1034All-cause mortality in Australia: impact of social isolation and living alone. (2nd September 2021)

• Record Type: Journal Article
• Authors: Barr, Margo; Kabir, Alamgir; Harris-Roxas, Ben; Comino, Elizabeth; Jackson, Tony; Crozier, Ann-Marie; Goodger, Brendan; Finch, Julie; Harris, Mark
• Is Part Of: International journal of epidemiology. Volume 50(2021)Supplement 1
• Publication Date: 2021-09-02
• Subjects: Epidemiology -- Periodicals; 614.4
• DOI: 10.1093/ije/dyab168.060
• Languages: English
• ISSNs: 0300-5771

Abstract

Abstract: Background: The prevalence of social isolation (SI), living alone (LA) and loneliness among older people in Australia is 17%, 25% and 19% respectively. In a meta-analysis of 70 prospective studies undertaken by Holt-Lunstad et al (2015) 1 they found, after controlling for covariates, a 29% increased all-cause mortality for SI, 32% for LA, and 26% for loneliness. But how accurate is this evidence for Australia considering that the meta-analysis only included three Australian studies; all only reporting on LA, and only one reported a significant effect of LA on total mortality. Methods: We conducted a record linkage study in NSW using the Social Economic and Environmental Factors sub-study questionnaire data (n = 60, 399) from the 45 and Up Study (n = 267, 153) and death registrations (Centre for Health Record Linkage). Overall, 20% were SI and 18% were LA. Univariate and multivariable generalised linear models were used to estimate crude and adjusted 7-year all-cause mortality rate ratios (MRR). Results: Crude MRR were 28% higher for SI and 98% higher for LA. When adjusted by age and sex they were still significantly higher (28% and 23% respectively). However, when controlled for covariates, SI was 9% higher [AdjRR(95%CI):1.09(1.01-1.18)] and LA had no effect on mortality [AdjRR(95%CI):1.05(0.94-1.18)]. Conclusions: As was the case for the Australian studies used in the meta-analysis we found no association between LA and all-cause mortality. We did however find a small significant effect for SI. Key messages: Applying the overall meta-analysis results to Australia overestimates the effect of SI and LA on increased all-cause mortality.