Emergency hospital admissions among older adults living alone in the community

In this study among Singaporean older adults, we found that living alone was associated with higher odds of ED admission, longer inpatient days and higher inpatient costs than living with others. Furthermore, while older adults living alone were at higher risk of ED admissions and higher inpatient costs regardless of multimorbidity, those living alone without multimorbidity were at higher risk of longer inpatient days than their counterparts living alone with multimorbidity.

The results are in line with previous studies showing that older people living alone are at higher risk of ED admissions [13‐15, 24]. Older adults living alone are a highly heterogeneous group that also includes those who prefer to live without company, are functionally independent and have an extended network of support to tap into. It has been reported that older adults living alone tend to have more diverse networks of support, while those living with others often have less social ties beyond family [45]. In fact, good health is often a precondition for older adults to live independently, making those living alone more likely to be healthier. In line with this, reverse causality and a wide range of social and psychological confounders could affect the associations between living alone and health outcomes [68]. However, the associations between living arrangements and health follow different mechanisms from the associations between living arrangements and emergency admissions. While we could expect older adults in poor health to be more likely to be hospitalized, the hospitalization route could differ by living arrangements: those living with family could be more likely to have planned admissions whereas those living alone may have more frequent emergency admissions. For example, a healthy older adult living alone could be at higher risk of an emergency admission in the event of a fall compared to another adult in poorer health who lives with a ‘gatekeeper family’ for such events. Moreover, for those living alone without the timely consultation and assistance from family members, the reasons for emergency admissions could go beyond health issues, including poor access to preventive care services, poor health behaviors, lack of health monitoring or lack of instrumental support.

Therefore, regardless of health status and social engagement, older adults living alone may not be able to draw support as immediately as those living with others. This immediate availability of support not only may avert costly ED admissions, but also provide regular assistance in managing age related conditions and prevent further complications [69]. Studies have indeed shown that having informal assistance at home can improve the management of chronic diseases and reduce the need for unplanned hospitalizations and ED visits [23, 70‐74]. We also found that older adults living alone are more likely to have longer inpatient days after adjusting for demographic, health and social factors. This suggests that having informal support at home could indirectly shorten inpatient days by ensuring post-discharge assistance.

The fact that multimorbidity did not moderate the association between living alone and risk of ED admission in this study highlights the commonalities of older Singaporeans living alone. Regardless of health status, those living alone tend to pride themselves on self-reliance and not being a burden for their families [47]. The remarkable resourcefulness, coping mechanisms and self-determination of older Singaporeans who live alone has been well documented [25, 47, 75]. This resilience and adaptation to old age adversities could buffer the influence of health problems on healthcare use and explain why multimorbidity status is not an aggravating factor in the association of living alone and ED admission risk.

Interestingly, we found that, compared to those living with others and without multimorbidity, the relative mean increase in inpatient days was 1.13 days for those living alone without multimorbidity and 0.73 days for those living alone with multimorbidity. While we would expect older adults living alone with multimorbidity to have the longest inpatient days, they may also be more health aware and fearful of emergency health expenditures and thus avoid longer hospitalizations related to traumatic injury such as falls [75]. Conversely, those without multimorbidity could be less risk averse and engage in more hazardous activities leading to longer hospitalizations [76, 77]. In addition, those living alone with multimorbidity may interact more often with tertiary care providers and thus be better monitored than their counterparts without multimorbidity. In fact, older Singaporeans who live alone with multimorbidity prefer using tertiary care services for continuity of care because of their established relationship with regular doctors at the hospital [25]. This familiarity could result in more frequent ED admissions related to multimorbidity complications but shorter inpatient days because the case is well known to care providers.

The conclusions presented here also call for adapting existing theories of healthcare use to the Asian population, where living arrangements and family support are key elements of the care continuum of older adults. In the Western literature, the individual typically comes into contact with healthcare systems through primary care for screening purposes or after becoming ill or suffering an injury [78]. In this context, the Andersen’s Behavioral Model of Health Services Use posits that social resources such as availability of family or friends are enabling factors that facilitate access to healthcare services [62]. However, in Asian countries where older adults are heavily supported by co-residing family, living arrangements may not just be an enabling factor but the first point of care for older adults. Co-residing family could therefore act as a gatekeeper of costly ED admissions by monitoring health status, ensuring medical adherence and facilitating access to primary care services. In the absence of support from other household members, older adults living alone could be less aware of their health, have worse adherence to medication regimes and have difficulties accessing primary care services on their own [37‐40]. Ultimately, this could aggravate age related complications and result in emergency situations requiring immediate medical care [69]. This pattern of resorting to acute tertiary services due to poorly managed chronic diseases could be even more salient in countries with fragmented primary care and support services for older people.

The results presented here hold policy relevant implications in terms of health service planning amid ageing demographics. Although living arrangements may be hard to modify, the delivery of healthcare can certainly be adapted to a rapidly growing number of older adults lacking informal support at home [42]. Previous research has shown that strengthening primary care services at the community level could fill the gaps in informal support and prevent ED admissions [10, 79, 80]. These services could be complemented with more frequent community nurse visits to monitor health status and manage chronic diseases. Our results also suggest that extending informal support after hospitalization discharge could potentially reduce inpatient days and inpatient costs, particularly among those living alone without multimorbidity. Previous research has also shown that programs facilitating the transition between hospital and community are effective in reducing future re-admissions [11, 81]. In addition, interventions should consider the heterogeneity of older adults and discern between social and medical needs. For example, those living with family may have their instrumental support covered but lack social engagement with people beyond their household, while those living alone may be socially active but lack instrumental support to manage their chronic diseases. As a result, community interventions could fill the gaps in support through befriending activities for those living with family, while providing instrumental assistance to those living alone.

The strengths of this study are the use of individual-level inpatient admission records, the wide range of hospitalization outcomes covered, the large sample size and the use of validated instruments to adjust for physical, social, cognitive and depressive factors among older adults. There are, however, some limitations. First, medical history was based on self-reported data, which is not equivalent to a clinical diagnosis. Similarly, the use of screening instruments such as the GDS, Lawton IADL or the MMSE, while useful for population-based studies, is not equivalent to in-depth clinical assessments. Second, our analysis is an under-estimate as it does not include data from specialist outpatient clinics, which comprise a large fraction of tertiary healthcare expenditure. Third, our independent variables were measured at a fixed point of time and therefore our analysis cannot capture changes in health status and healthcare needs over time. Fourth, the individual-level records could not single out preventable unplanned admissions, which are more likely to be affected by living arrangements than non-preventable admissions. Hence, future research should focus on understanding the social factors influencing preventable unplanned admissions to promote preventive care services among groups at risk.