, higher levels of social support were negatively associated with ED visits but only in older-old adults (Study III). In relation to the need for care factors, poor health status was associated with ED visits (Studies II-IV), ED revisits (Study I), and frequent ED use (Study IV). Concerning the utilisation of care, primary health care visits in the previous 12 months were associated with ED visits (Study II) and ED revisits (Study I). Hospital admissions and ED visits 12 months before the initial ED visit were associated with ED revisits (Studies I & II). Older-old adults admitted to inpatient care at the initial ED visits were 29% less likely to revisit an ED within 30 days of the initial ED visit than those discharged home (Study II). Older-old adults receiving home help for instrumental services and personal care were 148% more likely to visit an ED compared to those not receiving home help. This group with intensive home help also had a 30% higher likelihood of an ED revisit within 30 days of the initial ED visit (Study II). Conclusions: Contextual factors contribute to understanding ED care use in older adults. Our findings on poor health status suggest that the need for care determines ED care use in older adults. However, factors other than health status also explain the use of ED care. For example, social support indicates inequalities and suggests investing in public health resources to address these risk factors. Discharge to home from the ED and risk of an ED revisit could indicate that health care and post-discharge care are not meeting the needs of older patients. Findings on the home help receipt and ED care use illustrate the vulnerability of this group and highlight the importance of future research on self-reported unmet needs of home help and the effect of unmet needs on the use of ED care