Clinical rules don't predict osteoporosis in women

Clinical prediction rules used by doctors to identify patients who will develop Osteoporosis, which take into account various risk factors, are not useful in identifying women likely to develop the bone-thinning condition, new research shows. However, the rules are good at spotting women who are not likely to develop Osteoporosis and probably don't require bone density scans.

From a statistical standpoint, the rules have a low "positive predictive value," but a high "sensitivity," for predicting osteoporosis. This means that if the rules predict osteoporosis will develop, it may not. In contrast, if the rules predict osteoporosis won't develop, then it is a good bet that it won't.

The US Preventive Service Task Force currently recommends routine osteoporosis screening with bone testing for all women 65 years of age or older. For younger postmenopausal women, however, it is unclear exactly who should be screened with such testing.

Clinical prediction rules may offer a way to individualize bone testing in these women, allowing some to avoid X-ray evaluation, lead author Dr. Karen F. Mauck, from the Mayo Clinic in Rochester, Minnesota, and colleagues note.

The new findings, which appear in the Archives of Internal Medicine, are based on a study of 202 women who were evaluated with three clinical prediction rules: the Simple Calculated Osteoporosis Risk Estimation, Osteoporosis Risk Assessment Instrument, and the National Osteoporosis Foundation practice guidelines.

Basically, each of the rules assigns points for known Osteoporosis risk factors and then adds them together - the higher the score, the more likely the disease is to develop. In the present study, the scores were compared with actual bone test results.

The rules detected at least 98 percent of the women who were likely to develop osteoporosis. Unfortunately, they also predicted osteoporosis in a substantial number of women who were unlikely to develop the disease, based on their bone test results.

"Our results suggest that these clinical prediction rules do not perform well as a general screening method to identify postmenopausal women who are more likely to have osteoporosis," the investigators note. "However, (the rules) may be useful in identifying women who need not undergo testing, especially younger postmenopausal women."

SOURCE: Archives of Internal Medicine, March 14, 2005.