Economic costs of diabetes in the u.s 2012

Economic Costs of Diabetes in the U.S.in 2012AMERICAN DIABETES ASSOCIATION

OBJECTIVEdThis study updates previous estimates of the economic burden of diagnoseddiabetes and quantifies the increased health resource use and lost productivity associated withdiabetes in 2012.

RESEARCH DESIGN AND METHODSdThe study uses a prevalence-based approachthat combines the demographics of the U.S. population in 2012 with diabetes prevalence, ep-idemiological data, health care cost, and economic data into a Cost of Diabetes Model. Healthresource use and associated medical costs are analyzed by age, sex, race/ethnicity, insurancecoverage, medical condition, and health service category. Data sources include national surveys,Medicare standard analytical files, and one of the largest claims databases for the commerciallyinsured population in the U.S.

RESULTSdThe total estimated cost of diagnosed diabetes in 2012 is $245 billion, including$176 billion in direct medical costs and $69 billion in reduced productivity. The largest com-ponents of medical expenditures are hospital inpatient care (43% of the total medical cost),prescription medications to treat the complications of diabetes (18%), antidiabetic agents anddiabetes supplies (12%), physician office visits (9%), and nursing/residential facility stays (8%).People with diagnosed diabetes incur averagemedical expenditures of about $13,700 per year, ofwhich about $7,900 is attributed to diabetes. People with diagnosed diabetes, on average, havemedical expenditures approximately 2.3 times higher than what expenditures would be in theabsence of diabetes. For the cost categories analyzed, care for people with diagnosed diabetesaccounts for more than 1 in 5 health care dollars in the U.S., and more than half of that expen-diture is directly attributable to diabetes. Indirect costs include increased absenteeism ($5 billion)and reduced productivity while at work ($20.8 billion) for the employed population, reducedproductivity for those not in the labor force ($2.7 billion), inability to work as a result of disease-related disability ($21.6 billion), and lost productive capacity due to early mortality ($18.5billion).

CONCLUSIONSdThe estimated total economic cost of diagnosed diabetes in 2012 is $245billion, a 41% increase from our previous estimate of $174 billion (in 2007 dollars). Thisestimate highlights the substantial burden that diabetes imposes on society. Additional compo-nents of societal burden omitted from our study include intangibles from pain and suffering, re-sources from care provided by nonpaid caregivers, and the burden associated with undiagnoseddiabetes.

Diabetes Care 36:1033–1046, 2013

D iabetes imposes a substantial bur-den on the economy of the U.S. inthe form of increased medical costs

and indirect costs from work-related ab-senteeism, reduced productivity at work

and at home, reduced labor force partic-ipation from chronic disability, and pre-mature mortality (1,2). In addition to theeconomic burden that has been quanti-fied, diabetes imposes high intangible

costs on society in terms of reduced qual-ity of life and pain and suffering of peoplewith diabetes, their families, and friends.

Improved understanding of the eco-nomic cost of diabetes and its majordeterminants helps to inform policymakersand to motivate decisions to reduce di-abetes prevalence and burden. The pre-vious cost of diabetes study by theAmerican Diabetes Association (ADA) esti-mated that there were nearly 17.5 millionpeople living in the U.S. with diagnosedtype 1 or type 2 diabetes in 2007, at anestimated cost of $174 billion in highermedical costs and lost productivity (2).

The percentage of the populationwith diagnosed diabetes continues torise, with one study projecting that asmany as one in three U.S. adults couldhave diabetes by 2050 if current trendscontinue (3). In this updated cost of di-abetes study, we estimate the total na-tional economic burden of diagnoseddiabetes in 2012 reflecting continuedgrowth in prevalence of diabetes and itscomplications; changing health care prac-tices, technology, and cost of treatment;and changing economic conditions.

RESEARCH DESIGN ANDMETHODSdThis study follows themethodology used in the 2002 and 2007costs of diabetes studies by the ADA, withmodifications to refine the analyseswhere appropriate (1,2). A prevalence-based approach is used to estimate themedical costs by demographic group,health service category, and medical con-dition. One difference from earlier studiesis that for some analyses we now includerace/ethnicity as a demographic dimen-sion. We analyze the prevalence of diag-nosed diabetes, utilization and costsattributable to diabetes by age-group (un-der 18, 18–34, 35–44, 45–54, 55–59,60–64, 65–69, and over 70 years of age),sex, race/ethnicity (non-Hispanic white,non-Hispanic black, non-Hispanic other,and Hispanic), and insurance status (pri-vate; government including Medicare,Medicaid, Children’s Health InsuranceProgram, and other government-sponsoredcoverage; and uninsured). State-specific es-timates of prevalence and costs are pro-vided in Supplementary Table 11.

c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c

This report was prepared under the direction of the AmericanDiabetes Association byWenya Yang (The LewinGroup, Inc., Falls Church, Virginia); Timothy M. Dall (IHS Global Inc., Washington, DC); Pragna Halder(The LewinGroup, Inc.); Paul Gallo (IHSGlobal Inc.); Stacey L. Kowal (IHSGlobal Inc.); and Paul F. Hogan(The Lewin Group, Inc.).

Address correspondence to Matt Petersen, American Diabetes Association, 1701 N. Beauregard Street, Alex-andria, VA 22311. E-mail: [email protected].

DOI: 10.2337/dc12-2625This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10

.2337/dc12-2625/-/DC1.© 2013 by the American Diabetes Association. Readers may use this article as long as the work is properly

cited, the use is educational and not for profit, and thework is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

See accompanying commentary, p. 775.

care.diabetesjournals.org DIABETES CARE, VOLUME 36, APRIL 2013 1033

S C I E N T I F I C S T A T E M E N T

©

http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc12-2625/-/DC1

mailto:[email protected]



http://creativecommons.org/licenses/by-nc-nd/3.0/

http://creativecommons.org/licenses/by-nc-nd/3.0/

Major data sources analyzed includeNational Health Interview Survey (NHIS),American Community Survey (ACS), Be-havioral Risk Factor Surveillance System(BRFSS), Medical Expenditure Panel Sur-vey (MEPS), OptumInsight’s de-identifiedNormative Health Information database(dNHI), the Medicare 5% sample StandardAnalytical Files (SAFs), Nationwide Inpa-tient Sample (NIS), National AmbulatoryMedical Care Survey (NAMCS), NationalHospital Ambulatory Medical Care Survey(NHAMCS), National Nursing Home Sur-vey (NNHS), National Home and HospiceCare Survey (NHHCS), and Current Pop-ulation Survey (CPS). We use the most re-cent year’s data available for each of thesedata sources, though for certain analyseswe combine 3 years of data to achieve suf-ficient sample size. To estimate medicalcosts for less common health service cate-gories such as hospital inpatient care, emer-gency care, home health, and podiatry, wecombine 5 years of MEPS data to reducevariance in utilization and cost. The demo-graphics of the U.S. population in 2012with diabetes prevalence, epidemiologicaldata, health care cost, and economic dataare then combined into a Cost of DiabetesModel. Supplementary Table 1 describeshow these data sources are used, alongwith their respective strengths and limita-tions, pertinent to this study. All cost andutilization estimates are extrapolated to theprojectedU.S. population in 2012 (4),withcost estimates calculated in 2012 dollarsusing the appropriate components of themedical consumer price index or total con-sumer price index (5).

Estimating the size of the populationwith diabetesTo estimate the number of people withdiagnosed diabetes in 2012 we combinedU.S. Census Bureau population numberswith estimated prevalence of diabetes byage-group, sex, race/ethnicity, insurancecoverage, and whether residing in a nurs-ing home.

Combining the 2009, 2010, and2011 NHIS data produced a samplesufficient to estimate diabetes prevalenceby demographic and insurance coverage(n5 123,185). Prevalence is based on re-spondents answering “yes” to the ques-tion, “Have you EVER been told by adoctor or health professional that youhave diabetes or sugar diabetes?” We ex-clude gestational diabetes mellitus fromthe prevalence estimates. Previous re-search finds that self-report of a physi-cian’s diagnosis of diabetes is accurate in

estimating prevalence of diagnosed diabe-tes (6).

For the 2007 cost study, the esti-mated prevalence of diagnosed diabetesamong the institutionalized population(24%) came from an analysis of the 2004NNHS. There has been no update of theNNHS since 2004. Nearly one in three(32.8%) nursing home residents has di-agnosed diabetes based on a nationallyrepresentative study that analyzed medi-cal charts, minimum dataset records, andprescription claims files to identify peoplewith diabetes (7). On the basis of this up-dated information on diabetes prevalenceamongnursinghome residents,we estimateage-group–, sex-, and race/ethnicity–specific prevalence using the same distri-bution of the population demographicvariables as shown in the 2004 NNHSsurvey data among the 1.6million nursinghome residents in 2012. Few data existregarding the prevalence of diabetesamong the noncivilian population or theinstitutionalized populations other thanthose in nursing homes (e.g., in prisons).We assume that the noncivilian popula-tion and the institutionalized populationsother than those in nursing homes havediabetes prevalence similar to the nonin-stitutionalized population, controlling fordemographics, based on the limited evi-dence available (8,9).

Combining the NHIS and NNHSdata, we estimate the prevalence of di-agnosed diabetes among population sub-groups (by age-group, sex, race/ethnicity,and insurance coverage). SupplementaryTable 3 shows that prevalence of diabetesincreases with age, is somewhat higher formales than for females, and is highestamong non-Hispanic blacks. Reflectingthe high prevalence among the elderlypopulation, 13.4% of the populationwith government-sponsored medical in-surance (e.g., Medicare, Medicaid) has di-agnosed diabetes as compared with 4.6%among the privately insured and 3.7%among the uninsured populations.

State-specific estimates of diabetesprevalence (Supplementary Table 11)come from combing the 2010 ACS, the2009 and 2010 BRFSS, and the 2004NNHS. We applied a statistical matchingprocedure that randomly matches eachperson in the 2010 ACS with a similarperson either in the BRFSS (if not livingin a nursing home) or in the NNHS (ifliving in a nursing home). Each noninsti-tutionalized person in the ACS is matchedwith a person in the BRFSS in the samestate, sex, age-group (15 age-groups),

race/ethnicity, household income level(eight levels), and insured/uninsured sta-tus. Each person in the ACS in a nursing/residential facility is matched with a personin the NNHS in the same sex, age-group,and race/ethnicity. Our state prevalence es-timates are slightly different from those re-ported by the U.S. Centers for DiseaseControl and Prevention (CDC) for 2010,which are based solely on the BRFSS (10).

Estimating the direct medical costattributed to diabetesWe estimate health resource use amongthe population with diabetes in excess ofresource use that would be expected inthe absence of diabetes. Diabetes increasesthe risk of developing neurological, periph-eral vascular, cardiovascular, renal, endo-crine/metabolic, ophthalmic, and othercomplications (see Supplementary Table2 for a more comprehensive list of comor-bidities) (2). Diabetes also increases thecost of treating general conditions thatare not directly related to diabetes (2,11–13). Therefore, a portion of health careexpenditures for these medical conditionsis attributed to diabetes.

As elaborated in the 2007 study, theapproach used to quantify the increase inhealth resource use associated with di-abetes was influenced by four data limi-tations: 1) absence of a single data sourcefor all estimates, 2) small sample size insome data sources, 3) correlation of bothdiabetes and its comorbidities with otherfactors such as age and obesity, and 4)under-reporting of diabetes and its co-morbidities in certain data sources. Be-cause of these limitations we estimatediabetes-attributed costs using one oftwo approaches for each cost component.

For cost components estimated solelyfrom the MEPS (ambulance services,home health, podiatry, diabetic supplies,and other equipment and supplies), weuse a simple comparison of annual percapita health resource use for people withand without diabetes controlling forage, sex, and race/ethnicity. For nursing/residential facility use (which is not cap-tured in the MEPS) and for cost compo-nents that rely on analysis of medicalencounter data (hospital inpatient, emer-gency care, and ambulatory visits), we usean attributed risk methodology oftenused in disease-burden studies that relieson population etiological fractions (2,14).Etiological fractions estimate the excessuse of health care services among the di-abetic population relative to a similarpopulation that does not have diabetes.

1034 DIABETES CARE, VOLUME 36, APRIL 2013 care.diabetesjournals.org

Scientific Statement

©







Both approaches are equivalent under areasonable set of assumptions, but thefirst approach cannot be used with somenational data sources analyzed (e.g., NIS)that are visit/hospital discharge level files,which might or might not identify the pa-tient as having diabetes even if the patientdoes indeed have diabetes (2,14).

The attributable fraction approachcombines etiological fractions («) with to-tal projected U.S. health service use (U) in2012 for each age-group (a), sex (s), med-ical condition (c), and care delivery set-ting (H)dhospital inpatient, emergencydepartments, and ambulatory visits (phy-sician office visits combined with hospitaloutpatient/clinic visits):

Attributed health resource useH5∑

age∑sex

∑medicalcondition

«H;a;s;c 3 UH;a;s;c

The etiological fraction is calculatedusing the diagnosed diabetes prevalence(P) and the relative rate ratio (R):

«H;a;s;c5Pa;s 3

�RH;a;s;c 2 1

�

Pa;s 3�RH;a;s;c 2 1

�1 1

The rate ratio for hospital inpatientdays, emergency visits, and ambulatoryvisits represents how annual per capitahealth service use for the populationwith diabetes compares to the populationwithout diabetes:

RH;a;s;c

5annual per capita use for people with diabetesa;s;c

annual per capita use for people without diabetesa;s;c

Diabetes and its comorbidities are cor-related with other patient characteristics(e.g., demographics and body weight). Tomitigate bias caused by correlation, weestimate age/sex/setting–specific etiologi-cal fractions for each medical condition.The primary data sources for calculatingetiological fractions are OptumInsight’sdNHI data (a consolidation of the IngenixResearch Data Mart and MCURE databasesused in the 2007 study) and the 2010 5%sample Medicare SAFs. The dNHI datacontains a complete set of medicalclaims for over 23 million commerciallyinsured beneficiaries in 2011 and allowspatient records to be linked during theyear and across health delivery settings.This allows us to identify people with adiabetes ICD-9 diagnosis code (250.xx)in any of their medical claims during theyear. The Medicare 5% sample SAFs

contain claims data filed on behalf ofMedicare beneficiaries under both PartA and Part B, and like the dNHI we iden-tify people with diabetes based on dia-betes ICD-9 diagnosis codes. The largesize of these two claims databases enablesthe generation of age/sex/setting–specificrate ratios for each medical condition,which are more stable than rates estimatedusing the MEPS.

Unlike the MEPS, the dNHI data andMedicare 5% claims data do not containrace/ethnicity and select patient charac-teristics that could affect both patients’health status and health seeking behav-iors. For the 10 medical conditionsdcataract, cellulitis, conduction disordersand cardiac dysrhythmias, general medi-cal condition, heart failure, hypertension,myocardial infarction, other chronic is-chemic heart disease, renal failure andits sequelae, and urinary tract infectiondwhich are the largest contributors to theoverall cost of diabetes, we estimated twomultivariate Poisson regressions, usingdata from the MEPS, to determine the ex-tent to which controlling only for age andsex might bias the rate ratios. First, weestimated a naïve model that producesdiabetes-related rate ratios for hospital in-patient days, emergency visits, and ambu-latory visits controlling for age and sexonly. Then, we estimated a full modelthat includes diabetes status as the mainexplanatory variable and various knownpredictors of health service utilization in-cluding age, sex, education level, income,marital status, medical insurance status,and race/ethnicity as covariates. For thefull model our focus is not on the relation-ship between health care use and the co-variates (other than diabetes), but ratherthese covariates are included to controlfor patient characteristics not availablein medical claims data that could be cor-related with both medical conditions andhealth-seeking behavior. The full modelomits indicators for the presence of co-existing conditions or complications ofdiabetes (e.g., hypertension), since in-cluding such variables could bias lowthe estimated relationship between diabe-tes and health care use for each of the 10medical conditions. The rate ratio coeffi-cients for the diabetes flag variable in thenaïve and full models are then compared.The findings suggest statistically signifi-cant overestimates of the rate ratios foremergency visits when using the naïvemodel for five condition categories. Forinpatient days, we found significant over-estimates in the rate ratios for three

condition categories. For ambulatory vis-its, only hypertension was found to have asignificantly higher rate ratio by compar-ing the MEPS-based naïve model and thefull model.

To remedy the relative risk overesti-mation for these condition categories, wescaled the rate ratios estimated fromdNHI and Medicare 5% sample usingthe regression results from the MEPSanalysis by applying a scalar (with thescalar calculated as the full model rateratio divided by the naïve model rate ra-tio) (2). For emergency department visits,claims-based rate ratios were scaled downfor myocardial infarction (scale 5 0.94),other chronic ischemic heart disease(0.93), hypertension (0.71), cellulitis(0.72), and renal failure (0.95). For inpa-tient days, claims-based rate ratios werescaled down for hypertension (0.62), cel-lulitis (0.93), and renal failure (0.90).Physician office visits were scaled downfor hypertension (0.89). We did notfind a significant overestimate of the rateratios for general medical conditions forany of the three health service deliverysettings comparing the MEPS-based naïvemodel and the full model. However, acomparison of the claims-based rate ratioswith the rate ratios calculated from theMEPS-based naïve model found that theclaims-based rate ratios for general condi-tions were significantly higher than theMEPS-based rate ratios for emergencydepartment visits, hospital inpatientdays, and ambulatory visits, respectively.Therefore, to be conservative in our costestimates, we downward adjusted claims-based rate ratios for emergency departmentvisits (0.70), hospital inpatient days (0.68),and ambulatory visits (0.66) for the generalcondition group by applying a scalar calcu-lated as the MEPS-based naïve model rateratio divided by the claims-based rate ratio.

Estimates of health resource use at-tributed to diabetes were combined withestimates of the average medical cost perevent, in 2012 dollars, to compute totalmedical costs attributed to diabetes. Forhospital inpatient days, office visits, emer-gency visits, and outpatient visits, we useaverage cost per visit/day specific to themedical conditions modeled. We com-bined the 2008–2010 MEPS files to esti-mate the average cost per event, exceptthat for less common conditions or costcategories we combined the 2006–2010MEPS files to obtain a larger sample andthereby produce more precise cost esti-mates. Although the MEPS containsboth inpatient facility and professional


American Diabetes Association

©

expenditures and the NIS contains onlyfacility charges (which are converted tocosts using hospital-specific cost-to-chargeratios), the NIS has a much larger sample(n 5 ;8 million discharges in 2010) andalso contains 5-digit diagnosis codes.Therefore, we use the 2010 NIS to esti-mate inpatient facility costs and the com-bined 2008–2010 MEPS to estimatethe cost for professional services. The av-erage costs per event or day by medicalcondition are shown in SupplementaryTable 4.

Utilization of prescriptionmedication(excluding insulin and other antidiabeticagents) for each medical condition isestimated from medications prescribedduring physician’s office, emergency de-partment, and outpatient visits attributedto diabetes. The average number of med-ications prescribed during a visit for eachage-sex-race stratum was estimated from2008–2010 NAMCS and 2007–2009NHAMCS data. We calculated the totalnumber of people with diabetes that useinsulin and other antidiabetic agents bycombining diabetes prevalence and rateof use for these antidiabetic agents ob-tained from the 2009–2011 NHIS. Theaverage cost per prescription filled, insu-lin, and oral and other antidiabetic agentswere obtained from the combined MEPS2008–2010. We combined the utilizationof these medications with the average costper prescription to estimate the cost byage, sex, race/ethnicity, and insurance sta-tus. The average per capita cost for dia-betic supplies by age-sex-race stratumwas calculated from the MEPS 2008–2010. Over-the-counter medicationswere not included owing to the lack ofdata on whether diabetes increases theuse of such medications.

Consistent with the 2007 study, totalnursing/residential facility days attributedto diabetes were estimated by combiningthe average length of stay and the nursing/residential facility population. Using2004 NNHS, we calculated the numberof residents with diabetes in each age-sexstratum, which was adjusted using the32.8% diabetes prevalence estimateamong nursing home residents, obtainedfrom literature (7). Nursing/residential fa-cility use attributed to diabetes was esti-mated using an attributable risk approachwhere the prevalence of diabetes amongresidents was compared with the preva-lence of diabetes among the overall pop-ulation in the same age-sex stratum. Theanalyses were conducted separately forshort-stay, long-stay, and residential

facility residents to estimate total days ofcare. Similar to the 2007 study, cost perday was obtained from a geographicallyrepresentative cost of care survey for2012 (15).

Hospice days attributed to diabetesrepresents a combination of length of stayand diabetes prevalence among hospiceresidents. The 2007 NHHCS was used tocalculate the number of hospice residentswith diabetes and those that have a pri-mary diagnosis of diabetes along with theaverage length of stay for each age-sex-racestratum. Cost per resident per day obtainedfrom the Hospice Association of Americawas combinedwith hospice days attributedto diabetes to estimate the total cost ofhospice care attributed to diabetes.

The 2006–2010 MEPS files werecombined to increase the sample size toanalyze the use of home health, podiatry,ambulance services, and other equipmentand supplies. These cost components areestimated by comparing annual per capitacost for people with and without diabetes,controlling for age. Due to small samplesize, sex and race/ethnicity were not in-cluded as a stratum when calculatingcosts per capita.

Estimating the indirect costattributed to diabetesThe indirect costs associated with diabe-tes include workdays missed due tohealth conditions (absenteeism), re-duced work productivity while workingdue to health conditions (presenteeism),reduced workforce participation due todisability, and productivity lost due topremature mortality (16–18). Produc-tivity loss occurs among those in thelabor force as well as among the nonem-ployed population. To estimate thevalue of lost productivity, we calculatethe number of missed workdays result-ing from absenteeism, reduced workproductivity due to presenteeism, work-force participation reductions associatedwith chronic disability, and work yearslost resulting from premature mortalityassociated with diabetes. This approachmirrors the one used in the 2007 study,with the exception of adding race/ethnicityas a dimension. More recent data sourceswere used with per capita productivityloss calculated by combining the estimatesderived from the 2009–2011 NHIS andthe average annual earnings from the2011 CPS. Earnings were inflated to2012 dollars using the overall consumerprice index, and per capita estimateswere applied to the number of people

with diabetes by age-group, sex, andrace/ethnicity.

c Absenteeism is defined as the numberof workdays missed due to poor health,and prior research finds that peoplewith diabetes have higher rates of ab-senteeism than the population withoutdiabetes (16–18). Estimates of excessabsenteeism associated with diabetesrange from 1.8 to 7% of total workdays(17,19–22). Ordinary least squares re-gression with the 2009–2011 NHISshows that self-reported annual missedworkdays are statistically higher forpeople with diabetes. Control variablesinclude age-group, sex, race/ethnicity,diagnosed hypertension status (yes/no),and body weight status (normal, over-weight, obese, unknown). Diabetes isentered as a dichotomous variable (di-agnosed diabetes 5 1; otherwise 0), aswell as an interaction term with age-group. Controlling for hypertension andbody weight produces more conserva-tive estimates of the diabetes impact onabsenteeism as comorbidities of diabetesare correlated with body weight statusand a portion of hypertension is attrib-uted to diabetes. Workers with diabetesaverage three more missed workdaysthan their peers without diabetes, withexcess missed workdays varying by de-mographic group.

c Presenteeism is defined as reducedproductivity while at work, and isgenerally measured through workerresponses to surveys. These surveys relyon the self-reported inputs on thenumber of reduced productivity hoursincurred over a given time frame. Mul-tiple recent studies report that in-dividuals with diabetes display higherrates of presenteeism than their peerswithout diabetes (19,21,22). The rateof presenteeism among the populationwith diabetes exceeds rates for theircolleagues without diabetesdwith theexcess rates ranging from 1.8 to 38%of annual productivity (17,19–22).These estimates comparing presen-teeism for employees with diabetesversus those without diabetes, how-ever, fail to control for other factorsthat may be correlated with diabetes(e.g., age and weight status). Conse-quently, we model productivity lossassociated with diabetes-attributed pre-senteeism using the estimate (6.6%)from the 2007 study that controls forthe impact of factors correlated withdiabetes (2).



©



c Inability to work associated with di-abetes is estimated using a conservativeapproach that focuses on unemploy-ment related to long-term disability.The CDC estimates that roughly65,700 lower-limb amputations areperformed each year on people withdiabetes (23). These amputations andother comorbidities of diabetes canmake it difficult for some peoplewith diabetes to remain in the work-force or to find employment in theirchosen profession (22,24). To quantifydiabetes-related disability, we identifypeople in the 2009–2011 NHIS be-tween ages 18 and 65 years who receiveSupplemental Security Income (SSI)payments for disability. Using logisticregression, we estimate the relationshipbetween diabetes and the receipt ofSSI payments controlling for age-group,sex, race/ethnicity, hypertension, andweight. The results of this analysissuggest that people with diabetes have a2.4 percentage point higher rate of be-ing out of the workforce and receivingdisability payments compared withtheir peers without diabetes. The di-abetes effect increases with age andvaries by demographicdranging from0.7 percentage points for non-Hispanicwhite males aged 65–69 years to 7.4percentage points for non-Hispanicblack females aged 55–59 years. Mod-eling disability-related unemploymentis a conservative approach to modelingthe employment effect of diabetes; re-gression analysis of the NHIS suggeststhat people with diabetes have actuallabor force participation rates averag-ing approximately 10 percentage pointslower than their peers without diabetes.The average daily earnings for those inthe workforce are used as a proxy forthe economic impact of reduced em-ployment due to chronic disability. SSIpayments are considered transfer pay-ments and therefore are not included inthe social cost of not working due todisability.

c Reduced productivity for those notin the workforce is included in ourestimate of the national burden. Thispopulation includes all adults under 65years of age who are not employed(including those voluntarily or in-voluntarily not in the workforce). Thecontribution of people not in theworkforce to national productivity in-cludes time spent providing child care,household activities, and other ac-tivities such as volunteering in the

community. Prior estimates of reducedproductivity for those not in theworkforce were based on estimates of“bed days” (which is defined as a dayspent in bed because of poor health).The NHIS no longer collects data onbed days. Therefore, we use per capitaabsenteeism estimates for the workingpopulation as a proxy for reducedproductivity days among the non-employed population in a similardemographic. Whereas each work-day lost due to absenteeism is basedon estimated average daily earnings,there is no readily available measureof the value of a day lost for those not

in the workforce. Studies often useminimum wage as a proxy for thevalue of time lost, but this will un-derestimate the value of time. Usingaverage earnings for their employedcounterparts will overestimate thevalue of time. Similar to the 2007study, we use 75% of the averageearnings for people in the workforceas a productivity proxy for those un-der 65 years of age not in the laborforce (which is close to the midpointbetween minimum wage and the av-erage hourly wage earned by a de-mographic similar to the unemployedunder 65 years of age).

Table 1dHealth resource use in the U.S. by diabetes status and cost component, 2012(in millions of units)

Health resource

Population with diabetes

Incurred bypopulationwithoutdiabetes

U.S.total*

Attributedto diabetes

Incurred bypeople withdiabetes

Units% of U.S.total Units

% of U.S.total

Institutional careHospital inpatient days 26.4 15.7% 43.1 25.7% 124.9 168.0Nursing/residential facilitydays 101.3 16.4% 198.4 32.2% 418.0 616.4

Hospice days 0.2 0.3% 9.3 12.8% 63.1 dOutpatient care 1,026.7Physician office visits 85.7 8.3% 174.0 16.9% 852.8 128.7Emergency department visits 7.3 5.7% 15.3 11.9% 113.5 100.7Hospital outpatient visits 7.8 7.8% 15.0 14.9% 85.6 279.7Home health visits 25.7 9.2% 64.9 23.2% 214.7 72.4Medication prescriptions 361.4 11.8% 673.1 22.1% 2,377.9 3,051.1

Data sources: NIS (2010), NNHS (2004), NAMCS (2008–2010), NHAMCS (2007–2009), MEPS (2006–2010), and NHHCS (2007). *Numbers do not necessarily sum to totals because of rounding.

Table 2dHealth resource use attributed to diabetes in the U.S. by age-group and type ofservice, 2012 (in thousands of units)

Health resource

Age (years)

Total*(N 5 22.3 M)

,45(n 5 3.3 M)

45–64(n 5 10.2 M)

$65(n 5 8.8 M)

Institutional careHospital inpatient days 1,879 (,1%) 7,969 (37%) 16,535 (63%) 26,383Nursing/residential facility days 1,456 (,1%) 18,587 (20%) 81,288 (80%) 101,331Hospice days 0 (0%) 17 (9%) 168 (91%) 186

Outpatient carePhysician office visits 8,077 (9%) 28,437 (33%) 49,212 (57%) 85,726Emergency department visits 1,608 (22%) 2,589 (36%) 3,084 (42%) 7,280Hospital outpatient visits 1,233 (16%) 3,241 (41%) 3,342 (43%) 7,817Home health visits 3,249 (13%) 10,409 (40%) 12,076 (47%) 25,734Medication prescriptions 27,839 (8%) 118,493 (33%)215,105 (60%) 361,437

Data sources: NIS (2010), NNHS (2004), NAMCS (2008–2010), NHAMCS (2007–2009), MEPS (2006–2010), and NHHCS (2007). *Numbers do not necessarily sum to totals because of rounding.



©

c Premature mortality associated withdiabetes reduces future productivity(and not just the current year pro-ductivity). Ideally, to model the valueof lost productivity in 2012 associatedwith premature mortality one wouldcalculate the number and character-istics of all people who would havebeen alive in 2012 but who died priorto 2012 because of diabetes. Data lim-itations prevent using this approach.Instead, we estimate the number ofpremature deaths associated with di-abetes in 2012 and calculate the pres-ent value of their expected futureearnings.

To estimate the total number ofdeaths attributable to diabetes we ana-lyzed the CDC’s 2009 Mortality MultipleCause File to obtain mortality data by age,sex, and race/ethnicity for cardiovasculardisease, cerebrovascular disease, renalfailure, and diabetes. A literature reviewsupports the 2007 ADA report estimatethat ;16% of cardiovascular disease (ex-cluding cerebrovascular disease) deathscan be attributed to diabetes (1,2,25).To estimate the fraction of cerebrovascu-lar disease and renal failure deaths attrib-uted to diabetes, we used etiologicalfractions for emergency department useas a proxy for mortality etiological frac-tions (2). Our estimates suggest that;28% of deaths listing cerebrovasculardisease as the primary cause and ;55%of deaths listing renal failure as the pri-mary cause can be attributed to diabetes.The elderly represent the largest popula-tion group where deaths attributable todiabetes occur, with ;71% of deaths oc-curring among people aged $70 yearsand 8% of deaths occurring among peo-ple aged 65–69 years. To generate 2012estimates, we grow the 2009 CDC mor-tality data using the annual diabetic pop-ulation growth rate from 2009 to 2012 foreach age, sex, and race/ethnicity group.

Productivity loss associated withearly mortality is calculated by takingthe net present value of future productivity(PVFP) for men and women by ageand race/ethnicity using the same discountrate (3%), assumptions, and equationoutlined in the 2007 ADA report (2). Wecombined the average annual earningsfrom the CPS, expected mortality ratesfrom the CDC, and employment ratesfrom theCPS by age, sex, and race/ethnicityto calculate the net present value of futureearnings of a person who dies prematurely.Employment rates for 2007 (rather than

2012) are used to calculate PVFP as ratesfor 2007 are closer to the historical average(whereas rates for 2008–2012 are lowerthan average due to the recession). The re-sults incorporate U.S. Bureau of Labor Sta-tistics findings that many older workers aredelaying retirement because of the eco-nomic downturn (with ;15% employedat age 65 years and diminishing to ;5%employed at age 70 years), with this patternexpected to exist even after the economyrecovers (2).

RESULTSdIn 2012, an estimated 22.3million people in the U.S. were diagnosedwith diabetes, representing about 7% ofthe population. This estimate is higherthan but consistent with those publishedby the CDC for 2010 (23,26). The esti-mated national cost of diabetes in 2012 is$245 billion, of which $176 billion (72%)represents direct health care expendituresattributed to diabetes and $69 billion(28%) represents lost productivity fromwork-related absenteeism, reduced pro-ductivity at work and at home, unem-ployment from chronic disability, andpremature mortality.

Health resource use attributed todiabetesTable 1 shows estimates of health re-source utilization attributed to diabetesand incurred by people with diabetesas a percentage of total national utiliza-tion. For example, of the projected 168million hospital inpatient days in theU.S. in 2012, an estimated 43.1 milliondays (25.7%) are incurred by people withdiabetes of which 26.4 million days areattributed to diabetes. About one-thirdof all nursing/residential facility days areincurred by people with diabetes, andover half of those are attributed to diabe-tes. About half of all physician office vis-its, emergency department visits, hospitaloutpatient visits, and medication pre-scriptions (excluding insulin and otherantidiabetic agents) incurred by peoplewith diabetes are attributed to theirdiabetes.

Table 2 shows that the populationaged 65 years and older uses a substan-tially larger portion of services, especiallyhospital inpatient days, nursing/residentialfacility days, and hospice, comparedwith those under age 65 years. The signif-icant increase in nursing/residential daysattributed to diabetes from the 2007study reflects both the increasing costand the increased prevalence of diabetes T

able

3dHealthresource

useattributed

todiab

etes

intheU.S.by

medical

cond

itionan

dtype

ofservice,

2012

(inthousand

sof

units)

Medicalevent

Diabetes

Chron

iccomplications†

Generalmedical^

Total*

Neurological

Peripheralvascular

Cardiovascular

Renal

Metabolic

Ophthalmic

Other‡

Hospitalinp

atientdays

839(3%)

1,46

8(6%)

971(4%)

5,19

7(20%

)1,70

1(6%)

70(0%)

11(0%)

1,80

9(7%)

14,318

(54%

)26

,383

(100

%)

Physicianoffice

visits

28,150

(33%

)2,07

7(2%)

1,44

2(2%)

12,823

(15%

)3,83

1(4%)

1,21

1(1%)

6,47

3(8%)

1,51

1(2%)

28,207

(33%

)85

,726

(100

%)

Emergencydepartm

entvisits

379(5%)

209(3%)

95(1%)

747(10%

)42

4(6%)

46(1%)

22(0%)

414(6%)

4,94

4(68%

)7,28

0(100

%)

Hospitalo

utpatientvisits

3,03

8(39%

)15

5(2%)

341(4%)

1,25

1(16%

)24

1(3%)

49(1%)

283(4%)

267(3%)

2,19

2(28%

)7,81

7(100

%)

Datasources:NIS

(201

0),N

AMCS(200

8–20

10),andNHAMCS(200

7–20

09).†SeeSu

pplementary

Table2fordiagno

siscodesforeach

category

ofcomplications.‡Bacterem

ia,candidiasisof

skin

andnails,chron

icosteom

yelitisof

thefoot,other

andun

specified

noninfectious

gastroenteritis

andcolitis,impo

tenceof

organicorigin,infectiveotitisexterna,degenerative

skin

disorders,candidiasisof

vulvaandvagina,cellu

litis,

diabetes

withother

specified

manifestations,diabetes

with

unspecified

complication,

other

boneinvolvem

entin

diseaseclassified

elsewhere.^

Includ

esallother

health

care

usethat

isno

takn

owncomorbidity

ofdiabetes.*

Num

bersdo

notnecessarily

sum

tototalsbecauseof

roun

ding

.



©


(32.8%) in general, and among the elderlyin particular. Total utilization of prescrip-tionmedications attributed to diabetes hasmore than doubled from the estimate inthe 2007 study, reflecting a dramatic in-crease in the use of medications treatinggeneral conditions and diabetes comor-bidities among people with diabetes. Sup-plementary Table 5 shows the per capitahealth resource use by demographic.

Analysis of health resource use attrib-uted to diabetes by medical condition(Table 3), including diabetes, chroniccomplications of diabetes, and generalmedical conditions, shows that a largeportion of health resource use attributedto diabetesdparticularly hospital inpa-tient and emergency department visitsdisfor general medical conditions that arenot chronic complications of diabetes.As discussed in the 2007 cost of diabetesstudy, diabetes contributes to longerhospital length of stay regardless of thereason for admission (and controllingfor other factors that affect hospitallength of stay) (2). In addition to general

medical conditions, a substantial amountof attributed health resource use is forchronic complications of diabetes, partic-ularly cardiovascular diseases and renalcomplications. Finally, more than one-third of physician office visits and nearly40% of hospital outpatient visits have di-abetes listed as the primary reason forthe visit. Supplementary Table 8 showsthe proportion of total health resourceuse attributed to diabetes for each medicalcondition.

Health care expenditures attributedto diabetesHealth care expenditures attributed todiabetes reflect the additional expendi-tures the nation incurs because of diabe-tes. This equates to the total health careexpenditures for people with diabetesminus the projected level of expendituresthat would have occurred for those peo-ple in the absence of diabetes. Table 4summarizes the national expenditure forthe cost components analyzed, account-ing for over $1.3 trillion in projected

expenditure for 2012. Approximately$306 billion of the total is incurred bypeople with diabetes, reflecting 23%of the total health care dollars. Costs at-tributed to diabetes total $176 billion,or 57% of the total medical costs incurredby people with diabetes. For the costcomponents analyzed, more than 1 in ev-ery 10 health care dollars is attributed todiabetes.

National health-related expendituresare projected to exceed $2.8 trillion in2012, but slightly less than half of theseexpenditures are included in our analysis(27,28). These cost estimates omit na-tional expenditures (and any portion ofsuch expenditures that might be attributableto diabetes) for administering govern-ment health and private insurance pro-grams, investment in research andinfrastructure, over-the-counter medica-tions, disease management and wellnessprograms, and office visits to nonphysi-cian providers other than podiatrists(e.g., dentists and optometrists). Expendi-tures for health resources such as care inresidential mental retardation facilitiesare likewise excluded from the analysis.

More than 40% of all health careexpenditures attributed to diabetes comefrom higher rates of hospital admissionand longer average lengths of stay peradmission, constituting the single largestcontributor to the attributed medical costof diabetes. Of the projected $475 billionin national expenditures for hospital in-patient care (including both facility andprofessional services costs), approxi-mately $124 billion (or 26%) is incurredby people who have diabetes, of which$76 billion is directly attributed to theirdiabetes. Medications as a whole (pre-scription medications, insulin, and otherantidiabetic agents) represent over one-quarter (28%) of all health expendituresattributed to diabetes. Of the projected$286 billion in national cost for medica-tions, $77 billion (27%) is incurred bypeople with diabetes, of which $50 billionis attributed to their diabetes.

Approximately 59% of all health careexpenditures attributed to diabetes are forhealth resources used by the populationaged 65 years and older, much of which isborne by the Medicare program (Table 5).The population 45–64 years of age incurs33% of diabetes-attributed costs, with theremaining 8% incurred by the populationunder 45 years of age. The annual attrib-uted health care cost per person with di-abetes (Table 6) increases with age,primarily as a result of increased use of

Table 4dHealth care expenditures in the U.S. by diabetes status and type of service,2012 (in millions of dollars)

Cost component

Population with diabetes

Populationwithoutdiabetes Total*

Attributed todiabetes

Total incurred bypeople withdiabetes

Dollars% of U.S.total Dollars

% of U.S.total

Institutional careHospital inpatient 75,872 16% 123,726 26% 351,618 475,344Nursing/residential facility 14,748 17% 28,622 32% 59,744 88,366Hospice 32 0.3% 1,600 13% 10,889 12,489

Outpatient carePhysician office 15,221 8% 31,443 17% 155,226 186,669Emergency department 6,654 6% 14,119 12% 105,111 119,230Ambulance services 218 11% 453 23% 1,534 1,987Hospital outpatient 5,027 6% 11,354 13% 76,144 87,497Home health 4,466 9% 11,269 23% 37,264 48,533Podiatry 212 12% 458 25% 1,349 1,807

Outpatient medications and suppliesInsulin 6,157 100% 6,157 100% 0 6,157Diabetic supplies 2,296 100% 2,296 100% 0 2,296Other antidiabetic agents† 12,137 100% 12,137 100% 0 12,137Prescription medications 31,716 12% 59,067 22% 208,662 267,729Other equipment andsupplies‡ 1,063 4% 3,593 15% 20,076 23,669

Total 175,819 13% 306,293 23% 1,027,617 1,333,910

Data sources: NIS (2010), NNHS (2004), NAMCS (2008–2010), NHAMCS (2007–2009), MEPS (2006–2010), NHHCS (2007), and NHIS (2009–2011). †Includes oral medications and noninsulin injectable an-tidiabetic agents such as exenatide and pramlintide. ‡Includes, but not limited to eyewear, orthopedic items,hearing devices, prosthesis, bathroom aids, medical equipment, and disposable supplies. *Numbers do notnecessarily sum to totals because of rounding.



©




hospital inpatient and nursing facility re-sources, physician office visits, and pre-scription medications. Dividing the totalattributed health care expenditures by thenumber of people with diabetes, we esti-mate the average annual excess expendi-tures for the population aged under 45years, 45–64 years, and 65 years andabove, respectively, at $4,394, $5,611,and $11,825. Total health care expendi-tures are attributed to diabetes by sex andrace/ethnicity (Supplementary Table 6),insurance status (Supplementary Table9 and 10), and state (Supplementary Ta-ble 11).

Table 7 summarizes diabetes-attributedhealth care expenditures for those costcomponents modeled by medical con-dition. Hospital inpatient is the largestcomponent of attributed costs followedby physician office visit. Across differenthealth care delivery settings, general med-ical conditions and cardiovascular diseasecategories are the two largest contributorsof total health care expenditures attrib-uted to diabetes in addition to diabetesitself. Together, the general medical con-ditions and cardiovascular disease catego-ries are responsible for 78% of hospitalinpatient costs attributed to diabetes,

47% of the cost for physician office visits,82% of the cost for emergency departmentvisits, and 52% of the cost for hospitaloutpatient.

Figure 1 summarizes the proportionof medical expenditures attributed to di-abetes for each chronic complication overthe total U.S. health care expenditurecombining expenditures for hospital in-patient, hospital outpatient, emergencydepartment visits, physician office visits,and prescription medications. Over aquarter of expenditures, in five out ofthe eight conditions shown in the chart,are attributed to diabetes. In addition, 7,11, and 21% of national medical expendi-tures treating general conditions, endocrine/metabolic complications, and ophthal-mic complications are attributable todiabetes.

The population with diabetes is olderand sicker than the population withoutdiabetes, and consequently annual med-ical expenditures are much higher (onaverage) than for people without diabetes(Table 8). After adjusting for age-sex dif-ferences in these two populations, peoplewith diabetes have health care expendi-tures that are 2.3 times higher ($13,741vs. $5,853) than expenditures would be

expected for this same population in theabsence of diabetes. This suggests that di-abetes is responsible for $7,888 in excessexpenditures per year per person with di-abetes. This 2.3 multiple is unchangedfrom the 2007 study.

Indirect costs attributed to diabetesThe total indirect cost of diabetes isestimated at $68.6 billion (Table 9). Themajority of this burden comes from un-employment due to permanent disability($21.6 billion), presenteeism ($20.8 bil-lion), and premature mortality ($18.5 bil-lion). Workdays absent ($5.0 billion) andreduced productivity for those not in theworkforce ($2.7 billion) represent a rela-tively small portion of the total burden.

Our logistic regression analysis withNHIS data suggests that diabetes is asso-ciated with a 2.4 percentage point in-crease in the likelihood of leaving theworkforce for disability. This equates toapproximately 541,000 working-ageadults leaving the workforce prematurelyand 130 million lost workdays in 2012.For the population that leaves the work-force early because of diabetes-associateddisability, we estimate that their averagedaily earnings would have been $166 perperson (with the amount varying by de-mographic).

Presenteeism accounted for 30% ofthe indirect cost of diabetes. The estimateof a 6.6% annual decline in productivityattributed to diabetes (in excess of theestimated decline in the absence of di-abetes) equates to 113 million lost work-days per year. The average daily earningsare $185 for the employed populationwith diabetes, which equates to $20.8 bil-lion in annual cost attributed to diabetes(after factoring out absenteeism to pre-vent double counting).

The estimated number of deaths in2012 attributable to diabetes is 246,000(Table 10). For 73,000 deaths (30%), di-abetes is listed as the primary cause. Ofthe 687,000 deaths where cardiovasculardisease is listed as the primary cause, ap-proximately 110,000 (16%) are attribut-able to diabetes. Approximately 38,000cases where cerebrovascular disease islisted as the primary cause of death areattributable to diabetes, and 25,000 caseswhere renal disease is listed as the primarycause of death are attributable to diabetes.The average cost per premature death de-clines with age (reflecting fewer remain-ing expected working years), and acrossall premature deaths averaged approxi-mately $75,100 per case.

Table 5dHealth care expenditures attributed to diabetes in the U.S. by age-group andtype of service, 2012 (in millions of dollars)

Cost component

Age (years)

Total*(N 5 22.3 M)

,45(n 5 3.3 M)

45–64(n 5 10.2 M)

$65(n 5 8.8 M)

Institutional careHospital inpatient 4,924 (6%) 22,934 (30%) 48,015 (63%) 75,872Nursing/residential facility 211 (1%) 2,781 (19%) 11,757 (80%) 14,748Hospice 0 (0%) 3 (9%) 29 (91%) 32

Outpatient carePhysician office 1,334 (9%) 4,882 (32%) 9,005 (59%) 15,221Emergency department 1,435 (22%) 2,363 (36%) 2,856 (43%) 6,654Ambulance services 20 (9%) 169 (77%) 29 (13%) 218Hospital outpatient 679 (13%) 1,943 (39%) 2,405 (48%) 5,027Home health 564 (13%) 1,806 (40%) 2,096 (47%) 4,466Podiatry 43 (20%) 61 (29%) 108 (51%) 212

Outpatient medications and suppliesInsulin 1,102 (18%) 2,817 (46%) 2,239 (36%) 6,157Diabetic supplies 238 (10%) 1,003 (44%) 1,056 (46%) 2,296Other antidiabetic agents† 1,297 (11%) 5,767 (48%) 5,073 (42%) 12,137Prescription medications 2,443 (8%) 10,398 (33%) 18,875 (60%) 31,716Other equipment and supplies‡ 117 (11%) 309 (29%) 637 (60%) 1,063

Total 14,406 (8%) 57,235 (33%) 104,178 (59%) 175,819

Data sources: NIS (2010), NNHS (2004), NAMCS (2008–2010), NHAMCS (2007–2009), MEPS (2006–2010), NHHCS (2007), and NHIS (2009–2011). †Includes oral medications and noninsulin injectable an-tidiabetic agents. ‡Includes but not limited to eyewear, orthopedic items, hearing devices, prosthesis,bathroom aids, medical equipment, and disposable supplies. *Numbers do not necessarily sum to totalsbecause of rounding.



©






Figure 2 summarizes estimates ofPVFP if a person dies at that age. PVFPis the value in 2012 of expected futurelifetime earnings if the person had livedto the average age as the cohort born inthe same year. The differences in PVFP bydemographic reflect the differences in av-erage earnings, the propensity to be in theworkforce, and the number of years ex-pected to remain in the workforce.

The cost of missed workdays due toabsenteeism is estimated at $5.0 billion,representing 25 million days. If peoplenot in the workforce have similar ratesof days where they are unable to workdue to poor health as their employedpeers, this would equate to 20 million ex-cess sick days with the estimated produc-tivity loss valued at $2.7 billion. We donot count productivity loss for the popu-lation under age 18 years. While childrenconstitute a small proportion of the pop-ulation with diabetes, omitting produc-tivity loss associated with diabetesamong children will tend to bias low thecost estimates. For example, the eco-nomic cost associated with parents whotake time off from work to take their chil-dren to the doctor for diabetes-related vis-its is omitted from these cost estimates.

The average annual productivity lossper person aged 18 years or older withdiabetes is $3,100. Table 11 shows thatper capita estimates range from a high of$6,844 for men aged 45–54 years to a lowof $647 for women aged 70 years andolderdreflecting differences by demo-graphic in propensity to be in the work-force, average earnings, and mortalityrisk. Supplementary Table 7 shows theannual productivity loss per person withdiabetes by cause and race/ethnicity.

CONCLUSIONSdThis study foundthat there were more than 22.3 millionpeople (about 7% of the U.S. population)with diagnosed diabetes in the U.S. in2012. This is substantially higher than the2007 estimate of 17.5 million people,reflecting changing demographics, in-crease in the prevalence of risk factorsincluding obesity, decreasing mortality,and improvements in the detection ofdiabetes (29–32). Diabetes costs thenation a total of $245 billion, which in-cludes $176 billion in direct medical costand $69 billion in lost productivity.While the majority (59%) of direct med-ical cost is for the population aged 65years and over, about 88% of indirect

cost is borne by the population under65 years of age. We also found that afteradjusting for age and sex, annual per cap-ita health care expenditure is 2.3 timeshigher for people with diabetes than forthose without diabetes. Diabetes is espe-cially costly when it is associated withcomplications. While we were unable tocalculate diabetes-attributed cost by com-plication groups for every cost compo-nent across the major health caredelivery settings (hospital inpatient andoutpatient, physician office, and emer-gency department), from 25% (emer-gency department) to 45% (hospitalinpatient) of the diabetes-attributed med-ical expenditures were spent treatingcomplications of diabetes. Other studiesfound that people with uncontrolled di-abetes or with diabetes complications in-cur diabetes costs two to eight times morethan people with controlled or nonad-vanced diabetes (33,34).

For comparison, the $174 billion es-timate of the total burden for 2007 pub-lished previously is equivalent to $202billion when inflated to 2012 dollars us-ing the average general inflation rate of3%. The increase of $43 billion from the2007 estimate in 2012 dollars to the newestimate of $245 billion reflects 1) a 27%growth in diabetes prevalence, 2) chang-ing demographics of people with diabe-tes, 3) growth in the utilization of certaintypes of health care services for treatingdiabetes and its comorbidities such as in-creased use of prescription medicationsand advanced treatment for cardiovascu-lar disease, 4) rising prices for medicalgoods and services above the generalrate of inflation, and 5) refinements tothe data and methods used to calculatethe cost of diabetes.

We found that the proportions of to-tal national health services use attributedto diabetes and incurred by people withdiabetes both increased from the esti-mates in the 2007 study, including utili-zation of nursing/residential facility days,physician office visits, emergency depart-ment visits, hospital outpatient visits, andprescription medications. The number ofhospital inpatient days incurred by peo-ple with diabetes and those that are at-tributable to their diabetes have bothincreased from the 2007 level by about6 and 9%, respectively, although the na-tional utilization of hospital inpatientcare has decreased by about 10% from186 million days in 2007 to 168 milliondays in 2012 based on the analysis ofNIS data.

Table 6dAnnual per capita health care expenditures attributed to diabetes in the U.S.by age-group and type of service, 2012 (in actual dollars)

Cost component

Age (years)

All ages(N 5 22.3 M)

,45(n 5 3.3 M)

45–64(n 5 10.2 M)

$65(n 5 8.8 M)

Institutional careHospital inpatient 1,502 2,248 5,450 3,404Nursing/residential facility 64 273 1,334 662Hospice 0.01 0.29 3 1

Outpatient carePhysician office 407 479 1,022 683Emergency department 438 232 324 299Ambulance services 6 17 3 10Hospital outpatient 207 191 273 226Home health 172 177 238 200Podiatry 13 6 12 10

Outpatient medications and suppliesInsulin 336 276 254 276Diabetic supplies 73 98 120 103Other antidiabetic agents† 396 565 576 544Prescription medications 745 1,019 2,142 1,423Other equipment and supplies‡ 36 30 72 48

Total* 4,394 5,611 11,825 7,888

Data sources: NIS (2010), NNHS (2004), NAMCS (2008–2010), NHAMCS (2007–2009), MEPS (2006–2010), NHHCS (2007), NHIS (2009–2011), and the U.S. Census Bureau (2012). †Includes oral medicationsand noninsulin injectable antidiabetic agents. ‡Includes but not limited to eyewear, orthopedic items, hearingdevices, prosthesis, bathroom aids, medical equipment, and disposable supplies. *Numbers do not neces-sarily sum to totals because of rounding.



©


Additionally, even when using MEPSdata that have been shown to underesti-mate costs when compared with claimsdata, especially for the privately insured(35), we found that the price of medicalservices per event (visit or day) has in-creased by 5–17% over the rate of generalinflation from the 2007 level for hospitalinpatient, hospital outpatient, emergencydepartment, insulin, and other prescrip-tion medications. Due to the increase indiabetes prevalence, health resource utili-zation, and average per event cost of services,the $176 billion direct medical cost attrib-uted to diabetes in 2012 is 30% higherthan the general inflation-adjusted 2007direct medical cost of $135 billion.

The indirect cost estimate of $69 bil-lion for 2012 includes increased ab-senteeism ($5 billion) and reducedproductivity while at work ($20.8 billion)for the employed population, reducedproductivity for those not in the laborforce ($2.7 billion), unemployment as aresult of disease-related disability ($21.6billion), and lost productive capacitydue to early mortality ($18.5 billion).The $69 billion is only 3% higher thanthe inflation-adjusted 2007 estimate of$67 billion, despite the 27% growth indiabetes prevalence. Factors depressingthe 2012 estimate include the declinein the number of people participating inthe workforce in 2012 and the lowerdiabetes-attributed mortality estimates for2012. Including race/ethnicity as a study

dimension also depressed the national in-direct burden estimate relative to 2007, asHispanics and non-Hispanic blacks havehigher diabetes prevalence rates but lowerlabor force participation rates and loweraverage earnings. Since the 2007 study,the economic downturn has decreasedoverall rates of employment across all de-mographic groups regardless of diabetesstatus. A declining proportion of the adultpopulation in the workforce depresses theestimates of absenteeism and presenteeism,while increasing the estimates of diabetes-related productivity losses for the popula-tion not in the workforce.

Our estimate of $245 billion only rep-resents the economic cost of diagnoseddiabetes. An earlier study found that 6.3million U.S. adults have undiagnosed di-abetes with an associated cost of $18 bil-lion in 2007 (36). Furthermore, nearly 57million adults in that study were esti-mated to have prediabetes, a precursorto diabetes, costing an additional $25 bil-lion in higher medical spending (37,38).On the surface it appears that the financialburden of diabetes falls primarily on in-surers who pay a substantial portion ofmedical costs, employers who experienceproductivity loss, and the people with di-abetes and their families who incur higherout-of-pocket medical costs and reducedearnings potential or employment oppor-tunities. Ultimately, though, the burden ispassed along to all of society in the form ofhigher insurance premiums and taxes,

Table

7dHealthcare

expend

itures

attributed

todiab

etes

intheU.S.by

medical

cond

itionan

dtype

ofservice,

2012

(inmillions

ofdolla

rs)

Typ

eof

service

Diabetes

Chron

iccomplications†

Neurological

Peripheralvascular

Cardiovascular

Renal

Metabolic

Ophthalmic

Other‡

Generalmedical^

Total*

Hospitalinpatient

1,97

9(3%)

4,22

9(6%)

2,81

3(4%)

19,441

(26%

)3,80

7(5%)

175(0%)

28(,

1%)

4,00

2(5%)

39,399

(52%

)75

,872

Physicianoffice

4,13

6(27%

)41

3(3%)

449(3%)

2,19

6(14%

)1,00

7(7%)

176(1%)

1,48

3(10%

)29

5(2%)

5,06

5(33%

)15

,221

Emergency

department

301(5%)

161(2%)

86(1%)

805(12%

)32

4(5%)

64(1%)

16(,

1%)

262(4%)

4,63

5(70%

)6,65

4Hospitalo

utpatient

1,10

0(22%

)11

5(2%)

555(11%

)63

5(13%

)14

7(3%)

18(,

1%)

314(6%)

173(3%)

1,97

1(39%

)5,02

7

Datasources:N

IS(201

0),N

AMCS(200

8–20

10),NHAMCS(200

7–20

09),andMEPS

(200

6–20

10).†SeeSu

pplem

entary

Table2fordiagno

siscodesforeach

category

ofcomplications.‡Bacterem

ia,candidiasisofskin

andnails,chron

icosteom

yelitisofthefoot,otherandun

specified

noninfectiou

sgastroenteritisand

colitis,impo

tenceoforganicorigin,infective

otitisexterna,degenerative

skindisorders,candidiasiso

fvulva

andvagina,

cellu

litis,diabeteswithotherspecified

manifestations,diabeteswith

unspecified

complication,otherbo

neinvolvem

entindiseaseclassified

elsewhere.^Includ

esallotherhealth

careusethatisno

takn

owncomorbidity

ofdiabetes.*

Num

bersdo

notnecessarily

sum

tototalsbecauseof

roun

ding

.

Figure 1dPercent of medical condition–specific expenditures associated with diabetes. Datasources: NIS (2010), NAMCS (2008–2010), NHAMCS (2007–2009), and MEPS (2006–2010 or2008–2010). Note: See Supplementary Table 2 for diagnosis codes for each category of medicalcondition.



©



reduced earnings, and reduced standardof living.

The cost estimates presented mightbe conservative for several reasons:

c Due to data limitations, we omittedfrom this analysis the potential increasein the use of over-the-counter medi-cations and optometry and dentalservices. Diabetes increases the risk ofperiodontal disease, so one would ex-pect dental costs to be higher for peoplewith diabetes. We explored the MEPSdata for the feasibility of capturing op-tometry and dental costs, but the smallsample sizes prevented meaningfulanalyses. Also omitted from the costestimates are expenditures for the pre-vention programs targeted to peoplewith diabetes (e.g., diseasemanagementprograms), research activities (e.g., todevelop new drugs), and administra-tion costs (e.g., to administer theMedicare and Medicaid programs, toprocess insurance claims). Administra-tion costs for government health pro-grams and private insurers are ; $150billion per year. Public and private ex-penditures for medical research andhealth infrastructure total over $130billion per year (39). If a portion ofthese costs were attributed to diabetes,

the national cost of diabetes would bebillions of dollars higher than our esti-mate suggests.

c Also omitted from the cost estimatesare the intangible costs of diabetes such

as pain, suffering, and reduced qualityof life, as well as some of the non-medical costs attributed to diabetes.Specifically, diabetic patients with ad-vanced diabetic retinopathy, late-stage

Figure 2dNet present value of future lost earnings from premature death. Data sources: analysisof the NHIS (2009–2011), CPS (2011), and CDC mortality data.

Table 8dAnnual per capita health care expenditures in the U.S. by diabetes status, 2012 (in actual dollars)

Cost component With diabetes ($)

Unadjusted Adjusted for age and sex

Withoutdiabetes ($)

Ratio with towithout diabetes

Withoutdiabetes ($)

Ratio with towithout diabetes

Attributed todiabetes ($)

Institutional careHospital inpatient 5,551 1,196 4.6 2,147 2.6 3,404Nursing/residential facility 1,284 203 6.3 622 2.1 662Hospice N/A N/A N/A N/A N/A N/A

Outpatient carePhysician office 1,411 528 2.7 728 1.9 683Emergency 633 357 1.8 335 1.9 299Ambulance services 20 5 3.9 11 1.9 10Hospital outpatient and freestandingambulatory surgical center 509 259 2.0 284 1.8 226

Home health 506 127 4.0 305 1.7 200Podiatry 21 5 4.5 11 1.9 10

Outpatient medications and suppliesInsulin 276 NA NA NA NA 276Diabetic supplies 103 NA NA NA NA 103Other antidiabetic agents† 544 NA NA NA NA 544Prescription medications 2,650 710 3.7 1,227 2.2 1,423Other equipment and supplies‡ 161 68 2.4 113 1.4 48

Total 13,741 3,495 3.9 5,853 2.3 7,888

Data sources: NIS (2010), NNHS (2004), NAMCS (2008–2010), NHAMCS (2007–2009), MEPS (2006–2010), NHHCS (2007), NHIS (2009–2011), and the U.S.Census Bureau (2012). N/A, not available; NA, not applicable. †Includes antidiabetic agents such as exenatide and pramlintide. ‡Includes but not limited to eyewear,orthopedic items, hearing devices, prosthesis, bathroom aids, medical equipment, and disposable supplies.



©

renal complications, or lower-extremityamputations often require their homesand/or motor vehicles to be modified toaccommodate their daily activity needs.Diabetes is the leading cause of new casesof blindness among adults aged 20–74years (23), and the CDC estimates thatroughly 65,700 lower-limb amputationsare performed each year on people withdiabetes (23). The nonmedical cost asso-ciated with these disabilities could furtherincrease the total burden of diabetes.

c The lost productivity estimates are forthose individuals with diagnosed di-abetes and exclude lost productivityassociated with the care for familymembers with diabetes. For example,the productivity loss associated withadults who take time off from work tocare for a child or an elderly parent withdiabetes is not included in the cost es-timates. The value of informal caregiv-ing is excluded from our cost estimate.Time and costs associated with travel-ing to doctor visits and other medicalemergencies are omitted (except to theextent that such costs are partially

captured under ambulance costs andthe absenteeism estimate for those inthe workforce).

c Our estimate of lost productivity at-tributed to chronic disability from di-abetes is also likely to be conservativedue to three factors: 1) using SSI pay-ments to identify cases of disabilitylikely underestimates disability casesbecause the criteria for SSI eligibilityinclude requirements for documenta-tion of disability from a health pro-fessional and apply income limits; 2)these estimates omit the value of pro-ductivity loss that results in reducedearnings potential but does not preventworking; and 3) productivity loss as-sociated with early retirement is notincluded, and a longitudinal study us-ing the Health and Retirement Surveyfound that people with diabetes tend toretire ;1.2 years earlier than theirpeers without diabetes (40).

One challenge for this study was tocontrol for the correlation between di-abetes and the use of health resources for

reasons not directly attributed to diabetes.Health behavior that affects both thepresence of diabetes and the presence ofother comorbidities, unless controlledfor, could result in an overestimation ofthe link between diabetes and the use ofhealth resources. Controlling for age, sex,and race/ethnicity helps to control for thiscorrelation. In addition, for the top 10cost drivers, we conducted additionalanalysis controlling for other importantexplanatory variables using the MEPSdata. Based on the results, we reducedthe etiological fractions for several diabe-tes complications and for the generalmedical conditions group depending onthe setting of care. This potential limita-tion also applies to the estimates of in-direct costs attributed to diabetes,especially the estimated productivityloss due to presenteeism.

Other study limitations discussedpreviously include small sample size forsome data sources used, the use of a datasource (dNHI) that overrepresents thecommercially insured population for thepopulation younger than age 65 years,and the need to use different approachesto model different cost components be-cause of data limitations. Another limita-tion common to claims-based analysis isthe possibility of inaccurate diagnosiscodes. Claims data tend to be less accuratethan medical records in identifying pa-tients with specific conditions due toreasons such as rule-out diagnosis, cod-ing error, etc. The direction of such biason our risk ratio calculations is unknown,although it is anticipated to be small asthere is no reason to believe that thecoding of comorbidities would be signif-icantly different for people with andwithout diabetes.

Using a methodology that is largelyconsistent with our previous cost of di-abetes study in 2007 with updated na-tional survey and claims data fromprevious data sources, we estimated thetotal burden of diabetes in 2012. Theestimates presented here show that di-abetes places an enormous burden onsocietydboth in the economic terms pre-sented here and in reduced quality of life.The overall cost of diabetes estimates areconsistentwith earlier estimates after adjust-ing for the increasing prevalence of diabetesand price increases (though estimates forsome cost components and medical condi-tions differ from the earlier study).

A recent study estimates that preva-lence of diagnosed diabetes is likely to atleast double between 2010 and 2050, and

Table 9dIndirect burden of diabetes in the U.S., 2012 (in billions of dollars)

Cost component Productivity lossTotal cost attributable

to diabetes ($)Proportion ofindirect costs*

Workdays absent 25 million days 5.0 7%Reduced performance at work 113 million days 20.8 30%Reduced productivity days forthose not in labor force 20 million days 2.7 4%

Reduced labor force participationdue to disability 130 million days 21.6 31%

Mortality 246,000 deaths 18.5 27%Total 68.6 100%

Data sources: analysis of the NHIS (2009–2011), CPS (2011), CDC mortality data, and the U.S. CensusBureau population estimates for 2010 and 2012. *Numbers do not necessarily sum to totals because ofrounding.

Table 10dMortality costs attributed to diabetes, 2012

Primary cause of death

Total U.S.deaths

(thousands)*

Deaths attributed to diabetes

Deaths(thousands)

% of U.S.deaths incategory

Value of lostproductivity

(millions of dollars)

Diabetes 73 73 100.0% 7,147Renal disease 46 25 55.0% 2,004Cerebrovascular disease 136 38 28.0% 1,484Cardiovascular disease 687 110 16.0% 7,827Total N/A 246 N/A 18,462

*Data source: CDC National Vital Statistics Reports for total deaths in 2009 by primary cause of death, scaledto 2012 using the annual diabetic population growth rate from 2009 to 2012 for each age, sex, and race/ethnicity group (42).



©

the prevalence of total diabetes (diag-nosed and undiagnosed) may increasefrom the 2010 level of about one in nineadults to between one in five and one inthree adults in 2050 (3,41).

This study highlights the large eco-nomic burden of diabetes and its compli-cations on the individual and the healthcare system. Cost estimates from 2002,2007, and now 2012 show that theburden is increasingdeven after control-ling for population growth and inflation.Cost comparisons by age-group showthat the burden of diabetes increaseswith age. These trends underscore the im-portance of prevention and the efforts tomitigate the complications of diabetes.

AcknowledgmentsdNo potential conflicts ofinterest relevant to this article were reported.

References1. AmericanDiabetes Association. Economic

costs of diabetes in the U.S. in 2002. Di-abetes Care 2003;26:917–932

2. AmericanDiabetes Association. Economiccosts of diabetes in the U.S. in 2007. Di-abetes Care 2008;31:596–615

3. Boyle JP, Thompson TJ, Gregg EW, BarkerLE, Williamson DF. Projection of the year2050 burden of diabetes in the US adultpopulation: dynamic modeling of in-cidence, mortality, and prediabetes prev-alence. Popul Health Metr 2010;8:29

4. U.S. Census Bureau. Population Projec-tions: 2008National Population Projections[Internet], 2008. Available from http://www.census.gov/population/projections/data/national/2008.html. Accessed 16 Jan-uary 2013

5. U.S. Bureau of Labor Statistics. ConsumerPrice Index: CPI Databases [Internet].Available from http://www.bls.gov/cpi/data.htm. Accessed 16 January 2013

6. Okura Y, Urban LH, Mahoney DW,Jacobsen SJ, Rodeheffer RJ. Agreementbetween self-report questionnaires andmedical record data was substantial fordiabetes, hypertension, myocardial in-farction and stroke but not for heartfailure. J Clin Epidemiol 2004;57:1096–1103

7. Dybicz SB, Thompson S, Molotsky S,Stuart B. Prevalence of diabetes and theburden of comorbid conditions amongelderly nursing home residents. Am JGeriatr Pharmacother 2011;9:212–223

8. American Diabetes Association. Diabetesmanagement in correctional institutions(Position Statement). Diabetes Care 2011;34(Suppl. 1):S75–S81

9. Paris RM, Bedno SA, Krauss MR, Keep LW,Rubertone MV. Weighing in on type 2diabetes in the military: characteristics ofU.S. military personnel at entry who de-velop type 2 diabetes. Diabetes Care 2001;24:1894–1898

10. U.S. Centers for Disease Control andPrevention. CDC’s State SurveillanceData. Available from http://www.cdc.gov/nceh/lead/data/state.htm. Accessed 16January 2013

11. Osborn DP, Holt, R. Diabetes and mentalhealth. In Diabetes: Chronic Complica-tions. 3rd ed. Shaw KM, Cummings MH,Eds. Hoboken, NJ, John Wiley &Sons,Inc., 2012, p. 214–239

12. Vigneri P, Frasca F, Sciacca L, Pandini G,Vigneri R. Diabetes and cancer. EndocrRelat Cancer 2009;16:1103–1123

13. Egede LE, Zheng D, Simpson K. Co-morbid depression is associated with in-creased health care use and expendituresin individuals with diabetes. DiabetesCare 2002;25:464–470

14. Benichou J. A review of adjusted estima-tors of attributable risk. Stat MethodsMedRes 2001;10:195–216

15. Genworth Financial. Genworth 2012Cost of Care Survey [Internet], 2012.Genworth Financial, Inc., Richmond,VA. Available from www1.genworth.com/content/etc/medialib/genworth_v2/pdf/ltc_cost_of_care.Par.40001.File.dat/2012%20Cost%20of%20Care%20Survey%20Full%20Report.pdf. Accessed 16 January 162013

16. Cawley J, Rizzo JA, Haas K. The associationof diabetes with job absenteeism costsamong obese and morbidly obese workers.J Occup Environ Med 2008;50:527–534

17. Fu AZ, Qiu Y, Radican L, Wells BJ. Healthcare and productivity costs associatedwith diabetic patients with macrovascularcomorbid conditions. Diabetes Care2009;32:2187–2192

18. Lee LJ, Yu AP, Cahill KE, et al. Direct andindirect costs among employees with di-abetic retinopathy in the United States.Curr Med Res Opin 2008;24:1549–1559

Table 11dAnnual productivity loss per person with diabetes in the U.S. by age, sex, and cause, 2012 (in actual dollars)

Sex Age Absenteeism PresenteeismReduced productivity forthose not in labor force

Unemploymentfrom disability

Prematuremortality

Total annualburden

Male 18–34 170 1,147 61 769 2,408 4,55635–44 403 2,187 117 1,341 2,442 6,49045–54 811 1,691 336 1,416 2,591 6,84455–59 419 1,816 221 1,577 1,116 5,14960–64 211 1,530 188 1,413 463 3,80565–69 89 878 d 417 209 1,593701 d 305 d 503 68 876

Total 298 1,246 135 1,034 1,100 3,813Female 18–34 114 769 66 798 1,100 2,847

35–44 241 1,310 113 1,228 1,409 4,30145–54 436 908 297 1,241 1,340 4,22255–59 224 970 196 1,453 559 3,40160–64 93 679 142 1,224 256 2,39465–69 36 354 d 343 116 849701 d 132 d 469 46 647

Total 149 614 111 901 548 2,322

Data sources: analysis of the NHIS (2009–2011), CPS (2011), and CDC mortality data. Note: Age ,18 years is not included as no indirect costs are calculated forpersons under the age of 18. For the age 70 years and older population, the rate of labor force participation is low so indirect costs are relatively low for this populationdespite high prevalence of diabetes. The NHIS sample size of employed people over age 70 years is small, and regression analysis with the NHIS found that diabetes isnot associated with increased workdays absent for illness among the employed population aged 70 years and older. We conservatively assume that for the populationaged 65 years and older and not in the workforce there is no loss in societal productivity (e.g., from volunteer work) associated with diabetes.



©

http://www.census.gov/population/projections/data/national/2008.html



http://www.bls.gov/cpi/data.htm

http://www.bls.gov/cpi/data.htm

http://www.cdc.gov/nceh/lead/data/state.htm

http://www.cdc.gov/nceh/lead/data/state.htm

http://www1.genworth.com/content/etc/medialib/genworth_v2/pdf/ltc_cost_of_care.Par.40001.File.dat/2012%20Cost%20of%20Care%20Survey%20Full%20Report.pdf





19. DiBonaventura M, Link C, Pollack MF,Wagner J-S, Williams SA. The relation-ship between patient-reported tolerabilityissues with oral antidiabetic agents andwork productivity among patients havingtype 2 diabetes. J Occup Environ Med2011;53:204–210

20. Lamb CE, Ratner PH, Johnson CE, et al.Economic impact ofworkplace productivitylosses due to allergic rhinitis comparedwithselect medical conditions in the UnitedStates from an employer perspective. CurrMed Res Opin 2006;22:1203–1210

21. Loeppke R, Taitel M, Haufle V, Parry T,Kessler RC, Jinnett K. Health and pro-ductivity as a business strategy: a multi-employer study. J Occup Environ Med2009;51:411–428

22. Rodbard HW, Fox KM, Grandy S; ShieldStudy Group. Impact of obesity on workproductivity and role disability in in-dividuals with and at risk for diabetesmellitus. Am J Health Promot 2009;23:353–360

23. U.S. Centers for Disease Control andPrevention. Diabetes Public Health Re-source: 2011 National Diabetes Fact Sheet[Internet], 2011. Available from http://www.cdc.gov/diabetes/pubs/estimates11.htm. Accessed 16 January 2013

24. Wolf AM, Siadaty MS, Crowther JQ, et al.Impact of lifestyle intervention on lostproductivity and disability: improvingcontrol with activity and nutrition. J Oc-cup Environ Med 2009;51:139–145

25. Carnethon MR, Biggs ML, Barzilay J, et al.Diabetes and coronary heart disease asrisk factors for mortality in older adults.Am J Med 2010;123:556.e1–e9

26. U.S. Centers for Disease Control andPrevention. Diabetes Data & Trends:

Number (in Millions) of Civilian, Non-institutionalized Persons with DiagnosedDiabetes, United States, 1980–2010 [In-ternet], 2011. Available from http://www.cdc.gov/diabetes/statistics/prev/national/figpersons.htm. Accessed 17 January 2013

27. Poisal JA, Truffer C, Smith S, et al. Healthspending projections through 2016: mod-est changes obscure Part D’s impact. HealthAff (Millwood) 2007;26:w242–w253

28. Smith C, Cowan C, Heffler S, Catlin A.National health spending in 2004: recentslowdown led by prescription drug spend-ing. Health Aff (Millwood) 2006;25:186–196

29. U.S. Centers for Disease Control andPrevention. Overweight andObesity: AdultObesity Facts [Internet], 2012. Availablefrom http://www.cdc.gov/obesity/data/adult.html. Accessed 16 January 2013

30. Hoyert DL. 75 Years of Mortality in theUnited States, 1935–2010. NCHS databrief, no. 88. Hyattsville, MD, NationalCenter for Health Statistics, 2012

31. U.S. Preventive Services Task Force.Screening for Type 2 Diabetes Mellitus inAdults [Internet], 2008. Available fromhttp://www.uspreventiveservicestaskforce.org/uspstf/uspsdiab.htm. Accessed 16 Jan-uary 2013

32. Mokdad AH, Ford ES, Bowman BA, et al.Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA2003;289:76–79

33. Kim S. Burden of hospitalizations pri-marily due to uncontrolled diabetes: im-plications of inadequate primary healthcare in the United States. Diabetes Care2007;30:1281–1282

34. Brown JB, Pedula KL, Bakst AW. Theprogressive cost of complications in type 2

diabetes mellitus. Arch Intern Med 1999;159:1873–1880

35. Aizcorbe A, Liebman E, Pack S, Cutler DM,Chernew ME, Rosen AB. Measuring healthcare costs of individuals with employer-sponsored health insurance in the U.S.:a comparison of survey and claims data.Journal of the International Association forOfficial Statistics 2012;28:43–51

36. Zhang Y, Dall TM, Mann SE, et al. Theeconomic costs of undiagnosed diabetes.Popul Health Manag 2009;12:95–101

37. Zhang Y, Dall TM, Chen Y, et al. Medicalcost associated with prediabetes. PopulHealth Manag 2009;12:157–163

38. Dall TM, Zhang Y, Chen YJ, Quick WW,Yang WG, Fogli J. The economic burdenof diabetes. Health Aff (Millwood) 2010;29:297–303

39. Centers for Medicare & Medicaid Services.National Health Expenditure Projections2010–2020: Forecast Summary. Availablefrom https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/downloads/proj2010.pdf. Accessed17 January 2013

40. Vijan S, Hayward RA, Langa KM. Theimpact of diabetes on workforce partici-pation: results from a national householdsample. Health Serv Res 2004;39(6 Pt 1):1653–1669

41. NarayanKM,Boyle JP,Geiss LS, Saaddine JB,Thompson TJ. Impact of recent increasein incidence on future diabetes burden:U.S., 2005–2050. Diabetes Care 2006;29:2114–2116

42. Hoyert DL, Xu JQ. Deaths: PreliminaryData for 2011. National Vital StatisticsReports, vol. 61, no. 6. Hyattsville, MD,National Center for Health Statistics, 2012



©

http://www.cdc.gov/diabetes/pubs/estimates11.htm



http://www.cdc.gov/diabetes/statistics/prev/national/figpersons.htm



http://www.cdc.gov/obesity/data/adult.html

http://www.cdc.gov/obesity/data/adult.html

http://www.uspreventiveservicestaskforce.org/uspstf/uspsdiab.htm

http://www.uspreventiveservicestaskforce.org/uspstf/uspsdiab.htm

https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExp




Economic costs of diabetes in the u.s 2012

Documents

Transcript of Economic costs of diabetes in the u.s 2012