There is little understanding of short-term ("imminent") risk of bone breaks. Clinical tools for calculating long-term (10-year) fracture risk exist, but preliminary work suggests that they don't work to asses if a patient is likely to have a fracture at high short-term (1 and 2-year).
We aim to find out which patient characteristics could be used to predict if a patient is likely to suffer a short-term fracture and to then create an equation to predict short-term fracture risk.
To do this, we will study electronic medical records of thousands of patients that had recent fractures, or that are having anti-fracture treatment.
At the end of the study, we will provide clinicians with a new tool to target patients at high risk of short-term fracture/s, so they can receive appropriate treatment to prevent them.
OBJECTIVES
1. To identify predictors of 1- and 2-year fracture risk (IFR) amongst recently fractured or new users of oral bisphosphonates
2. To combine key risk factors to derive IFR prediction tools
METHODS
- Data sources: CPRD GOLD linked to ONS mortality and HES inpatient
- Participants: two cohorts: 1.newly diagnosed with fracture, and 2.new users of oral bisphosphonate/s in 2000-2018, with 1+ year of data available.
- Measurements: Known predictors of IFR (ISAC17_128R) will be obtained.
- Outcomes: fracture of hip, major osteoporotic, any (excluding skull and digits)
- Statistics: Lasso regression will be used to identify key predictors of IFR, to be included in prediction formulae. Discrimination and calibration will be assessed using AUC and observed/expected plots respectively. All analyses will be stratified by cohort.
Primary outcomes: Major fracture: hip, spine, non-hip non-spine, hip/humerus/wrist
Daniel Prieto-Alhambra - Chief Investigator - University of Oxford
Daniel Prieto-Alhambra - Corresponding Applicant - University of Oxford
Antonella Delmestri - Collaborator - University of Oxford
Emese Toth - Collaborator - UCB Biopharma SRL - Belgium Headquarters
Sara Khalid - Collaborator - University of Oxford
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation