Perimenopausal bone density screening--will it help prevent osteoporosis?

OBJECTIVE To estimate the potential efficacy and cost-effectiveness of hormone replacement therapy (HRT) in the prevention of osteoporotic fractures, with and without the assistance of perimenopausal bone mineral density (BMD) screening. METHOD Residual lifetime fracture experience of a hypothetical cohort of 100,000. British women aged 45 years at baseline, modelled using prevailing UK mortality and fracture rates. Appropriate fracture risk gradients were used to estimate the distribution of future fragility fractures (distal forearm, proximal femur and clinically diagnosed vertebral fractures) according to quarters of baseline bone density measured at fracture specific sites. We assumed that 72% of the population could be contacted and would attend for HRT counselling, with or without bone densitometry, that 10 years of continuous HRT use would reduce fracture rates by 50%, and that compliance with HRT might vary between 10% and 50%. Universal recommendation of HRT was compared to selective treatment protocols offering HRT to those women whose BMD fell below the 25th, 50th or 75th percentile of BMD at the lumbar spine, femoral neck or distal forearm, measured either singly or in combination. RESULTS The proportion of future fractures averted was closely related to compliance with therapy, but for any given level of compliance, universal treatment always achieved the greatest reduction in fractures. If compliance was 10% universal HRT was also the most cost-effective strategy, but if compliance was higher or if the unit cost of HRT increased, selective strategies were often more cost-effective. The sensitivity of BMD screening in identifying women at risk of future fracture could be increased by relaxing the BMD decision threshold, or expanding the number of skeletal sites measured, or both. However increments in test sensitivity were always accompanied by reductions in specificity. CONCLUSIONS If BMD measurement does not influence compliance, then universal treatment with HRT is likely to prevent more fractures, at a similar or lower average cost per fracture averted, than selective therapy. However, if BMD screening leads to increased compliance, or if more expensive forms of treatment were used, then our model suggests a favourable impact of screening on the numbers and/or net cost of fractures prevented.