Racial disparities in osteoporosis prevention in a managed care population. (Original Article).

Background: Osteoporosis in black women may result in increased disability, longer hospital stays, and higher mortality compared with white women. However, it is unknown whether osteoporosis treatment or bone mineral density (BMD) measurement is different in these women, particularly in those at highest risk.

Methods: To examine differences and determinants of osteoporosis preventive interventions among white and black women in a large regional health maintenance organization, women 50 years of age and older were surveyed (n = 8,909) to determine their receipt of BMD testing and medical therapies for osteoporosis prevention.

Results: After adjusting for potential confounders, black women had two- to threefold lower odds of BMD test or osteoporosis prescription treatment. Even among women with a previous fracture, blacks still had a significantly lower likelihood of both BMD testing and prescription therapy.

Conclusion: Compared with whites, black women reported significantly less BMD testing and prescription and nonprescription osteoporosis therapy. This disparity was not fully explained by other demographic or risk factor differences.

**********

Key Points

* Black women reported significantly less osteoporosis prevention than white women.

* Even among women with a previous fracture, blacks reported significantly fewer osteoporosis interventions than whites.

* Disparities in osteoporosis prevention among black and white women were not fully explained by demographic or risk factor differences.

The incidence of osteoporosis and related fractures among lack women is reported to be approximately half that of white women. (1,2) This finding has been attributed to differences in bone geometry (3) and higher peak bone mass in black women. (4,5) Despite a lower prevalence of osteoporosis, fractures still occur commonly among black women, although at a slightly later age. In comparison with white women, these fractures result in increased disability, longer hospital stays, and higher mortality. (2,6,7)

To our knowledge, there have been no prior comparative studies among different ethnic groups of nonhormonal osteoporosis treatment or bone mass measurement receipt. To address this knowledge gap, we compared the self-reported prevalence of osteoporosis risk factors and care in white and black women aged 50 years and older. The women we studied were enrollees of a regional managed care organization in the state of Alabama Motivated by known racial disparities in care received for other chronic disorders, (8-10) we investigated differences in osteoporosis preventive care in a group of women receiving health care through a large regional managed care organization. In addition, we conducted subgroup analysis of race differences among women at highest osteoporosis risk, namely, those respondents who reported they had sustained a previous fracture.

Methods

In April 2000, a two-page osteoporosis questionnaire was mailed to all women aged 50 or older in Alabama who received their health care coverage through a large regional managed care organization (n = 29,775 members). Women were asked to return the survey in a stamped, self-addressed envelope. A $10 gift certificate to a large national retail store was offered as an incentive to return the survey. Because this survey was conducted principally as part of a needs assessment for the managed care organization, additional reminders or second mailings were not sent to nonrespondents.

The survey included the validated Simple Calculated Osteoporosis Risk Estimation evaluation, (11-14) which contains questions on age, race, weight, fracture history, and use of estrogen. The survey also contained questions on measures of osteoporosis care, including 1) bone mineral density (BMD) testing within 2 years; 2) past and current use of estrogen replacement therapy (ERT), bisphosphonates, calcitonin, and raloxifene; and 3) past and current use of nonprescription therapies potentially beneficial for osteoporosis such as supplemental calcium, vitamin D, and a multivitamin. Additional information was collected about menopausal status, hysterectomy, oophorectomy, and other osteoporosis risk factors including past and current use of glucocorticoids, smoking, and exercise.

Using health maintenance organization (HMO) administrative data, we obtained the participant's type of insurance plan. Plan members had either a Medicare+Choice or other insurance plan through the HMO. Using 1990 U.S. Census data, we determined the median household income of the participant's ZIP code as a measure of socioeconomic status. In addition, we used the county of residence to assign a validated rural-urban residential continuum code. (15,16) On the basis of the relative geographic skewness of respondents, these residency continuum codes were later dichotomized as either metropolitan (Beale Codes 0-3) or nonmetropolitan (Beale Codes 4-9). We excluded women with unclear race, unclear age, and age younger than 50 years.

To determine whether our response rate was a representative sample, we compared age, geography of residence, and insurance characteristics of respondents with nonrespondents. To evaluate the precision of our sampling estimate, we also examined the margin of error with Cochran's sampling technique. (17) Finally, we sought to estimate the extent to which nonresponse bias may affect any findings of racial disparities. To do this, we conducted sensitivity analyses for varying proportions of nonrespondent blacks that received treatment.

Descriptive analyses examined osteoporosis care among black and white women. Categorical variables were analyzed using the [chi square] statistic and continuous variables were compared by two-tailed t test. We performed a modified stepwise logistic regression, forcing the clinically important covariates of age, sex, and weight to remain in all models. The regression algorithm required a bivariate P value of 0.25 to enter the model and a P value of 0.15 to remain in the model. Additional multivariable analyses were conducted on the subset of participants with a history of any fracture and those with a specific fracture of the hip, wrist, or rib since age 45 (sites most likely to be accurately reported and to represent osteoporotic fractures). One-way interactions between race and fractures and between race and age were examined in the major models to exclude effect modification by race. Goodness of fit and model calibration were assessed using the Hosmer-Lemeshow and c statistics, respectively. (18,19) All an alyses were conducted with SAS software (SAS Institute, Inc., Cary, NC).

Results

A total of 8,909 United Health Care members responded to the survey, yielding a 30% overall response rate. From administrative data, there were only small differences between respondents and nonrespondents. Women who responded were slightly older (68.7 [+ or -] 10.1 yr) than those who did not respond (67.9 [+ or -] 11.3 yr) and were more likely to have the HMO's Medicare product (72 versus 63%, P < 0.0001). Respondents also lived in ZIP code areas with a higher median household income ($26,705 [+ or -] $10,178) than nonrespondents ($24,800 [+ or -] $10,417).

Using Cochran's sampling technique (17) of an available population of 29,775, our sample yielded a margin of error of 1% for the main effect. We were not able to examine the differences between respondents and nonrespondents by ethnicity because the HMO does not routinely collect racial or ethnic information. The small proportion of respondents who reported their race as other than black or white or did not provide their race (n = 308 [3.5%]) were excluded from further analyses because of insufficient statistical power to analyze trends. Women who reported their age as <50 or who did not provide their age were also excluded (n = 98 [1.1%]).

The demographic characteristics of the participants stratified by ethnicity are shown in Table 1. Compared with white women, the black respondents (22%) were significantly older, reported greater body weight, had a significantly lower median household income based on ZIP code area, and were more likely to participate in the HMO's Medicare product.

More than 95% of all the women were postmenopausal and had seen a doctor within the past 2 years. White women were more likely to report having had hysterectomy, yet a higher proportion of black women reported that they had undergone oophorectomy. Blacks reported a lower proportion of prednisone and cigarette use than whites.

Overall, 37% of all women reported a history of fracture. Of these women who reported a fracture, 40% reported a fracture of the hip, rib, or wrist after age 45. Whites reported significantly higher proportions of rib and wrist fractures than blacks. However, the proportions of self-reported hip fractures after age 45 did not differ by ethnicity.

Black participants had approximately two-thirds lower odds of receiving BMD testing and all forms of prescription osteoporosis therapy (Table 2). Differences by race were most notable for the use of bisphosphonate medications (such as alendronate and risedronate). Among nonprescription therapies, whites reported significantly higher odds of using calcium and vitamin D or multivitamins.

After multivariable adjustment for confounding factors such as age, weight, median household income, menopausal status, smoking status, use of glucocorticoids, having seen a physician in the last 2 years, and history of fracture (Table 3), the association between ethnicity and receipt of prescription therapies was largely unchanged. Namely, black women still had nearly 50% lower odds of receiving BMD testing than white women.

Women with greater body weight, lower median household income (as determined by ZIP code area), and current smokers also had lower odds of reporting receipt of both BMD testing and all forms of treatment. In contrast, women who were receiving glucocorticoids or had seen their physician in the past 2 years had higher odds of receiving testing and treatment. In the models for BMD testing and any treatment, there was no interaction between race and age or race and previous fractures.

Women who reported having had an oophorectomy had three times higher odds of receiving ERT (odds ratio [OR], 3.2; 95% confidence interval [CI], 2.9-3.6) but lower odds of taking raloxifene (OR, 0.6; 95% CI, 0.4-0.8). Among nonprescription therapies, black women still had significantly lower odds of reporting over-the-counter calcium (OR, 0.45; 95% CI, 0.4-0.5) and vitamin D/multivitamin usage (OR, 0.7; 95% CI, 0.6-0.8), although the magnitude of the effect was attenuated somewhat by multivariable analyses (data not shown).

Even in women with a previous fracture, a significantly lower proportion of black women received BMD testing (n = 68 [14%] versus 713 [28%] for whites, P < 0.001) or any prescription therapy (n = 191 [38%] versus 1,713 [67%] for whites, P < 0.001). After adjustment for potential confounders, the likelihood of black women to receive osteoporosis testing or treatments was still 50 to 65% lower than that for white women (Fig. I). Restricting the analysis to only women who reported a hip, wrist, or rib fracture after age 45 (n = 1,071), the effect persisted, with black women still showing consistently lower odds of receiving testing (OR, 0.52; 95% CI, 0.32-0.84) or any prescription therapies (OR, 0.37; 95% CI, 0.24-0.56) (data not shown). In terms of nonprescription therapies, black women who had previously fractured bones had significantly lower odds than white women of reporting use of calcium and vitamin D or multivitamin. For our sensitivity analysis to quantify the effect of nonresponse bias, we assumed that the nonrespondents' racial distribution (78% white) and fracture rate (37%) were the same as that of the respondents. Specifically, we assumed that black nonrespondents had a fracture rate similar to that of whites (37%) and that white nonrespondents had a rate of treatment (67%) similar to that of respondents. Given these assumptions, over 73% of black nonrespondents with fractures (nearly twice the treatment rate reported by the black respondents) would need to have received antiosteoporosis treatment for our main result to become nonsignificant.

Discussion

We examined self-reported osteoporosis risk factors and care in a large, statewide, managed care population. In our sample, black women reported significantly lower use of both prescription and nonprescription antiosteoporotic therapies and less BMD testing than white women. These effects were noted even among a high-risk subset of women who had previously experienced a fracture, and persisted after adjustment for confounding factors. To our knowledge, no other community-based study has previously reported these findings.

Multiple reasons are likely to contribute to black women receiving less osteoporosis care. Black women in our study were less likely to have seen a physician in the past 2 years, and less frequent physician visits are directly associated with a lower rate of treatment and bone mass measurement. Other factors such as menopausal status and surgical oophorectomy were also associated with differing levels of osteoporosis testing and treatment. Black women in our sample were older than white women, but the self-report of menopause was higher among whites. If this difference were true and not a reporting error, whites would have been more likely to receive ERT than blacks on the basis of their more common hormonal deficiency. However, when these and other important cofactors were included in multivariable analyses, ethnicity remained an independent predictor of difference in care receipt.

Katz et al (20) found that black Medicare beneficiaries with rheumatoid arthritis and systemic lupus erythematosus were 60% as likely to see a rheumatologist as whites. Other studies have shown that rheumatologists order more BMD testing and prescribe more antiosteoporotic therapies than generalists and other types of specialists. (21-23) Although we do not know the relative proportion of women in our sample that saw rheumatologists or other osteoporosis specialists, it is possible that disparities in medical specialist referral may account for some of the observed differences.

Health services investigations in other chronic disorders have shown disparities between whites and blacks that may be relatively analogous to disparities in the process and outcomes of osteoporosis care. For example, Asch et al (24) found that blacks were significantly less likely to receive preventive care and more likely to develop avoidable adverse outcomes than whites in a sample of Medicare beneficiaries aged 65 years or older. Indeed, during the past 20 years, numerous studies have consistently documented that blacks are less likely than whites to receive a wide range of medical services. (9,25) Multiple factors such as patient social characteristics, provider and patient behavioral factors, and the responsiveness of the health care system have been attributed to these differences in past studies.

Although we do not know the true impact of less osteoporosis care on patient outcomes in this study, lower osteoporosis health care use, lower intensity of physical/occupational therapy, and less use of estrogen replacement therapy have previously been associated with poorer outcomes among blacks, including increased disability, longer hospital stays, and higher mortality. (2,6,7,26,27) Therefore, the ethnic disparities in osteoporosis preventive measures, which we have shown to be present even for women at high risk of osteoporosis, are especially troublesome.

In addition to patterns and predictors of health care quality, we also studied the epidemiology and determinants of osteoporosis prevalence among our population. As anticipated on the basis of prior reports, (2,6,7) white worn en in our study reported a significantly higher prevalence of fractures than black respondents. Despite the fact that rib and wrist fractures followed the expected trend of a higher prevalence among whites, hip fracture was reported equally in both racial groups. This finding conflicts with present literature, which shows a higher incidence of hip fracture among whites. (28,29) White women have also been shown to fracture at an earlier age than black women. (29) Although black women respondents in our study were older than their white counterparts, adjustment for age did not significantly attenuate our findings. Administrative claims data confirmed that the ratio of women who had a diagnostic code for a fracture event divided by the fracture rate reported on the survey differed by less than 5% between blacks and whites.

Our investigation offers some potential insights into why black women may have a lower risk of both osteoporosis and, in turn, a lower receipt of diagnostic and therapeutic modalities. Similar to other reports, we found that on average blacks reported higher body weight than white women and reported lower rates of smoking. Lower body weight and tobacco smoking are known risk factors for osteoporosis. (4,30-32) Lack of these risk factors could decrease physician concern about osteoporosis among their black patients. However, this potentially higher physician threshold to initiate osteoporosis evaluation and prevention would be less appropriate in the high-risk group of women with prior fractures. Our finding of similar ethnic disparities in this subgroup and in the overall population adds a sense of urgency to the need to understand and address these disparities.

A major limitation of our study is that we relied on self-report. The accuracy of the information provided may depend on several factors, such as age of the patient, educational level, satisfaction with health care, medical comorbidities, and the provision of an incentive for the return of the questionnaire. For the development of fractures, self-report may be acceptably reliable because fractures are traumatic events that are easily remembered. Several studies have shown high accuracy of self-report of hip fracture, with true-positive rates ranging from 82 to 95%. (33-36) Indeed, self-report of fractures may constitute a better "gold-standard" than either medical record review or analysis of claims data, which can both underdocument past events.

Another limitation of this study is the low response rate. However, only small differences were noted in demographics and income for respondents and nonrespondents. Although our response rate was adequate to ensure high statistical confidence in our results, it is possible that respondents may have differed from nonrespondents in ways that we could not measure. However, our calculations reveal that it is highly unlikely that bias from nonresponse could have reversed or mitigated our findings.

There may also be a selection bias, because we were not able to determine the racial composition of the nonrespondents. However, women responding to our questionnaire appear to have a slightly higher income level and are likely to be more concerned about their health than nonrespondents. Thus, our findings may underestimate the true burden on illness and receipt of preventive care. Although we have partially accounted for socioeconomic status through geographic proxies for individual income level, there may be residual confounding by unmeasured socioeconomic factors such as educational attainment, wealth, or occupation. Other factors not available from this short survey such as osteoporosis knowledge, health beliefs, and racial concordance between physician and patient (8) must be studied in the future to determine their effects on osteoporosis prevention and treatment. In particular, more research on attitudes, beliefs, and behaviors is needed to extrapolate strategies to control osteoporosis gleaned from in vestigations of predominately white populations to black groups. (5,37) Although the sample comes from a large regional HMO, our findings may not generalize to other populations. The choice of the health insurance provider and the geographic residence of our respondents' osteoporosis care may have influenced our findings. Despite these limitations, it is important to note that all respondents to this survey had similar levels of medical insurance coverage allowing reasonable access to at least primary care providers.

Conclusion

Our study documents racial disparities between postmenopausal black and white women in the prevalence, risk factors, testing, and prevention of osteoporosis. Even black women with a history of previous fracture received less care than white women. Although osteoporosis risk varies with race and ethnicity, our data suggest that physicians may be inappropriately less suspicious of osteoporosis and less aggressive in osteoporosis management for black women, even for those who are at highest risk. Future work is needed in other populations to better understand disparities in osteoporosis care and to identify effective strategies for making the latest research findings available to all women.

[FIGURE 1 OMITTED]

Table 1

Characteristics of managed care beneficiaries who responded to an
osteoporosis survey

                                            No. (%) or mean [+ or
                                                    -] SD

Patient characteristics                            Black

Age (yr) (n = 8,503)                         69.8 [+ or -] 9.9
Weight (lb) (n = 8,248)                     178.5 [+ or -] 42.9
Median household income from ZIP code (a)  18,015 [+ or -] 7,186
  (S) (n = 8,065)
Type of health plan--Medicare (n = 8,456)       1,826 (84)
Metropolitan residence (b) (n = 8,206)          2,025 (98)
Current cigarette smoker (n = 8,318)              213 (10)
Saw primary care physician in past 2 yr         1,931 (93)
  (n = 8,282)
Postmenopausal (n = 8,499)                      2,055 (94)
Hysterectomy (n = 8,466)                        1,094 (50)
Surgical oophorectomy (n = 8,267)                 694 (33)
Ever use of glucocorticoids (n = 7,650)           149 (8)
Any fracture (n = 8,304)                          511 (24)
Fractures at common osteoporotic sites            194 (9)
  (since age 45) (n = 8,206)
 Hip fracture (n = 8,201)                          72 (3)
 Rib fracture (n = 8,181)                          53 (3)
 Wrist fracture (n = 8,182)                        98 (5)

                                            No. (%) or mean [+ or
                                                    -] SD

Patient characteristics                            White

Age (yr) (n = 8,503)                        67.8 [+ or -] 10.1
Weight (lb) (n = 8,248)                     155.5 [+ or -] 33.6
Median household income from ZIP code (a)  29,670 [+ or -] 9,324
  (S) (n = 8,065)
Type of health plan--Medicare (n = 8,456)       4,210 (67)
Metropolitan residence (b) (n = 8,206)          5,876 (96)
Current cigarette smoker (n = 8,318)              730 (12)
Saw primary care physician in past 2 yr         6,019 (97)
  (n = 8,282)
Postmenopausal (n = 8,499)                      6,133 (97)
Hysterectomy (n = 8,466)                        3,530 (56)
Surgical oophorectomy (n = 8,267)               1,906 (31)
Ever use of glucocorticoids (n = 7,650)           682 (12)
Any fracture (n = 8,304)                        2,558 (41)
Fractures at common osteoporotic sites            877 (14)
  (since age 45) (n = 8,206)
 Hip fracture (n = 8,201)                         184 (3)
 Rib fracture (n = 8,181)                         308 (5)
 Wrist fracture (n = 8,182)                       527 (9)

                                           No. (%) or
                                           mean [+ or
                                             -] SD

Patient characteristics                    P value

Age (yr) (n = 8,503)                       <0.001
Weight (lb) (n = 8,248)                    <0.001
Median household income from ZIP code (a)  <0.001
  (S) (n = 8,065)
Type of health plan--Medicare (n = 8,456)  <0.001
Metropolitan residence (b) (n = 8,206)     <0.001
Current cigarette smoker (n = 8,318)        0.024
Saw primary care physician in past 2 yr    <0.001
  (n = 8,282)
Postmenopausal (n = 8,499)                 <0.001
Hysterectomy (n = 8,466)                   <0.001
Surgical oophorectomy (n = 8,267)           0.040
Ever use of glucocorticoids (n = 7,650)    <0.001
Any fracture (n = 8,304)                   <0.001
Fractures at common osteoporotic sites     <0.001
  (since age 45) (n = 8,206)
 Hip fracture (n = 8,201)                   0.268
 Rib fracture (n = 8,181)                  <0.001
 Wrist fracture (n = 8,182)                <0.001

(a)Determined on basis of 1990 census data.

(b)Defined using rural-urban continuum codes: metropolitan counties
(0-3) and nonmetropolitan counties (4-9). (15, 16)

Table 2

Osteoporosis testing and treatment among survey participants


                                  Black (%)  White (%)

Bone mineral density testing      222 (10)   1,577 (25)
  (n = 8,321)
Any prescription therapy (a)      798 (37)   4,162 (66)
  (n = 8,463)
 Estrogen replacement therapy     754 (36)   3,879 (62)
  (n = 8,319)
 Alendronate (n = 7,650)           36 (2)      500 (9)
 Raloxifene (n = 7,650)            21 (1)      234 (4)
 Calcitonin (n = 7,650)            19 (1)      167 (3)
Supplemental calcium (n = 7,650)  579 (32)   3,288 (56)
Vitamin D or multivitamin         910 (50)   3,870 (66)
  (n = 7,650)

                                         Odds Ratio
                                  (95% Confidence interval)

Bone mineral density testing          0.34 (0.30-0.40)
  (n = 8,321)
Any prescription therapy (a)          0.34 (0.30-0.37)
  (n = 8,463)
 Estrogen replacement therapy         0.30 (0.27-0.33)
  (n = 8,319)
 Alendronate (n = 7,650)              0.22 (0.16-0.31)
 Raloxifene (n = 7,650)               0.24 (0.14-0.41)
 Calcitonin (n = 7,650)               0.36 (0.23-0.58)
Supplemental calcium (n = 7,650)      0.37 (0.33-0.41)
Vitamin D or multivitamin             0.52 (0.47-0.58)
  (n = 7,650)

(a)Includes hormone replacement therapy, bisphosphonate therapy
(alendronate, etidronate, risedronate, pamidronate), raloxifene, and
calcitonin.

Table 3

Multivariable predictors of bone mineral density testing and
prescription osteoporotic therapy (a)

                                     Adjusted OR (95% CI)

                                Bone mineral density testing
                                        (n = 6,818)

Black                                 0.47 (0.39-0.58)

Age (yr)

 <60                                        REF
 60-68                                1.25 (1.07-1.47)
 69-74                                1.14 (0.96-1.35)
 [greater than or equal to]75         0.88 (0.74-1.05)

Weight (lbs)

 <135                                       REF
 135-154                              1.01 (0.86-1.18)
 155-179                              0.72 (0.61-0.85)
 [greater than or equal to]180        0.62 (0.52-0.73)

Median income (S)

 <20,634                                    REF
 20,634-26,079                        1.09 (0.91-1.32)
 26,080-31,643                        1.25 (1.04-1.50)
 [greater than or equal to]
 31,644                               1.44 (1.21-1.73)
Medicare plan                                NS (c)
Current smoker                        0.58 (0.48-0.71)
Seen physician past 2 yr              2.54 (1.65-3.92)
Postmenopausal                        2.26 (1.41-3.60)
Surgical oophorectomy                 0.88 (0.78-1.00)
Ever use of glucocorticoids           1.30 (1.09-1.55)
Any fracture                          1.30 (1.15-1.46)
C statistic (18)                           0.653
Goodness-of-fit statistic.
(19)                                     P = 0.877

                                     Adjusted OR (95% CI)

                                Any prescription therapy (b)
                                      (n = 6,965)

Black                               0.35 (0.30-0.41)

Age (yr)

 <60                                      REF
 60-68                              0.78 (0.64-0.96)
 69-74                              0.46 (0.37-0.58)
 [greater than or equal to]75       0.27 (0.22-0.34)

Weight (lbs)

 <135                                     REF
 135-154                            1.08 (0.92-1.26)
 155-179                            0.99 (0.85-1.17)
 [greater than or equal to]180      0.71 (0.60-0.83)

Median income (S)

 <20,634                                  REF
 20,634-26,079                      1.12 (0.95-1.31)
 26,080-31,643                      1.27 (1.07-1.50)
 [greater than or equal to]
 31,644                             1.33 (1.12-1.58)
Medicare plan                       0.50 (0.42-0.61)
Current smoker                      0.72 (0.60-0.85)
Seen physician past 2 yr            3.24 (2.39-4.39)
Postmenopausal                      45.6 (24.1-86.2)
Surgical oophorectomy               2.61 (2.30-2.95)
Ever use of glucocorticoids         2.04 (1.68-2.49)
Any fracture                               NS
C statistic (18)                         0.781
Goodness-of-fit statistic.
(19)                                   P = 0.351

(a)NS, not significant; REF, reference; OR, odds ratio; CI, confidence
interval.

(b)Includes estrogen, bisphosphonates (alendronate, etidronate,
risedronate, pamidronate), raloxifene, and calcitonin.

(c)Variables did not meet the criteria to enter (P < 0.25) or remain (P
< 0.15) in the model for stepwise logistic regression.

Accepted August 6, 2002.

References

(1.) Jacobsen SJ, Goldberg J, Miles TP, Brody JA, Stiers W, Rimm AA. Hip fracture incidence among the old and very old: A population-based study of 745,435 cases. Am J Public Health 1990;80:871-873.

(2.) Kellie S, Brody J. Sex-specific and race-specific hip fracture rates. Am J Public Health 1990;80:326-328.

(3.) Theobald TM, Cauley JA, Gluer CC, Bunker CH, Ukoli FA, Genant HK. Black-white differences in hip geometry: Study of Osteoporotic Fractures Research Group. Osteoporos Int 1998;8:61-67.

(4.) Bohannon A. Osteoporosis and African-American women. J Womens Health Gender Based Med 1999;8:609-615.

(5.) Aloia JF, Vaswani A, Yeh JK, Flaster E. Risk for osteoporosis in black women. Calcif Tissue Int 1996;59:415-423.

(6.) Jacobsen SJ, Goldberg J, Miles TP, Brody JA, Stiers W, Rimm AA. Race and sex differences in mortality following fracture of the hip. Am J Public Health 1992;82:1147-1150.

(7.) Furstenberg A, Mezey M. Differences in Outcome between black and white elderly hip fracture patients. J Chronic Dis 1987;40:931-938.

(8.) Fiscella K, Franks P, Gold MR, Clancy CM. Inequality in quality: Addressing socioeconomic, racial, and ethnic disparities in health care. JAMA 2000;283:2579-2584.

(9.) Epstein A, Ayanian J. Racial disparities in medical care. N Engl J Med 2001;344:1471-1473 (editorial).

(10.) Chen J, Rathore SS, Radford MJ, Wang Y, Krumholz HM. Racial differences in the use of cardiac catheterization after acute myocardial infarction. N Engl J Med 2001;344:1443-1449.

(11.) Lydick E, Cook K, Turpin J, Melton M, Stine R, Byrnes C. Development and validation of a simple questionnaire to facilitate identification of women likely to have low bone density. Am J Manag Care 1998;4:37-48.

(12.) Von Muhlen D, Visby Lunde A, Barrett-Connor E, Bettencourt R. Evaluation of the simple calculated osteoporosis risk estimation (SCORE) in older Caucasian women: The Rancho Bernardo study. Osteoporos Int 1999;10:79-84.

(13.) Cadarette S, Jaglal S, Murray T. Validation of the simple calculated osteoporosis risk estimation (SCORE) for patient selection for bone densitometry. Osteoporos Int 1999;10:85-90.

(14.) Cadarettc SM, Jaglal SB, Murray TM, McIsaac WJ, Joseph L, Brown JP; Canadian Multicentre Osteoporosis Study. Evaluation of decision rules for referring women for bone densitometry by dual-energy x-ray absorptiometry. JAMA 2001;286:57-63.

(15.) Butler MA, Beale CL. Rural-Urban Continuum Codes for Metro and Nonmetro Counties, 1993 (Staff Rep. No. AGES 9425). Washington, DC, U.S. Department of Agriculture, Economic Research Service, Agricultural and Rural Economy Division, 1994.

(16.) Butler M. Rural-Urban Continuum Codes for Metro and Nonmetro Counties, 1990 (Staff Rep. No. 9028). Washington, DC, U.S. Department of Agriculture, Economic Research Service, Agricultural and Rural Economy Division, 1990.

(17.) Cochran WG. Sampling Techniques. New York, Wiley, 1977, ed 3.

(18.) Lemeshow S, Hosmer J. A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol 1982; 115:92-106.

(19.) Harrell FE Jr, Lee KL, Califf RM, Pryor DB, Rosati RA. Regression modelling strategies for improved prognostic prediction. Stat Med 1984; 3:143-152.

(20.) Katz JN, Barrett J, Liang MH, Kaplan H, Roberts WN, Baron JA. Utilization of rheumatology physician services by the elderly. Am J Med 1998; 105:312-318.

(21.) Mudano A, Allison J, Hill J, Rothermel T, Saag K. Variations in glucocorticoid induced osteoporosis prevention in a managed care cohort. J Rheumatol 2001;28:1298-1305.

(22.) Buckley LM, Marquez M, Feezor R, Ruffin DM, Benson LL. Prevention of corticosteroid-induced osteoporosis: results of a patient survey. Arthritis Rheum 1999;42:1736-1739.

(23.) Aagaard EM, Lin P, Modin GW, Lane NE. Prevention of glucocorticoid-induced osteoporosis: Provider practice at an urban county hospital. Am J Med 1999;107:456-460.

(24.) Asch SM, Sloss EM, Hogan C, Brook RH, Kravitz RL. Measuring underuse of necessary care among elderly Medicare beneficiaries using inpatient and outpatient claims. JAMA 2000;284:2325-2333.

(25.) Weinick R, Zuvekas S, Cohen J. Racial and ethnic differences in access to and use of health care services, 1977 to 1996. Med Care Res Rev 2000;57(Suppl 1):36-54.

(26.) Hoenig H, Rubenstien L, Kahn K. Rehabilitation after hip fracture: Equal opportunity for all? Arch Phys Med Rehabil 1996;77:58-63.

(27.) Ganesan K, Teklehaimanot S, Norris K. Estrogen replacement therapy use in minority postmenopausal women. Ethn Dis 2000;10:257-261.

(28.) Villa M, Nelson L. Race, ethnicity and osteoporosis, in Marcus R, Feldman D, Kelsey J (eds): Osteoporosis. San Diego, Academic Press, 1996, pp 435-447.

(29.) Maggi S, Kelsey JL, Litvak J, Heyse SP. Incidence of hip fractures in the elderly: A cross-national analysis. Osfeoporos Int 1991;1:232-241.

(30.) Turner L, Wang M, Fu Q. Risk factors for hip fracture among southern older women. South Med. J 1998;91:533-540.

(31.) Slemenda C. Cigarettes and the skeleton. N Engl J Med 1994;330:430-431 (editorial).

(32.) Baron JA, Farahmand BY, Weiderpass E, Michaelsson K, Alberts A, Persson I, et al. Cigarette smoking, alcohol consumption, and risk of hip fracture in women. Arch Intern Med 2001;161:983-988.

(33.) Nevitt MC, Cummings SR, Browner WS, Seeley DG, Cauley JA, Vogt TM, et al. The accuracy of self-report of fractures in elderly women: Evidence from a prospective study. Am J Epidemiol 1992;135:490-499.

(34.) Paganini-Hill A, Chao A. Accuracy of recall of hip fracture, heart attack, and cancer: A comparison of postal survey data and medical records. Am J Epidemiol 1993;138:101-106.

(35.) Bergmann MM, Byers T, Freedman DS, Mokdad A. Validity of self-reported diagnoses leading to hospitalization: A comparison of self-reports with hospital records in a prospective study of American adults. Am J Epidemiol 1998;147:969-977.

(36.) Ismail AA, O'Neill TW, Cockerill W, Finn JD, Cannata JB, Hoszowski K, et al. Validity of self-report of fractures: Results from a prospective study in men and women across Europe--EPOS Study Group: European Prospective Osteoporosis Study Group. Osteoporos Int 2000;11:248-254.

(37.) Luckey MM, Wallenstein S, Lapinski R, Meier DE. A prospective study of bone loss in African-American and white women: A clinical research center study. J Clin Endocrinol Metab 1996;81:2948-2956.

From the Department of Medicine and Center for Education and Research on Therapeutics of Musculoskeletal Disorders, Division of Continuing Medical Education, Center for Outcomes and Effectiveness Research and Education, University of Alabama at Birmingham, and United Healthcare of Alabama, Department of Quality Improvement, Birmingham, AL.

This work was supported by Grant HS 10389 from the Agency for Healthcare Research and Quality, Washington, DC, and in part by a grant from Merck & Co., Inc., Whitehouse Station, NJ.

Reprint requests to Kenneth G. Saag, MD, MSc, Division of Clinical Immunology/Rheumatology, University of Alabama at Birmingham, 1530 Third Avenue S., FOT 820, Birmingham, AL 35294-3408. Email: [e-mail]

Copyright [c] 2003 by The Southern Medical Association 0038-4348/03/9605-0445