Gallagher TC, Geling O, Comite F. Missed Opportunities for Prevention of Osteoporotic Fracture. Arch Intern Med. 2002;162(4):450-456. doi:10.1001/archinte.162.4.450
Copyright 2002 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2002
Osteoporotic fracture is a growing public health problem burden to society. Despite the importance of physician practices in preventing it, relatively little is known about the osteoporosis-related practices of US physicians.
A total of 1500 female members of a Connecticut independent practice association model health plan (aged 40-69 years) were surveyed to identify women's receipt of osteoporosis-related services (eg, prevention counseling, bone mineral density [BMD] testing, and communication about treatment options). These findings were compared with recommendations of the US Preventive Services Task Force and the National Osteoporosis Foundation. We received 1007 completed questionnaires, for a response rate of 69%.
Only 49% of the sample reported that a health care provider ever discussed osteoporosis with them. In multivariate analyses, women with multiple risk factors were not more likely than other women to have been counseled about osteoporosis and its prevention, although those with an osteopenia/osteoporosis diagnosis were. In contrast to National Osteoporosis Foundation recommendations, only a small minority of high-risk women (12%-34%) had their BMD tested. Although most women with an osteopenia/osteoporosis diagnosis reported receiving information on estrogen replacement therapy, calcium, and weight-bearing exercise, fewer reported receiving information on pharmaceutical alternatives to estrogen (33%) and vitamin D (20%).
The main trigger to physician counseling of women about osteoporosis and its prevention is an osteopenia/osteoporosis diagnosis. Women with multiple risk factors for osteoporosis are not being identified for preventive counseling interventions or BMD testing.
OSTEOPOROSIS IS a growing public health problem resulting in an enormous burden to society. Unless comprehensive programs of prevention and treatment are initiated, the direct medical costs for osteoporotic fractures alone are expected to more than double (from an estimated annual $14 billion in 19951) during the next 50 years.2,3 Osteoporotic fractures also have significant social costs (chronic pain, disability, increased dependence on others, need for nursing home care, deformity, and death), and it is estimated that more than 243 000 older US adults have an osteoporosis-related disability (unpublished data, 2001).
Greater utilization of available interventions to prevent, diagnose, and treat osteoporosis can dramatically reduce the rate of osteoporotic fracture and subsequent sequelae. Several associations, including the US Preventive Services Task Force (USPSTF) and the National Osteoporosis Foundation (NOF), have published evidence-based recommendations for physician practices with regard to the prevention, detection, and treatment of osteoporosis in women.4,5 However, relatively little is known about physician practices in reference to these recommendations. We surveyed a large sample of female members of a managed care network to identify women's receipt of osteoporosis-related services and compared the results with recommendations of the USPSTF and the NOF. Because this study was conducted a year before publication of the NOF recommendations in 1999, this study should not be viewed as an assessment of the impact of these recommendations on clinical practice. Instead, it provides a benchmark of physicians' osteoporosis-related practices before dissemination of the NOF recommendations.
Osteoporosis-related services examined here include physician-patient discussion of osteoporosis and preventive measures, receipt of bone mineral density (BMD) testing, and physician-patient communication about treatment options (among women with osteopenia/osteoporosis only). The NOF recommends that physicians counsel all women on the risk factors for osteoporosis. Both the USPSTF and the NOF recommend that physicians counsel all women about universal preventive measures related to fracture risk (eg, take in adequate amounts of calcium and vitamin D, perform regular weight-bearing exercise, avoid tobacco smoking, and keep alcohol intake moderate), and the USPSTF recommends that all perimenopausal and postmenopausal women be counseled about estrogen replacement therapy (ERT) for osteoporosis prevention. The NOF identifies high-risk groups that should be referred for testing: all women 65 years or older, regardless of additional risk factors; all postmenopausal women younger than 65 years who have 1 or more additional risk factors for osteoporosis (besides menopause); postmenopausal women who present with fractures; women who are considering therapy for osteoporosis (if BMD testing would facilitate the decision); and women who have undergone hormone replacement therapy for prolonged periods. The NOF recommends that postmenopausal women with low BMD or established osteoporosis be offered a pharmaceutical therapy (estrogen, alendronate, calcitonin, and raloxifene were Food and Drug Administration approved at the time of the study), and that all patients being considered for drug treatment of osteoporosis also be counseled on the importance of calcium, vitamin D, and exercise as part of any pharmacologic treatment program.
Data are from a self-administered mail survey of a randomly selected sample of 1500 female members (aged 40-69 years) of a Connecticut managed care network in early 1998. Members of this independent practice association model health plan are required to select a primary care physician from a list of participating physicians at the time of enrollment; are encouraged but not required to rely on their primary care physician to coordinate their care; and do not need a referral from their primary provider to see a specialist. Participants in the plan's Medicaid and Medicare risk plans were excluded. The initial survey mailing was followed with a postcard reminder to the entire sample, and 2 follow-up mailings to nonrespondents. Thirty-one questionnaires were undeliverable or could not be completed owing to respondent illness or lack of English proficiency. One thousand seven completed questionnaires were received, for an effective response rate of 69% (1007/[1500 − 31]); 1004 surveys arrived in time to be entered into the database.
We identified the percentage of women reporting that a health professional ever talked with them about osteoporosis and selected preventive measures. Each respondent was asked whether a doctor or other health professional ever talked with her about "osteoporosis or brittle bones," calcium in her diet, weight-bearing exercise, "hormone replacement therapy to prevent or treat osteoporosis," or ever advised her to quit smoking (smokers only).
We report also the prevalence of osteoporosis discussions by women's stage of menopause. Stage of menopause was coded as premenopause, perimenopause, or postmenopause (surgical or nonsurgical), and was determined on the basis of survey questions on length of time since last menstrual period, recent change in cycle regularity, and hysterectomy. (Gallagher et al6 provide more detail on study measures.)
Using multivariate analysis, we determined whether women with selected osteoporosis risk factors5 were more likely to have discussed osteoporosis and (in separate analyses) specific preventive measures with their physicians. Risk factors as measured in this survey include early menopause (menopause before age 45 or bilateral ovariectomy); low body mass index (BMI) (<21 kg/m2, calculated using self-reported height and weight); current cigarette smoking status; personal history of fracture (ever fractured wrist, spine, or hip); family history of fracture (wrist, spine, or hip fracture in parent, sibling, or grandparent); and long-term use of corticosteroids (self-report of having used steroid medicines at least 1 month of the year for at least 5 years). All measures were based on survey questions. Diagnosed osteopenia/osteoporosis was measured by affirmative response to a question asking whether bones were "below normal density" on BMD testing or, "Have you ever been told by a doctor that you had osteoporosis or brittle bones?"
Other variables included in the multiviarate analyses predicting osteoporosis-related discussions were demographic characteristics of women (age, education, household income), menopause-related characteristics (stage/type of menopause, menopausal symptoms, medical consult for menopausal symptoms), self-perceived health status, and provider characteristics. Respondents were asked to indicate whether they had "any of these symptoms related to menopause" (hot flashes, difficulty sleeping, bladder problems, poor memory, vaginal dryness, irritability/mood swings), and to rate the severity of their experience with each symptom on an 8-point scale, ranging from "never" to "severe." A total symptom score was created by summing the score on these 6 items (Cronbach α = .79), and symptom experience was categorized as none or slight (0-6), mild (7-13), moderate (14-20), or severe (21+). Medical consult for menopausal symptoms was measured by a question asking whether the respondent ever telephoned or visited a doctor or other health care provider for any of the above symptoms. Measures of provider characteristics included sex and specialty of a woman's primary care physician and the pattern of physician(s) that a woman used for regular care (family practitioner or internist and no obstetrician/gynecologist [ob/gyn]; ob/gyn but no family practitioner/internist; family practitioner or internist and ob/gyn).
We considered all of the above measures for entry into the regression models predicting discussion of osteoporosis and (in separate models) individual preventive measures. The stepwise method was used to progressively enter variables into the models, and the fits of the resulting nested models were compared using likelihood ratio tests. For clinical risk factors such as smoking and low BMI, we tested separately models that included the individual risk factors and models that included multiple risk factors. In examining the effects of potentially correlated factors (eg, osteoporosis risk factors and diagnosed osteopenia/osteoporosis), we also tried fitting models that included these factors separately. The resulting nonnested models were compared using the Akaike Information Criterion. The models discussed are the most parsimonious models chosen, using the above procedures.
We identified the percentage of women in the entire sample and the percentage of high-risk women (as outlined in the NOF recommendations) reporting BMD testing. Respondents were provided a description of BMD testing and asked whether they ever had a BMD test. High-risk women were identified through survey questions. We did not have a measure of "postmenopausal women who present with fracture" and approximated this with "postmenopausal women with a lifetime fracture history." "Women who are considering therapy for osteoporosis" was measured with a question asking whether respondent had ever considered taking hormone replacement therapy "to treat or prevent bone thinning, bone loss, or osteoporosis." Use of hormone replacement therapy for prolonged periods was measured by self-report of hormone replacement therapy use for 5 or more years.
We identified the percentage of women with self-reported osteopenia (low BMD on BMD testing) or osteoporosis ("Have you ever been told by a doctor that you had osteoporosis or brittle bones?") who reported receiving information about recommended treatment options. Respondents were asked whether a doctor or other health professional ever talked with them about "medicines or supplements to prevent or treat osteoporosis or brittle bones," and if so to indicate which of the following (ERT, raloxifene, alendronate, calcitonin, calcium supplements, vitamin D supplements) were discussed.
Most respondents were white and non-Hispanic with a fairly high household income (Table 1). Almost half were in their forties, 36% in their fifties, and 15% in their sixties. Most (58%) viewed themselves as in excellent or very good health, and very few (8%) as in fair or poor health. Almost one third (28%) had a family history of fracture, and 8% a personal history of fracture. Only 8% of the sample reported osteopenia (low BMD on BMD testing) or physician diagnosis of osteoporosis. Almost all had a primary care provider, and 26% saw a female primary care provider. Fifty-nine percent utilized both a family practitioner or internist and an ob/gyn for their regular care.
The rate of receipt of osteoporosis-related practices reported by women in this sample varied according to the type of intervention, with very low BMD testing rates, "middling" rates of counseling about osteoporosis and its prevention, and high rates of communication about treatment options among those with osteopenia/osteoporosis (Table 2). Only 49% of women in the sample reported that a health care provider ever discussed osteoporosis with them. The overall BMD testing rate in the sample was quite low (9%), and only somewhat higher (12%-34%) in the high-risk groups. Most women with osteopenia or osteoporosis reported receiving information about various treatment options (ERT, calcium, weight-bearing exercise), but fewer reported communication about pharmaceutical alternatives to ERT (33%) or vitamin D (20%). Only 2% reported receiving information about the full range of treatment options.
As outlined in Table 3, the percentage of women reporting having ever discussed osteoporosis with a health care provider was very low among premenopausal women (28%) and increased with each stage of menopause, but was still less than universal (63%) among postmenopausal women. Discussion of specific preventive measures was also lowest among premenopausal women (21%-44%, with the exception of a high rate for smoking-cessation counseling among smokers), somewhat higher among perimenopausal women, and still less than universal among postmenopausal women (57%-63%).
In multivariate analyses predicting an osteoporosis discussion with a health care professional (Table 4), women with single or multiple risk factors for osteoporosis were not more likely to have ever discussed osteoporosis with a health care provider, although those with osteopenia or osteoporosis were (odds ratio, 27.4; 95% confidence interval, 6.46-116.55). Also more likely to report an osteoporosis discussion were women who were white, in their 50s, postmenopausal, in very good to excellent health, or who consulted a health care provider regarding menopausal symptoms. Respondent education, menopausal symptoms, and provider characteristics were not related to the likelihood of an osteoporosis discussion.
In multivariate analyses (not shown) separately predicting discussions of specific preventive measures (dietary calcium, weight-bearing exercise, ERT to prevent osteoporosis), results were very similar to those predicting a general osteoporosis discussion. In general, women with an osteopenia/osteoporosis diagnosis were more likely to report having discussed each of these measures with a health professional (odds ratio range, 2.15-5.04), while those with single or multiple risk factors were not.
The rate of osteoporosis discussions found here (49%) represents a striking improvement since 1991, when only 25% of women aged 45 to 75 years reported having ever discussed osteoporosis with a physician.7 Counseling rates were high among postmenopausal women, suggesting particular awareness of the importance of osteoporosis prevention in the postmenopausal period. These findings are consistent with other studies indicating that women are most likely to be targeted for osteoporosis interventions in the postmenopausal period or from age 50 years onward.8- 10
However, counseling of premenopausal and perimenopausal women about osteoporosis and its prevention could be dramatically increased: only 28% and 43% of these groups, respectively, reported an osteoporosis discussion. A significant portion of the age-related decrease in bone mass in women occurs before menopause,11,12 and many premenopausal and perimenopausal women may have low bone mass. Preventive measures such as ensuring adequate intake of calcium and vitamin D may help to maintain bone that has been acquired and counteract the process of age-related bone loss that occurs before menopause.13
It is surprising that only 8% of this sample report having been identified as having osteopenia or osteoporosis, possibly warranting intervention. Based on the epidemiology of osteopenia/osteoporosis, one would expect to see a much higher rate, as the combined prevalence of osteopenia and osteoporosis among non-Hispanic white women 50 years and older is estimated at 50% to 68%.14 The low rate of diagnosed osteopenia/osteoporosis observed here is especially notable because the sample is a relatively advantaged one with generally excellent access to medical care. Most women in the sample utilized 2 physicians for regular care (a family practitioner or internist plus an ob/gyn); their managed care plan imposed few constraints on access to specialty care; and the sample members lived in a geographic area anchored by a prestigious academic medical center with a plethora of well-trained physicians and modern health services. However, the low rate of diagnosed osteopenia/osteoporosis observed here is consistent with the low rate of BMD testing in this sample (9%) and other research samples indicating very low rates of BMD testing and diagnosis of osteopenia/osteoporosis in women.15- 20
The lack of targeting of women with multiple osteoporosis risk factors for prevention counseling and the low BMD testing rates in this sample indicate that high-risk women are not being identified early enough in the disease to benefit from preventive measures. Other research also indicates that women with multiple risk factors do not perceive themselves as being at risk for osteoporosis, are not more likely than those at lower risk to worry about osteoporosis, and are not more likely to utilize preventive measures.21- 25 An osteopenia or osteoporosis diagnosis, according to our findings and those of other studies, seems to be the cue to osteoporosis interventions; women are most likely to discuss osteoporosis with a health care professional and implement preventive measures after an osteoporosis diagnosis, and are most likely to receive osteoporosis medications after a fracture.22,24,25
Interestingly, women seeing a female primary care physician and those utilizing an ob/gyn in addition to an internist or family practitioner for their regular care were not more likely to report having been counseled about osteoporosis and its prevention. This contrasts with a long line of research indicating that women seeing female providers (if internists) or ob/gyns, and those seeing 2 providers for regular care (family practitioner or internist plus ob/gyn) are more likely to receive female-specific clinical preventive services.26- 31 Only a handful of studies have specifically explored practice variations in osteoporosis care: female providers are more likely to refer women for BMD testing,15 and gynecologists are more likely to treat for osteoporosis,32,33 particularly with ERT.33
Many of the gaps in osteoporosis-related practices identified here seem to be in areas where the clinical interventions, scientific knowledge, or practice recommendations are very new. For example, low counseling rates of women before menopause about osteoporosis prevention may be related to the relatively weak state of scientific evidence on the efficacy of preventive measures early in the life cycle. Low BMD testing rates among high-risk women may be related to the relative newness of BMD testing as a diagnostic technology and the recency of published detailed clinical recommendations outlining who should be tested. The low rates of communication reported among women with osteopenia/osteoporosis about pharmaceutical alternatives to ERT and about vitamin D may be related to the very recent introduction of some of these medications (raloxifene having been approved only several months before this study was implemented) and the newness of scientific evidence on the efficacy of vitamin D in reducing the risk of osteoporotic fracture.
Greater diffusion of BMD testing and osteoporosis practice guidelines, including dissemination of available scientific evidence on the efficacy of current osteoporosis interventions (including universal preventive measures) may help to address the gaps in practice identified here. New clinical interventions may also help to increase the diagnosis and treatment rates of osteoporosis, including the availability of convenient methods to measure BMD and expansion of the repertoire of pharmacologic treatment options well beyond ERT. The cost of central (axial) skeleton bone densitometry devices is decreasing, which should increase their utilization, and newer, less expensive devices that measure peripheral bone density are becoming available. Since the time of this study, several other bisphosphonates have been approved for osteoporosis treatment, and anabolic agents that stimulate new bone formation34 will become available.
However, there may be significant barriers to osteoporosis prevention counseling, diagnosis, and treatment that may not quickly self-resolve. A national survey of primary care physicians would provide us with useful information on current osteoporosis-related practices in a large, representative sample of physicians and allow us to identify variations in practice by organizational and provider characteristics. This can be followed with surveys and focus groups to identify the barriers to more proactive osteoporosis-related practices. General barriers might include attitudes that osteoporosis is of lower priority than other medical conditions or insufficient knowledge of osteoporosis and the appropriate use of available interventions (ie, "when to treat?"). Barriers to osteoporosis prevention counseling may include competing demands on physician time, lack of reimbursement to physicians for this activity, and questions about the efficacy of preventive measures. Barriers to greater utilization of BMD testing may include limited reimbursement of such by managed care and insurance companies, organizational policies that restrict the use of BMD testing, or controversy over the cost-effectiveness of osteoporosis diagnosis and intervention strategies that rely heavily on BMD testing. Potential barriers to greater use of osteoporosis-related pharmaceutical agents might include policies and practices of managed care organizations designed to limit the prescribing of these medications, limited reimbursement by insurance policies, and patient compliance.
Physician lack of awareness or acceptance of current osteoporosis practice recommendations might also be a barrier to more widespread implementation of recommended practices. It is important to identify physicians' awareness, acceptance, and use of osteoporosis practice recommendations to determine the efficacy of these recommendations in changing current practices. More health services research on the effectiveness and cost-effectiveness of recommended practices should also help guide future practices. At present, all of the necessary data are not available, and some guidelines rely heavily on the opinion of experts. This research will become even more important in guiding practices as the number of available osteoporosis clinical interventions multiplies.
We used women's self-reports to measure the delivery of osteoporosis clinical services. Women may not always "register" the delivery of information on health conditions by their physician or recall this information in a health survey, so we may have underestimated the delivery of services. Our sample was limited to women with employment-based health insurance, and so low-income and older women are underrepresented. However, our estimates are similar to those obtained in a recent national survey, where 61% of women 45 years or older reported having discussed osteoporosis with their physician.35 In support of generalizability, the study was conducted in an independent practice association model managed care organization, and the study participants received care from an almost open-ended list of providers throughout the state of Connecticut. Using physician-patient communication about osteoporosis as an indicator of women's awareness of osteoporosis and preventive measures may underestimate such awareness because much of this information is being disseminated through the popular media, particularly in female-specific media such as women's magazines.
There is room for improvement in current osteoporosis-related practices to more closely approximate those recommended by current practice guidelines. This may require a change from viewing osteoporosis as an illness that begins at the point of fractures to one that should be actively assessed and treated before fractures occur. Much work remains to be done to address the growing and preventable burden of osteoporotic fracture in society.
Accepted for publication July 2, 2001.
This study was supported in part by a grant from the University of Illinois Research Board, Champaign (Dr Gallagher), and an unrestricted educational grant from Eli Lilly & Co, Indianapolis, Ind (Dr Comite). M.D. Health Plan, North Haven, Conn, provided material support.
This study was presented in part at the World Congress on Osteoporosis, Chicago, Ill, June 16, 2000.
Corresponding author and reprints: Florence Comite, MD, Yale University School of Medicine, 40 Temple St, Suite 7H, New Haven, CT 06510.