• Osteoporosis;
  • Glucocorticoids;
  • Guidelines;
  • Bone densitometry;
  • Randomized controlled trial;
  • Education


  1. Top of page
  2. Abstract


To assess the effectiveness of a multifaceted intervention to improve the management of glucocorticoid-induced osteoporosis (GIOP).


Of 21 rheumatologists, 11 were randomly assigned to a 3-part intervention consisting of a lecture and discussion regarding optimal management of GIOP, a confidential doctor-specific audit regarding management of GIOP, and a reminder mailing including concise pharmacologic recommendations. The remaining 10 rheumatologists received no special education. Patients with rheumatoid arthritis (RA) taking oral glucocorticoids seen in the 2 months after the intervention were followed for 6 months. Medical records were assessed to determine the proportion undergoing bone mineral density testing or receiving pharmacologic interventions for GIOP during the 6 months before and 6 months after the intervention.


There were 373 patients with RA taking oral glucocorticoids whose records were assessed. Patients in both arms of the trial were similar with respect to age, sex, menopausal status, glucocorticoid dosage and duration, duration of RA, disease-modifying antirheumatic drug use, and the proportion with comorbid conditions. At baseline, there was no significant difference between the patients with respect to osteoporosis medication use (intervention 32% versus control 34%) or bone densitometry use (intervention 9% versus control 5%). After the intervention and a 6-month followup period, there were no differences in treatment (intervention 33% versus control 38%) or bone densitometry use (intervention 8% versus control 8%). Adjusting for patient and physician characteristics did not significantly change these results.


A multifaceted intervention for GIOP, including doctor education, practice audit, and treatment suggestions, had no significant benefit on testing or treatment by rheumatologists over a 6-month followup period. Other intervention approaches need to be tested.


  1. Top of page
  2. Abstract

Glucocorticoid-induced osteoporosis (GIOP) represents a major adverse effect for patients with inflammatory diseases (1, 2). Rheumatoid arthritis (RA) and glucocorticoids are independently associated with bone demineralization and clinical osteoporosis (3, 4), as well as fragility fractures (1, 2, 5). Several pharmacologic treatments have been shown to be effective for improving bone density and reducing clinical fractures (6–8). Based on the recognized epidemiology of GIOP and the existence of proven treatment, management guidelines for GIOP have been developed by a committee of the American College of Rheumatology (ACR) and the British Consensus Group on Glucocorticoids (9–11). Despite these guidelines, GIOP goes undiagnosed and untreated in many patients in both the US and Europe (12–16).

The long-range goal of our work is improving management of GIOP. Prior quality improvement literature suggests that traditional continuing medical education does not result in significant behavior change, but interactive, multifaceted programs may be more effective (17, 18). Based on this knowledge, we developed and tested a 3-part intervention for GIOP in a randomized controlled trial.


  1. Top of page
  2. Abstract

Participating rheumatologists and intervention.

All practicing rheumatologists at 1 large academic arthritis practice were invited to participate. To create a balanced study patient population, we stratified rheumatologists based on their patient volume and randomly assigned them to either intervention or control groups. Written informed consent was obtained from all rheumatologists before randomization, and all aspects of the study were approved by the appropriate Institutional Review Board.

The intervention consisted of 3 parts delivered over 6 weeks. Initially, we conducted a 90-minute dinner meeting facilitated by 3 of the authors (DHS, JNK, JSC) that consisted of a lecture on the epidemiology and treatment of GIOP, a review of baseline data on GIOP from the arthritis practice (15), followed by an open discussion of methods for improving GIOP management. The meeting differed from a typical continuing medical education session in that data were presented from the arthritis practice and physicians were engaged about methods for improving management of GIOP. The meeting was attended by all but 1 of the rheumatologists randomized to the intervention group. This rheumatologist was given copies of all the materials discussed at the meeting.

Three weeks after this meeting, each rheumatologist in the intervention arm received a confidential list of their patients with RA who were noted to be taking oral glucocorticoids. We chose to restrict the study population to patients diagnosed with RA because a previous study from our institution showed that this group of patients was not being managed according to the ACR guidelines for GIOP (9, 15). The list of patients contained information regarding whether an individual had undergone bone densitometry and if a medication for osteoporosis was listed in the patient's medical record. These data came from the previous study on GIOP conducted in this same practice (15). Three weeks later, a 1-page reminder sheet entitled “A Concise Guide to GIOP Pharmacotherapy” was sent to intervention doctors (available upon request). Doctors in the control arm received none of these interventions.

Patient population and data collection.

Eight months after sending out the third part of the intervention, we reviewed the billing records of all patients in the rheumatology practice. Patients with at least 1 visit in the 2 months after completion of the intervention with a diagnosis code for RA (International Classification of Diseases code 714) were eligible for inclusion. The charts of all eligible patients were reviewed by a trained research assistant (AMLT). The first visit during the 2 months after completion of the intervention was considered the index visit. Patients were included if all of the following criteria were met: 1) the impression given in the medical record for the index visit noted RA, 2) the note for this visit described use of oral glucocorticoids, 3) the patient was not in an investigational drug trial, and 4) the patient was seen on at least 1 other occasion by the treating rheumatologist in the 6 months after the index visit.

Six months of medical records, beginning with the index visit, were reviewed for all patients who met inclusion criteria. One of the authors (AMLT) abstracted the medical records using a structured review form (available upon request). The form included patient age; sex; menopausal status; duration of RA; rheumatoid factor status; the preparation and dosage of oral glucocorticoids; the use of medications for osteoporosis, such as bisphosphonates, calcitonin, hormone replacement therapy, raloxifene, and calcium and vitamin D supplements; the use of bone densitometry; prescriptions for disease-modifying antirheumatic drugs (DMARDs); prior fractures; and the presence of comorbid medical conditions. We assessed preintervention and postintervention time periods for osteoporosis medication and bone densitometry use. The preintervention period assessed for the use of medications was 6 months. We extended the preintervention assessment period to 5 years for bone densitometry, and all radiology records from the study center were reviewed. Postintervention management was assessed over the 6 months after the intervention.


There were 2 primary endpoints of interest: the use of bone densitometry and prescription medications for osteoporosis in the 6 months after all 3 parts of the intervention had been completed. (We decided to include only prescription medications in the analysis because we were not confident that medical records accurately reflected calcium and vitamin D use.) The proportions of patients with these endpoints were compared for rheumatologists in the intervention arm versus those in the control arm. Comparisons of these proportions were made using chi-square statistics. As well, the effect of the intervention was tested in 2 separate multivariate logistic models (PROC GENMOD in SAS Statistical Software, Cary NC, version 8.0), one for osteoporosis medication use and another for bone densitometry. In these models, assignment to intervention or control was included as an indicator variable; the covariates included the baseline proportion of patients receiving an osteoporosis medication or undergoing bone densitometry; patient age, sex, and menopausal status; mean glucocorticoid dosage; duration of RA; rheumatoid factor status; use of a DMARD; number of comorbid conditions; the physician's sex, years from fellowship, and the number of half-day clinical sessions per week. Because the intervention was directed at rheumatologists and patients were clustered within rheumatologists' practices, we adjusted the analyses for treating rheumatologist by including a term for the doctor in the multivariate model using the generalized estimating equation (GEE). P values < 0.05 were considered statistically significant.

Before the trial, we performed a power calculation, assuming an α of 0.05 with a one-sided test of significance, and an expected absolute improvement in osteoporosis management from 30% to 45% in the intervention group. To attain a power of 88%, we required 175 patients in each arm. After the trial, we assessed the power to rule-out a significant difference based on the methods suggested by Localio et al (19). This method allows one to estimate the confidence limits around the intervention's effect after adjusting for the observed degree of intraphysician correlation clustering in the GEE.


  1. Top of page
  2. Abstract

All 32 rheumatologists practicing at the study site were randomized, but only 21 saw patients that fit the inclusion criteria. There was an imbalance in the number of patients seen in each arm of the trial, however, because 1 rheumatologist with a large practice went on a leave of absence after randomization had occurred (see Table 1). In most other regards, the intervention and control patients were quite similar. The mean age of patients in the intervention arm was 59 years and 60 in the control arm; the majority of patients were women in both arms and a similar proportion was postmenopausal. Disease duration was ∼10 years in both arms and the vast majority of patients in both arms were taking at least 1 DMARD. During the 1-year study period, 6 months before and 6 months postintervention, the mean glucocorticoid dosage was ∼7 mg (prednisone equivalents) and the mean duration of use was 7 months in both arms. Approximately half of all patients had at least 1 comorbid medical condition, and the proportions with a prior fracture, thyroid disease, or tobacco use were not statistically different.

Table 1. Patient and physician characteristics*
  • *

    RA = rheumatoid arthritis; DMARD = disease-modifying antirheumatic drug.

  • Mean glucocorticoid dosage is reported as milligram equivalents of prednisone. The mean dosage and duration include only data from the 1-year study period.

 Age, mean ± SD, years59 ± 1760 ± 16
 Sex, % female8378
 Postmenopausal, % of women6367
 Duration of RA, mean ± SD, years11 ± 910 ± 9
 Rheumatoid factor present, % positive7061
 DMARD use, % yes9187
 Glucocorticoid dosage, mean ± SD, mg7.4 ± 4.66.4 ± 3.8
 Glucocorticoid duration, mean ± SD, months6.4 ± 3.67.4 ± 3.7
 Comorbid conditions, % with at least 15453
 Prior fracture, % yes2028
 Thyroid disease, % yes1514
 Tobacco use, % current106
 Time since fellowship, mean ± SD, years15 ± 1112 ± 9
 Sex, % female3045
 Number of half-day sessions per week, mean ± SD4.9 ± 3.33.7 ± 3.2

The time since fellowship was slightly longer and there were fewer female doctors among rheumatologists in the intervention arm versus the control arm. The number of half-day sessions was slightly more among rheumatologists in the intervention arm.

During the 6 months before the intervention, prescription medications for osteoporosis were used by 32% of patients in the intervention arm and 34% in the control arm (Figure 1). After the intervention, the proportion receiving an osteoporosis medication did not significantly increase in either group and was not statistically different from each other (P = 0.3). In the multivariate regression analyses with prescription osteoporosis medication as the dependent variable, the indicator term for the trial assignment (intervention versus control) was not significant after adjusting for the intradoctor clustering (odds ratio [OR] = 1.0, 95% confidence interval [95% CI] 0.4–2.3, P = 0.9).

thumbnail image

Figure 1. Effect of educational intervention on osteoporosis management. A, Proportion who were prescribed a medication for osteoporosis. B, Proportion who underwent bone densitometry. P values are from chi-square tests. Note that the scales are different for each graph. Black bar = intervention group; white bar = control group.

Download figure to PowerPoint

Similar analyses were run for bone densitometry use. In the 5 years prior to the intervention, <10% of patients underwent testing in either group (Figure 1). After the intervention, the proportion of patients undergoing bone densitometry was not statistically different between the groups (P = 0.9). Because bone densitometry is only recommended once per year even in a high risk population, we also examined the proportion of patients undergoing bone densitometry before or after the intervention; these were similar between groups (P = 0.4) (Figure 1). In the regression analyses with bone densitometry as the dependent variable, the indicator term for the intervention group was not statistically significant in the multivariate analyses that accounted for physician clustering (OR = 0.5, 95% CI 0.2–1.5, P = 0.2).

Table 2 gives a breakdown of osteoporosis medication use before and after the intervention for both the intervention and control groups. There were no statistically significant differences before or after the intervention for any of the specific drugs.

Table 2. Medication use before and after the intervention*
MedicationSix months prior to interventionSix months after intervention
  • *

    Data presented as no. (%). P value is for the unadjusted chi-square test. N = 168 for intervention and 205 for control groups. HRT = hormone replacement therapy.

Alendronate29 (17.3)33 (16.1)0.828 (16.7)39 (19.0)0.6
Risedronate4 (2.4)2 (1.0)0.34 (2.4)3 (1.5)0.5
Pamidronate2 (1.2)1 (0.5)0.43 (1.8)2 (1.0)0.5
Etidronate1 (0.6)1 (0.5)0.90 (0.0)2 (1.0)0.19
HRT16 (9.5)30 (14.6)0.1416 (9.5)29 (14.2)0.18
Raloxifene8 (4.8)5 (2.4)0.28 (4.8)5 (2.4)0.2
Calcitonin5 (3.0)7 (3.4)0.86 (3.6)8 (3.9)0.9
Calcium43 (25.6)69 (33.7)0.0945 (26.8)68 (33.2)0.12
Vitamin D37 (22.0)58 (28.3)0.1740 (23.8)50 (24.4)0.9


  1. Top of page
  2. Abstract

Based on a previous study (15), we developed a multifaceted intervention for rheumatologists hoping to improve management of GIOP. The intervention included an interactive continuing medical education conference, a confidential practice audit, and a reminder about pharmacotherapy for GIOP. The intervention was tested in 1 large academic rheumatology practice in a randomized controlled trial. We found no effect of the intervention over a 6-month observation period. As with our prior study (15), we observed that most patients with RA taking oral glucocorticoids did not receive medications for osteoporosis and did not undergo bone densitometry.

Although carefully conducted reviews have suggested that traditional continuing medical education is not effective at changing doctors' behavior, multifaceted programs that involve an interactive component have been found effective (17, 18). We developed an interactive educational program, coupled it with information about specific patients that might be good candidates for bone densitometry and/or osteoporosis medication, and added a reminder about treatment options. This 3-part intervention was not associated with any improvement in GIOP management. Several potential reasons may explain the lack of improvement. First, discussing GIOP is difficult to fit into a busy office visit during which DMARDs were often the focus. Second, introducing the potential need for a new medication to manage potential adverse effects of long-term oral glucocorticoid use may have been a difficult topic for physicians to introduce. Finally, local inconveniences to scheduling bone densitometry may have also been a factor.

Limitations to our trial design include a relatively short followup, inclusion of only 1 rheumatology practice, reliance on medical records as a source of information, the possibility that medication intolerance limited the use of antiresorptive therapies, and reduced power because of intradoctor correlation. It is possible that a longer followup might have shown some benefits to the intervention, but we expected the most dramatic effect to be earliest. The wait time for bone densitometry can be ∼1 month, but this should not have limited the chance of seeing an intervention effect over the 6-month followup. The use of other rheumatology practices would have allowed us to discuss whether these findings can be generalized. We encourage others to test interventions for GIOP in other settings. The medical record may not always reflect all medications and radiology tests, thus there may have been some under-reporting. However, this misclassification should be similar across intervention and control groups. Although medication intolerance to the bisphosphonates and hormonal therapies is reported, intranasal calcitonin and intravenous pamidronate are well tolerated. We therefore doubt that patients would not be able to tolerate any osteoporosis medication. Furthermore, medication intolerance should have been randomly distributed across the groups, making bias unlikely.

Finally, the power to detect a significant difference in osteoporosis management was less than we had originally estimated because of intradoctor clustering. In regression models using the GEE that adjusts the confidence limits for intradoctor clustering, there was still no significant effect of the intervention. However, the confidence limits around the ORs were wide. Thus, we cannot rule out that there was a difference between control and intervention groups as large as 2.3 times the proportion of osteoporosis medication use in the control group. Similarly, for bone densitometry, the difference between groups could have been as large as 1.5 times the proportion in the control group.

In conclusion, we conducted a randomized controlled trial of a multifaceted intervention for GIOP. This intervention was not associated with any increase in use of osteoporosis medications or bone densitometry over a 6-month postintervention observation period. Based on several potential explanations for these findings, we are considering future interventions that rely on direct-to-patient educational mailings as well as the use of rheumatology nurses to recognize at-risk patients and initiate a diagnostic workup.


  1. Top of page
  2. Abstract
  • 1
    Michel BA, Bloch DA, Wolfe F, Fries JF. Fractures in rheumatoid arthritis: an evaluation of associated risk factors. J Rheumatol 1993; 20: 16669.
  • 2
    Hooyman JR, Melton IIILJ, Nelson AM, O'Fallon WM, Riggs BL. Fractures after rheumatoid arthritis: a population based study. Arthritis Rheum 1984; 27: 135361.
  • 3
    Haugeberg G, Uhlig T, Falch JA, Halse JI, Kvien TE. Bone mineral density and frequency of osteoporosis in female patients with rheumatoid arthritis. Arthritis Rheum 2000; 43: 52230.
  • 4
    Laan RFJ, van Riel PLCM, van de Putte LBA, van Erning LJ, van't Hof MA, Lemmens JA. Low-dose prednisone induces rapid reversible axial bone loss in patients with rheumatoid arthritis: a randomized controlled trial. Ann Intern Med 1993; 119: 9638.
  • 5
    Van Staa TP, Leufkens HGM, Abenhaim L, Zhang B, Cooper C. Use of oral corticosteroids and risk of fractures. J Bone Miner Res 2000; 15: 9931000.
  • 6
    Saag KG, Emkey R, Schnitzer TJ, Brown JP, Hawkins F, Goemaere S, et al, Glucocorticoid-Induced Osteoporosis Intervention Study Group. Alendronate for the prevention and treatment of glucocorticoid-induced osteoporosis. N Engl J Med 1998; 339: 2929.
  • 7
    Adachi J, Cranney A, Goldsmith CH, Bensen WG, Bianchi F, Cividino A, et al. Intermittent cyclic therapy with etidronate in the prevention of corticosteroid induced bone loss. J Rheumatol 1994; 21: 19226.
  • 8
    Cohen S, Levy RM, Keller M, Boling E, Emkey RD, Greenwald M, et al. Risedronate therapy prevents corticosteroid-induced bone loss: a twelve-month, multicenter, randomized, double-blind, placebo-controlled, parallel-group study. Arthritis Rheum 1999; 42: 230918.
  • 9
    American College of Rheumatology Task Force on Oteoporosis Guidelines. Recommendations for the prevention and treatment of glucocorticoid-induced osteoporosis. Arhritis Rheum 1996; 39: 17911801.
  • 10
    American College of Rheumatology Ad Hoc Committee on Glucocorticoid-Induced Osteoporosis. Recommendations for the prevention and treatment of glucocorticoid-induced osteoporosis: 2001 update. Arthritis Rheum 2001; 44: 1496503.
  • 11
    Eastell R, Reid DM, Compston J, Cooper C, Fogelman I, Francis RM, et al. A UK consensus group on management of glucocorticoid-induced osteoporosis: an update. J Intern Med 1998; 244: 27192.
  • 12
    Aagaard EM, Lin P, Modin GW, Lane NE. Prevention of glucocorticoid-induced osteoporosis: provider practice at an urban hospital. Am J Med 1999; 107: 45660.
  • 13
    Yood RA, Harrold LR, Fish L. Prevention of glucocorticoid-induced osteoporosis: experience in a managed care setting. Arch Intern Med 2001; 161: 13227.
  • 14
    Mudano A, Allison J, Hill J, Rothermel T, Saag K. Variations in glucocorticoid-induced osteoporosis prevention in a managed care cohort. J Rheumatol 2001; 28: 12981305.
  • 15
    Solomon DH, Katz JN, Jacobs JP, La Tourette AM, Coblyn J. Management of glucocorticoid-induced osteoporosis in patients with rheumatoid arthritis: rates and predictors of care in an academic rheumatology practice. Arthritis Rheum 2002; 46: 313642.
  • 16
    Haugeberg G, Orstavik RE, Uhlig T, Falch JA, Halse JI, Kvien TK. Bone loss in patients with rheumatoid arthritis: results from a population-based cohort of 366 patients followed up for two years. Arthritis Rheum 2002; 46: 17208.
  • 17
    Davis D, O'Brien MAT, Freemantle N, Wolf FM, Mamanian P, Taylor-Vaisey A. Impact of formal continuing medical education: do conferences, workshops, rounds, and other traditional continuing education activities change physician behavior or health outcomes? JAMA 1999; 282: 86774.
  • 18
    Oxman AD, Thomson MA, David DA, Haynes RB. No magic bullets: a systematic review of 102 trials of interventions to improve professional practice. Can Med Assoc J 1995; 153: 142331.
  • 19
    Localio AR, Berlin JA, Ten Have TR, Kimmel SE. Adjustments for center in multicenter studies: an overview. Ann Intern Med 2001; 135: 11223.