Methods: Fracture risk was assessed using an algorithm developed for the World Health Organization. This estimates an individual's risk of fracture based on age, sex, bone mineral density (BMD), and on boolean indicators for corticosteroid, alcohol, prior fracture, family history of hip fracture, smoking and rheumatoid arthritis status. The female population of England and Wales was categorised by combinations of the determined clinical risk factors in five year age bands. The distribution of the risk of fracture was estimated assuming that BMD was normally distributed. An updated version of the model previously used by the National Institute for Clinical Excellence determined the expected costs and QALYs gained through osteoporosis interventions. As women without prior fracture cannot be identified without cost, the costs for GP consultation and BMD must be included. For each age group and combination of risk factors, three alternative treatment strategies was evaluated to determine the optimal strategy. These alternatives were (1) offer a BMD scan to all women and treat those whose expected cost per QALY is below the cost-effectiveness threshold (2) treat all women without recourse to a BMD scan and (3) offer neither a BMD scan nor treatment. The first option will have the highest sensitivity and specificity for detecting patients who can be treated cost-effectively. The second option has lower specificity but incurs no cost for BMD scanning. The third will have zero sensitivity but no cost implications.

Results: Assuming a cost-effectiveness threshold of £20,000 per QALY, cost-effective identification strategies were found for women aged 70 years and over. These ranged from offering BMD scans to 30% of women without prior fracture at 70 years of age, to immediately initiating treatment in 16% of women without prior fracture at 80 years of age and offering BMD scans to the remainder. Increasing the cost-effectiveness threshold to £30,000 per QALY resulted in cost-effective identification strategies being found for women aged 65 - 69 years.

Conclusions: Selective identification strategies to identify women without prior fracture can be employed cost-effectively in England and Wales.