Osteoporosis International

, Volume 16, Issue 9, pp 1079–1085

Improving osteoporosis management in patients with fractures


  • S. L. Johnson
    • Endocrinology SectionMcGuire Veterans Affairs Medical Center
    • Virginia Commonwealth University School of Pharmacy
  • V. I. Petkov
    • McGuire Research Institute Inc.
    • Department of Preventive Medicine and Community HealthVirginia Commonwealth University School of Medicine
  • M. I. Williams
    • Endocrinology SectionMcGuire Veterans Affairs Medical Center
    • Virginia Commonwealth University School of Pharmacy
  • P. S. Via
    • Endocrinology SectionMcGuire Veterans Affairs Medical Center
    • Endocrinology SectionMcGuire Veterans Affairs Medical Center
    • Department of Preventive Medicine and Community HealthVirginia Commonwealth University School of Medicine
    • Department of Internal MedicineVirginia Commonwealth University School of Medicine
    • McGuire VAMC (111P)
Original Article

DOI: 10.1007/s00198-004-1814-z

Cite this article as:
Johnson, S.L., Petkov, V.I., Williams, M.I. et al. Osteoporos Int (2005) 16: 1079. doi:10.1007/s00198-004-1814-z


A history of fracture is an independent risk factor for future fractures, but patients who have sustained a fracture are rarely evaluated for osteoporosis (OP). The objective of this study was to determine if a simple intervention in a general orthopedic clinic would lead to more fracture patients receiving evaluation and treatment for OP. Patients with a history of fracture visiting a weekly orthopedic clinic during a 6-month intervention period were educated about OP, and a bone mineral density (BMD) test was offered. The number of BMD tests performed and other OP-specific actions taken as a result of the intervention were compared with a 6-month pre-intervention period. The prevalence of OP in those who underwent BMD testing was examined. In the pre-intervention period, only 12.7% (16 of 126) had a BMD test as compared with 62.5% of the 136 intervention-period subjects (odds ratio [OR] 11.5, 95% confidence interval [CI] 6.1, 21.4). Based on BMD test results, 11.9% of the pre-intervention patients, and 41.9% of the intervention patients received OP-specific recommendations (OR 5.3, 95% CI 2.8, 10.1). The intervention led to more patients being treated for low bone mass (9.5% vs 23.5%); OR 2.9, 95% CI 1.4, 5.9. Low bone mass was common among all types of fracture patients: 20% had osteoporosis and 41%, osteopenia. BMD testing in patients with fractures should identify those at risk for future fractures, leading to appropriate treatment.


Bone densitometryFracturesOsteoporosis


A history of a fragility fracture has been reported consistently to be an independent risk factor for future fractures [14]. In a cohort study of 672 healthy postmenopausal women, a previous fragility fracture showed the strongest (odds ratio [OR] 3.3, confidence interval [CI] 1.8, 5.7) association with any prospective fragility fracture, compared with six other independent predictors [age, hip bone mineral density test (BMD), past falls, left grip strength, maternal history of fragility fracture, and low physical activity] [1]. In 9,704 female participants age 65 or older, prevalent vertebral deformities were associated with a fivefold increased risk for new vertebral fractures, 2.8-fold risk for hip fractures and 1.9-fold risk for any non-vertebral fracture [2]. Although limited in number, studies in men have demonstrated a similar magnitude of risk [5]. Any fracture, including traumatic, is associated with an increased risk of a fragility fracture. Fractures of the tibia and ankle early in life were an independent predictor of osteoporotic fractures later in both women and men [6]. In a cross-sectional study of postmenopausal women, those who had premenopausal fractures between ages 20 and 50 were more likely to experience postmenopausal fragility fractures (74% increase) [7]. Klotzbuecher et al. [8] performed a meta-analysis of studies discussing the above association. Their pooled estimates showed a twofold [1.8, 2.1] increased risk for any subsequent fracture in peri-menopausal or postmenopausal women with any prior fracture. For women of all ages and men, the pooled risk estimate was 2.2 [1.9, 2.6].

Despite the increased risk for new fractures, few patients who have sustained a fracture are evaluated for osteoporosis or low bone mass [911]. After a distal forearm fracture, only 18% of postmenopausal women (45 and older) received any pharmacologic osteoporosis (OP) intervention during the next 12 months [9]. Even more worrisome is the fact that older women who are at greatest risk are less likely to receive any intervention [10]. Remarkably, patients who have sustained a hip fracture are unlikely to undergo osteoporosis evaluation. A medical records review of hip fracture patients enrolled in four Midwestern-US health systems revealed dual energy X-ray absorptiometry (DXA) use in 12–24%, calcium and vitamin D supplementation in 1–25%, and antiresorptive treatment in 7–37% [11]. The situation was even worse in a tertiary teaching hospital in the Northeast, where a DXA scan was ordered for 3% of hip fracture patients. In the same study, only 4% received calcium, 2% vitamin D, and 1% alendronate [12]. Despite fewer barriers for OP management in the Veterans Affairs health-care system, veterans who sustained a hip fracture rarely received adequate evaluation and treatment (12% BMD test, 12% calcium and multivitamins, 0% bisphosphonates) [13].

Fractures, especially of the hip, cause increased morbidity and mortality, which ultimately lead to increased health-care costs [1415]. To reduce the cost and complications associated with fractures, populations at greatest risk need to be identified and evaluated [16]. Veterans enrolled in a general orthopedic clinic have a varied history of traumatic and non-traumatic fractures. Many may be at risk for osteoporosis. The objective of this prospective study was to determine if a simple intervention in a general orthopedic clinic would lead to more fracture patients receiving evaluation and treatment for osteoporosis.


After approval by a local IRB, patient recruitment was conducted weekly in a general orthopedic clinic at a single Veterans Affairs Medical Center (VAMC) over a 6-month period (October 2002 to March 2003). Subjects who had sustained fractures at any time were identified through review of the computerized medical records. The chart review focused on each patient’s problem list (a section which includes both active and inactive medical conditions), orthopedic clinic consultation requests and/or follow-up notes, and imaging reports. Patients who had a previous BMD test or were currently undergoing osteoporosis evaluation or treatment were excluded. Due to weight limitations of the bone densitometer (Hologic QDR Delphi, Waltham, MA, USA) subjects weighing more than 300 pounds were also excluded.

Recruiters worked closely with the orthopedic clinic staff to identify and contact fracture patients on the date of their visit to the clinic. A study team member met with each of the identified fracture patients, described the study, obtained consent and, using a pre-structured format, explained the importance of diagnosing and treating osteoporosis. The interview provided a definition of OP, explanation of potential risk factors, and a brief description of the BMD test. This information was reinforced by issuing an educational leaflet. A BMD test by DXA was offered as a part of the intervention. Those patients who consented to the study underwent BMD testing. An endocrinologist certified by the International Society of Clinical Densitometry (ISCD) reviewed and interpreted the results in the context of the patient’s medical history and made recommendations. Full evaluation was recommended for patients with osteoporosis by BMD, or osteopenia and other risk factors such as glucocorticoid use or previous fragility fracture. Conservative treatment (calcium and vitamin D supplementation, and weight-bearing exercise) was suggested for patients with osteopenia. No recommendation was made for a normal BMD test. The physician’s BMD note was placed in the patient’s electronic medical record. Additionally, if the BMD test showed low bone mass, the report was sent electronically to the patient’s primary care provider (PCP) for implementation of OP-specific recommendations. After 7 months, the records of patients who underwent BMD testing were examined to determine the number of recommendations followed. This time interval was chosen to account for the lag time of patients’ follow-up with their PCP or osteoporosis clinic (veterans are usually seen by their PCP every 6 months, and the waiting time for the osteoporosis clinic is 2–3 months).

The outcomes measured were the proportions of fracture patients who had a BMD test, received OP-specific recommendations, and those who followed the recommendations. A comparison was made with fracture patients (identified using the same search criteria as the study patients) evaluated in the same orthopedic clinic during a 6-month pre-intervention period (October 2001 to March 2002).

Statistical analysis included descriptive characteristics, proportional distributions, Chi-square test for comparing proportions, and odds ratio (OR) along with 95% confidence intervals. The prevalence of low bone mass in patients who underwent DXA testing was reported by site as well as a combination of several sites (hip only, Looker et al. [17]; hip and spine, ISCD guidelines from 2001 [18], and hip, spine and total forearm, modified ISCD guidelines, 2003 [19]). Osteoporosis was defined as a T -score ≤−2.5 and osteopenia as a T -score between −1 and −2.4. The Hologic densitometer utilizes gender- and race-specific normative databases for T -score calculation. The NHANES III normative database was used for hip T -score and the manufacturer’s database for spine and forearm. An attempt was made to find independent clinical predictors of low BMD in fracture patients. Each patient’s age, weight, body mass index (BMI, kg/m2), glucocorticoid treatment ≥ 7.5 mg daily for 3 months, opiate use ≥ 12 months, alcohol consumption (yes/no; as recorded on patient’s problem list and last PCP note), smoking history (yes/no per problem list and last PCP note) and type of fracture (traumatic vs minimal trauma, recent vs historical, and fracture location) were recorded. Univariate analysis was used to assess the relationship between bone density at the lumbar spine, femoral neck, total hip and distal forearm and osteoporosis risk factors. Multiple regression was utilized to build predictive models. Analysis was done using SPSS version 10.0. Significance was determined using a two-sided alpha level of 0.05.


During a 6-month study intervention period 136 fracture patients were seen at the orthopedic clinic; 103 (75.7%) agreed to participate and signed an informed consent. Eighteen (13.2%) declined participation, and 15 (11.1%) were not contacted due to scheduling conflicts. Study outcomes in these subjects were compared with 126 fracture patients from the pre-intervention period. Characteristics of the study groups are shown in Table 1. There was no significant difference in any of the examined characteristics between the two groups. The mean age was 60.5 years for the pre-intervention group and 59.1 years for the intervention group. The majority of the study subjects were Caucasian males.
Table 1

Characteristics of pre-intervention and intervention study groups




N =126

N =136

Age (mean, SD)

60.5 (15.2)

59.1 (12.7)

Weight in kg (mean, SD)

88.2 (22.6)

89.3 (21.0)

Gender (% males)



Race (% Caucasians)



BMD tests performed

Only 16 (12.7%) subjects from the pre-intervention period had a BMD test as compared with 85 (62.5%) during the intervention period ( p <0.0001) (Fig. 1).
Fig. 1

Study outcomes [bone mineral density ( BMD) tests performed, osteoporosis/osteopenia ( OP) specific recommendations for management, and specific osteoporosis/osteopenia action taken] in the pre-intervention and intervention groups

Fracture patients educated about OP and provided with an opportunity for a BMD test were 11.5 times more likely to have the test than controls (95% CI [6.1, 21.4]). Very few fracture subjects had a BMD test in the pre-intervention period, resulting in rare diagnosis of osteoporosis (7.1%, 9 of 126) and osteopenia (4.8%, 6 of 126). In the intervention group, these numbers were 12.5% (17 of 136) and 25.7% (35 of 136), respectively. The proportional distribution of BMD test results (osteoporosis: T -score ≤−2.5, osteopenia: T -score between −1 and −2.4, and normal: T -score >−1 by 2001 ISCD recommendations [18]) were 56% (9 of 16), 38% (6 of 16), and 6% (1 of 16), respectively, in the pre-intervention group and 20% (17 of 85), 41.2% (35 of 85), and 38.8% (33 of 85), respectively, in the intervention group.

Osteoporosis-specific recommendations

During the intervention period, 41.9% (57 of 136) were identified for OP/osteopenia treatment as compared with 11.9% (15 of 126) from the control period ( p <0.0001). The crude OR was 5.3, 95% CI [2.8, 10.1]. Table 2 shows the distribution of pre-intervention and post-intervention subjects according to recommendations made based on bone density and other clinical factors. Out of 16 patients who had a BMD test in the pre-intervention period, 56% received recommendations for full OP evaluation or referral to an osteoporosis clinic, 38% for conservative treatment (calcium, vitamin D, and weight-bearing exercise), and 6% had no recommendations due to a normal BMD test. These proportions were 31%, 36%, and 33%, respectively, in 85 intervention subjects who had a DXA scan. Fig. 2 shows the frequency distribution of the study groups by recommendations made.
Table 2

Distribution of study patients according to recommendations made by an endocrinologist

Recommendations made



N (%)

N (%)

Full evaluationa

9 (7.1)

26 (19.1)

Conservative treatmentb

6 (4.8)

31 (22.8)

No recommendations (normal BMD)

1 (0.8)

28 (20.6)

No BMD test

110 (87.3)

51 (37.5)

aReferral to metabolic bone clinic

bRecommended calcium, vitamin D /multivitamin, weight-bearing exercise, and follow-up BMD in 2 years

Fig. 2

Frequency distribution of study subjects by group and recommendations made

OP-specific actions taken

The intervention led to more patients being treated for low bone mass (23.5% vs 9.5%, p =0.002), OR 2.9, 95% CI [1.4, 5.9], illustrated by Fig. 1. Pre-intervention patients who had recommendations were somewhat more likely to have them followed up than were study patients: 77.8% of pre-intervention patients for whom full osteoporosis evaluation was recommended actually received specific OP measures/treatment as compared with 65.4% ( p -value not significant) in the intervention group. For conservative treatment recommendations, the percentages were 83.3% and 48.4%, respectively ( p =0.03).

Determinants of bone mass in the intervention patients

At the time of this report, 85 patients had DXA tests out of 103 who were scheduled (17.5% no-show rate for BMD test). Of potential predictors of low BMD at the examined sites, only weight/BMI and age were significantly associated. The type of fracture (minimal trauma vs traumatic), fracture occurrence (recent vs old), fracture after age of 40, and other risk factors (race, smoking, alcohol, opiate, and glucocorticoid use) were not associated with significantly lower BMD. For example, 22.2% of the patients who had minimal trauma fracture were osteoporotic as compared with 21.2% of subjects with traumatic fracture. In multiple regression models, only age and weight were independent predictors of low BMD, accounting for 0.38 of the variance. Specific characteristics of the tested patients are reported in Table 3. They were predominantly Caucasian men, with a mean age of 58.9 years. Twenty two percent had recent fractures (defined as a fracture for which the patient was currently being treated in the orthopedic clinic). Minimal trauma fractures (falling from a standing position) were described for 31.8% and traumatic for 40.0%. For 28.2% of the subjects the nature of the fracture could not be determined from the medical records. Fracture locations were as follows: 47.1% lower extremity, 30.6% upper extremity, 7.1% hip, 5.9% spine, and 9.4% other. The majority of the subjects experienced fracture after age 40.
Table 3

Characteristics of intervention patients who had a bone mineral density (BMD) test


Mean (SD)


(min. to max.)

Age (years)

58.9 (12.9)

23 to 83

Weight (kg)

90.8 (18.4)

46 to 136

Body mass index (kg/m2)

29.4 (5.4)

18.4 to 42.8

T -score lumbar spine

−0.3 (1.8)

−3.7 to 5.9

T -score total hip

−0.6 (1.1)

−3.2 to 1.9

T -score femoral neck

−1.1 (1.0)

−4.0 to 1.6

T -score trochanter

−0.5 (1.2)

−3.2 to 2.3

T -score total forearm

−0.9 (1.7)

−6.8 to 3.4


Number ( N)

Percentage (%)

Gender (males)



Race (Caucasians)



Alcohol use






Opiate use



Glucocorticoid use



Low bone mass was common among the 85 study patients who underwent BMD testing. Fig. 3 shows the percentage distribution of patients according to T -score at various BMD sites. Assessing spine and hip only, 20% had osteoporosis and 41% osteopenia. If forearm DXA is added, 28.2% had at least one T -score ≤ −2.5 and 39% had osteopenia (Fig. 4). Thus, almost 75% of the fracture patients needed at least some intervention for low bone mass.
Fig. 3

Percentage distribution of intervention subjects according to T -score at the lumbar spine ( LS), total hip ( TH), femoral neck ( FN), trochanter ( Troch), total forearm ( TFA), and any site

Fig. 4

Distribution of intervention subjects among the osteoporosis group ( T -score ≤−2.5), osteopenic group ( T -score between −1 and −2.4), and normal group ( T -score >−1.0) according to sites used for classification ( hip femoral neck, trochanter, intertrochanter, and total hip; LS lumbar spine; TFA total forearm)


Osteoporosis is recognized as a public health problem. At the same time it is a widely under-diagnosed and under-treated condition. In the past, a similar situation was observed regarding cardiovascular events and the diagnosis and treatment of hypertension and hyperlipidemia. Much research and clinical work were necessary to improve cardiovascular prevention. Apparently, a similar effort is needed to improve osteoporosis detection and management [16].

In our study, we tested a simple intervention (an OP educational leaflet and an offer for a BMD test by DXA) in fracture patients evaluated in an orthopedic clinic. The intervention dramatically increased the number of patients who had BMD testing (crude OR 11.5 [6.1, 21.4]), received specific recommendations (crude OR 5.3 [2.8, 10.1]), and initiated osteoporosis/osteopenia treatment (crude OR 2.9 [1.4, 5.9]) as compared with the pre-intervention period. Review of the literature reveals very few studies whose objective was to test different approaches for improving osteoporosis management. In a similar pre-intervention, post-intervention design, Hawker et al. focused their intervention only on patients with recent fragility fractures [20]. The orthopedic surgeon informed the patients about the nature of their fracture and osteoporosis as a possible cause, and provided them with a letter to their PCP. The intervention significantly increased follow-up with a physician (adjusted OR 1.85, p =0.02) and recommendations for BMD testing (adjusted OR 5.22, p <0.0001) but not OP-specific recommendations (adjusted OR 2.07, p =0.07) as compared with a pre-intervention period. Our intervention suggested somewhat better results. The reasons could be several. Our intervention provided a direct link for BMD testing. The BMD reports were placed in each patient’s electronic medical record, and, in addition, a report was e-mailed to the PCP. Finally, the nature of the Veterans Affairs health-care system may play a role because of the provision of primary, secondary and tertiary care within the same facility. In general, there are few costs to veterans and few barriers to having several clinic visits or test appointments in one day. A randomized controlled trial in women reported results similar to ours [21]. Women who sustained an osteoporotic fracture and had not had BMD evaluation and treatment were assigned to no intervention, electronic patient-specific clinical guidelines to the PCP, or the guidelines plus a letter to the patient. Compared with 6% in the group with no intervention, the first intervention resulted in 51% of the women receiving a BMD test or OP treatment. An observational study of an osteoporosis disease management program (consisting of provider and community education as well as BMD testing of women aged 55 and older enrolled in a health-care system) resulted in a $7.8 million savings from the reduction in hip fractures [22].

Along with the present study in the orthopedic clinic, we have launched a randomized clinical trial with the objective of testing interventions for improving osteoporosis diagnosis and treatment in primary care practices. The interventions utilize the Osteoporosis Self-Assessment Tool (OST) as a screening for BMD test referral [23]. Our preliminary results showed that education and electronic clinical reminders based on OST dramatically increased osteoporosis diagnosis in men as compared with provider education only or education plus an OST desktop calculator [24]. For the orthopedic clinic patients, we believe it is crucial that the PCPs be notified of the DXA results.

Although a significantly greater number of patients initiated osteoporosis/ osteopenia treatment during the intervention period, the proportion of those who followed up with the recommendations is less than during the pre-intervention period (OP/osteopenia: 65.4%/ 48.4% vs 77.8%/ 83.3%). A potential explanation is that the study team ordered the BMD, not the patient’s PCP, perhaps making PCPs less likely to comply with recommendations. Furthermore, studies have suggested that more PCP education about osteoporosis management is needed [25, 26].

In contrast to studies that focused only on recent fragility fractures, we decided to target patients with any type of fracture. This decision was based on reports that any fracture increases the risk for future fractures [68]. In fact, we did not observe a significant difference in BMD between patients with minimal trauma and traumatic fractures. Similar proportions of patients were diagnosed with osteoporosis in both groups. This suggests that BMD evaluation may be indicated for all patients with a history of a fracture. However, this needs further investigation, because our study did not have adequate power to detect differences in BMD or OP/osteopenia rates.

We have tried to compare the prevalence of low bone mass in our study patients with widely cited prevalence estimates. Looker et al. [17] have estimated the prevalence of OP in US men 50 years and older to be 3–6%, using any T -score ≤ −2.5 in the hip (femoral neck, trochanter, intertrochanter and total hip). Applying the same criteria, we found osteoporosis in 16.7%. Using spine in addition to hip BMD, the overall prevalence of osteoporosis was 20%. When forearm DXA was added, the prevalence increased to 28%, supporting the recommendation of Vallarta-Ast et al. that forearm DXA should be routinely used, particularly in older men [27]. In the present study, we used total forearm rather than the 1/3 radius site as the latest ISCD guidelines suggest [19]. The reasons are as follows: The study was completed and analyzed prior to ISCD guidelines from 2003. The Hologic BMD report gives total forearm by default. Melton et al. reported that the total forearm was the best predictor of osteoporotic fractures in men [28]. Recent studies exploring different forearm regions for osteoporosis diagnosis did not show superiority of the 1/3 radius [29, 30]. In our patients, at least two-thirds had osteopenia, which is usually treated with calcium and vitamin D. In addition, a future BMD test is suggested for patients with osteopenia.

Our study has limitations. The study design of pre-testing and post-testing is not as strong as a randomized controlled trial. Some patients with a history significant for fractures may not have been included due to lack of information in their electronic medical records. However, there is no reason to believe that the number of these patients is different in the two study groups. Finally, there might be fracture misclassification due to incompleteness of medical records. The study might have limited generalizability due to the specifics of the Veterans Affairs health-care system. However, a similar intervention (education about OP) among school teachers resulted in a threefold increase of BMD tests ordered at the next PCP visit [31].

In summary, low bone mass is common in a predominantly male population with a current or previous fracture. The addition of forearm DXA measurements identified more patients with low bone mass who may have otherwise been undiagnosed. Performing DXA tests on patients with fractures should identify those at risk for future fractures, many of which may be prevented with appropriate treatment.


We thank the staff of the McGuire VAMC Wednesday Orthopedic Clinic for their help and support. This study was supported in part by an unrestricted grant from Procter & Gamble

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2004