This is the first study to assess GQOL in patients with low-energy wrist and hip fractures and to compare the scores with age-and sex-matched controls. The hip fracture patients reported lower GQOL before the fracture occurred compared with controls. Adjusting for known covariates of GQOL decreased these differences substantially, but the differences between the hip fracture group and controls remained significant. However, unadjusted and adjusted GQOL scores before the fracture did not differ between the wrist patients and controls.
Adjusting for well-known predictors of QOLS such as age, sex, education level, marital status, clinical characteristic, and health-focused QOL reduced the differences between the hip patients and controls in our study, but the differences remained significant [48–56]. We expected these adjustments to eliminate or reduce the differences between the hip patients and controls substantially more than what we found. The remaining differences might be explained by more co-morbidity and lower physical function caused by aging and age-related diseases in the hip group, which were not captured by the SF-36.
The variance in QOLS and the three sub-dimensions explained by health-focused QOL was substantial, especially the mental component. A strong association between mental health and GQOL has been reported by others [53, 55–59]. In a meta-analysis of the QOL literature that distinguished between QOL and health status, Smith at al.  found that patients give greater emphasis to mental health than physical functioning when rating GQOL. Our findings seem to be consistent with the meta-analysis by Smith at al. .
Wilson and Cleary  proposed a model to classify different measures of health outcomes. They divided the outcomes on a continuum comprising five levels: biological and physiological factors, symptoms, functioning, general health perception, and overall QOL. Patients' preferences and emotional or psychological factors play important roles at several points in the model and are particularly important in understanding general health perceptions and GQOL. In addition, perceptions of health appear to be more important than objective health in terms of their effects on GQOL . Although we did not include measures of patients' preferences and emotional factors in our analysis, our data seem to coincide with the pattern described by Wilson and Cleary. The associations proposed in their model may explain the strong correlation between the health-focused QOL and GQOL and the weak correlation between clinical fracture characteristics and GQOL in our study. Both Osoba  and Ferrans et al  present adjusted Wilson and Cleary  models, emphasizing the bidirectional relationship between health- focused QOL and GQOL (and the other health outcomes in the model), which is also seen in our study. However health-focused and global-QOL are distinct as health-focused QOL centres on the individual's experience of general state of health, such as physical, social, and mental well-being, while GQOL focuses on the individual's satisfaction with life as a whole [17, 18, 60].
Our study has some limitations, which should be considered when interpreting the findings. The patients were asked to evaluate their "pre-fracture" GQOL after the fracture had occurred. Changes in health, such as experiencing a fracture, might cause a shift in how the patients judged their GQOL (selective reporting bias and response shift) . On the other hand, patients who have experienced a recent change in health are more likely to make accurate responses [5, 16]. Furthermore, have a short time span since events shown be important to report more accurate QOL [62, 63]. The questionnaire was designed with a clear instruction that the patients should think of the period before the fracture, and in most of the patients, GQOL was assessed within the first two weeks after the fracture. It seems unlikely that the patients were unable to recall their GQOL immediately before and at the time of the fracture. Furthermore, the method used to in our study to assess GQOL the week before fracture, seems to be the most realistic and appropriate alternative.
The patients were asked to describe their GQOL at the time before the fracture, whilst health-focused QOL was more specifically restricted to the 4 weeks before the fracture [21, 35, 43, 60]. The restricted time span with regard to health-focused QOL assessment could raise doubts regarding, the prudence of measuring GQOL and health-focused QOL within the same time before the fracture. Studies have shown that patients tend to think of the time before the event regardless of the instructions specifying "the time before" the event (fracture) or "the four weeks before" the event (fracture) [63–65]. Furthermore, both questionnaires were followed by the instruction to relate to the time before the fracture occurred [16, 62, 63].
We chose to use imputation techniques with regard to missing values in the QOLS questionnaire when at least 80% of the items had valid response. Some doubts have been raised regarding this technique, because of the underlying assumptions. However, it should be emphasized that failing to impute missing data also involves making assumptions and may have negative consequences. Patients failing to respond one or more items are then deleted as non-responders in furthur analyses, thereby reducing statistical power and possibly biasing the sample being analyzed .
All patients included in the study were identified at the hospital, which is the only referral centre for orthopaedic trauma in the region. Hence, the external validity of the study should be satisfactory. A high number of the hip fracture patients (n = 271) did not fulfil the inclusion criteria. Closer examination showed that most of these patients were nursing home residents who suffered from dementia, confusion, or severe diseases, and they were older than the participants. Hence, it is likely that the excluded hip fracture patients had more impaired health than those included in the present study and that the results for the hip fracture patients may be generalized only to people residing in their own homes. The patients unwilling to participate in the study were older than the participants were. Younger patients might be more aware of the benefits of participating in a study like this. The older age of the patients who were unwilling to participate might also be related to aging and age-related diseases in this group, and we probably reached the most healthy fracture patients .
The findings in our study are based on fewer participants less in the hip group than in the wrist group, and hip patients are slightly older than wrist patients. Even thought both wrist and hip fractures are strongly associated with objective health factors like osteoporosis and falls, we found that wrist and hip fracture patients are quite different with regard to demographics and clinical variables. However, when comparing wrist fracture patients versus controls and hip fracture patients versus controls with regard to GQOL, known covariates of GQOL like age, sex, education, marital status, clinical variables and health-focused QOL were adjusted for in the multivariate analysis. Such adjustments allows for a more meaningful comparison of GQOL between fracture patients and controls by removing the possible effects of "confounders" (common underlying causes) of GQOL and group membership . Rather than aiming for a study population with "balanced" comparison groups with the same number of participants in each, we included all eligible participants, thus decreasing confidence intervals and increasing statistical power .
Hip fracture patients had a lower GQOL even before the fracture occurred, and they seemed to be less satisfied with life as a whole. GQOL assessment seems to add knowledge to the complexity of the conditions prior to fracture, and decreased GQOL in elderly seem to be an independent associate of low energy hip fracture. Decreased GQOL have been identified as an associate of other diseases and conditions as well . However, our findings suggest that by identifying patients with low GQOL, in addition to other known risk factors for hip fracture, may rise the probability to target preventive health care activities. Preventive programmes might include efforts to help reduce the tendency to fall, improve the patient's diet and help him or her stop smoking, increase physical activity , and promote better GQOL.
It is unknown how low GQOL before a fracture occurs influences rehabilitation after the fracture, and prospective studies are needed to answer this question. This knowledge would help healthcare providers develop and initiate prevention and rehabilitation efforts.