An indirect method for the estimation of osteoporotic fractures from injury and fracture profiles

Study objective was to develop a valid epidemiological method for the estimation of osteoporotic fracture risk, using administrative databases and accounting for variable baseline risks of injury. Design is the secondary analysis of inpatient and outpatient utilization data. A baseline injury risk was estimated by the incidence of primary utilization of medical services for soft tissue injuries (ICD-9 diagnostic codes 910-929), and the risk profile was compared after normalization with the overall primary utilization rate for fractures (ICD-9 diagnostic codes 800-829). The setting is a county with approximately 100,000 inhabitants in the former East Germany. Participants were all inhabitants of the county who had a physician contact (inpatient or outpatient) during 1987-1988, as well as hospital inpatients for all of Germany in 1989. The number of fractures increased with age, especially in women, when compared to the number of fractures expected from the incidence of soft tissue injury. Similar patterns were identified in hospitalization data from East and West Germany. Estimating the prevalence of osteoporosis directly from certain "osteoporotic" fracture types associated with higher age is potentially biased, since it neglects the underlying risk of injury. Our model distinguished the osteoporotic fracture risk as the excess risk over an expected injury-related fracture risk for a given age and sex, and may allow a more valid quantification of osteoporotic fractures in different populations.