| Summary: | The introductory section of this thesis outlines the importance of obtaining accurate and reliable skeletal fracture pattern data. The difficulties commonly encountered by injury epidemiologists in obtaining fracture data are discussed, and a review of the currently available literature is presented. The substantial variation in reported fracture incidence in the literature is further investigated in Section 2. Using a cohort of fracture patients from the Edinburgh population, this study aimed to investigate the effect of employing two different methods of obtaining numerator fracture data on the number and patterns of fractures reported. The results illustrate the marked difference that exists between numerator data obtained from emergency department (ED) sources and data obtained from the orthopaedic trauma unit (OTU). The positive predictive value (PPV) of a correct ED outpatient fracture diagnosis was only 74%, meaning that for every four fractures diagnosed and recorded in the ED, one of these will have been coded incorrectly. Fractures of the carpus, proximal tibia, proximal radius, calcaneus, talus and midfoot were miscoded more frequently than the average, while those of the clavicle, proximal humerus, metacarpus, metatarsus and distal radius were coded with greater accuracy. These results suggest that epidemiological fracture research using ED numerator data is likely to overestimate adult fracture incidence when compared to studies obtaining data from orthopaedic sources. Section 3 examines the range and pattern of acute fractures of the upper limb, lower limb and pelvis that occur in Edinburgh adults. The results of a 12 month longitudinal observational study are presented, including a detailed review of the patterns of fracture types encountered and the modes of injury involved in fracture occurrence. Where possible a comparison is made with historical data from the Edinburgh, as well as the existing fracture epidemiology literature. The socioeconomic status of patients is an important factor in the distribution of morbidity and disability in many areas of medicine. Section 4 of this thesis explores the association between deprivation, as measured by the Scottish Index of Multiple Deprivation (SIMD) and the incidence of fractures in Edinburgh adults. Logistic regression analysis was used to control for the influence of a number of variables known to have an effect on fracture patterns, such as the age and gender of the patient, the injury mode involved and the fracture type. The results show that increasing socioeconomic deprivation correlates with increasing fracture incidence, even after controlling for confounding variables. A stronger correlation exists in men than in women, and the fracture incidence seen in the most affluent decile is only 50% of that seen in the most deprived group. A stronger correlation exists for certain fracture types, namely fractures of the metacarpus, distal radius, proximal humerus and ankle. The final section contains a discussion of the relevance of the findings presented in Sections 2, 3 and 4 and how they relate to the existing fracture epidemiology literature. This section also considers which fracture types should now be considered as fragility fractures. The significant strengths and considerable limitations of this thesis, including issues related to the numerator, the denominator, causation and multiplicity in epidemiological fracture research, are discussed in detail. Suggestions for future work are presented.
|
|---|
