Modeling changes in health perception following hip fracture

Background and objective: Health perception is an important outcome associated with health-related quality of life. Various correlates of perceived health have been identified, but attempts at multivariate modeling have often failed to mirror the complex system of causal pathways surrounding this concept. The objective of this study was to develop a multivariate model to improve our understanding of how physical, social, and psychological factors interact to influence health perception following a hip fracture. Methods: Patients were interviewed in-hospital during the week following the fracture to obtain information on in-hospital mental status and pre-fracture function, pre-fracture health perception, and pre-fracture social support. A follow-up interview was conducted by telephone to assess function, health perception, and social support 3 months post-fracture. Comorbidities and demographic information were obtained from medical records. Results: This study included 222 hip fracture patients aged 65 + and fracturing a hip between 10 July 1996 and 30 August 1997. Our estimated model was theory-based, developed from existing research identifying the following correlates of health perception: physical function, comorbidities, socioeconomic status, social support, age, and prior health perception. In addition to these correlates, our model was unique in incorporating a variable to assess whether the gap between pre- and post-fracture functioning acted as a separate cause on post-fracture health perception. Initially, the fit between the data and the model was poor, however minor modifications to the model corrected this. While there was a good fit between the data and the estimated model, only about 25% of the variation in both pre- and post-fracture health perceptions were explained. The number of comorbidities and physical function were found to have the strongest influence on health perception. Post-fracture health perceptions were also improved by increased post-fracture social support, higher income/status, and higher pre-fracture health perception. The network of effects indicates that the causal sources of health perception may differ in the pre- and post-fracture periods. This suggests that the causal sources of health perception following major health changes may differ substantially from the causal sources of health perception operating during normal times. Conclusions: The findings suggest that interventions aimed at improving pre-fracture function and post-fracture social support could increase health perception following hip fracture. In interpreting our results, it is important to remember that while we attempted to adjust for measurement error in the analysis, the study may be limited due to the retrospective nature of the questions. Future research should focus on improving the model by including other components of the patient's emotional health. This revised version was published online in June 2006 with corrections to the Cover Date.