An evaluation of the NICE guidance for the prevention of osteoporotic fragility fractures in postmenopausal women
- First Online:
- Cite this article as:
- Kanis, J.A., McCloskey, E.V., Jonsson, B. et al. Arch Osteoporos (2010) 5: 19. doi:10.1007/s11657-010-0045-5
- 130 Downloads
The National Institute for Health and Clinical Excellence (NICE) in the UK issued guidance based on a health economic assessment of interventions for the primary and secondary prevention of osteoporosis. The recommendations in the guidance are unworkable in clinical practice and the foundation on which they are based is insecure.
The NICE in the UK recently issued final appraisal documents on the health economic assessment of interventions for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. The majority of interventions were considered to be cost-ineffective except at very low T scores for bone mineral density (BMD). Concerns have been raised with respect to the construct and assumptions that populate the model used by NICE and the feasibility of implementing the subsequent guidance.
The application of the NICE guidance to primary care is problematic. Intervention thresholds are based on a complex array that includes the agent to be used, age, T scores and the presence of different categories of risk factors. Alendronate is the first-line treatment, but women who cannot take or tolerate alendronate may have to wait till their T score deteriorates before they qualify for treatment. The guidance takes no account of women with a T score > −2.5 SD, of glucocorticoid-induced disease or of men. Newer interventions, such as ibandronate and zoledronic acid, are not included. The development of guidelines by the National Osteoporosis Guideline Group (NOGG) avoids many of these problems and unlike the NICE guidance, can be used with FRAX®, the WHO-supported fracture risk assessment tool. NOGG provides intervention thresholds based on fracture probabilities computed from clinical risk factors for fracture with or without information on BMD that are readily accessed by primary care physicians for the assessment of all postmenopausal women and men over the age of 50 years. The NICE guidance is based on a health economic assessment of several interventions. The model used to assess cost-effectiveness is based on Gaussian regression functions which were derived from an individual state transition model. Since the source individual state transition model is not available, the Gaussian functions cannot be evaluated. Moreover, neither the internal nor external validity of the model is established, and the model is not accessible for such an evaluation. Although the NICE model incorporates the clinical risk factors (CRFs) used in FRAX, it neglects the impact of CRFs on the death hazards giving estimates of fracture probability that differ from those using FRAX®. The estimates of cost-effectiveness differ from reference models for reasons that relate in part to the model construct and in particular to the assumptions used to populate the model.
The guidance provided by NICE is cumbersome and cannot be readily used in the setting of primary care. The model on which the guidance is based is opaque. The authors do not support the view of NICE that there are no issues which cause it to doubt the validity of the model or that raise justifiable doubts about the appropriateness of the use of the model to inform its guidance.
KeywordsFRAX Economic models Fracture probability Clinical risk factors Intervention thresholds
Body mass index (computed as kg/m2)
Bone mineral density (in this report at the femoral neck measured by dual energy X-ray absorptiometry)
Clinical risk factors
Decision support unit
Final appraisal document
Algorithms that assess the probability of fracture related to any combination of clinical risk factors with or without BMD
Health technology assessment
Incremental cost-effectiveness ratio
National Institute for Health and Clinical Excellence
National Osteoporosis Guideline Group
Quality-adjusted life years
Royal College of Physicians, London
School of Health and Related Research, University of Sheffield, UK
- T score
The deviation in SD units of measured BMD from the mean of the young adult female reference range
World Health Organization
Willingness to pay