Bone Vibration Analysis as a Novel Screening Tool for Long Bone Fractures

The aim of this study is to reduce the number of X-ray scans taken to detect fractures, by developing a bone fracture screening system. When assessing a bone injury, doctors need to decide whether the injury has resulted in a fracture or a sprain so that they can provide appropriate treatments. The current way to differentiate between these is by an X-ray scan. In 2011, the 46,000 children attending Sheffield Children’s Hospital Emergency Department had 10,400 X-rays, mostly to help diagnose bone fractures. Roughly half the X-ray scans taken indicate that the injury is sprain. Unnecessary X-ray scan means raising costs and exposing patients to ionising radiation. Vibration analysis is a well-established technology for condition monitoring for defect detection in industries however; its use in the medical field is still evolving. In bone vibration analysis, periodic or aperiodic oscillations or oscillating signals are introduced, and subsequent responses are recorded followed by using mathematical methods to reach a conclusion. In this study, a computer-controlled mechanism induces a mild vibration and successive responses are recorded via a piezoelectric sensor. To demonstrate the method's feasibility, a preliminary study was carried out on five blocks of wood of different density, with the same dimensions. The tests indicated a significant reduction in the blocks' vibration frequency following their fracture. After obtaining National Health Services (NHS) Research Approval, appropriate number of bone vibration responses was recorded from adults’ wrists and children’s wrists and ankles who attended local hospitals following wrist or ankle injuries. Suitable signal processing and pattern recognition techniques were developed on the basis of vibration responses from bones at various stages to interpret the recorded signals. Data were acquired from healthy participants at a local school which were compared with the data acquired from hospital participants to verify the methods. Currently, this study differentiates around 80% of the injuries accurately. Additionally, both the data acquisition program and the device have been modified to improve the developed procedures. This study made some promising discoveries and the resulting techniques can be used for further explorations.