This is the first study to quantify change in HRQoL following an acute low-moderate energy fracture in Australian older adults and the associated loss in quality-adjusted life years. We estimated a mean loss of 18 % in QALYs across all fractures over 12 months, which is equivalent to an average loss of 65 days in full health for each fragility fracture. This represents a substantial health impact. Further analysis by fracture site found that the pattern associated with HRQoL assessed within 2 weeks of the fracture, then at 4, 12 and 18 months was similar for all fracture sites; however, the magnitude of the change differed according to fracture site. Of importance, these data show that the HRQoL for non-hip, non-vertebral fractures does not reach pre-fracture levels until 12 to 18 months post-fracture, while the post-fracture HRQoL decrement associated with hip and vertebral fractures is sustained and remains 13 and 11 % lower respectively, at 18 months.
The study included a follow-up period of 18 months and a large number of participants from several study centres across Australia. The cohort reflects the typical profile of fracture incidence statistics associated with osteoporosis, with three quarters of participants being female and a mean age around 70 years . The cohort was unique in collecting HRQoL data from all fracture sites including separate estimates for hip, vertebral, wrist, ankle and humeral fractures in addition to other fractures grouped. Previous post-fracture HRQoL work has been largely confined to single fracture sites including hip [1, 4, 10–15], vertebral [1, 4, 10, 11, 14, 16], distal forearm [10, 16, 17], proximal humerus [15, 16], pelvic and rib fractures . Prospective cohort studies measuring HRQoL over time after an acute event are not common. Longitudinal studies that have assessed HRQoL following an acute event include stroke , acute myocardial infarction  and total joint replacement . However, the emphasis is more on understanding the predictors of HRQoL rather than quantifying the amount.
In the immediate acute fracture period, quality of life declined by a mean of 51 %, ranging from almost 70 % post-hip fracture to 36 % post-wrist fracture. Notably, at 4 months post-fracture, the decline in HRQoL for ankle fracture was similar to that for hip fracture, suggesting that factors such as mobility that impact on HRQoL are equally relevant in both hip and ankle fractures at 4 months. At 12 months post-fracture, HRQoL for ankle fractures was 8 % lower than the pre-fracture level. The mean decline of all non-hip, non-vertebral fractures at 12 months post-fracture was 6 % lower than pre-fracture level with attainment of the pre-fracture level by 18 months.
Interim analysis of international hip, vertebral and wrist data from the parent ICUROS study also found no significant difference in loss of HRQoL at 4 months between participants with and without prior fracture  and reports that the strongest and most consistent predictor of HRQoL at 4 months post-fracture was baseline HRQoL. Our Australian data confirm this and demonstrate that those with a higher pre-fracture HRQoL experienced a greater decline in HRQoL over the 18 months following fracture.
The decrement in HRQoL did not differ by sex while age (70 + years) conferred only a minor extra decline in HRQoL compared with participants’ aged 50 to 69 years. A previous fracture in the past 5 years was not associated with QALY loss at 12 months. Hospitalisation as a consequence of the acute fracture resulted in an additional 5 % loss of QALYs for all fractures, and 3 % loss for non-hip, non-vertebral fractures. This is probably due to hospitalisation being a potential marker of either fracture severity, or poorer health. The effects of education and income were mixed with respect to direction and fracture type.
At 1 year post-fracture, there was a 22 % loss in QALYs among our Australian cohort of 150 hip fracture participants. Similarly, the Belgian Hip Fracture Study Group, using the SF-36 tool rather than the EQ-5D, reported a 24 % functional decline . QALY loss in the first 12 months after hip fracture was estimated to be 0.47 mainly due to hospital and nursing home stay . In contrast for wrist fracture, our Australian estimate of 0.10 QALY loss at 12 months is higher than the 0.05 reported by the National Osteoporosis Foundation .
There are limited published data on the quality of life decline associated with humeral fractures  despite it being the fourth most common site for fracture among Australians aged 50 years and over (0.4 % per year (p.a). for women and 0.1 % p.a. for men ). In our Australian cohort, the decline in HRQoL was greater than for wrist fracture (QALY: 18 vs 11 % at 12 months) while Hallberg et al. using the SF-36 tool reported that humeral fractures had a similar pattern to that of wrist fracture with less decline compared with vertebral fracture . These Australian data show that the decline in quality of life for humeral fractures was greater in the immediate fracture period than for vertebral fractures (59 vs 47 %, respectively) although, in contrast to vertebral fracture, HRQoL returned to pre-fracture levels by 18 months.
While participants with ankle fracture were on average 3 years younger than those with wrist fracture (63.6 and 66.6 years, respectively), the cumulative loss of QALYs at 12 months was double that of wrist fractures (24 vs 11 %, respectively). The decline in quality of life over 12 months was similar to that following vertebral fracture. However, at 18 months post-fracture, HRQoL had returned to the pre-fracture level.
A potential limitation of this study is that participants were asked for their pre-fracture HRQoL after their fracture. The participant may therefore perceive their pre-fracture status as better than it was, potentially overestimating the loss in HRQoL. However, Marsh et al. have reported that older patients can accurately recall their preoperative health status 6 weeks following hip arthroplasty . Another limitation is the relatively low number of vertebral fractures compared with other fracture sites (92/915; 10 %). In Australia, vertebral fracture patients not involved in a high trauma event do not routinely present at a hospital emergency department. They are more likely to present to a general practitioner, but this may not be within 2 weeks of their fracture, as required for this study. To improve recruitment of people with a vertebral fracture, financial bonuses were offered to site coordinators. As the majority of fracture participants were recruited through attendance at an emergency department, the vertebral fracture cohort is likely to be those who present with a sudden and acute onset of pain. These patients may therefore represent those who experience a large and immediate decline in HRQoL and may not be representative of the general vertebral fragility fracture population. A third limitation is that deaths during the follow-up period were excluded in further analysis, therefore suggesting that our QALY estimates may be conservative.
Quality of life is reported using a generic questionnaire (EQ-5D) which can be expressed as a single index that enables comparison of HRQoL across diseases or populations, for example, in priority setting. The assessment of QALY loss associated with fragility fracture allows the burden of osteoporosis to be compared with that of other diseases and can be used to drive change in health care policy. The reported estimates of HRQoL can also be used in cost-effectiveness analyses of different treatment scenarios, as well as the potential value of directing community health resources towards primary and secondary fracture prevention.
In summary and conclusion, fractures reduce quality of life. These novel data highlight the neglected and significant decline in HRQoL associated with fractures at sites other than the hip, vertebrae and wrist. They also show that the HRQoL for non-hip, non-vertebral fractures does not reach pre-fracture levels until 12 to 18 months post-fracture. On the other hand, the HRQoL decrement associated with hip and vertebral fractures is sustained at 18 months post-fracture. Previous estimates of the full impact on HRQoL associated with fragility fractures are likely to represent an underestimate as the observational period does not generally extend past 12 months and are not inclusive of all fracture sites.