Background Self-perceived risk of fracture (SPR) has previously been shown to be associated with personal fracture history, falls, education about osteoporosis, and lifestyle, although previous studies have also reported that many individuals underestimate their fracture risk. In this study of community-dwelling older adults, we considered whether the presence of social isolation, self-efficacy, or mental health status influenced the relationship between SPR and fracture.

Methods Participants were recruited from the Hertfordshire Cohort Study, an established cohort study of UK community-dwelling adults. SPR, compared to individuals of the same sex and age, was assessed via questionnaire and categorised into ‘lower’, ‘similar’ and ‘higher’. Fractures since age 45 were self-reported. Social isolation was assessed using the six-item Lubben Social Network Scale. Self-efficacy was assessed using a shortened General Self-Efficacy Scale (GSE). Mental health status was assessed using the anxiety/depression item from the EuroQoL questionnaire. Associations between SPR and previous fracture were examined using logistic regression with adjustment for sex and age.

Results 146 men and women (median age 83.3 [IQR 81.5–85.5] years) participated. More than half of the participants (n=79 [54.1%]) reported a lower than average SPR, and the majority (n=109 [74.7%]) reported no previous fractures. As expected, SPR was associated with increased odds of previous fractures (OR 1.72, 95% CI 1.03–2.87, per higher band of SPR) when adjusting for sex and age only. The association was unaffected by further individual adjustment for social isolation (1.73, 1.04–2.89), self-efficacy (1.71, 1.02–2.85), and mental health (1.77, 1.05–2.96). However, the association between SPR and fracture history was removed by adjustment for BMI, smoking, alcohol consumption, physical activity, diet quality, number of comorbidities, and use of bisphosphonates.

Conclusion Higher SPR was indeed related to prior fracture since age 45 in a cohort of UK community-dwelling older adults. Adjustment for social isolation, self-efficacy or mental health status did not affect this association although adjustment for BMI, lifestyle, comorbidity and use of bisphosphonates did remove it. These observations suggest that prior medical history and lifestyle factors are the major determinants of SPR, while psychological factors appeared less important in this sample.