As fracture risk assessment is a basis for treatment decisions, accurate risk assessments on bone mineral density (BMD) reports are important. Over 50 % of sampled BMD reports for Ontarians with fracture histories underestimated fracture risk by a single category. Risk assessments in Ontario may not accurately inform treatment recommendations.
Introduction: The shifting emphasis on fracture risk assessment as a basis for treatment recommendations highlights the importance of ensuring that accurate fracture risk assessments are present on reading specialists' BMD reports. This study seeks to determine the accuracy of fracture risk assessments on a sample of BMD reports from 2008 for individuals with a history of fracture and produced by a broad cross section of Ontario's imaging laboratories.
Methods: Forty-eight BMD reports for individuals with documented history of fragility fracture were collected as part of a cluster randomized trial. To compute fracture risk, risk factors, and BMD T-scores from reports were abstracted using a standardized template and compared to the assessments on the reports. Cohen's kappa was used to score agreement between the research team and the reading specialists.
Results: The weighted kappa was 0.21, indicating agreement to be at the margin of "poor to fair." More than 50 % of the time, reported fracture risks did not reflect fracture history and were therefore underestimated by a single category. Over 30 % of the reports containing a "low" fracture risk assessment were assessed as "moderate" fracture risk by the research team, given fracture history. Over 20 % of the reports with a "moderate" fracture risk were assessed as "high" by the research team, given fracture history.
Conclusions: This study highlights the high prevalence of fracture risk assessments that are underestimated. This has implications in terms of fracture risk categorization that can negatively affect subsequent follow-up care and treatment recommendations.