Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirect

Journal of Crohn's and Colitis (2014) 8, 288–295

National estimates of the burden ofinflammatory bowel disease among racialand ethnic groups in the United States

Geoffrey C. Nguyena,b,c,⁎, Christopher A. Chongd, Rachel Y. Chongd

a Mount Sinai Hospital Centre for Inflammatory Bowel Disease, University of Toronto, Toronto, Ontario, Canadab Institute for Health Policy Management and Evaluation, University of Toronto, Canadac Division of Gastroenterology and Hepatology, Johns Hopkins School of Medicine, USAd Department of Medicine, Lakeridge Health, Oshawa, Ontario, Canada

Received 18 July 2013; received in revised form 11 August 2013; accepted 2 September 2013

⁎ Corresponding author at: Mount SinOntario, Canada M5G 1X5. Tel.: +1 41

E-mail address: geoff.nguyen@utor

1873-9946/$ - see front matter © 2013http://dx.doi.org/10.1016/j.crohns.2

KEYWORDS:Inflammatory bowel disease;Crohn's disease;Ulcerative colitis;Prevalence;Hospitalization;Race

Abstract

Background: The epidemiology of inflammatory bowel disease (IBD) is poorly characterized inminorities in the U.S. We sought to enumerate the burden of IBD among racial and ethnic groups usingnational-level data.Methods: Data from the National Health Interview Survey was used to calculate prevalence andincidence of IBD among adults (≥18 years) in 1999. The Nationwide Inpatient Sample wasqueried to ascertain rates of IBD-related hospitalizations and the Underlying Cause of Death

Database was accessed to quantify IBD-related mortality.Results: An estimated 1,810,773 adult Americans were affected by IBD yielding a prevalenceof 908/100,000, which was higher in Non-Hispanic Whites (1099/100,000) compared withNon-Hispanic Blacks (324/100,000), Hispanics (383/100,000), and non-Hispanic Other (314/100,000). Relative to Non-Hispanic Whites, the odds ratios for having a diagnosis of IBDassociated with being Non-Hispanic Black, Hispanic, and Other Non-Hispanic race afteradjusting for age, sex, and geographic region were 0.33 (95% CI: 0.19 – 0.57), 0.45 (95% CI:0.26 – 0.77), and 0.34 (95% CI: 0.12 – 0.93), respectively. IBD incidence was similarly lowerin Non-Hispanic Blacks (24.9/100,000) and Hispanics (9.9/100,000) compared toNon-Hispanic Whites (70.2/100,000). The ratio of IBD hospitalizations to prevalence wasdisproportionately higher among Non-Hispanic Blacks (7.3%) compared with Non-HispanicWhites (3.0%) and Hispanics (2.7%). Similarly, the ratio of IBD-related mortality was greaterin Non-Hispanic Blacks (0.061%) compared to Non-Hispanic Whites (0.036%) and Hispanics(0.026%).

ai Hospital Centre for Inflammatory Bowel Disease, 600 University Avenue, Suite 437, Toronto,6 586 4800x2819; fax: +1 416 586 5971.onto.ca (G.C. Nguyen).

European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved.013.09.001

289National Estimates of the Burden of IBD Among Racial and Ethnic Groups in the U.S.

Conclusions: IBD disease burden is lower in ethnic minorities compared to Non-Hispanic Whites.However, IBD-related hospitalizations and deaths seem disproportionately high in Non-HispanicBlacks.© 2013 European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved.

1. Introduction

The prevalence of inflammatory bowel diseases (IBD)comprising Crohn's disease (CD) and ulcerative colitis(UC) appears to be rising worldwide, with the greatestprevalence and incidence in North America and Europe,while it is less common in South America, Africa, andAsia.1 Even in low-prevalence regions, there is emerg-ing incidence of IBD in India and Southeast Asiancountries where it was not previously described.2 Therelative contributions of genetic and environmentalfactors to the geographic variation in IBD prevalenceand incidence remain unclear. Moreover, it remainslargely unknown whether racial and ethnic minorities inthe United States share a similar burden of IBD as theirwhite counterparts.

A series of hospital-based epidemiological studies fromBaltimore in the 1960s and 1970s suggest that theincidence of IBD among Blacks increased sharply fromone-fifth to three-quarters that of Non-Hispanic Whitesin the time interval between the two studies. In thelast study in the series, the incidence of CD amongBlack females actually surpassed that of Non-HispanicWhite females and males.3 A mail survey of KaiserPermanente health plan members suggests similarhospitalization rates between Non-Hispanic Whites andBlacks, which were much higher than that of Hispanicsand Non-Hispanic Asians.4 In the same study, theprevalence of CD among Non-Hispanic Whites was nearly50% higher than that of Blacks and 8- to 10-fold higherthan that of Hispanics and Asians. The hospitalization-based studies in both Baltimore and California, however, likelycaptured only more severe IBD cases, while the prevalencestudy was limited by its geographic scope and absence ofepidemiological data on UC.

The need to understand the nationwide burden ofinflammatory bowel disease in diverse ethnic populations isunderscored by the prediction that by 2050, 54% of the U.S.population will comprise minority populations, with His-panics and Non-Hispanic Blacks, the two largest racial andethnic groups, expected to contribute to 45% of the U.S.population.5 Moreover, the health of America's minoritygroups are threatened by disparities in healthcare accessthat have been described across a spectrum of chronicdiseases.6 More robust and generalizable estimations of IBDdisease burden among Hispanics and Non-Hispanic Blacks arewarranted in order to properly allocate and optimizehealthcare delivery. Given the dearth of IBD epidemiologicaldata among racial and ethnic minorities, we implementednationally derived data to characterize IBD prevalence,incidence, hospitalizations, and mortality in these under-served populations.

2. Materials and methods

2.1. Data sources

2.1.1. National Health Interview StudyPrevalence estimates were ascertained from the NationalHealth Interview Study (NHIS) conducted in 1999, which hasbeen previously described for health utilization and laboreconomic studies in IBD.7,8 The NHIS is an annual surveyconducted by U.S census bureau employees on behalf ofthe National Center for Health Statistics (NCHS) underthe Centers for Disease Control (CDC). It comprises across-sectional sample of all non-institutionalized civiliansin the U.S., excluding military personnel, prisoners, andresidents of long-term facilities. Selected households wereinvited to participate in a computer-assisted interviewconducted by U.S. Census personnel, and the householdresponse rate was 87.6% in 1999. One adult (18 years andolder) and one child from each household were randomlyselected to answer an adult and child questionnaire, respec-tively, that queried information on demographics; income;healthcare access, utilization, and insurance; presence ofmedical and psychiatric conditions, health status and limita-tions; and satisfaction with health services. Race and ethnicitywere self-reported in all respondents and further categorizedinto mutually exclusive categories: Non-Hispanic White, Non-Hispanic Black, Hispanic, and Non-Hispanic Other. While corecomponents of the survey were repeated each year, thesupplemental component changed annually. In 1999 only, adultsurvey respondents were asked, “Have you EVER been told by adoctor or other health professional that you had Crohn's diseaseor ulcerative colitis?” Those responding affirmatively werefurther asked, “How old were you when you were first told youhad Crohn's disease or ulcerative colitis?” and “DURING THEPAST 12 MONTHS, have you had symptoms of Crohn's disease orulcerative colitis?” Disease duration was derived from thedifference between attained age and age at diagnosis. Incidentcases were defined as individuals with disease duration of1 year or less with active IBD symptoms within the past12 months.

2.1.2. Nationwide Inpatient SampleHospitalization data were extracted from the NationwideInpatient Sample (NIS) from 1999. The NIS is a 20% stratifiedsample of non-federal, acute-care hospitals in the UnitedStates and is maintained as part of the Healthcare Cost andUtilization Project (HCUP) sponsored by the Agency forHealthcare Research and Quality (AHRQ). Each dischargeabstract includes a unique identifier, demographic variables(defined as age, gender, and race/ethnicity, median incomefor ZIP code), source of admission, discharge disposition,

290 G.C. Nguyen et al.

primary and secondary diagnoses (up to 15), primary andsecondary procedures (up to 15), primary insurance payers,total hospital charges, and length of stay. The NIS has beenextensively used to characterize IBD hospitalizations andIBD-related health outcomes.9–13 For this study, we identi-fied all hospital discharges in 1999 with the followingcriteria: 1) were 18 years or older; 2) had a principaldiagnosis of inflammatory bowel disease (UC or CD) identifiedby the Clinical Modification of the International Classificationof Diseases, 9th Revision (ICD-9-CM) code of 555.xx and556.xx. IBD hospitalizations for less than 24 h were excludedbecause many of these may have been for inpatient infusionsof infliximab. Hospital discharges in the states (GA, IL, ME, OR,WA) which did not collect race and ethnicity data were alsoexcluded. Major abdominal surgery (i.e. bowel resection)undertaken during a hospitalization for IBD was identifiedusing ICD-9-CM procedural codes as previously described.9,10

2.1.3. Underlying Cause of Death databaseDeaths caused by CD or UC in 1999 were identified throughthe Underlying Cause of Death database spanning from 1999to 2010. These mortality data are derived from deathcertificates from the fifty states and the District of Columbiaand reported to the National Center for Health Statisticsthrough the Vital Statistics Cooperative Program. The databasecontains demographic information (age, sex, ethnicity, race,state residence) and primary cause of death. We identifiedIBD-related deaths using ICD-10 codes for IBD (K50.x andK51.x). Ethnicity- and race-specific death rates due to IBDwerecalculated by accessing the CDC WONDER Online Database,released 2012 on Mar 9, 2013 at http://wonder.cdc.gov/ucd-icd10.html.

2.2. Statistical analyses

Data were analyzed using the Stata 10.0 SE software package(Stata Corp LP, College Station, Texas). Because of thecomplex two-stage sampling design of NHIS data, we usedStata's SVY command to take into account individual samplingweights for variance estimation. Point prevalence for IBD wasdetermined by estimating from the NHIS sample, the totalnumber diagnoses of IBD in the US divided by the total numberof adult non-institutionalized citizens in the US. Sex-specificand region-specific point prevalence of IBD was calculated.Estimates for IBD point prevalence and their 95% confidenceintervals were calculated for the following mutually exclusiveracial/ethnic groups: Non-HispanicWhite, Non-Hispanic Black,Hispanic, and Non-Hispanic Other. Similarly, incidence wascalculated by dividing new cases of IBD diagnosedwithin 1 yearby individuals at risk for IBD, which was the US Censuspopulation minus prevalent cases of IBD. Descriptive statisticswere performed to compare other clinical factors among racialgroups. Using NIS data, race-specific rates of IBD hospitaliza-tion were calculated by dividing the national estimate of IBDhospitalizations for each racial/ethnic group by the 1999 U.S.census population for that respective racial group, excludingtheweighted contribution of states which did not collect racialdata in the NIS. The rate of IBD-related surgery was similarlycalculated by estimating the national counts of IBD-relatedsurgery within each racial/ethnic group by the respective USCensus population.

To determine whether the rates of IBD-related hospitaliza-tions, surgeries, and deaths were proportional to the preva-lence of IBD in each racial/ethnic group, we calculated theratio of race-specific hospitalization rates, surgery rates, anddeath rates to the point prevalence of IBD for each respectiveracial/ethnic group.

Survey-based multiple logistic regression was to determinethe association between IBD diagnosis and race/ethnicitywhile adjusting for age, sex, and geographic region.

2.3. Ethical considerations

The NHIS and NIS are publically accessible datasets andcontain completely de-identified data, and the UnderlyingCause of Death database reported data only in aggregate.Thus, use of this database was exempt from research ethicsapproval consistent with the policies at the Johns HopkinsMedical Institutions (Baltimore) and Mount Sinai Hospital(Toronto).

3. Results

3.1. Prevalence and incidence

There were 271 individuals among the 30,801 adult respon-dents who reported having either UC or CD, yielding anationwide IBD prevalence of 908 per 100,000 (95% CI: 798 –1033 per 100,000). There were an estimated 1,810,773 adultsin the United States with IBD in 1999. Sex-specific prevalenceof IBD was greater among females compared with males (1154vs. 640 per 100,000, P b 0.0001). Therewas significant regionalvariation in IBD prevalence (Fig. 1), with the prevalence beingnearly 60% higher in the Northeast and Midwest compared tothe South and West.

Prevalence of IBD also differed among racial/ethnicgroups. It was highest among Non-Hispanic Whites estimatedat 1099 per 100,000 (95% CI: 958 – 1261 per 100,000). Theprevalence was significantly lower in Non-Hispanic Blacks(324/100,000; 95% CI: 190 – 552/100,000), Hispanics(383/100,000; 95% CI: 236–620/100,000), and Non-HispanicOther (314/100,000; 95% CI: 115–856/100,000). Fig. 2 showsthe age-specific prevalence among racial/ethnic groups, andNon-Hispanic Whites consistently exhibited higher prevalencethan Non-Hispanic Blacks and Hispanics in all groups. Relativeto Non-Hispanic Whites, the odds ratios for having a diagnosisof IBD associated with being Non-Hispanic Black, Hispanic, andOther Non-Hispanic race after adjusting for age, sex, andgeographic region were 0.33 (95% CI: 0.19 – 0.57), 0.45 (95%CI: 0.26 – 0.77), and 0.34 (95% CI: 0.12 – 0.93), respectively.

The overall incidence of IBD was 58.7 per 100,000 andwas similar between males and females (63.4/100,000 vs.54.4/100,000). Stratified by ethnicity and race, the incidenceof IBD for Non-Hispanic Whites, Blacks, and Hispanics was70.2/100,000, 24.9/100,000, and 9.9/100,000, respectively.

Table 1 shows clinical characteristics of the adult respon-dents who reported having a diagnosis of IBD. The mean age atIBD diagnosis was 38.7 years and themean disease durationwas13.6 years and did not significantly vary by ethnicity and race.Hispanics with IBD were more likely to have never smoked thanNon-Hispanic Whites (78% vs. 40%, P = 0.04). About half ofrespondents reported having active IBD symptoms and some

Figure 1 Prevalence of inflammatory bowel disease within each of the four geographic regions of the United States is shown in thisbar graph with error bars reflecting 95% confidence intervals.

291National Estimates of the Burden of IBD Among Racial and Ethnic Groups in the U.S.

functional limitation in the prior 12 months, and these pro-portions were not significantly different amongst racial andethnic groups.

3.2. Hospitalization and surgery rates

In 1999, the nationwide hospitalization rate for IBD was 30.5per 100,000 with variation by race and ethnicity as shown inFig. 3. Rates of hospitalization were higher among Non-Hispanic Whites (32.9/100,000) compared with Non-HispanicBlacks (23.8/100,000), Hispanics (10.4/100,000) and otherNon-Hispanic racial groups (18.3/100,000). Similar trends in

Figure 2 Age-specific prevalence of inflammatory boweldisease is shown stratified by race and ethnicity. The overallage-specific prevalence is also shown.

hospitalization rates were observed when stratified by CDand UC (Fig. 3). However, the ratio of hospitalizations toprevalence was higher among non-Hispanic Blacks (7.3%)compared with Non-Hispanic Whites (3.0%) and Hispanics(2.7%).

The overall nationwide surgery rate for IBD was 5.4 per100,000. Among Non-Hispanic Whites, the rate of surgerywas 6.2/100,000 and was higher than that for Non-HispanicBlacks (2.9/100,000), Hispanics (1.2/100,000), and Non-Hispanic Other (2.8/100,000). However, the ratio of surgeryto prevalence was higher among Non-Hispanic Blacks (0.9%)compared to Non-Hispanic Whites (0.56%) and Hispanics(0.32%).

3.3. Mortality

In 1999, the mortality rate attributable to IBD in the U.S.was 0.4/100,000. The IBD-related rate of death was highestamong Non-Hispanic Whites at 0.5/100,000. The IBD mor-tality rate was 0.2/100,000 among Non-Hispanic Blacks whileit was 0.1/100,000 among Hispanics (Table 2). The ratio ofIBD-related mortality to prevalence was highest amongNon-Hispanic Blacks (0.061%) compared to Non-HispanicWhites (0.036%), Hispanics (0.026%) and Non-Hispanic Other(0.032%).

4. Discussion

Our study characterizes at a national level the burden of IBDamong racially and ethnic diverse populations. Our datasuggest that IBD remains a condition that predominantlyaffects Non-Hispanic Whites with a point prevalence 3-foldhigher than that of other racial and ethnic groups. However,when compared with other racial groups, Non-Hispanic

Table 1 Characteristics of survey respondents with inflammatory bowel disease stratified by race/ethnicity.

Overall IBD population Race/Ethnicity group

Non-Hispanic White Non-Hispanic Black Hispanic Non-Hispanic Other

Nationwide estimate of numberof IBD cases

1,810,773 1,635,533 72,760 78,513 23,967

Age at diagnosis 38.7 38.7 44.1 33.6 38.0Disease duration (yr) 13.6 14.0 9.0 13.1 2.1Body mass index 28.3 28.3 27.6 28.2 27.6Smoking status ⁎

Never 50% 49% 40% 78% 21%Former 30% 30% 29% 14% 79%Current 20% 21% 31% 8% 0%Education levelDid not graduate HS 18% 16% 24% 40% 31%HS graduate or GED 60% 61% 54% 55% 18%Bachelor’s degree 14% 14% 18% 5% 21%Advanced degree 8% 9% 4% 0% 30%Regular health provider 90% 91% 84% 75% 82%IBD Symptoms in past 12 months 51% 50% 63% 42% 100%Functional limitation 53% 52% 56% 56% 79%

HS = high school.GED = general educational development.⁎ P = 0.04, Hispanics compared with Non-Hispanic White.

292 G.C. Nguyen et al.

Blacks may have experienced IBD-related hospitalizationsand mortality that were disproportionately high relative totheir prevalence. We estimated that there were more than150,000 Non-Hispanic Blacks and Hispanics with IBD, and thisnumber will continue to rise as the minority populationscontinue to rapidly expand over the next several decades.

Figure 3 The rate of hospitalizations for inflammatory bowel diseHispanics, and Non-Hispanic Other. The contributions of hospitalizahospitalization rate are indicated by the shaded and black bars, res

Interestingly, our overall estimate of IBD prevalence washigher than that reported in the literature.

We report an IBD prevalence of 908 per 100,000 which isnearly 2-fold higher than the highest estimates previouslyreported worldwide.1 The discrepancy is partly due to theinclusion of only those 18 years and older in our study

ase (IBD) is shown for Non-Hispanic Whites, Non-Hispanic Blacks,tions for Crohn's disease and ulcerative colitis to the overall IBDpectively.

Table 2 Rate of death from inflammatory bowel disease among adults in the United States stratified by ethnicity and race.

Ethnicity and race No. deaths Census population Death rate (95% CI) (per 100,000)

Hispanic 23 22,036,194 0.1 (0.1 – 0.2)Non-HispanicWhite 713 151,819,063 0.5 (0.4 - 0.5)Black 53 23,590,803 0.2 (0.2 - 0.3)Other 4 9,648,059 0.04 (0.0 - 0.1)Total 770 185,057,925 0.4 (0.4 - 0.4)

293National Estimates of the Burden of IBD Among Racial and Ethnic Groups in the U.S.

because IBD-related questions were only administered toadults in the NHIS. Age-specific prevalence is low in thoseunder 18 years and increases with older age groups becauseit increases with disease duration for chronic diseases wheremortality is low. Both Kappelman et al. and Herrinton et al.have reported rising age-specific IBD prevalence with olderage groups in US populations.14,15 Thus, the exclusion ofchildren is expected to raise the overall IBD prevalence aswell as the average age of diagnosis, which we also observedin our study. The overall IBD prevalence in the Kappelmanstudy was 504/100,000, which was still lower than ourestimates.14 However, this study, like two other nationwideepidemiological studies, identified prevalent IBD casesthrough encounters with the health system using insuranceclaims data that spanned only a few years.14,16 It is possiblethat these studies undercounted IBD prevalent cases byfailing to capture individuals in long-term disease remissionwho did not receive regular IBD healthcare. The use ofclinic-based registries to ascertain IBD prevalence may alsopartly explain why epidemiological studies from OlmstedCounty showed a contrasting decline in age-specific preva-lence with older age groups. If IBD became quiescent manydecades after onset, prevalence of IBD, including individualswith inactive disease, may be undercounted in older agecategories. Moreover, because the study captured onlyindividuals in Olmsted County, prevalence may be decreasedwith age if older individuals, especially those who are elderly orretired, emigrate out of the region. In contrast to previouspublished studies, prevalent IBD cases were identified byself-report in the NHIS survey, which sampled the entire USpopulation and did not rely on regular interactions with thehealthcare system—capturing both active and inactive IBD. Thelatter is particularly important since half of IBD respondents inthe NHIS survey reported having no IBD-related symptoms in theprior 12 months. Consequently, it may be more useful tocompare the figures from our study with peak age-specificprevalence from the Loftus study, which may circumventpotential issues arising from emigration and undercounting ofinactive IBD cases. In that study, the peak prevalence of IBDwas992/100,000 (493/100,000 for UC and 499/100,000 for CD),which was comparable to the estimate of 908/100,000 based onthe NHIS survey.17

Our current study also demonstrated substantial geo-graphic variation in prevalence of IBD, which has beenpreviously reported in the U.S.16,18 Consistent with thereported literature, we found that IBD prevalence was lowerin the South and the West. These findings emphasize thepotential limitations of extrapolating epidemiological data

from single counties or even states and the importance ofstudying geographically diverse populations to estimate IBDburden. Though these geographical variations may be due toenvironmental influences, they may also partly reflect racialdifferences in prevalence of IBD. The South and the Westhave relative higher concentration of Non-Hispanic Blacks,Hispanics, and other minority groups, in whom we showedIBD to be substantially less prevalent.

The lower prevalence of IBD in Non-Hispanic Blacks andHispanics is likely multifactorial arising from both biologicaland socioeconomic factors. Compared with Non-HispanicWhites, Non-Hispanic Blacks have been shown to havesubstantially lower allele frequency of genetic polymor-phisms, such as those in the NOD2 gene that predispose toCD.19–21 However, the lower prevalence observed in minor-ity groups may be also be due to under-diagnosis due tosuboptimal access to healthcare. Having Medicaid insurancehas been shown to be associated with lower prevalence ofIBD compared to being privately insured and may also be anindicator of differential access to specialists and diagnosticprocedures.16

Though the prevalence of IBD among minority groupswas low, there was a disproportionately higher rate ofIBD-associated hospitalization and mortality among Non-Hispanic Blacks compared to Non-Hispanic Whites that wasnot observed in other minority ethnic and racial groups.These racial variations in clinical outcomes may also reflectboth biological and socioeconomic mechanisms. We havepreviously shown that there may be differences in diseasephenotype between Non-Hispanic Whites and Blacks with thelatter more likely to experience upper GI and perianal Crohn'sdisease.22 Non-Hispanic Blacks also had lower utilization ofgastroenterology care and immunomodulators and biologicseven in the presence of chronic steroid use.23 Moreover, evenwith access to specialist care, Non-Hispanic Blacks exhibitedlower adherence to medications and trust in their physicians,which in turn may erode the patient–physician relationship.24

These above factors may contribute to suboptimal outpatientcare that may result in increased hospitalizations, surgery,and even mortality.

Themain limitation of the NHIS component of our analysis isthe ascertainment of IBD diagnosis which was by participantself-report and did not distinguish between CD and UC. It ispossible that participants mistakenly over-reported an IBDdiagnosis which may have partly explained a higher IBDprevalence than previously reported. We should reiteratethat the questionnaire specifically queried “Crohn's disease”and “ulcerative colitis” and not “IBD” which could easily be

294 G.C. Nguyen et al.

mistaken for “IBS.” Furthermore, the inability to stratify ouranalysis by IBD subtype may obscure potential differences inthe influence of demographic characteristics on the epidemi-ology of CD and UC. Moreover, the lack of clinical data ondisease phenotype and severity precluded us from differenti-ating potential mechanisms for potential racial and ethnicaldifferences in hospitalization and surgery. These disparitiesmay arise from racial differences in disease characteristics aswell as from sociocultural factors.22,25 The lack of pediatricdata in the NHIS dataset also leads to an important gap in ourknowledge of racial variations in IBD disease burden particu-larly in an age group that is experiencing an appreciable rise inincidence globally.26,27 Lastly, the IBD sample size in the NHISis modest and limited our ability to ascertain precise estimatesof incidence and to perform extensive subgroup analyseswithin racial and ethnic groups.

Despite these limitations, the prevalence, hospitalization,and mortality data from our analysis provide a glimpse of theburden of IBD among different racial and ethnic groups in theU.S., which has not been previously well characterized at anational level. These estimates support previous observationsthat there is less overall IBD disease burden among Non-HispanicBlacks and Hispanics. These figures have implications forhealthcare planning and strategies to allocate resources tooptimize IBD medical care as minority populations rapidlyexpand in the U.S. While we acknowledge that these data werecollected nearly a decade and a half ago, the most frequentlycited epidemiological data on IBD among Blacks and Hispanicsare even older from the 1960s to early 1980s and derived fromlocalized populations.3,4 Data from the NHIS survey is not onlymore recent but reflects the entire United States population.Further research is warranted to confirm and understand whyIBD-related hospitalizations surgeries, and mortality may bedisproportionately high among Non-Hispanic Blacks. Data fromthe NHIS also suggest that the prevalence of IBD in adults ishigher than previously reported, and further epidemiologicalstudies should be undertaken to delineate whether thisdiscrepancy is due to overestimation by self-report or underes-timation by insurance claims databases that rely on regularhealthcare contact and misclassification-prone administrativedefinitions of IBD. The NHIS survey provides the ideal infra-structure for the forementioned endeavor. Additional clinicaland medication-related questions may further support thedefinition of IBD and an introduction of a mechanism to confirmIBD diagnosis with medical records would augment the validityof IBD diagnosis. Moreover, analysis of IBD data over multipleyears would enable a larger sample size and consequently moreextensive subgroup and time trend analyses. The NHIS surveystudy is due for a decadal revamp in its study design in the nextseveral years. Involvement of IBD epidemiologistswould providean invaluable opportunity to characterize IBD disease burden inthe U.S. and mechanisms of potential racial disparities inoutcomes.

Funding

G.C.N. is a recipient of New Investigator Awards from theCanadian Institutes of Health Research (CIHR), the CanadianAssociation of Gastroenterology, and the Crohn's and ColitisFoundation of Canada.

Conflicts of interest

The authors (G.C.N., R.Y.C., and C.A.C.) report no conflictsof interest.

Author contributions

G.C.N., R.Y.C., and C.A.C. conceptualized and designed thestudy. G.C.N. performed all analyses with assistance fromR.Y.C. and drafted the manuscript. All authors contributedto the critical review of the manuscript.

References

1. Molodecky NA, Soon IS, Rabi DM, Ghali WA, Ferris M, Chernoff G,et al. Increasing incidence and prevalence of the inflammatorybowel diseases with time, based on systematic review.Gastroenterology 2012;142:46–54.

2. Prideaux L, Kamm MA, De Cruz PP, Chan FK, Ng SC.Inflammatory bowel disease in Asia: a systematic review. JGastroenterol Hepatol 2012;27:1266–80.

3. Calkins BM, Lilienfeld AM, Garland CF, Mendeloff AI. Trends inincidence rates of ulcerative colitis and Crohn's disease. Dig DisSci 1984;29:913–20.

4. Kurata JH, Kantor-Fish S, Frankl H, Godby P, Vadheim CM. Crohn'sdisease among ethnic groups in a large health maintenanceorganization. Gastroenterology 1992;102:1940–8.

5. From an older and more diverse nation by midcentury.Available at: http://www.census.gov/newsroom/releases/archives/population/cb08-123.html [Accessed: March 13,2013].

6. Smedley BD, Stith AY, Nelson AR, Institute of Medicine, editors.Unequal treatment: confronting racial and ethnic disparities inhealth care. Washington, D.C.: National Academies Press; 2005.p. 1–79.

7. Longobardi T, Bernstein CN. Health care resource utilization ininflammatory bowel disease. Clin Gastroenterol Hepatol 2006;4:731–43.

8. Longobardi T, Jacobs P, Bernstein CN. Work losses related toinflammatory bowel disease in the United States: results fromthe National Health Interview Survey. Am J Gastroenterol2003;98:1064–72.

9. Nguyen GC, Bayless TM, Powe NR, Laveist TA, Brant SR. Raceand health insurance are predictors of hospitalized Crohn'sdisease patients undergoing bowel resection. Inflamm BowelDis 2007;13:1408–16.

10. Nguyen GC, Laveist TA, Gearhart S, Bayless TM, Brant SR. Racialand geographic variations in colectomy rates among hospital-ized ulcerative colitis patients. Clin Gastroenterol Hepatol2006;4:1507–13.

11. Nguyen GC, Tuskey A, Dassopoulos T, Harris ML, Brant SR. Risinghospitalization rates for inflammatory bowel disease in theUnited States between 1998 and 2004. Inflamm Bowel Dis2007;13:1529–35.

12. NguyenGC, Sam J, Murthy SK, Kaplan GG, Tinmouth JM, Laveist TA.Hospitalizations for inflammatory bowel disease: profile of theuninsured in the United States. InflammBowel Dis 2009;15:726–33.

13. Nguyen GC, Steinhart AH. Nationwide patterns of hospitaliza-tions to centers with high volume of admissions for inflamma-tory bowel disease and their impact on mortality. InflammBowel Dis 2008;14:1688–94.

14. Kappelman MD, Moore KR, Allen JK, Cook SF. Recent trends inthe prevalence of Crohn's disease and ulcerative colitis in acommercially insured US population. Dig Dis Sci 2013;58:519–25.

295National Estimates of the Burden of IBD Among Racial and Ethnic Groups in the U.S.

15. Herrinton LJ, Liu L, Lewis JD, Griffin PM, Allison J. Incidenceand prevalence of inflammatory bowel disease in a NorthernCalifornia managed care organization, 1996–2002. Am JGastroenterol 2008;103:1998–2006.

16. Kappelman MD, Rifas-Shiman SL, Kleinman K, Ollendorf D,Bousvaros A, Grand RJ, et al. The prevalence and geographicdistribution of Crohn's disease and ulcerative colitis in theUnited States. Clin Gastroenterol Hepatol 2007;5:1424–9.

17. Loftus CG, Loftus Jr EV, Harmsen WS, Zinsmeister AR, TremaineWJ, Melton III LJ, et al. Update on the incidence and prevalenceof Crohn's disease and ulcerative colitis in Olmsted County,Minnesota, 1940–2000. Inflamm Bowel Dis 2007;13:254–61.

18. Sonnenberg A, McCarty DJ, Jacobsen SJ. Geographic variation ofinflammatory bowel disease within the United States. Gastro-enterology 1991;100:143–9.

19. Bonen DK, Cho JH. The genetics of inflammatory bowel disease.Gastroenterology 2003;124:521–36.

20. Kugathasan S, Loizides A, Babusukumar U, McGuire E, Wang T,Hooper P, et al. Comparative phenotypic and CARD15 mutationalanalysis among African American, Hispanic, and White Childrenwith Crohn's disease. Inflamm Bowel Dis 2005;11:631–8.

21. Wang MH, Okazaki T, Kugathasan S, Cho JH, Isaacs KL, Lewis JD,et al. Contribution of higher risk genes and European admixtureto Crohn's disease in African Americans. Inflamm Bowel Dis2012;18:2277–87.

22. Nguyen GC, Torres EA, Regueiro M, Bromfield G, Bitton A,Stempak J, et al. Inflammatory bowel disease characteristicsamong African Americans, Hispanics, and non-Hispanic Whites:characterization of a large North American cohort. Am JGastroenterol 2006;101:1012–23.

23. Nguyen GC, LaVeist TA, Harris ML, Wang MH, Datta LW, BrantSR. Racial disparities in utilization of specialist care andmedications in inflammatory bowel disease. Am J Gastroenterol2010;105:2202–8.

24. Nguyen GC, LaVeist TA, Harris ML, Datta LW, Bayless TM, BrantSR. Patient trust-in-physician and race are predictors ofadherence to medical management in inflammatory boweldisease. Inflamm Bowel Dis 2009;15:1233–9.

25. Hou JK, El-Serag H, Thirumurthi S. Distribution and manifesta-tions of inflammatory bowel disease in Asians, Hispanics, andAfrican Americans: a systematic review. Am J Gastroenterol2009;104:2100–9.

26. Benchimol EI, Fortinsky KJ, Gozdyra P, Van den Heuvel M, VanLJ, Griffiths AM. Epidemiology of pediatric inflammatory boweldisease: a systematic review of international trends. InflammBowel Dis 2011;17:423–39.

27. Benchimol EI, Guttmann A, Griffiths AM, Rabeneck L, Mack DR,Brill H, et al. Increasing incidence of paediatric inflammatorybowel disease in Ontario, Canada: evidence from healthadministrative data. Gut 2009;58:1490–7.