Practice Variation in the Diagnosis and Treatment of Osteoporosis: A Case for More Effective Physician Education in Primary Care

Abstract

Bone mineral density (BMD) performed by dual energy x-ray absorptiometry (DEXA) has been used at our institution as a screening test for osteoporosis since 1986. Of 2789 bone densitometry tests done between 1992 and 1996 on women aged 51-75 years, 1743 (62.5%) were ordered by general internists (GIM), endocrinologists (ENDO), rheumatologists (RHEUM), and a metabolic bone disease specialist (MBDS). We compared the percentage of densitometry tests ordered by GIM, ENDO, RHEUM, and MBDS resulting in one of three possible diagnoses (normal, osteopenia, or osteoporosis). Applying the World Health Organization's (WHO) definition of normal (<1 standard deviation [SD] below the mean for young, adult women), osteopenia (1SD-<2.5 SD below the mean), and osteoporosis (2.5 SD below the mean), we found that 34% of patients tested between 1992 and 1996 were osteoporotic, 42% were osteopenic, and 24% had normal bone density results. The rate of osteoporosis diagnosis was highest in the MBDS cohort (chi2 = 9.19, p = 0.002) compared with patients in the other cohorts. To explore trends in management of this condition, a random sample of osteoporotic women aged 51-75 who had densitometry in 1996 (n = 82) was obtained. Review of medical records revealed that 73% were on some form of osteoporosis treatment (bisphosphonate, estrogen, or calcitonin, with or without calcium and vitamin D supplementation). Treatment rates differed significantly, however, by the ordering physician specialty (96% for MBDS, 63% for ENDO, 75% for RHEUM, and 53% for GIM, chi23df = 11.37, p = 0.01). There were no significant differences in selected clinical or demographic characteristics between patients treated by GIM and MBDS. This variation in treatment rates suggests that an opportunity to enhance primary care physicians' recognition and treatment of osteoporosis exists. Making osteoporosis management an educational focus may help narrow differences in practice and improve the effectiveness of a larger number of physicians treating patients with this problem.