Editor—Hodson and Marsh say that ultra-sound may be a promising screening or diagnostic test for detecting osteoporosis.1 Clarification is necessary before their findings can be accepted. Their population, selected on the basis of perceived risk, has a higher risk (16%) than would normally be expected in primary care (4%).2 Interpretation of diagnostic data is better presented as positive likelihood ratios, allowing estimates of post-test probability.3 Ultrasound alone gives a post-test probability of less than 50% and is less impressive if a prior of 4% is used (table).
Table 1.
Test accuracy of risk factors, ultrasound examination, and their combination. Adapted from Hodson and March1
| Assessment | Positive likelihood ratio (95% Cl) | Post-test probability of osteoporosis (95% Cl) if prior probability 16%1 | Post-test probability of osteoporosis (95% Cl) if prior probability 4%2 |
|---|---|---|---|
| Risk factors present | 1.2 (0.9 to 1.5) | 18.6 (14.8 to 23.1) | 4.7 (3.6 to 6.0) |
| Ultrasound T score <-1.7 | 4.8 (2.8 to 6.3) | 44.9 (35.1 to 55.1) | 14.8 (10.4 to 20.8) |
| Risk factors and ultrasound T score <-1.7 combined | 1.5 (1.2 to 1.7) | 22.0 (19.3 to 25.1) | 5.7 (4.9 to 6.7) |
Hodson and Marsh have reported the combined diagnostic value of nine risk factors. This is likely to be misleading. When osteoporosis is the target disorder, individual risk factors have very different diagnostic test properties—a history of steroid use has a positive likelihood ratio of 12, age greater than 70 has a positive likelihood ratio of 5.5.2,4 The diagnostic value of each separate risk factor or combined risk factors in each person is more informative. Furthermore, the number of risk factors cited in the text (n = 169) and in the table (n = 113) do not add up.
A receiver operating characteristics curve for different cut-off points of ultrasound could be constructed. A more informative interpretation of ultrasound can then be made as a screening or diagnostic test.3
The introduction of ultrasound as either a screening or diagnostic test should not be considered until randomised trials have shown effectiveness and cost effectiveness in prevention of fractures and improved quality of life.
Competing interests: None declared.
References
- 1.Hodson J, Marsh J. Quantitative ultrasound and risk factor enquiry as predictors of postmenopausal osteoporosis: comparative study in primary care. BMJ 2003;326: 1250. (7 June.) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Jarvik JG, Deyo RA. Diagnostic evaluation of low back pain with emphasis on imaging. Ann Intern Med 2002;137: 586-97. [DOI] [PubMed] [Google Scholar]
- 3.Black ER, Bordley D, Tape TG, Panzer RJ, eds. Diagnostic strategies for common medical problems. Philadelphia: American College of Physicians, 1999.
- 4.Deyo RA, Rainville J, Kent DL. What can the history and physical examination tell us about low back pain? JAMA 1992;268: 760-5. [PubMed] [Google Scholar]