Racial and Ethnic Disparities in Bone Health and Outcomes in the United States

Abstract

Osteoporosis is a bone disease classified by deterioration of bone microarchitecture and decreased bone strength, thereby increasing subsequent risk of fracture. In the United States, approximately 54 million adults aged 50 years and older have osteoporosis or are at risk due to low bone mass. Osteoporosis has long been viewed as a chronic health condition affecting primarily non-Hispanic white (NHW) women; however, emerging evidence indicates racial and ethnic disparities in bone outcomes and osteoporosis management. The primary objective of this review is to describe disparities in bone mineral density (BMD), prevalence of osteoporosis and fracture, as well as in screening and treatment of osteoporosis among non-Hispanic black (NHB), Hispanic, and Asian adults compared with NHW adults living on the US mainland. The following areas were reviewed: BMD, osteoporosis prevalence, fracture prevalence and incidence, postfracture outcomes, DXA screening, and osteoporosis treatments. Although there are limited studies on bone and fracture outcomes within Asian and Hispanic populations, findings suggest that there are differences in bone outcomes across NHW, NHB, Asian, and Hispanic populations. Further, NHB, Asian, and Hispanic populations may experience suboptimal osteoporosis management and postfracture care, although additional population-based studies are needed. There is also evidence that variation in BMD and osteoporosis exists within major racial and ethnic groups, highlighting the need for research in individual groups by origin or background. Although there is a clear need to prioritize future quantitative and qualitative research in these populations, initial strategies for addressing bone health disparities are discussed.

Introduction

Osteoporosis is characterized by deterioration of bone microarchitecture and decreased bone strength, thereby increasing subsequent risk of fracture. In the United States, approximately 54 million adults aged ≥50 years have osteoporosis or are at risk due to low bone mass.^(1,2) This silent condition frequently remains undiagnosed until complicated by fracture, with nontraumatic hip and spine being the most prevalent (66% of all osteoporosis-related fractures). Over 2 million US adults experience osteoporosis-related fractures annually,⁽³⁾ imposing significant functional, economic, and societal burden.^(3,4) Incident fractures are a significant public health problem and are projected to increase from >2 million to >3 million each year, with an increase in fracture-related costs estimated from $17 billion to $25.3 billion by 2025.⁽⁵⁾ Recent Medicare data indicated that mean all-cause medical costs for hip fracture patients was ~$22,000 before fracture and ~$70,000 12 months postfracture, highlighting the substantial economic impact.⁽⁶⁾ In addition to economic implications, fractures related to osteoporosis, particularly of the hip, decrease quality of life, functional status, and independence, and increase risk of institutionalization and mortality.^(3,7) Approximately 21% to 30% of individuals who experience a hip fracture will die within 1 year.⁽⁸⁾ The US population is aging, with nearly one in five residents estimated to be ≥65 years by 2030.⁽⁹⁾ Thus, the growing problem of osteoporosis and its complications are an important national public health concern.

Osteoporosis has long been viewed as a chronic condition affecting primarily non-Hispanic white (NHW) women; however, evidence continues to emerge highlighting racial and ethnic disparities in osteoporosis outcomes and management. A current review of differences in bone health and management among major racial and ethnic populations in the United States is needed to inform future research and recommendations. The primary objectives of this review are to: (i) describe disparities in bone mineral density (BMD), osteoporosis, and osteoporosis management (eg, screening and treatment) among US non-Hispanic black (NHB), Hispanic, and Asian adults compared with NHW adults; (ii) identify research gaps; and (iii) provide future research directions for bone research among racial and ethnic groups in the United States.

Racial and Ethnic Differences in BMD and Prevalence of Osteoporosis

There is mounting evidence that BMD varies across racial and ethnic groups, although the directionality of these differences is not consistent across all population groups (Table 1). NHB adults have higher BMD as compared with Hispanic, NHW, and Asian adults.^{(15,17,18,21,24,26,28,35,38)} NHB adults also have lower rates of bone loss compared with NHW adults (Table 1).^(16,17) Prevalence of osteoporosis is also lower among NHB compared with NHW, Hispanic, and Asian adults (Table 1).^(1,32,34) Based on most recent National Health and Nutrition Examination Survey (NHANES 2013–2014) data, the age-adjusted prevalence of low bone mass and osteoporosis at the femoral neck and/or lumbar spine was lowest among NHB adults (women: 40.4% with low bone mass, 8.2% with osteoporosis; men: 25.7% with low bone mass, 1.9% with osteoporosis).⁽¹⁾

Table 1.

Summary Published Literature Evaluating Bone Mineral Density and Prevalence of Osteoporosis in United States

Author; Year	Study design and data source	# Participants	Racial distribution and characteristics	Bone measurement (site and tool)	T-score reference group	Conclusions
Bone Mineral Density and Prevalence of Osteoporosis
Ross PD; 1996(10)	Cross-sectional; Early Postmenopausal Interventional Cohort (EPIC) study	1609	Early postmenopausal women Caucasian: n = 1367 Asian: n = 162	DXA Hip, spine, lateral spine, wrist and forearm	N/A	BMD was approximately 4–6% lower in Asian women at most sites versus Caucasian women BMD at the wrist in Asian women was significantly greater than Caucasian women ↓BMD in Asian women except at wrist sites
Looker AC; 1998(11)	Cross-sectional; NHANES I Epidemiologic Follow-up Study, 1988–1994	14,646	Adult men and women 20+ years (NHANES) Non-Hispanic White: n = 6181 Non-Hispanic Black: n = 4021 Mexican American: n = 3858	DXA Femoral neck, trochanter, intertrochanter, wards triangle and total hip	N/A	BMD tended to be lower in NHW in the southern U.S. by approximately 1–4% Men mean total femur BMD: NHW 0.993 g/cm2, NHB 1.103 g/cm2, 1.103 g/cm2 Women mean total femur BMD: NHW 0.868 g/cm2, NHB 0.977 g/cm2, MA 0.926 g/cm2 ↓BMD in adults from southern U.S.
Taaffe DR; 2000(12)	Cross-sectional	116	Postmenopausal women 60–86 years Non-Hispanic Caucasian: n = 62 Mexican American: n = 54	DXA Femoral neck, trochanter, wards triangle, whole body	N/A	Once adjusted for height, MA women have significantly higher BMD at all anatomical site compared with NHC ↑Mexican American women ↓ NHC women
Nelson DA; 2000(13)	Cross-sectional; Women’s Health Initiative (WHI)	371	Women entering the Women’s Health Initiative clinical trials in Detroit ages 50–79 years Non-Hispanic White: n = 235 Non-Hispanic Black: n = 136	DXA Proximal femur	N/A	NHB women had significantly higher BMD at both neck and shaft sites compared to non-Hispanic whites Mean femoral neck (g/cm2): NHW 0.69, NHB 0.76, % diff 10.1%*** ↑NHB, ↓ NHW
Morton DJ; 2003(14)	Cross-sectional; San Diego County, CA	803	Postmenopausal women 50–69 years Non-Hispanic White: n = 354 Filipina: n = 285 Hispanic/Latino: n = 164	DXA Femoral neck, total hip, lumbar spine, total body	N/A	General lack of differences in BMD between ethnic groups Filipina woman had lower BMD at lumbar spine versus NHW women when accounting for body size (BMI, BF%) (model 1: 0.896 mg/cm2 versus 0.939 mg/cm2)***, (model 2: 0.892 mg/cm2 versus 0.944 mg/cm2)*** and at the total body when 0.896 mg/cm2 versus 0.873 mg/cm2 but only once anthropometrics and health behaviors were adjusted for
George A; 2003(15)	Cross-sectional; Baltimore metropolitan area, 2000–2001	694	Older men (65+ years) White: n = 503 Black: n = 191	DXA Femoral neck, Lumbar spine, total body	N/A	Older Black men have significantly higher BMD (0.09 mg/cm2 femoral neck, 0.07 mg/cm2 lumbar spine,0.06 mg/cm2 total body) than older White men even after adjustment for factors associated with BMD ↑older Black men ↓older White men
Cauley JA; 2005(16)	Longitudinal; Study of Osteoporotic Fracture, 2000–2002	6489	Postmenopausal women Caucasian: n = 6007 African American: n = 482	DXA Total hip and femoral neck	N/A	The average rate of BMD loss was twice as great in Caucasian as compared to African-American women (Total Hip: −0.574%/year versus −0.334%/year, Femoral neck:−0.515%/year versus −0.203%/year)*** ↓ BMD in Caucasian, j African American women
Tracy JK; 2005(17)	Longitudinal; The Baltimore Men’s Osteoporosis Study, 2000–2001	468	Older men (65+ years) White: n = 349 Black: n = 119	DXA Femoral neck, total hip	N/A	Older black men seem to lose BMD at a lower rate than older white men (Femoral neck: 0.017 g/cm2/year versus 0.010 g/cm2/year*, Total hip: 0.010 g/cm2/year versus +0.001 g/cm2/year**) ↓ bone loss with age in Black men
Cauley JA; 2005(18)	Cross-sectional; The Osteoporotic Fractures in Men (MrOS) Study, 2000–2002	5995	Men 65+ living in the U.S. Caucasian: n = 5632 African American: n = 244 Asian: n = 191 Hispanic/Latino: n = 127 American Indian: n = 5 Native Hawaiian/ Pacific Islander: n = 7 Multi-racial: n = 59	DXA Lumbar spine (L1–L4) Total hip (femoral neck/trochanter)	N/A	African American men had 6 to 11% higher BMD than Caucasian men independent of multiple factors Family history of low BMD was a strong independent predictor of BMD in older men
Barrett-Connor E; 2005(19)	Cross-sectional; The National Osteoporosis Risk Assessment (NORA), 1997–2003	197,848	Community-dwelling postmenopausal women Caucasian: n = 179,471 African American: n = 7784 Asian: n = 1912 Hispanic/Latino: n = 6973 Native Americans: n = 1708	Heel (SXA) Finger (pDXA)	Young normal reference population	Ethnic differences in BMD were strongly related to body weight BMD T-scores: Asian −1.22, Hispanic/Native −1.13, NHW −0.89, NHB −0.39 Osteoporosis prevalence: 11.9% Native Americans, 10% Asian, 9.8% Hispanics, 7.2% Caucasian, 4.2% African American There were no significant differences in respect to bone decline with age by ethnicity ↑ African American, Similar in Caucasian, Asian and Hispanic, ↓Asian
Walker MD; 2006(20)	Cross-sectional, New York City, 2002–2004	359	Healthy Chinese American women Ages 20–90 years	DXA Lumbar spine (L1–L4), Total hip	Non-Hispanic whites	Chinese American BMD values were significantly lower than Caucasian Peak BMD was 3.3–6.1% lower in Chinese Americans compared to non-Hispanic white database but was only significant at femoral neck** ↓ BMD Chinese American women
Araujo AB; 2007(21)	Cross-sectional; The Boston Area Community Health (BACH) Survey, 2002–2005	1209	Men ages 30–79 years old White: n = 451 Hispanic = 401 Black: n = 367	DXA Whole body, total hip, lumbar spine and forearm	N/A	Steeper decline in BMD for Hispanic men with age Change in whole body BMD (Hispanic: −0.20 g/cm2, NHW: −0.06 g/cm2, Black: −0.04 g/cm2)* ↓BMD in older Hispanic men
Finkelstein JS; 2008(22)	Longitudinal; The Study of Women’s Health Across the Nation (SWAN), 1996–2008	3302	Community based groups of women Caucasian: n = 1550 African American: n = 935 Japanese: n = 281 Chinese: n = 250 Hispanic: n = 286	DXA Lumbar spine and total hip	N/A	Body weight is a major determinant of BMD, whereas ethnicity was not Lumbar spine bone loss was most rapid in Chinese and Japanese women, moderate in Caucasian and slowest in African American in late perimenopause and post menopause***
Marshal LM; 2008(23)	Cross-sectional; The Osteoporotic Fractures in Men (MrOS) Study, 2000–2002	3305	Older men ages 65+ White: n = 2924 Black: n = 170 Asian/ Pacific Islander: n = 121 Hispanic/ Latino: n = 90	DXA Total hip and subregions	Non-Hispanic white	NHB and Asian men had higher BMD at hip sites Femoral neck BMD (g/cm2): White 0.783, Black 0.881***, Asian 0.750, Hispanic 0.799 Cortical BMD (g/cm3): White 1.014, Black 0.997*, Asian 1.052**, Hispanic 1.048 ↑ Black/Asian BMD compared to White/Hispanic
Looker AC; 2009(24)	Cross-sectional; NHANES, 1994–2004	6532	Adults 20+ years w/complete DXA measurements Non-Hispanic whites: n = 3276 Non-Hispanic black: n = 1329 Mexican Americans: n = 1438	DXA Total body (subregions: lumbar spine, pelvis, right leg, left arm)	N/A	Mexican Americans had lower BMD than NHW Male mean total body BMD (g/cm2): NHW 1.184, NHB 1.252, MA 1.142 Female mean total body BMD (g/cm2): NHW 1.083, NHB 1.148, MA 1.061 Age differences by sex but not race/ethnicity
Peacock M; 2009(25)	Cross-sectional; Indiana, 1990–2004	1437	Healthy American men (23–57 years) White: n = 612 Black: n = 164 Healthy American women (20–63 years) White: n = 492 Black: n = 169	DXA DPXL (men) Prodigy (women) Right femoral neck	N/A	Whites have lower BMD than Blacks Mean femoral neck aBMD (mg/cm2): White men 1.09, Black men 1.17 Mean femoral neck aBMD (mg/cm2): White women 1.02, Black women1.11 Women had higher BMD than men but smaller geometry variables
Looker AC; 2010(26)	Cross-sectional; NHANES III: 1988–1994 NHANES 2005–2006	NHANES III: 6401 NHANES 05–06: 1614	NHANES III: Non-Hispanic white: n = 3603 Non-Hispanic black: n = 1342 Mexican American: n= 1225 NHANES 2005–2006: Non-Hispanic white: n = 967 Non-Hispanic black: n = 325 Mexican American: n= 250	NHANES II: DXA Femoral neck NHANES 2005–2006: DXA QDR-4500A Femoral neck	29-year-old NHW women	Decrease in low BMD at the femur between NHANES III and NHANES 2005–2006 but low BMD among older adults remains high Femoral neck BMD (g/cm2): NHW 0.803**, NHB 0.889*, MA 0.807* Total hip BMD (g/cm2): NHW 0.986**, NHB 1.072**, MA 0.984** Prevalence of OP at femoral neck (M:2%, F:10%), Total hip (M:1.4%, F:8.3%)
Travison TG; 2011(27)	Cross-sectional; The Boston Area Community Health (BACH) Survey, 2002–2005	1209	Community dwelling men ages 30–79 years White: n = 426 Black: n = 353 Hispanic/Latino: n = 387	DXA Femoral neck, distal radius, whole body	N/A	Much of the racial/ethnic differences in BMD can be accounted for by differences in body composition, diet and socioeconomic factors 29% difference in femoral neck BMD between Balck (0.14 g/cm2), and White (0.10 g/cm2) ↑ Black, ↓ White Weakened after adjustment
Nelson DA; 2011(28)	Cross-sectional; Women’s Health Initiative (WHI), 1992–2007	10,563	Post-menopausal women ages 50–79 years Non-Hispanic white: n = 8156 African American: n = 1466 Mexican American: n = 702 Native American: n = 124 Asian/ Pacific Islander: n = 36 Multi-racial: n = 76	DXA Femoral neck, total hip	N/A	Differences in proximal femur geometry are consistent with lower hip fractures among Mexican and Asian Americans Femoral neck mean BMD (g/cm2): NHW 0.712, AA 0.792***, MA 0.722*, NA 0.742** BMD in Mexican American= BMD in NHW
Looker AC; 2012(29)	Cross-sectional; NHANES III, 1988–1994) and NHANES 2005–2008	6022	Participants from NHANES survey between 2005 to 2008 Non-Hispanic white: n = 3279 Non-Hispanic black: n = 1415 Mexican American: n = 1328	DXA Lumbar spine and femoral neck	30-year-old white females	The inclusion of lumbar spine BMD into FRAX screening tool would improve correct classification Male adjusted mean lumbar spine BMD (g/cm2): NHW 1.055***, NHB 1.120***, MA 1.020a Female adjusted mean lumbar BMD (g/cm2): NHW1.021***, NHB 1.078***, MA 0.991* Male age adjusted prevalence of OP: NHW 3.9%, NHB 0%, MA 6.3% Female age adjusted prevalence of OP: NHW 15.3%, NHB 9.0%, MA 25.8%
Khandewal S; 2012(30)	Longitudinal, 1997–2003	7184	Women from North America ages 50–85 years White: n = 4490 Chinese: n = 2245 South Asian: n = 449	DXA Femoral neck	White women	Chinese and South Asian women had significantly lower BMD compared to White women Prevalence of OP: South Asian 8.9%, Chinese 11.9%, White 6.5% ↓Chinese/South Asian, ↑White women
Danielson ME; 2013(31)	Longitudinal; The Study of Women’s Health Across the Nation (SWAN), 1996–1997	1942	Women from five study centers across U.S. ages 42–52 years Caucasian: n = 966 African American: n = 517 Chinese: n = 220 Japanese: n = 239	DXA Proximal femur, lumbar spine	Non-Hispanic whites	African American and Japanese women have structural advantages at the hip over Caucasian and Chinese women resulting in lower hip fracture incidence Adjusted mean femoral neck BMD (g/cm2): Caucasian 0.82, AA 0.90, Chinese 0.83, Japanese 0.83 Adjusted mean trochanter BMD (g/cm2): Caucasian 0.93, AA 0.98, Chinese 0.84, Japanese 1.11 ↑ BMD in AA and Japanese women
Wright NC; 2014(32)	Cross-sectional; NHANES 2005–2010	99,048,838	Non-institutionalized adults Non-Hispanic white: n = 75,272,609 Non-Hispanic black: n = 9,830,977 4,595,535 Mexican American	DXA Proximal femur, lumbar spine	NHW women ages 20–29	Overall prevalence of low bone mass was 43.6%, most were NHW Estimated prevalence of OP for men: NHW 3.9%, NHB 1.3%, MA 5.9% Estimated prevalence of OP for women: NHW 15.8%, NHB 7.7%, MA 20.4%
Zhou B; 2014(33)	Cross-sectional; Columbia University Medical Center (CUMC)	66	Postmenopausal women 33 Caribbean Latino women age matched with 33 non-Hispanic white women	DXA Lumbar spine, total hip, femoral neck, 1/3 distal radius, ultradistal radius	Young normal non-Hispanic whites	Caribbean Latino women had lower bone microarchitecture than NHW Only significant difference between NHW and Hispanics was at lumbar spine (0.959 g/cm2 for NHW versus 0.884 g/cm2 for Hispanics)** ↓ Caribbean Latino BMD, ↑ NHW women
Looker AC; 2015(34)	Cross-sectional; NHANES 2005–2010	2476	Non-institutionalized adults ages 65+ years Non-Hispanic white: n = 1705 Non-Hispanic black: n = 450 Mexican American: n = 321	DXA Femoral neck, lumbar spine	Healthy NHW women	Prevalence of OP for males age adjusted to 2010: NHW 5.2%, NHB 0%, MA 11.2% Prevalence of OP for females age adjusted to 2010: NHW 24.7%, NHB 15.8%, MA 36.9% ↓ NHW/Mexican American, ↑ NHB
Xu Y; 2018(35)	Cross-sectional; NHANES 2005–2014 without 2011–2012 cycle (BMD was not measured)	14,188	Men and women in the U.S. ages 20+ years Non-Hispanic white: n = 7005 Non-Hispanic black: n = 2683 Hispanic/Latino: n = 3575 Multi-racial: n = 925	DXA Femoral neck and lumbar spine	Young normal NHW women	BMD decreased in both men and women overall There were no differences between ethnicity and race
Noel SE; 2018(36)	Cross-sectional, 2004–2009	953	Puerto Rican adults living in the Greater Boston area ages 45–75 years	DXA Hip and lumbar spine	30-year-old non-Hispanic white females	When compared to prevalence of OP of NHW from NHANES, Puerto Ricans had a similar if not slightly higher prevalence of OP compared to NHW Prevalence of OP from BPROS compared to NHANES for women: NHW 10.1%, MA 16%, PR 10.7%, NHB 3.8% For men: NHW 2.3%, MA 3.9%, PR 8.6%, NHB 1.3%
Lo JC; 2020(37)	Cross-sectional/Retrospective; northern California healthcare system, 1998–2017	139,632	Asian/ non-Hispanic White women ages 50–79 years Non-Hispanic White: n = 115, 318 Filipino: n = 11,147 Chinese: n = 10,648 Japanese: n = 2519	DXA Femoral neck	Non-Hispanic white women	Filipino, Chinese and Japanese women had significantly (30–50%) lower BMD than NHW ↓Filipino, Chinese and Japanese ↑ NHW

↓ = lower; ↑ = higher; AA = African American; BMD = bone mineral density; BMI = body mass index; BF= body fat; DXA = dual-energy X-ray absorptiometry; MA = Mexican American; NHANES = The National Health and Nutrition Examination Survey; NHC = Non-Hispanic Caucasian; OP = Osteoporosis.

^aRace and ethnicity as reported in the manuscript.

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

Differences in mean BMD between Hispanic and NHW adults, however, vary across studies, with some showing higher,^(12,26) similar,^(14,26) and in some cases, lower BMD⁽³⁸⁾ among Hispanic compared with NHW adults (Table 1). It has also been noted that the directionality of these differences may vary by skeletal site. For example, NHANES III (1988–1994) data showed higher mean BMD at the proximal femur for Mexican American men and women compared with NHW men and women,⁽¹¹⁾ but based on NHANES 1999–2004 data, Mexican Americans had lower mean BMD for total body and other sites compared with NHW adults aged ≥20 years.⁽²⁴⁾ In addition to different skeletal sites being scanned, observed differences between NHANES cycles may be due differences in populations due to the cross-sectional nature of the study, or to other factors, but comparison of demographic, acculturation, and anthropometric variables for Mexican Americans between the two cycles this did not explain this discrepancy.^(24,26) Thus, reasons for these differences remains unknown and longitudinal studies of bone loss are needed. Hispanic adults have been shown to have similar,^(1,14,36,39) and in some cases higher, prevalence of osteoporosis⁽³⁸⁾ as compared with NHW adults (Table 1). The majority of research on BMD among Hispanics in the United States has focused on Mexican Americans likely given their majority status in the US population. However, studies on bone health among Hispanic origin groups (eg, 15 largest US Hispanic groups by origin include Mexicans, Puerto Ricans, Salvadorans, Cubans, Dominicans, Guatemalans, Colombians, Hondurans, Spaniards, Ecuadorians, Peruvians, Nicaraguans, Venezuelans, Argentines, and Panamanians)⁽⁴⁰⁾ are limited, with few,^(21,36) but not all,⁽¹⁹⁾ indicating variation in BMD and prevalence of osteoporosis across individual origin groups.

Data on BMD in US Asian adults are also limited, with few studies indicating lower BMD,^(1,11,20,37) and others suggesting similar BMD^(19,31) in Asian populations compared with NHW adults (Table 1). Similar to Hispanic adults, there are few data among Asian populations by origin group.⁽¹⁴⁾ Most recently, there was less than a 3% difference in BMD at the femoral neck between Filipino, Chinese, and Japanese women aged 50 to 79 years, which was significantly lower than BMD of NHW women; however adjustment for height attenuated associations by 30% to 40%.⁽³⁷⁾ As for bone loss, findings from the Study of Women’s Health Across the Nation (SWAN) study suggest that Japanese and Chinese perimenopausal and postmenopausal women experienced bone loss at a faster rate than NHW and NHB women.⁽²²⁾ According to the most recent NHANES data, Asian women and men had the highest age-adjusted prevalence of low bone mass and osteoporosis (women: 47.0% and 40%; men: 47.7% and 7.5%, respectively).⁽¹⁾

There are many studies demonstrating differences in BMD and prevalence of osteoporosis across the broader racial and ethnic groups; however, the majority of this work has been cross-sectional with few longitudinal studies examining bone loss across racial and ethnic groups, particularly for Hispanic and Asian populations.⁽²²⁾ Although there are several longitudinal cohort studies (ie, Framingham Osteoporosis Study,⁽⁴¹⁾ Nurses’ Health Study,⁽⁴²⁾ and Study of Osteoporotic Fractures,⁽⁴³⁾ Women’s Health Initiative,^(44,45) and Osteoporotic Fractures in Men Study)⁽⁴⁶⁾ that have longitudinal data, populations included are primarily NHW (ranging from 86% to 97%). This limits the ability to not only understand differences in BMD, but also risk factors for loss of BMD and fracture across racial and ethnic groups. There is a need for large population-based, multiethnic studies examining how BMD changes over time among specific racial and ethnic minority groups.

It also remains unclear as to which reference database to use for defining osteoporosis and low bone mass in persons of color. As shown in Table 1, the majority of studies defined osteoporosis using WHO T-score thresholds based on the NHANES young, normal NHW reference population. The use of one reference database for all populations was supported as T-scores calculated with the young-normal NHW reference were similar to those calculated using sex- and race/ethnic-specific reference data.⁽⁴⁷⁾ However, data from other countries suggests that using only a NHW female database may result in overdiagnosis or underdiagnosis of osteoporosis in specific populations due to differences in skeletal characteristics among racial and ethnic groups.^(48–51) This requires further attention and additional prospective studies of BMD and facture risk in order to be able to make recommendations.

Another limitation of current studies is that the majority of bone research has been conducted in among broader racial and ethnic groups in the United States. This is despite the fact that this is a social construct created to categorize populations and there are known heterogeneity across populations, particularly Hispanic/Latino and Asian culture, for many factors such as sociodemographics, biological risk factors, and health behaviors in the United States.^(52,53) Therefore, aggregation of the broader populations likely masks meaningful differences in bone outcomes. It is also difficult to develop recommendations for prevention and treatment among racial and ethnic origin groups without studies showing differences and risk factors for each of these groups.

Last, the field needs to make methodological decisions with respect to measurement of outcomes, such as included DXA measurement sites, and significant risk factors such as age, body size, and lifestyle factors to adjust for in models. With respect to DXA site, the National Osteoporosis Risk Assessment Study (NORA) study measured BMD of the heel using single X-ray absorptiometry and of the finger using peripheral DXA,⁽¹⁹⁾ whereas the majority of other studies measured DXA of hip and spine sites, which can influence findings, particularly when compared by race and ethnicity. With respect to risk factors, studies have shown that adjusting for body weight explains some of the variation in BMD across racial and ethnic groups, including Hispanics,^(19,27) whereas adjusting for factors such as height and other anthropometric variables have shown to attenuate racial differences between NHW and Asian populations.⁽³⁷⁾ However, we have limited information on the effect of social determinants of health, lifestyle factors (eg, smoking, alcohol, physical activity), and comorbidities on bone health in racial and ethnic groups, further elucidating the need for longitudinal studies to develop effective public health strategies for improving bone health in diverse populations.⁽¹⁴⁾

Racial and Ethnic Differences in Fractures

Table 2 summarizes studies that have evaluated incidence or prevalence of fractures by race and ethnicity. Most studies have shown that fracture prevalence is lower in NHB compared to NHW adults.^{(19,39,55,57,59,60,62,66,67,69)} More recently, Wright and colleagues⁽⁵⁹⁾ showed that among NHB and NHW postmenopausal female Medicare beneficiaries who experienced fractures between 2010 and 2015, NHB postmenopausal women had significantly higher prevalence of femur and hip fractures than NHW postmenopausal women; however, radius/ulna and clinical vertebral fractures were significantly more prevalent in NHW than NHB women. However, earlier studies exclusively evaluating morphometric vertebral fractures have found no differences in prevalence between black and white women.^(54,56) To our knowledge, there are no studies comparing fracture prevalence among Hispanic and Asian populations to other race and ethnic groups.

Table 2.

Summary of Published Literature Evaluating Racial Differences in Fracture Prevalence, Incidence, and/or Trends over Time^a

Author; Year	Study design and data source	# Participants	Gender	Racial distribution and characteristics	Fracture sites and assessment	Conclusions
Fracture Prevalence
Mui, LW; 2003(54)	Cross-sectional; Jacobi Medical Center, NY, 200–2002	106	Female	Age: 65 (8) years Hispanic: n = 45 Black: n = 42 White: n = 9 Other: n = 10;	Vertebral Outpatient lateral chest X-rays graded using Genant’s semiquanitative (SQ) assessment.	Overall prevalence of moderate vertebral Fx: was 24%; Only 1% (1/106) had a severe vertebral Fx Prevalence by race: Hispanic: n = 27% Black: n = 21% White: n = 11% Other: n = 40% No significant difference in prevalence by race/ethnicity
Tracy, JK; 2006(55)	Cross-sectional; Baltimore Men’s Osteoporosis Study, 2001–2003	542	Male	White: n = 415 Black: n = 127	Vertebral Chest X-rays, graded using binary SQ grading scheme and further defined using vertebral dimensions	Overall prevalence (95% CI): 5.8% (4.0%, 8.3%). Prevalence by race: White: 7.3% Black: 0.9% (Fishers’ exact p = 0.01)
Vokes, TJ; 2007(56)	Cross-sectional; University of Chicago Medial Center BMD center, 2001–2005	521	Female	White: n = 345 Black: n = 176; Black women were significantly: Older Had more years since menopause Had less HRT use More steroid use Weighed more Fewer women diagnosed with OP at the hip or spine	Vertebral Vertebral Fx assessment, vertebrae a Fx were graded using Genant’s SQ scale.	Prevalence by race: White: 21% Black: 21% No significant difference in prevalence
Cauley JA; 2008(57)	Longitudinal cohort - SOF	8332	Female	White: n = 7860 Black: n = 472	Vertebral Vertebra was classified as having a prevalent fracture on the baseline radiograph if ratios were >3 SD below the trimmed normal mean for that vertebral level	White: 19.1% Black: 10.6% No comparison of Fx prevalence by race reported.
Lansdown, D; 2011(58)	Cross-sectional; first 600 MRNs from 2005 and 2006 from Chicago-based hospital	1011	Female	Caucasian: n = 238 African American = n = 773 no difference in age African American women were significantly more likely to have ESRD, smoke, have PCP at hospital Caucasian women were more likely to have cancer	Vertebral Chest X-rays graded using Genant’s SQ assessment	Prevalence by race: Caucasian: 13.0% African American: 10.4% No significant difference in prevalence (p = 0.26)
Wright, NC; 2020(59)	Longitudinal Cohort - Medicare Postmenopausal Osteoporosis (PMO) cohort, 2010–2015	399,000	Female	White: n = 387,832 Black: n = 11,168 No significant difference in age on PMO date Black women with Fx had: Slightly higher proportion of women in the 65–75 (youngest) age group higher proportion of women with a Charlson score 2+	Hip, pelvis, femur, radius/ulna, humerus, clinical vertebral ICD-9 &10 diagnosis code, HCPCS Fx repair codes, and imaging codes (vertebral only)	The prevalence of femur and hip Fx were higher in Black than White women [Femur: 9.0% versus 3.9%, p < 0.001; Hip: 30.7% versus 28.0%, p < 0.001] Prevalence of radius/ulna and clinical vertebral Fx was lower in Black than White women [Radius/ulna: 14.2% versus 17.0%, p < 0.001; Clinical Vertebral: 28.8% versus 33.5%, p < 0.001] No racial differences in prevalence of other Fx sites
Fracture Incidence
Baron, JA; 1996(60)	Longitudinal Cohort - Medicare 5% Sample, 1986–1990	94,672 Fx	Male and Female	No racial data breakdown of population provided Lower limb Fx: n = 59,630 Upper limb Fx: n = 35,042	Lower limb (hip, other parts of the femur, patella, ankle, other parts of the tibia/fibula, or pelvis), Upper limb (proximal humerus, other parts of the humerus, proximal radius/ulna, shaft of the radius/ulna, and the distal radius/ulna) ICD-9 diagnosis codes and CPT Procedure Codes	Rate Ratios (95% CI) by Fx site (Black versus White): Proximal Humerus – Men: 0.49 (0.29, 0.83) Women: 0.25 (0.18, 0.33) Distal Forearm – Men 0.45 (0.36, 0.57) Women: 0.30 (0.27, 0.33) Hip – Men: 0.60 (0.51, −0.70) Women: 0.38 (0.34, 0.42) Tibia/Fibula – Men: 1.09 (0.80, 1.49) Women: 0.72 (0.58, 0.88)
Lauderdale, DS; 1998(61)	Longitudinal; Medicare Beneficiaries, 1992–1993	897,786	Male and Female	50% sample of Hispanic and “Other” 10% sample of White, Black, Native American, and Unknown	Hip ICD-9 Codes	Yearly age-adjusted hip Fx rates per 1000 (data only provided in figure form) Men – White: ~4 Mexican American: ~4 Cuban American: ~3 Puerto Rican: ~2.2 Black: ~3 No significant difference in incidence by Hispanic origin Puerto Rican and Black men had significant lower incidence than White men No difference between Black and Hispanic men Women – White: ~10 Mexican American: ~9 Cuban American: ~7.8 Puerto Rican: ~6 Black: ~4 Puerto Ricans women had significantly lower incidence than Mexican and Cuban women No difference in hip Fx incidence between Mexican and Cuban women Puerto Rican and Black women had significant lower incidence than White women Black women had significantly lower incidence that White and Hispanic women
Fang, J; 2004b (62)	Longitudinal cohort - New York (NY) Statewide Hospitalization Data, 1988 – 2002		Male and Female		Hip ICD-9 Codes	Annual age-adjusted hip Fx incidence hospitalization rates per 100,000 Women – White: 459 Black: 137 Hispanic: 143 Asian: 174 Men – White: 230 Black: 109 Hispanic: 87 Asian: 104 For both genders, and all race/ethnicity groups, the rates increased sharply with age
Barrett-Connor, E; 2005(62)	Longitudinal cohort - NORA	197,848	Female	White: n = 179,470 Black: n = 7784 Asian: n = 1912 Hispanic: n = 6973 Native American: n = 1708; lower proportion of Native American women in younger age group Higher proportion of Native American women in 80+ age group	Clinical hip, spine, forearm, wrist, or rib Self-report	All Fx - p < 0.001 White: 1.5% Black: 0.8% Asian: 0.7% Hispanic: 1.7% Native American: 1.8% Hip Fx - p = 0.217 White:0.3% Black:0.2% Asian: 0.2% Hispanic: 0.4% Native American: 0.5% Wrist/arm - p = 0.006 White: 0.7% Black: 0.5% Asian:0.2% Hispanic: 0.7% Native American: 1.2% Fx Rates by Age Black and Asian women had thelowest Fx rates within each age group Black, White, and Hispanic women 80 years of age had markedlyhigher Fx rates than women in younger age groups Age effect was less obvious among the oldest Asian and Native American women
Cauley JA, 2005(63)	Longitudinal cohort - SOF	7970	Female	White: n = 7334 Black: n = 636 Black women were: Older Higher BMI Less likely to report a Fx since age 50 Lower current use of calcium supplements	Non-vertebral Fx Self-report and adjudicated by radiographic report	Incidence by site over 6.1 years of follow-up: Hip or pelvis, Black: 12 (19.7%) White: 347 (20.3%) Rib Black: 8 (13.1%) White: 141 (8.2%) Wrist Black: 7 (11.5%) White: 306 (17.9%) Arm, shoulder, or elbow Black: 5 (8.2%) White: 247 (14.4%) Leg, ankle, or knee Black: 12 (19.6%) White: 248 (14.5%) No comparison of the incidence by race reported
Cauley JA, 2007(64)	Longitudinal cohort - WHI	159,579 Observational Study: n = 92,368 Clinical Trials: n = 67,211	Female	White - not of Hispanic origin Black - not of Hispanic origin Hispanics/Latinos - Mexican, Cuban, Puerto Rican, Central American, or South American Asians or Pacific Islanders - Chinese, Indo-Chinese, Korean, Japanese, Pacific Islander, Vietnamese American Indian - including Alaskan Native Other Did not provide numeric breakdown in manuscript	All Fx, except for fingers, toes, face, skull, or sternum. Self-report followed by local and central review of radiology reports confirmed hip Fx. Other Fx sites were confirmed by radiographic report in CT women.	Incidence of Fx over average 8.0 years of follow-up: White: 21,083 (15.8%) Black: 1112 (7.6%) Hispanic: 580 (8.9%) Asian: 391 (9.3%) American Indian: 164 (14.6%) Annualized (%) rate of Fx: White: 2.0 Black: 0.9 Hispanic: 1.3 Asian: 1.2 American Indian: 2.0
Mackey DC, 2007(65)	Longitudinal cohort - Health ABC	1446	Male and Female	Black men: n = 268 White men: n = 435 Black women: n = 374 White women: n = 369	Non-spine Fx, except for fingers, toes, face, skull Self-report at annual clinic visits and at intermediate 6-month interviews and confirmed by medical documentation	Total Fx incidence after 6.4 years of follow-up Overall population: 16.6 (13.9, 19.2) per 1000 person-years Black men: 5.8 (2.4, 9.2) White men: 10.9 (7.0, 14.8) Black women: 16.8 (11.4, 22.2) White women: 31.7 (24.2, 39.3) No difference in incidence of Fx between men; however, White women had significantly higher Fx rate.
Pressley, JC; 2011(66)	Longitudinal cohort - New York (NY) Statewide Hospitalization Data, 2000 – 2002	158,351	Male and Female	White: n = 138,763 Black: n = 19,588 Blacks were: Younger More frequently male Less likely to be insured More likely to be on Medicaid	Skull, facial, vertebral, hip, all lower extremity, all upper extremity ICD-9 diagnosis codes	Overall: Blacks exhibit a higher cumulative incidence of Fx through age 50. During the 50-year to 54-year age interval, a crossover occurs and White incidence exceeds that of Blacks By Race and Gender: A racial crossover is observed for Black and White males around age 62. Among women, incidence rates diverge around age 46, with total Fx incidence rate being higher and occurs at earlier ages in Whites than Blacks. In Black women, increases were more moderate and lag those observed for White women by 5 years to 10 years. Both White and Black adult men experience a plateau in Fx incidence during their adult years, while their similarly aged female counterparts demonstrate a gradual increase in Fx incidence rate with age
Khandewal, S; 2012(30)	Longitudinal cohort - Kaiser Permanente Northern California (CA), 1997–2003	7184	Female	South Asian (Indian, Pakistani, Sri Lankan): n = 449 Chinese: n = 2245 White: n = 4490 No difference in age due to matching Prior Fx after age 45 – Whites: 9.6% South Asian: 7.1% Chinese: 4.5% Osteoporosis based on DXA – Chinese: 11.9% South Asian: 8.9% White: 6.5%	Hip, wrist, and humerus ICD-9 diagnosis codes	Incidence Rates (95% CI) per 100,000 person-years Hip – South Asian: 0 (0.0, 104.4) Chinese: 45.0 (19.4, 88.7) White: 86.9 (59.0, 123.3) Wrist – South Asian: 286.5 (137.4, 5.26.8) Chinese: 129.9 (82.3, 194.9) White: 303.0 (248.4, 355.2) Humerus - South Asian: 85.2 (17.6, 249) Chinese: 56.3 (27.0, 103.5) White: 197.1 (153, 249.1) All Fx - South Asian: 373.8 (199.1, 639.3) Chinese: 232.2 (166.6, 315.0) White: 571.8 (495.3, 656.8)
Lo, JC; 2014(67)	Longitudinal cohort - Kaiser Permanente Northern CA, 2006–2012	10,948	Female	Hip Fx racial distribution: White: 81.2% Hispanic: 6.5% Asian: 5.4% Other: 3.9% Black: 2.9% Diaphyseal Fx racial distribution: White: 47.7% Asian: 34.7% Hispanic: 8.3% Other: 4.7% Black: 4.7%	Hip and diaphyseal Fx ICD-9 diagnosis codes	Age-adjusted incidence of hip Fx per 100,000 (p < 0.05 for all comparisons) White: 288 Hispanic: 198 Asian: 148 Black: 87 Age-adjusted incidence (per 100,000) of diaphyseal Fx (p < 0.001 comparing Asians to other race/ethnicities, with no significant differences among other race and ethnic groups) Asian: 27 Black: 10 Hispanic: 6 White: 5
Hip Fracture Trends
Zingmond, DS; 2004(39)	Longitudinal assessment of annual CA’s hospital Patient Discharge Database, 1983–2000		Male and Female	85+ - Women: ~35% Men: ~29% Mean Charlson score: Women: ~0.8 Men: ~1.1	Hip ICD-9 diagnosis codes and surgical hip repairs codes including hip arthroplasty and/or bone pinning	Change in annual hip Fx rates by sex and race/ethnicity: Women: NH-White: −1.0 (01.5, −0.6)* Black: 0.3 (−0.1, 0.7) Hispanic: 2.6 (2.2, 3.1)* Asian: −0.3 (−0.8, 0.2) Men: NH-White: 0.1 (0.0, 0.3) Black: 0.0 (−0.3, 0.4) Hispanic: 0.8 (0.7, 1.1)* Asian 0.0 (−0.2, 0.3) By race and age: Significant decrease in annual hip Fx rates in NH-Whites 55+ men and women 55–64 men and women 65–84 women Significant increase in rates in Hispanic: 55+ men and women 55–64 men 65–84 men and women 85+ men and women Significant decrease in rates in Asian: 55–64 women 85+ women
Wright, NC; 2012(68)	Cohort, 5% Medicare Sample	109,072	Male and Female	Overall: White: 88.3% Black: 8% Asian: 1.4%, Hispanic: 1.6% Hip Fx: Women – White: 94.6% Black: 3.5% Asian: 0.8% Hispanic: 1.1% Men- White: 93.4% Black: 4.6% Asian: 0.7% Hispanic: 2.0%	Hip ICD-9 diagnosis and HCPCS repair codes	Average change in age-adjusted hip Fx incidence rate (per 100,000 person-years) from 2000–2001 to 2009–2009: Women – White: −11.6 (−22.8, −0.4) Black: −10.7 (−18.5, −2.9) Asian: −11.4 (031.3, 8.5) Hispanic: −2.9 (−25.8, 20.1) Men – White: −3.6 (−6.6, −0.5) Black: −4.3 (−14.7, 6.1) Asian: −7.1 (−31.2, 17.0) Hispanic: 2.8 (−7.6, 13.1) Greater declines in IR for those 75+ years, White men and women Black women No significant change in hip Fx IR in those <75 years by race and ethnicity.
Lo, JC; 2014(67)	Longitudinal cohort - Kaiser Permanente Northern CA, 2006–2012	10,948	Female	Hip Fx racial distribution: White: 81.2% Hispanic: 6.5% Asian: 5.4% Other: 3.9% Black: 2.9% Diaphyseal Fx racial distribution: White: 47.7% Asian: 34.7% Hispanic: 8.3% Other: 4.7% Black: 4.7%	Hip and diaphyseal Fx ICD-9 diagnosis codes	Hip Fx incidence rates significantly declined in White and Black women 50+ years Hip Fx rates plateaued in 2011 for Hispanic and Asian women. Diaphyseal Fx incidence rates increased slightly for all race and ethnic groups Significant increases observed starting in 2008 for Asian women, and 2010 for Hispanic women, and 2011 for White and Black women
Sullivan, KJ; 2016(69)	Population based review of CA’s non-federal hospital admissions, 2000–2011	317,677		White: 78.8% Hispanic: 9.8% Black: 2.8% Asian: 5.3% Native American: 0.1% Other: 1.5%	Hip ICD-9 procedure codes for the treatment of hip Fx	Overall - Hip Fx rates significantly decreased over time. Whites had the highest incidence rates of hip Fx and observed significant declines over time. Rates in other race and ethnic groups appeared to have plateaued around 2009, with increases (though probably not significant) observed in Native American women and men

BMD = bone mineral density; DXA = dual energy X-ray absorptiometry; MRN = medical record number; HRT = hormone replacement therapy; OP = osteoporosis; ESRD = end-stage renal disease; PCP = primary care provider; ICD = International Classification of Disease; HCPCS = Healthcare Common Procedure Coding System; CPT = Current Procedural Terminology; NH = non-Hispanic; Fx = fracture; SOF = Study of Osteoporotic Fractures; NORA = National Osteoporosis Risk Assessment Study; WHI = Women’s Health Initiative; Health ABC = Health Aging and Body Composition Study.

^aRace and ethnicity as reported in the manuscript.

^bData derived from abstract.

^*p < 0.01.

Few studies have evaluated fracture incidence rates by race and ethnicity.^{(19,30,39,54–69)} As summarized in Table 2, using New York State inpatient hip fracture data from 1988 to 2002, Fang and colleagues⁽⁶²⁾ reported that the annual age-adjusted hip fracture incidence rate per 100,000 was 459, 137, 143, and 174, in NHW, NHB, Hispanic, and Asian women, respectively; and 230, 109, 87, and 104, in NHW, NHB, Hispanic, and Asian men, respectively. The National Osteoporosis Risk Assessment (NORA) study showed similar fracture rates per 100 person-years for Hispanic and NHW women aged 50–59 and 60–69 years. However, fracture rates diverged between NHW and Hispanic women as age increased, with slightly higher rates observed in Hispanic women aged 70–79 years, and significantly higher fracture rates in Hispanic women ≥80 years compared to their NHW women counterparts.⁽¹⁹⁾ Black and Asian women had significantly lower fracture rates than NHW and Hispanic women at all age groups, especially significantly lower incidence of wrist/arm fractures in Asian women (Table 2).⁽¹⁹⁾

Very few studies have examined fracture for the individual Hispanic origin groups (Table 2). Medicare data collected from 1992–1994 showed that Mexican American men and women had the highest yearly age-adjusted hip fractures rate, followed by Cuban, and then Puerto Rican beneficiaries.⁽⁶¹⁾ Though only portrayed graphically with confidence intervals, data indicated that there were no significant differences in hip fracture incidence by Hispanic origin among the men; however, Puerto Rican and black men had significant lower incidence than white men. Among women, the Puerto Ricans women had significantly lower incidence than their Mexican and Cuban counterparts, whereas there were no differences in incidence between Mexican and Cuban women (Table 2).⁽⁶¹⁾

Similarly, few studies have evaluated fracture rates by for Asian groups by origin. Although there have been studies that have compared US Asians to their international counterparts,^(70–72) to our knowledge, only one study examined incidence rates for Asian groups by origin (Table 2). Using Kaiser Permanente Northern California (KPNC) data, Khandewal and colleagues⁽³⁰⁾ found that South Asian women (Indian, Pakistani, Sri Lankan) women had a significantly higher wrist fracture incidence rates per 100,000 (95% confidence interval [CI]) person-years than that of Chinese women (South Asian: 286.5 [95% CI, 137.4–5.26.8] versus Chinese: 129.9 [95% CI, 82.3–194.9], p < 0.05); higher (albeit not significant) incidence rates for any fracture (South Asian: 373.8 [95% CI, 199.1–639.3] versus Chinese: 232.2 [95% CI, 166.6–315.0]) and humerus fracture (South Asian: 85.2 [95% CI, 17.6–249] versus Chinese: 56.3 [95% CI, 27.0–103.5]); whereas Chinese women had higher, again not significant, hip fracture incidence rates than South Asian women (Chinese: 45.0 [95% CI, 19.4–88.7] versus South Asian: 0 [95% CI, 0.0–104.4]).

Trends in hip fracture incidence rates

Although trends in incidence rates of hip fractures have been widely evaluated since the advent of bisphosphonates, only a handful of studies have examined trends by race and ethnicity. As summarized in Table 2, hip fracture rates declined over time in all studies reviewed, with significant declines consistently being observed in NHWs and NHBs.^(39,67–69) Hip fracture trends in the Hispanic population have been inconsistent; some studies have shown a decreasing rate of hip fractures,⁽⁶⁸⁾ whereas others have shown significant increasing annual hip fracture rate ranging from 0.1 (per 100,000 population year (pys)) in the 55–64 age group, to 55.9 (per 100,000 pys) in those ≥85 years.⁽³⁹⁾ A more recent study using 2000 to 2011 California statewide data, did not quantify rate of decline, but visually portrayed a slight increase in rates between 2008 and 2009 in both NHW and Hispanic populations; but after 2009, NHW women experienced a decline in rates, whereas rates appeared to stabilize in Hispanic women from 2009 to 2011.⁽⁶⁹⁾

Overall, trend data in the Asian population has shown a decline in rates,^(39,68) but often had small sample sizes that were underpowered to observe significant associations. In the 1983–2000 California state data, the largest decrease in annual hip fracture rates of any race and age group was among Asian women ≥85 years⁽³⁹⁾; however, in more recent California state data, Asian adults experienced a slight increase in hip fracture incidence between 2009 and 2011.⁽⁶⁹⁾ Of interest, a study utilizing 2006–2012 KPNC data evaluated hip and diaphyseal femur fracture, one of the fractures used to identify atypical femur fractures in administrative data, by race and ethnicity among female beneficiaries. The authors reported age-adjusted incidence of hip fracture was highest among NHW, followed by Hispanic, Asian, and NHB women. For diaphyseal fractures, however, age-adjusted incidence rates (per 100,000 pys) were highest in Asian (27), followed by NHB (10), Hispanic (6), and NHW (5) women, with the incidence rate being significantly higher among Asian women.⁽⁶⁷⁾

Overall, data show that NHW adults have the highest burden associated with fragility fractures, NHB have the lowest, with Hispanic and Asian populations somewhere in the middle. Differences in BMD, geometry, and microarchitecture could potentially explain these differences in fracture incidence and trends. The data also show that national heritage plays a role in fracture rates. Other than the in NORA and SWAN studies,^(73,74) using a cohort of premenopausal women just coming of age for significant number of fracture outcomes, current longitudinal studies providing fracture incidence information are from majority NHW cohorts (eg, Study of Osteoporotic Fractures [SOF], Osteoporotic Fractures in Men [MrOS], Framingham Heart Study [FHS], Women’s Health Initiative [WHI]), supporting the need for multiracial and ethnic longitudinal studies, particularly with the ability to evaluate by country of origin rather than by the global race and/or ethnicity categories. Data also show that fracture experience over time has not been equal in all race and ethnic groups, further highlighting the need for studies to understand factors associated with fractures by race and ethnicity (Table 2).

Racial and Ethnic Differences in Fracture Outcomes

Several national and international studies have shown that on average, 25% of hip fracture patients die within 1 year of fracture.⁽⁸⁾ It has been documented as early as 1964, that NHB men and women having significantly higher mortality following a hip fracture than their NHW counterparts.^(91–94) There are fewer studies evaluating differences in postfracture mortality among Hispanic and Asian and NHW populations. For example, being non-white and NHB was associated with 60% higher odds of mortality compared with being NHW.⁽⁹⁵⁾ However, other studies have found that after adjusting for potential confounders, Hispanics and Asians have 15% and 35% lower odds of mortality 1 year after hip fracture compared to NHW.^(96–98) To our knowledge, only one study evaluated 1-year mortality within Asians by origin and found no significant differences between Chinese (14.0%), Japanese (15.0%), and Filipina (14.3%) women.⁽⁹⁹⁾

NHB women have also been shown on average a 49% and 225% higher debility (long-term nursing home placement), and destitution (becoming Medicaid-dependent), respectively, than their NHW counterparts 1 year following major fragility fracture,⁽⁵⁹⁾ further highlighting disparities in fracture outcomes in NHB women. Many studies have also shown that people of color have higher rates of being discharged without rehabilitative care (eg, physical therapy) after fracture.^(100,101) Other studies have shown that compared to NHWs, NHB (odds ratio [OR] 1.3–2.0),^(102–104) Hispanic (OR 1.5–3.2)^(102–104) and Asian adults (OR 1.4–2.1)^(102,103) had higher odds of being discharged home following acute hospitalization due to hip fracture. Although causal inferences cannot be determined, the data show that compared to NHWs, postfracture outcomes such as mortality in NHB, Hispanic, and Asian adults are worse. There are also differentials in outcomes related to quality of life such as postfracture rehabilitation, which could be from biases at the provider or systems level, or lack of adherence by the patient, based on knowledge and/or cultural practices. We need to understand where breakdowns are occurring in order to develop potential interventions that could mitigate these disparities.

Racial and Ethnic Differences in Osteoporosis Management

DXA screening

Primary fracture prevention and reduction of disparities in fracture outcomes require significant improvements in osteoporosis management. Osteoporosis management includes screening and appropriate treatment plans based on risk levels. Central DXA scans remain the gold standard for screening and clinically diagnosing osteoporosis.⁽¹⁰⁵⁾ Screening for osteoporosis has significant benefits in reducing fracture risk through the implementation of early lifestyle modification and/or pharmacological treatment.⁽¹⁰⁶⁾ Guidelines for osteoporosis screening vary across organizations.⁽¹⁰⁷⁾ For example, the National Osteoporosis Foundation (NOF) states that BMD screening is recommended for any person who has sustained a hip or clinical vertebral fracture, women ≥65 years and men aged ≥70 years, or individuals over the age of 50 years with risk factors (eg, frequent falls, low body mass index, low calcium intake, vitamin D insufficiency),⁽¹⁰⁸⁾ whereas, the US Preventive Services Task Force BMD screening recommendations include: (i) women aged ≥65 years and (ii) women <65 years who have been through menopause and are at increased risk for osteoporosis as assessed by clinical tools to identify fracture risk assessment (eg, Fracture Risk Assessment Tool [FRAX]).⁽⁸⁾

Unfortunately screening for osteoporosis has been shown to be suboptimal across all population groups.^(75,77,86) The Healthcare Effectiveness Data and Information Set (HEDIS) osteoporosis screening measures, developed by the National Committee for Quality Assurance (NCQA) to objectively measure, report, and compare quality across health plans, reported that 76% of women in Medicare received a BMD test in 2017.⁽¹⁰⁹⁾ However, a retrospective analysis of insurance claims data within a 2-year continuous enrollment period indicated that despite universal screening recommendations for women aged 65–79 years, only 26.5% underwent screening, further highlighting gaps in osteoporosis management (Table 3).⁽⁸⁶⁾

Table 3.

Summary of Published Literature Evaluating Osteoporosis Screening and Treatment in the U.S (publication period 2005–2019)^a

Author; Year	Study design and data source	# Participants	Sample characteristics	Screening and treatment rate	Conclusions
Osteoporosis Screening and Treatment
Wilkins CH; 2004(75)	Cross sectional/ Retrospective; Washington University Medical Center chart review, 2000	252	African American women: n = 214 Caucasian women: n = 35 other women: n = 3	Only 11.5% had received OP screening Statistically significant difference in breast cancer screening versus OP screening in this population, especially for African American across all age groups 39% had Vit D and Calcium intake discussions with provider	Ethnicity should not be considered exclusion criteria for osteoporosis screening Low BMD occurred at considerable rates in female African American patients
Miller RG; 2005(76)	Cross sectional/Retrospective medical chart review, 2000	205	Women ages ≥65 years weighing ≤127 pounds African American women: n = 103 White women: n = 102	43% of at risk women referred to DXA Significantly less AA women (32%) referred for DXA compared to White women (55%) No racial difference in compliance for those who were referred Physicians less likely to discuss calcium and vitamin D supplementation and mention osteoporosis in medical record with AA patients	At risk AA women are significantly less likely to be referred for DXA screening compared to at risk White women
Neuner JM; 2006(77)	Cross sectional/ Retrospective; Medicare administrative claims: Florida, Illinois, and New York, 1999–2001	43,802	Female medicare recipients aged 66–90 White women: n = 36,032 Black women: n = 3443 Asian women: n = 1927 Hispanics women: n = 1121 Native American: n= 114 Other: n = 1065	22.9% were screened for Osteoporosis during study period Bone density tests decreased with increased age; 27.2% of patients ≤70 years, 25.6% of those aged 71–75 years, and −4 to 6% for each 5 year increment after 75 years Bone density testing more common in White and Asian woman compared to other race and ethnicities	Bone density testing rates decreased with advanced age, despite those groups having higher risk of fracture
Hamrick I; 2006(78)	Cross sectional/ Retrospective; 1998–2002	531	Women ages 50+ years Caucasian women: n = 452 African American women: n = 79	Only 14.5% of African American women received DXA screening versus 82.8% of Caucasian women, despite making up 45.9% of primary care clinic population Prevalence of osteoporosis was found to be similar for both populations, 20.1% in Caucasian women & 21.5% in African American women	Unexplained disparities exist in BMD screening between Caucasian and African American women 50+ years in this primary care facility
Neuner JM; 2007(79)	Retrospective cohort study; Florida, Illinois, and New York 2001–2003	35,681	Female Medicare recipients 65–89 years w/ hip fractures during study period White women: n = 34,039 Black women: n = 1044 Hispanics women: n = 598	In total, 20.7% underwent BMD testing in the 2 years prior to FRAX and another 6.2% underwent testing in the 6 months after When compared to White women, Black women were 48%, and Hispanic women 34% less likely to undergo screening before FRAX and remained less likely (RR 0.66 [0.50, 0.88] and 0.58 [0.39, 0.87], respectively) to undergo testing after fracture	Racial disparities in OP screening persisted even after Medicare patients experienced fractures
Curtis JR; 2009(80)	Cross-sectional; The REasons for Geographic And Racial Differences in Stroke (REGARDS), 2003–2007	24,783	Men and women aged ≥45 years African American: n = 9825 Caucasian: n = 14,958	Estimated fracture risk highest in Caucasian women Among highest risk group (Caucasian women with 10 years hip fracture risk), only 26% received treatment Likelihood of receipt of osteoporosis medications was lower among African Americans when compared with Caucasians [odds ratio (OR) = 0.44, 95% confidence interval (CI) 0.37, 0.53] and for men compared to women (OR = 0.08, 95% CI 0.06–0.10), even after adjustments for socioeconomic and risk factors for fracture	A notable gap exists between 2008 NOF treatment guidelines and use of osteoporosis prescription medication, this gap was further noted among African Americans and men
Navarro RA; 2011(81)	Observational Study; Kaiser Permanente Southern California (CA), 2008–2009	13, 412	Enrollees aged 60+ years who experienced fragility fracture For women: White = 64% Hispanic = 14% African American = 11.9% Asian = 10% For men: White = 64% Hispanic = 16.6% African American = 9.6% Asian = 9.8%	After fragility fracture, women more likely to be treated than men (92.1% versus 75.2%) Before fracture, osteoporosis treatment was more likely in White men and women and Hispanic men when compared to other groups Treatment rate after fragility fracture did not differ greatly among racial/ethnic groups in either sex (women 87.4–93.4% and men 69.3–76.7%)	In this population, there was low variation in osteoporosis care after fragility fracture across racial/ ethnic groups when using EMR data
Gordon NP; 2012(82)	Cross-sectional; Kaiser Permanente Northern California (CA)	16,049	Survey respondents aged 25–85 non-Hispanic White: n = 52% Black: n = 7% Latino/a: n = 15.5% Filipino/a: n = 7.6% Chinese: n = 5.6%	Of those <50 years, 40% of women and 54% of men get no vitamin D from dietary supplements Of those >50+ years 24% of women and 53% of men get no vitamin D from dietary supplements Black and Latina women aged 25–85 years and Filipina women ≤ 50 were significantly less likely to be getting vitamin D from dietary supplements, compared to NHW	In 2008, a large percentage of women and even larger percentage of men on a Northern California health plan, did not obtain vitamin D from dietary supplements Black, Latino, and obese adults at higher risk of Vitamin D deficiency are the least likely to get any vitamin D from dietary supplements.
Hamrick I; 2012(83)	Longitudinal; University of Wisconsin primary care clinics, 1998–2009	1, 000	Women aged 60+ years African American women: n = 500 Caucasian women: n = 500	29.8% of AA women referred to DXA compared to 38.4% Caucasian AA women less likely to receive medication compared to Caucasian women	Despite eligibility many women are not being screened and treated for OP in primary care, and even fewer AA women
Powell H; 2012(84)	Retrospective EMR review; University of Washington Medical Center, 2006–2008	1363	Women aged ≥65	Adherence to screening recommendations varied from 33–100% for individual providers 70% of women had a DEXA study but	In women aged ≥65 years, good levels of adherence to the USPSTF 2002 Guidelines for osteoporosis screening were found in a large urban academic medical center Significantly different screening rates were found and varied by site and not provider gender
Liu SK; 2013(85)	Retrospective observational cohort study; U.S. Medicare administrative inpatient, outpatient (2003–2010) and prescription (2006–2010) data	61,832	Medicare patients 68+ years who have had at least 1 hip, radius, or humerus fracture White: n = 88.5% Black: n = 2.6%	Proportion of those receiving care increased from 2006–2009 Black race, male sex, and an upper extremity fracture (versus hip) was associated with significantly lower likelihood of receiving testing, pharmacotherapy, or both	Post fracture osteoporotic care was uncommon, but especially in Black and male participants Care increased over time but incidence of fracture remained inadequate in initiating secondary prevention
Gillespie CW; 2017(86)	Retrospective Longitudinal Study; OptumLabs Data Warehouse -Medicare Advantage enrollees data, 2008–2014	1,638,454	Women ages ≥50 years living in the United states with no prior history of OP diagnosis, OP drug use, or hip fracture non-Hispanic White: n = 1,178,543 non-Hispanic Black: n = 174, 756 non-Hispanic Asian: n = 45, 102 Hispanic/Latino: n = 112, 781 Other/ not reported: n = 127,272	Women who initiated some form of osteoporosis therapy, underwent screening (78.7%), but only 24.2% of women experiencing hip fracture underwent screening During study period, few women ages 65+ years underwent screening (26.5% and 12.8% among women ages 65–79 and 80+ years, respectively) Hispanic and Non-Hispanic Asian women in the 50–79 age categories had the highest odds of screening Compared with other racial/ethnic groups assessed, NHB women across every age group was less likely to receive screening (18.2%)	Vast deficiencies in osteoporosis screening exist in women 65+ years despite notable changes to utilization of osteoporosis screening among women 50–64 years and 80+ years There is a disconnect between evidence-based guidelines for osteoporosis screening versus real world implementation
Rooney MR; 2018(87)	Retrospective/ Cross-sectional; NHANES analysis, 1999–2014	42,038	Adults aged 20+ years 50% female non-Hispanic White: 46.9%	Calcium supplementation usually 3–4 fold in women versus men Between 1999–2014, Calcium intake ≥EAR wavered, increasing until 2007–2012 and later decreasing Calcium supplementation rates higher among NHW when compared to NHB, Hispanic, and Asian American	Only ~5% of U.S adults meet daily calcium requirements from supplemental calcium Calcium supplementation most common in women, NHW and older adults
Nayak S; 2018(88)	Review of 43 RCTs, 2004–2017	46–13,455	Men and women that participated in RCTs that evaluated an osteoporosis quality improvement strategy	Fracture liaison service/case management, multifaceted interventions targeting providers and patients, orthopedic surgeon or fracture clinic initiation of osteoporosis evaluation or management, and patient education and/or activation were noted as efficient strategies for increasing osteoporosis screening and treatment	For those with recent or prior fracture, several types of interventions were found to be efficacious for increasing osteoporosis treatment and or/BMD testing Summary estimates of risk difference were commonly larger for the outcome of BMD/DXA testing than for osteoporosis treatment
Noel SE; 2019(89)	Cross Sectional	45	Caribbean Latino adults aged ≥ 50 years	90% of participants had heard of osteoporosis but most were not able to accurately describe the condition Despite reports of feeling rushed during visits, healthcare providers were perceived as most trustworthy source for health information	Despite having some knowledge of osteoporosis and importance of nutrition for prevention, Caribbean Hispanic adults experienced confusion between osteoporosis and other bone conditions Culturally specific interventions are of critical necessity for this population
Weaver FM; 2019(90)	National retrospective cohort study; United States Veterans Health Administration (VA) national databases. 2005–2015	11,048	Veterans with traumatic SCI who received VA health care Female: n = 299 Male: n = 10,749	During study period, less than 10% received pharmacological therapy for osteoporosis Of those who received pharmacological therapy for osteoporosis, they were more likely to be older (p = 0.0009), female (p < 0.0001), have a lower BMI (p < 0.0001), have a paraplegic-level injury (p = 0.0008) and have a complete injury (p < 0.0001) Significant decline in the number of filled prescriptions for osteoporosis medications; 13.0% in 2005, 10% by 2010, and only 2.2% in 2015 (Z = −6.91; p < .0001)	Pharmacological therapies for osteoporosis are not commonly prescribed for Veterans with a SCI Among those with SCI, Clinicians rely on factors that predict fracture risk in this population (information from DXA measurements, prior fracture history and use of medications associated with bone loss) in addition to clinical risk factors, to target treatments with pharmacological therapies for osteoporosis

BMD = bone mineral density; DXA = dual-energy X-ray absorptiometry; EMR = electronic medical record; NHANES = The National Health and Nutrition Examination Survey; NHB = non-Hispanic black; NHW = non-Hispanic white; SCI = spinal cord injury; VA = Veteran’s Administration.

^aRace and ethnicity as reported in the manuscript.

In addition to inadequate implementation of screening for osteoporosis in the general population, certain groups are even less likely to receive screening. NHB women, in particular, are least likely to receive screening compared with NHW adults.^{(76,78,79,83)} For example, Hamrick and colleagues⁽⁷⁸⁾ found that 14.5% of NHB women were screened versus 45.9% of NHW women. Another study found that NHB and Hispanic women were 48% (relative risk [RR] 0.52; 95% CI, 0.43–0.62) and 34% (RR 0.66; 95% CI, 0.54–0.80) less likely to receive bone density testing prior to fracture than NHW women, respectively, over 2 years.⁽⁷⁹⁾ These discrepancies persisted postfracture. NHB women are less likely to be referred for osteoporosis screening or to have had a DXA scan compared to their NHW counterparts from the same clinic.^(75,76,78) An interesting study showed that between 2008 and 2014, although screening disparities based on socioeconomic status tapered, gaps in osteoporosis screening persisted for those of advanced age and those who were NHB.⁽⁸⁶⁾

Unfortunately, data on screening for osteoporosis among Hispanic and Asian populations compared to NHW populations is lacking. In the most recent study, Hispanic and Asian women aged 50–64 years and 65–79 years were more likely to be screened for osteoporosis than NHW women.⁽⁸⁶⁾ This is in contrast to an older study of female Medicare recipients aged 66–99 years that showed less screening for osteoporosis among Hispanic women (23% lower).⁽⁷⁷⁾ Prior to hip fracture, NHB and Asian adults have been found to have lower rates of screening when compared with NHW and Hispanic adults.^(77,81) Based on the limited evidence for Hispanic and Asian populations, and the lack of screening information among the individual origin groups, it is difficult to draw conclusions on the current state of screening practices in these populations. A mixed method study of Caribbean origin (those identifying as Puerto Rican, Dominican, or Cuban) adults aged ≥50 years reported that most had not communicated with their doctors about osteoporosis, despite their medical care providers being viewed as their most reliable and trustworthy sources of health information.⁽⁸⁹⁾ Thus, there is a need for more quantitative and qualitative data to understand differences in screening rates and barriers and facilitators to screening in diverse populations, which may be specific to individual communities based on available resources.⁽⁷⁹⁾

Likely in part to its lower prevalence, osteoporosis in men has been understudied in the past. Studies such as MrOS have increased our knowledge on bone health in men, but we still observe lower screening and treatment provided for men, even in older age.⁽¹¹⁰⁾ Unfortunately, there are little data on screening rates by race and ethnicity in men, identifying another need in the field. Differences in screening may be due to sex of providers, such that female providers are more likely to refer patients for screening compared with male providers.^(83,84) A recent systematic review and meta-analysis identified several strategies for increasing screening through BMD/DXA testing for those with a fracture such as fracture liaison service/case management interventions and other intervention, including patient and provider education, notification, and reminders⁽⁸⁸⁾; however, it remains unclear how effective these strategies would be among diverse populations.

Fracture risk prediction

It is well documented that up to 50% of fractures occur in people without DXA-defined osteoporosis^(111–113); thus, it is imperative to identify individuals at risk of fracture based on BMD and other clinical risk factors. FRAX^® is the most widely used fracture risk prediction tool in the United States; it predicts the 10-year risk of hip and major osteoporotic fractures based on age, sex, smoking and drinking status, race/ethnicity, parental history of fractures, and certain medications and medical conditions (eg, previous fractures, rheumatoid arthritis, and conditions requiring long-term steroid use).⁽¹¹⁴⁾ Like discussions regarding the reference population for determining the prevalence of osteoporosis using DXA, questions have recently emerged regarding the appropriateness of having a race and ethnicity correction in the FRAX tool.⁽¹¹⁵⁾ It is clear that there are racial differences in bone density and structure that need to be accounted for with respect to fracture risk; however, given the lack of longitudinal studies on risk factors for fractures by race and ethnicity, there is potential room for improvement in the tool, but it is out of the scope of this work to discuss the appropriateness of such a race correction and/or the value assigned to the correction.

Osteoporosis treatment

Based on BMD and fracture risk, treatment plans can range from dietary modification (calcium and vitamin D supplementation) and encouraging physical activity for those at lower risk, to pharmacological therapies for those at higher risk. Given that individuals of color, particularly NHB, are less likely to be screened for osteoporosis and are perceived to be at lower fracture risk, there are racial disparities throughout the treatment plan spectrum. However, given the lack of longitudinal studies on risk factors for fractures by race and ethnicity, more studies are needed to better understand racial and ethnic differences use of nonpharmacological osteoporosis therapies.

With respect to pharmacologic therapy, racial differences in treatment rates have been reported, but to our knowledge there are no data on racial differences in treatment regimens or adherence. Several studies have reported that osteoporosis treatment is lower in NHB women.^(80,81,85) In the Reasons for Geographical and Racial Differences in Stroke study, 13.6% and 8.4%, respectively, of NHS and NHB women who were considered at high risk of fracture based on FRAX score had prescriptions for osteoporosis medications at their baseline in-home interview, and being NHB was associated with 64% lower odds of osteoporosis medications compared to NHW (OR 0.36; 95% CI, 0.31–0.42) after adjusting for age, sex, and FRAX score.⁽⁸⁰⁾ Using 2006–2010 Medicare data, NHB women who sustained a hip fracture were significantly 18% less likely to have claims for osteoporosis medication following the fracture than NHW women.⁽⁸⁵⁾ In a comprehensive study of the Kaiser Permanente Southern California (KPSC) Healthy Bones Model of Care program, an interdisciplinary osteoporosis prevention and management program started in 2000, authors evaluated racial differences in the proportion of patients who used osteoporosis medications overall and the proportion of hip fracture patients who used medications from 2008–2009.⁽⁸¹⁾ In women, overall treatment rates ranged from 87.4% in NHB women to 93.4% in NHW women, with no significant difference in treatment rate by race/ethnicity.⁽⁸¹⁾ However, among women who sustained a hip fracture, only 54.9% of NHB women received therapy, compared to 63.9%, 72.0%, and 75.6% Asian, Hispanic, and NHW women, with a significant difference between NHB and NHW women.⁽⁸¹⁾ The overall proportion of osteoporosis medication use in men was lower (69.3–76.7%), but again not significantly different by race and ethnicity.⁽⁸¹⁾ Among men with hip fractures there were two notable findings: (i) the treatment rates in men were significantly lower than women (11.1–35.0%), and (ii) there were significant racial differences in treatment rates among men with hip fractures, with NHB and Asian men having significantly lower treatment rates (11.1% and 13.3%) than Hispanic (22.5%) and NHW (35.0%) men.⁽⁸¹⁾

Although the KPSC study is not generalizable because it was performed in a closed health system and included an intensive bone health model of care, it still revealed racial disparities in treatment rates, including further evidence on the low treatment rates in NHB, interesting data on the low use in Asian females despite the high prevalence of osteoporosis in this group, and the extremely low treatment rates in all men, particularly men of color. Again, this further highlights the need for more data from multiracial and ethnic populations for epidemiologic studies regarding bone outcomes and osteoporosis management for the improvement of bone health in all.

Discussion and Implications

Overall, NHB adults have higher BMD, lower prevalence of osteoporosis, and lower rates of fracture compared with NHW adults. Research on Hispanic adults, however, is less clear, with conflicting evidence regarding BMD, osteoporosis, and fractures. Although Asian populations generally show lower BMD, higher prevalence of osteoporosis, and lower fracture rates compared with NHW adults, data are limited. It is evident that there is an urgent public health need for multiethnic population-based studies using similar designs and methodology (eg, consistent adjustment of covariates such as body composition) to clarify differences in bone outcomes, particularly in Hispanic and Asian populations. Understanding these differences, as well as risk factors for osteoporosis and fracture in these underrepresented population groups, is imperative for developing strategies to reduce bone health disparities.⁽¹¹⁶⁾ Further, conflicting findings in bone health outcomes between major groups may be due, in part, to the fact that there is considerable variation within these groups based on origin for genetic, lifestyle, social, cultural, and environmental factors.⁽¹¹⁷⁾ In particular, examining traditional between-group differences, rather than specific groups by origin or background, prevents understanding and development of effective prevention methods to meet the specific needs of these groups. Thus, there is a need for bone health research in Hispanic and Asian groups by origin, which has been poignantly discussed as early as 1994.⁽¹¹⁸⁾

Although overall screening for osteoporosis is underutilized in the United States, there are clear disparities in screening, particularly for NHB adults.⁽⁷⁸⁾ Unfortunately, data on referrals and adherence for screenings for osteoporosis in Hispanic and Asian populations is limited, and has been understudied. Although there is some evidence that osteoporosis treatment is lower in NHB women, data in other racial and ethnic minority populations and in men is lacking. Mixed methods studies designed to understand barriers and facilitators to adherence to screening and treatment recommendations for providers and patients are needed to inform effective and culturally relevant interventions. However, based on the existing clinical literature the following strategies should be considered.

Strategies for Addressing Racial and Ethnic Bone Disparities

Patient-level strategies

Studies have also shown that diverse populations, particularly NHB and Hispanic groups, perceive reduced risk of osteoporosis,⁽⁸³⁾ which may lead to lower adherence of retaining DXA screening appointments and or treatment recommendations. This is supported by a mixed method study comparing osteoporosis knowledge, attitudes, and beliefs between NHW and NHB women living in the southern United States (n = 48)⁽¹¹⁹⁾ that found NHB women had lower osteoporosis knowledge than NHW women.⁽¹¹⁹⁾ Lower knowledge highlighted lower activation around osteoporosis in the discussion groups with NHB women. A similar mixed methods study in Caribbean origin Hispanic adults (n = 45) in the northeastern United States found that although participants had heard the term osteoporosis and that the majority agreed or strongly agreed that osteoporosis was a serious chronic disease, most participants confused osteoporosis with other bone and joint conditions.⁽⁸⁹⁾ This is consistent with other studies that have shown confusion with respect to severity and management of osteoporosis.⁽¹²⁰⁾ Lower screening rates in racial and ethnic minority populations may be due to misconceptions of risk and importance of osteoporosis in these groups.⁽⁷⁵⁾

Patient education, particularly to increase awareness of severity of fracture and how to advocate for screening for osteoporosis, may be an important strategy for increasing screening rates.⁽¹²¹⁾ However, in underserved communities, tailoring of educational messages to the respective culture and/or needs of the group is essential The NOF not only translated their materials in Spanish but made a concerted effort to work to ensure that the language was affirming and activating for women in the Hispanic community. American Bone Health has recently developed targeted osteoporosis educational programs for NHB and Asian communities. Their “Why Healthy Bones Matter” series for NHB women showed an increase in osteoporosis knowledge over the 60-minute program.⁽⁵⁹⁾ Although there is some evidence that patient and/or provider education-based initiatives alone do not improve diagnosis or treatment of osteoporosis,⁽¹²²⁾ interventions that can include multiple strategies, such as provider feedback and follow-up⁽¹²³⁾ or activation tools⁽¹²⁴⁾ to enable patients to talk with providers about testing, may improve screening rates. A low active role in interactions with physicians has been reported among some ethnic minority populations,^(89,125,126) which may be due to language barriers, varying levels of acculturation, and/or difficulty navigating the health care system.^(125,126) Thus, interventions and programs that promote patient engagement must account for culture specific barrier to accessing care.

Healthcare provider–level strategies

Disparities in screening across populations may be due to providers’ perceptions of osteoporosis risk in ethnic and minority populations. Healthcare providers’ attitudes toward diagnosis and treatment of osteoporosis have been documented as major factors impacting recommendations for screening.⁽¹²⁷⁾ In a breakout session of the 2010 American Academy of Orthopaedic Surgeons/Orthopaedic Research Society/Association of Bone and Joint Surgeon Musculoskeletal Health Disparities Research Symposium, two of three factors identified that contribute to disparities in osteoporosis care among ethnic minority populations and men included: (i) “lack of awareness among healthcare providers and the public regarding the need for osteoporosis screening and prevention among non-white and male patients, and (ii) “gaps in knowledge regarding the effectiveness of differing approaches to screening and medical management of osteoporosis, particularly among non-white and male patients.”⁽¹²⁸⁾ This work highlights the need for greater and effective education strategies for providers on bone health and management strategies, particularly with a culturally relevant lens.⁽¹²⁸⁾ Initiatives by NOF and the American Orthopaedic Association (Bone ECHO) have incorporated more information for providers on racial differences in bone health, but more is needed, particularly at the primary care level, where osteoporosis is primarily managed.

Wright and colleagues⁽¹¹⁹⁾ identified that trust was a key component needed in the healthcare setting. NHB women also relied more on external forces with respect to osteoporosis management (eg, clinicians and religion), whereas NHW women were more internally driven by osteoporosis management (eg, family history, knowledge of disease).⁽¹¹⁹⁾ Likewise, in the focus groups among Caribbean Hispanic adults, language barriers and ethnic concordance influenced quality of care.⁽⁸⁹⁾ Studies have found that personal characteristics of the provider may influence osteoporosis management care.⁽¹²⁹⁾ For example, a study found that male providers referred fewer women for DXA compared with female providers (21.7% versus 27.7%).⁽⁸³⁾ Having more female, NHB, and Hispanic doctors could improve the patient-provider relationship to achieve optimal osteoporosis care. Further, strategies for providers aimed at improving and encouraging rapport with patients with a focus on cultural sensitivity and creating a caring and supportive environment may be most successful particularly for specific population groups.^(89,130)

Conclusion

This article highlights the current knowledge of racial and ethnic differences in bone health and fractures and disparities in management and fracture outcomes. Many may attribute these disparities in fractures and outcomes to higher prevalence of comorbidities that are associated with adverse bone health outcomes; however, even after adjusting for demographic, socioeconomic, comorbidities, and other bone related factors, the social construct of race remained a significant factor.⁽¹³¹⁾ Thus, there is a need to be vigilant in understanding and identifying the roles that not only biology and ancestry, but also the roles that social determinants of health, cultural practices, and other specific risk factors play in racial differences in bone health and its management, particularly given that these populations the are among the fastest growing segments of the US population (the Asian population aged ≥65 years is estimated to almost triple⁽¹³²⁾ and the Hispanic population aged ≥65 years to double by 2060⁽¹³³⁾). There is also a need for innovative and targeted multilevel interventions and programs to educate patients and providers about the importance of bone health, even in communities perceived as being lower risk. The messages around bone health also need to extend to the research community, because studies are needed to inform the design of interventions to reduce bone disparities. This research can then be utilized to inform policies that account for cultural differences. Studies citing racial and ethnic disparities in bone health have been published over the past 50 years, but have not been fully addressed at a clinical or systems level given the lack of longitudinal data on bone health in populations of color. Given the aging US population and the known disparities in bone health, now is the time for patient advocacy groups, researchers, public health officials, clinicians, and policy makers to collaborate to address these important disparities.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.