Cost of Illness of Chronic Angina

Background: Angina pectoris is one of the principalmanifestations of coronary artery disease (CAD).Chronic angina is a debilitating condition that affectsmillions of people in the United States.

Objective: The objective of the study is to estimate,from a societal perspective, the direct costs ofchronic angina in the year 2000.

Methods: Data on medical utilization related tochronic angina were extracted from National Centerfor Health Statistics public-use databases and fromIMS databases on medications (nitrates, beta-blockers,and calcium channel blockers). National averageMedicare reimbursement rates were used to estimatecosts. We identified medical utilization related tochronic angina based on International Classification ofDiseases, Ninth Revision (ICD-9) codes. When ICD-9codes that do not explicitly identify angina are used inmedical databases, people with chronic angina maybe coded as having CAD only. To address this, wedeveloped upper- and lower-boundary estimates ofthe costs of chronic angina. The lower-boundary estimateis based on diagnoses that narrowly define thepresence of chronic angina, and is termed "narrowlydefined chronic angina." The upper-boundary estimateis based on diagnoses of CAD.

Results: The lower boundary on the cost ofchronic angina is the estimated direct medical costof narrowly defined chronic angina ($1.9 billionwhen it is the first-listed diagnosis and $8.9 billionwhen it is listed in any position). The upper boundaryon the cost of chronic angina is the estimatedtotal direct medical cost of CAD, which is $33 billionwhen it is the first-listed diagnosis and $75 billionwhen it is listed in any position.

Conclusion: These analyses capture the range ofdirect costs that might be attributed to the care ofchronic angina in the United States for the year2000. Some components of care were not available,and estimated costs will be significantly higher if privatepayer reimbursement rates are used.

(Am J Manag Care. 2004;10:S358-S369)

Angina is a clinical syndrome, usuallyassociated with coronary artery disease(CAD). It is often termed "chestpain," but the classical description of anginais a sensation of pressure, heaviness, or adull band.^1,2 Anginal symptoms can be localizedin the chest or radiate to adjacent areas.Chronic angina results from a fixed coronaryartery lesion that restricts the supply ofoxygen to the myocardium during times ofincreased demand.³ Chronic angina episodestypically last several minutes and are oftenprecipitated by exertion or emotionalstress.⁴

There are no published studies of the costof illness of chronic angina and there is limitedliterature related to the cost of illness ofCAD. One paper that purports to reviewcosts of chronic angina only does so for limitedinterventions.⁵ Three previous studiesestimated the 5- and 10-year costs of CAD; 2developed models^6,7 and 1 estimated thecosts from a single medical center,⁸ but noneestimated annual cost. Two publishedpapers present comprehensive estimates ofthe total US annual expenditures for a numberof diseases, but the disease category forheart disease is broader than it is forCAD.^9,10

International Classification ofDiseases, Ninth Revision

The objective of the current study is toestimate, from a societal perspective, thedirect costs of chronic angina. Becausechronic angina is a manifestation of CAD,and because patients may not always presentwith chronic angina during medicaltreatment, patients may be coded with CADwhen (ICD-9) codes thatdo not also specifically indicate chronicangina are used. As a result, the costs ofchronic angina are underestimated whenthey are based solely on medical encountersidentified by using ICD-9 codes specific tochronic angina. The true cost of chronicangina will likely lie between the costs fornarrowly defined chronic angina when thisdisease is specifically identified as the principalcause for rendered services and theestimated cost for CAD (a broader diseasecategory that may be used to code forpatients with stable angina). We provide anestimate of the total cost of CAD when it isthe principal cause for the rendered services(ie, CAD is the first-listed diagnosis) and anestimate of the cost when it is either theprincipal or a contributing cause for renderedservices (ie, CAD is listed as a diagnosisin any position). Thus, we examined thecost of chronic angina based on a rangedefined by costs for narrowly defined chronicangina and costs for CAD.

To examine the range of costs for chronicangina, we estimated the lower and upperboundaries of the cost. The lower boundaryuses the ICD-9 codes that are associatedwith a narrow definition of chronic angina.The upper boundary uses a wider range ofICD-9 codes that are associated with CAD. Arange is used to capture all patients whohave narrowly defined chronic angina aswell as patients who are coded with CAD butmay have angina, too.

Datasets Used to Identify Utilization

A variety of public and private databaseswere examined to determine the range ofcost of angina both for narrowly definedchronic angina and CAD. Use of medicalresources was identified for the followingcomponents of care: hospitalizations, emergencydepartment visits, outpatient visits tophysicians⁹ offices, nursing home stays,home healthcare, hospice care (for CADonly), and prescription medications. Allcomponents of care except medicationswere matched with national public-usedatasets available from the National Centerfor Health Statistics of the Centers forDisease Control and Prevention. Theseincluded the National Hospital DischargeSurvey (NHDS), the National AmbulatoryMedical Care Survey (NAMCS), the emergencydepartment and outpatient componentsof the National Hospital AmbulatoryMedical Care Survey (NHAMCS), theNational Nursing Home Survey (NNHS), andthe home healthcare and hospice componentsof the National Home and HospiceCare Survey (NHHCS). All of these datasetsuse a nationally representative sample ofinstitutions and include patient weights,which were used to extrapolate the sampleto levels representative of the entire nation.Analyses were conducted using calendaryear 2000 data in all cases except the NNHS,which was last conducted in 1999.

Medications were analyzed using dataavailable from the IMS National SalesPerspectives^™ (NSP) database and the IMSNational Disease and Therapeutic Index^™(NDTI) for calendar years 2000 and 2001.The NSP database (formerly called the Retailand Provider Perspective database) was usedto obtain estimates of the total sales of classesof drugs in the United States. The NSPdatabase contains reports of the total numberof units (ie, pills, patches, etc) sold bynational drug code number and the totaldollars paid by outlets for those products,whether purchased directly from a manufactureror indirectly via a wholesaler. Althoughprompt-payment discounts and bottom-lineinvoice discounts exist, they are not reflectedin the dollar purchase amounts.However, invoice line item discounts arereflected. The NDTI was used to determinethe office visits during which drugs were prescribed.The NDTI data are obtained from a1% sample of physicians who record all prescriptionsissued on 2% of days annually.

For the purpose of the current analysis,which depended on identifying narrowlydefined chronic angina and CAD in utilizationdatasets, specific ICD-9 codes wereidentified for each condition. Utilization wasassociated with narrowly defined chronicangina if the ICD-9 code was 413.XX (ie, anyICD-9 code beginning with 413). Thesecodes include 413.0X (angina decubitus),413.1X (Prinzmetal angina), and 413.9X(other and unspecified angina pectoris).Utilization was also associated with narrowlydefined chronic angina when the diagnosesincluded 786.5 (chest pain) inconjunction with any diagnosis of CAD (asdefined below).

Utilization was associated with CAD if theICD-9 code included any of the following:413.XX (angina pectoris), 410.XX (acutemyocardial infarction), 411.0X (postmyocardialinfarction syndrome), 411.8X (otheracute and subacute forms of ischemic heartdisease), 412.XX (old myocardial infarction),414.XX (other forms of chronicischemic heart disease), 429.2X (cardiovasculardisease, unspecified), 429.5X (ruptureof chordae tendineae), 429.6X (rupture ofpapillary muscle), 429.7X (certain sequelaeof myocardial infarction, not elsewhereclassified), 996.03 (complications due tocoronary artery bypass graft), V45.81 (postsurgicalstatus after aortocoronary bypass),or V45.82 (postsurgical status after percutaneoustransluminal coronary angioplasty[PTCA]). Utilization was also associatedwith CAD when the diagnoses included786.5 (chest pain) in conjunction with anydiagnosis of CAD. The ICD-9 code for unstableangina (411.1) was excluded from ouranalysis of CAD because we aimed to estimatethe cost of chronic forms of angina andthe conditions that often arise in people withchronic angina, specifically differentiatedfrom unstable angina.

Any ICD-9 codes associated with narrowlydefined chronic angina as the first-listeddiagnosis and any codes listed in any positionavailable in the dataset were analyzed.In addition, first-listed CAD and any-listedCAD were analyzed separately.

Strategy Used to Value Utilization

Very few public-use datasets include dataon charges and none include actual costdata. One national database, the MedicalExpenditure Panel Survey, captures paymentsfor several categories of care, but hastoo few patients with narrowly definedchronic angina to provide reliable estimates.Because national cost data are unavailableand because charges are known to be variablyinflated,¹¹ Medicare reimbursementrates were used to value all components ofcare except medications. Medications werevalued at average wholesale revenues.Medicare reimbursement rates were used forvaluing utilization for other components ofmedical care for a number of reasons. First,there is no public-use database of costs orcharges for all components of healthcare.Second, Medicare rates are available to thepublic free of charge; private insurance companiesconsider their cost information proprietary.Third, Medicare collects considerabledata to estimate provider costs and paysproviders according to those estimates.Fourth, about 60% of the patients with anginaand CAD are older than age 65¹²; therefore,Medicare is the primary payer for theirhealthcare. Overall, Medicare reimbursementrates represent a reasonable proxy forthe long-run opportunity costs of medicalcare and are one accepted approach to valuinghealthcare in these types of cost-of-illnessanalyses.^13-15 Throughout this paper weuse the term cost to mean the above-mentionedvaluation approach.

Hospitalizations

Hospitalizations were estimated usingthe NHDS. Hospitalizations were identifiedwhen any of the ICD-9 codes associated withnarrowly defined chronic angina were listedin the first diagnosis position or in any of the7 diagnosis positions available in the dataset.This analysis identified 82 500 hospitalizationsfor which narrowly defined chronicangina was first listed and 543 000 hospitalizationsfor which narrowly defined chronicangina was listed in any position. The principalcontributors to hospitalizations withnarrowly defined chronic angina as the first-listeddiagnosis were 73 200 hospitalizationsfor diagnosis-related group (DRG) 140 (anginapectoris), 5863 hospitalizations for DRG125 (circulatory disorder excluding acutemyocardial infarction [AMI] with cardiaccatheterization and without a complex diagnosis),and 1137 hospitalizations for DRG116 (other permanent cardiac pacemakerimplant or PTCA with stent).

This analysis identified 2 027 400 hospitalizationsfor which CAD was the first-listeddiagnosis and 5 308 000 hospitalizations forwhich CAD was listed in any position. Theprincipal contributors to hospitalizationswith CAD as the first-listed diagnosis were423 075 hospitalizations for DRG 116 (otherpermanent cardiac pacemaker implant orPTCA with stent), 259 469 hospitalizationsfor DRG 132 (atherosclerosis with complicatingcondition), 229 746 hospitalizationsfor DRG 121 (circulatory disorder with AMIand cardiovascular complications dischargedalive), 194 527 hospitalizations forDRG 122 (circulatory disorder with AMIwithout cardiovascular complications dischargedalive), 144 718 hospitalizations forDRG 124 (circulatory disorder excludingAMI with cardiac catheterization and complex diagnosis), 131 220 hospitalizations forDRG 107 (coronary bypass with cardiaccatheterization), and 127 360 hospitalizationsfor DRG 109 (coronary bypass withoutcardiac catheterization). Other DRGs with58 000 to 96 000 hospitalizations were DRG125 (circulatory disorder excluding AMI,with cardiac catheterization without complexdiagnosis), DRG 133 (atherosclerosiswithout complicating condition), DRG 140(angina pectoris), DRG 112 (percutaneouscardiovascular procedures), and DRG 123(circulatory disorder with AMI, expired).All other DRGs had fewer than 19 000hospitalizations.

Hospitalizations obtained via the NHDSwere valued using the national averageMedicare Part A DRG reimbursements andestimates of the reimbursement ratesunder Part B for these hospitalizations. Weused national average Medicare Part A DRGreimbursement rates for fiscal year 2000 toestimate the facility component of hospitalizations.¹⁶Medicare reimburses physiciansfor inpatient stays under Part B, which isseparate from hospital reimbursementsunder Part A. The Centers for Medicare andMedicaid Services do not compile statisticson Part B reimbursements by DRG. Mostrecently, the Urban Institute studied Part Breimbursements associated with hospitalizationsby DRG by using the 1987 Medicaredatabase.¹⁷ Miller and Welch calculated thenumber of relative value units (RVUs) associatedwith inpatient care by DRG and valuedthem using reimbursements per RVU in1987. We updated those reimbursements byrevaluing the reimbursement rates for RVUsfor physician services to fiscal year 2000reimbursement rates. However, physicianpractices have changed since 1987; therefore,projections of Part B reimbursementsfrom 1987 to 2000 may be inaccurate.Estimated costs of hospitalizations for narrowlydefined chronic angina and CAD areshown in Table 1.

Emergency Department Visits

Data from the NHAMCS-EmergencyDepartment (ED) component were used toestimate the number of emergency departmentvisits, and the number and type of proceduresperformed during those visits.Three separate cost components were calculatedfor emergency department visits: (1) baseMedicare reimbursement for emergencydepartment visits, (2) Medicare reimbursementfor separately billable procedures, and(3) ambulance costs (at Medicare reimbursementsof $294 per visit requiring thisservice).^18,19 The base Medicare reimbursementsfor emergency department visits were$64.97, $102.31, $154.67, and $416.99 forambulatory payment classification (APC)group numbers 610, 611, 612, and 620,respectively.²⁰ APCs code for visits thatinvolve a problem-focused history or examination,or medical decision making that isstraightforward or low to moderately complex.Other APCs code for visits rangingfrom more comprehensive examinationsand highly complex medical decisions to visitsinvolving critical care evaluation andmanagement of critically ill or injuredpatients requiring constant physician attendance.All procedures that had ICD-9 procedurecodes were cross-referenced to thecorresponding Current Procedural Terminology(CPT) or Health Care FinancingAdministration (HCFA) Healthcare CommonProcedure Coding System (HCPCS)codes,²¹ and total Medicare reimbursements(ie, facility and physician components)were calculated.²⁰ These includedprocedures such as computed tomographyscans, magnetic resonance imaging, x rays,ultrasound, electrocardiograms, lumbarpunctures, endotracheal intubation, cardiopulmonaryresuscitation, nasogastrictube/gastric lavage, plus coded ambulatorysurgery procedures taking place in theemergency department. Minor procedures inthe NHAMCS-ED dataset that were nonspecific(ie, did not have ICD-9 procedurecodes, such as "blood tests"), were not separatelycosted because of the uncertainty ofmapping them to CPT/HCPCS codes andbecause many of these procedures areincluded in the APC reimbursement forthe visit.

Medicare does not separately reimbursefor emergency department visits if thepatient is admitted to the hospital. In suchcases, payment is considered part of theDRG-based reimbursement. In 2000,312 800 emergency department visits withnarrowly defined chronic angina were listedas the first diagnosis. Of these, 169 000(54%) patients were admitted and 143 800(46%) patients were not admitted to the hospital.There were 724 500 emergency departmentvisits with CAD listed as the firstdiagnosis. Of these, 515 700 (71.5%) patientswere admitted and 205 800 (28.5%) patientswere not admitted to the hospital. We havecalculated the cost of emergency departmentvisits separately, depending onwhether the patient was admitted to the hospital.The results from this analysis areshown in Table 2.

Office Visits

The NAMCS and the (NHAMCS-OfficeVisits (OV) were used to quantify the numberof office visits for narrowly definedchronic angina and CAD. NAMCS capturesoffice visits to physicians in private practiceor group practices, and NHAMCS capturesoffice visits to physicians in hospital-basedclinics. Estimates from both databases werecalculated and summed. The analyses indicatedthat there were 2.01 million office visitsfor first-listed narrowly defined chronicangina and 3.14 million office visits whennarrowly defined chronic angina was listedas any of the 3 possible diagnoses. Therewere 13.55 million office visits for first-listedCAD and 22.66 million office visits for any-listedCAD. For these visits, Medicare reimbursementswere calculated based onCPT/HCPCS codes, which consider the numberof minutes that the patient spent with thephysician and whether the patient was newor established.²² All separately reimbursableprocedures that had ICD-9 procedure codeswere cross-referenced to the correspondingCPT/HCPCS codes,²¹ and total Medicarereimbursements (ie, facility and physiciancomponents) were calculated.²⁰ We used thesame approach to valuing outpatient visitswhether they were hospital-based or not.The results are shown in Table 3.

Nursing Home Stays

The NNHS information, including approximately8200 current residents, was collectedduring a 1-year period and is weighted torepresent the number of nursing home residentson a typical day. The analysis indicatedthat there were 3918 person-years with aprimary diagnosis of narrowly definedchronic angina and 36 242 person-yearswith any diagnosis of narrowly definedchronic angina in 2000. Similarly, therewere 56 166 person-years with a primarydiagnosis of CAD and 283 848 person-yearswith any diagnosis of CAD.

In 1998, Medicare began paying for nursinghome care using resource utilizationgroups (RUGs) based on the patient'sexpected service needs. The NNHS did notinclude data to determine RUG, but did containinformation concerning the payer fornursing home services and the daily ratepaid for those services by level of service.The average Medicare reimbursement rateswere $177.44 for skilled care, $122.05 forintermediate care, $57.81 for residentialcare, and $199.91 for unspecified level ofcare. Medicare typically reimburses nursinghome stays for just the first 100 days afterhospitalization. Medicaid is a frequent payerof nursing home stays extending beyond 100days. An analysis of discharged patients withnarrowly defined chronic angina indicatedthat Medicare was the payer of 40% of stays,Medicaid was the payer of 35% of stays, andprivate sources paid for the remaining 25%.Analyses also indicated that the Medicarereimbursement rate was similar to the ratefor private sources. Thus, we have used theMedicare average reimbursement rates andapplied these to the number of nursing homeyears for patients with narrowly definedchronic angina or CAD. The results areshown in Table 4.

Home Healthcare

Data from the NHHCS are collected duringa 1-year period and weighted to the numberof home healthcare residents on atypical day. Medicare has recently introduceda prospective payment system forhome healthcare, but at the time of this surveyMedicare reimbursed on the basis of thetype of services received. In 2000, the averagereimbursement per visit and number ofvisits per discharge for these services was$153.59 per visit for 0.32 medical socialservice visits, $113.26 per visit for 0.18speech pathology service visits, $104.76 pervisit for 0.53 occupational therapy servicevisits, $104.05 per visit for 3.05 physicaltherapy service visits, $94.96 per visit for14.08 skilled nursing service visits, and$41.75 per visit for 13.4 home health aidservice visits.²³ Multiplying the cost per visitby the average visits per discharge and summingover the different services we obtainan average reimbursement of $2339 per dischargedhome healthcare patient. From thedischarge dataset, we computed the averagelength of stay—truncated at 60 days to correspondto Medicare reimbursement policies,which pay in 60-day blocks ofservice—at 29.08 days. Thus, the averagereimbursement per day for the services listedabove was $80.43. Applying this dailyrate to the number of days of home healthcareper year (not capped at 60 days), weobtain the approximate costs of homehealthcare in the year 2000.

The analysis indicated that there were1.44 million days of home healthcare per yearfor patients with a primary diagnosis of narrowlydefined chronic angina and 8.87 milliondays per year for any diagnosis ofnarrowly defined chronic angina. The analysisshowed 15.78 million days per year forfirst-listed CAD and 48.85 million days peryear for any-listed CAD. The analysis indicatedthat the cost of home healthcare was $116million for patients with first-listed narrowlydefined chronic angina and $714 million forpatients with any-listed narrowly definedchronic angina using approximate dailyMedicare reimbursement rates. The costs ofCAD were $1 269 million for first-listed CADand $3 929 million for any-listed CAD.

Hospice Care

Data from the NHHCS are collected duringa 1-year period and weighted to the numberof hospice care residents on a typicalday. Costs for hospice care for narrowlydefined chronic angina and CAD were calculatedusing daily Medicare reimbursementrates of $101.84 for routine in-home care,$594.41 for continuous in-home care,$105.35 for inpatient respite care, and$453.04 for inpatient general care.²⁴ Almostall patients (92.5%) were receiving routinein-home care on any given day, with anadditional 7.4% receiving inpatient respite orgeneral care, and only 0.1% receiving continuousin-home care.

The NHHCS database contained only 7patients who had any diagnosis of narrowlydefined chronic angina, projecting to 52 600days of hospice care costing $10.8 million.The small number of patients results in thisestimate having a standard error larger thangenerally accepted. (National Center forHealth Statistics [NCHS] guidelines for thisdatabase recommend that estimates bebased on 60 or more observations, or on asfew as 30 observations if the standard errorof estimate is 30% or less.) However, evenwith a large standard error this estimate issufficiently accurate to be able to concludethat, with 95% confidence, the total cost ofhospice care for any diagnosis of narrowlydefined chronic angina is less than $55 million,which is still a negligible proportion ofall costs for treatment of this condition (ie, <1% of total costs).

The NHHCS database also contains 28patients with a primary diagnosis of CAD and128 patients with any diagnosis of CAD, projectingto 217 800 days of hospice care for aprimary diagnosis of CAD (costing $24.4 million)and 1.9 million days per year for any-listedCAD (costing $215.3 million). Theestimates for first-listed CAD do not meetthese criteria, but total costs can be shown tobe less than $45 million with 95% confidence.The estimates for any-listed CAD meet NCHSguidelines for estimation. Again, these costsare a negligible proportion of total costs.

Prescription Medications

Recently updated American College ofCardiology/American Heart Association clinicalpractice guidelines4 for the treatment ofchronic angina recommend 3 classes of prescriptionmedication—beta-blockers, calciumchannel blockers, and nitrates.²⁵

Thus, we estimated medication usage andcosts for beta-blockers, calcium channelblockers, and nitrates by analyzing IMSdatabases. The nitrate medications includedboth short- and long-acting forms. The IMSNSP database was used to obtain estimatesof the total sales of these 3 classes of drugsin the United States. We calculated theannual average wholesale revenue and salesvolume by drug class. Because of discountsgiven to some purchasers, average wholesalerevenue may be less than average wholesaleprice. Data from the IMS NDTI were thenused to estimate the total number of officevisits during which beta-blockers, calciumchannel blockers, or nitrates were prescribed(regardless of diagnosis).

The results of this analysis showed thatbeta-blockers were prescribed in 62 235 000visits, calcium channel blockers were prescribedin 61 541 000 visits, and nitrateswere prescribed in 18 081 000 visits. Thesevalues were divided into the revenue forthese 3 classes of medications to obtainwholesale revenue per prescription over alltypes of purchasers ($18.55 for beta-blockers,$74.37 for calcium channel blockers,and $27.41 for nitrates). This cost estimatealso includes the cost of all refills.

The IMS NDTI dataset was also used toallocate a share of the sales reported in theNSP to narrowly defined chronic angina orCAD. The database contained 4-digit ICD-9codes for a primary diagnosis and a secondarydiagnosis. Allocation for a specific drugclass was accomplished by calculating theratio of the number of physician visits duringwhich a drug in that class was prescribedand narrowly defined chronic angina (orCAD) was indicated as a diagnosis to thetotal number of physician visits duringwhich a drug in that class was prescribed.Total sales volume and revenue was thenmultiplied by that ratio to estimate the proportionattributable to narrowly definedchronic angina or CAD. Table 5 shows theallocation of prescriptions for beta-blockers,calcium channel blockers, and nitrates usingthe diagnosis codes in the NDTI.

Total Direct Costs

Estimated total direct costs of narrowlydefined chronic angina and CAD were calculatedby summing the cost for each componentof care. Table 6 shows the total directcost of narrowly defined chronic anginawhen it is the first-listed diagnosis or listedas a diagnosis in any position, and of CADwhen it is the first-listed diagnosis or listedin any position. Table 6 also includes thetotal estimated cost of emergency departmentcare. Although Medicare does notreimburse separately for emergency departmentcare (other than ambulance costs) forpatients who are admitted to the hospital,we have included those costs in Tables 6and 7 to permit comparisons with otherstudies when data from other payers areused (a breakdown of emergency departmentcosts is shown in Table 2).

As can be seen in Table 6, the contributionof each component of care varied acrossthe 4 columns. As shown in Figure 1, forfirst-listed narrowly defined chronic angina,the greatest contributor to cost was outpatientvisits (38%), followed by hospitalizations(16%), prescription medications (15%),emergency department visits (12%), andnursing home stays (12%). Hospitalizationwas the largest contributor to cost for any-listednarrowly defined chronic angina(53%), followed by nursing home stays (22%).

The average cost per component of utilizationis shown in Table 7. Each componentof utilization was calculated by dividingthe total US direct cost by the number ofpeople in the United States who had thattype of utilization. For example, 82 000 peoplewere hospitalized with first-listed narrowlydefined chronic angina in the UnitedStates in the year 2000 and their averagecost per hospitalization was $3744.

The range of estimated costs for chronicangina was calculated in this paper. Thetotal direct medical cost of narrowly definedchronic angina was $1.9 billion when it waspositioned as the first-listed diagnosis and$8.9 billion when it was listed in any position.The total direct medical cost of CADwas $33.0 billion when it was the first-listeddiagnosis and $74.8 billion when it was listedin any position. Thus, a lower boundaryon the cost of chronic angina (based on adiagnosis of narrowly defined chronic angina)is between $1.9 billion and $8.9 billion,and an upper boundary of the total medicalcosts for people with chronic angina (basedon a diagnosis of CAD) is between $33.0 billionand $74.8 billion.

Two previous studies have calculatedthe annual direct cost of heart diseaseusing methodologies somewhat similar tothe current analysis for first-listed diagnoses.Neither study examined narrowlydefined chronic angina or CAD specifically.Both estimated costs for heart disease(broadly defined) and one also estimatedcosts for CAD (using ICD-9 codes 410-414). Hodgson and Cohen¹⁰ estimated thetotal personal healthcare expenditures forCAD to be $38.7 billion and $75.9 billionin 1995 dollars ($53.8 billion and $105.4billion in Consumer Price Index [CPI]-adjusted 2000 dollars). Cohen andKrauss⁹ estimated the total expendituresfor heart disease to be $57.5 billion in1997 dollars ($71.1 billion in CPI-adjusted2000 dollars). The estimate from ouranalysis of $33.0 billion for first-listedCAD and $74.8 billion for any-listed CADis similar to the estimate by Cohen andKrauss of $57.5 billion.

In the current analysis, for first-listed narrowlydefined chronic angina, the greatestcontributor to cost was outpatient visits,followed by hospitalizations, prescriptionmedications, emergency department visits,and nursing home stays. In contrast, for any-listednarrowly defined chronic angina, hospitalizationwas the largest contributor tocost followed by nursing home stays. ForCAD, hospitalization was by far the largestcontributor to cost for first-listed CAD. Thepercentages for components of care in thecurrent analysis are similar to percentagesreported in a previous study that examinedexpenditures for heart disease.⁹ The expendituresfor hospitalization in the currentstudy accounted for 74% of total costs and ina previous study the expenditures for hospitalizationwere responsible for 66% of thetotal costs.⁹

The finding that hospitalization was amuch larger contributor to cost for first-listedCAD (Figure 2) than for first-listednarrowly defined chronic angina is likelydue to the pattern of expensive cardiovascularprocedures and surgeries (eg, PTCAwith stent or coronary artery bypass graft)that are more frequently performed forCAD, plus hospitalizations for AMI that arecaptured under CAD. Hospitalization fornarrowly defined chronic angina, in contrast,was largely for medical managementor diagnostic evaluation rather than for surgicalinterventions.

The current analysis has several limitationsthat should be considered when interpretingand generalizing the findings. Theprimary limitation is the absence of data onseveral components of care for chronic anginathat results in an underestimation ofcosts. The components of care that were notincluded in this study were: (1) other prescriptiondrugs; (2) nonprescription medications;(3) preventive care; (4) cardiac rehabilitation;and (5) durable medical equipment.

A second potential limitation arises fromthe challenges of identifying people withangina by using ICD-9 codes to extract casesfrom databases. Because angina is a manifestationof CAD, there may be instanceswhen the CAD but not the angina is coded.The number of cases missed by using thenarrow definition of chronic angina isimpossible to estimate, but probably resultedin undervaluing the estimates of costs.The use of CAD to provide an upper boundaryon the costs of chronic angina probablyoverstates costs because not all patients withCAD will experience angina. In an attemptto control for this study limitation, we presentedthe lower (narrowly defined chronicangina) and upper boundaries (CAD) ofchronic angina and posit that the actualdirect cost for chronic angina lies withinthis range.

A third potential limitation to the currentanalysis is the reliance on Medicare reimbursementrates to value almost all componentsof care. To the degree that Medicarereimbursement rates underestimate whatother payers reimburse for comparable services,the current estimates will be significantlylower than estimates derived usingprivate payer valuations.

In conclusion, this study captured therange of direct costs attributed to the care ofchronic angina in the United States for theyear 2000, with a lower boundary of $1.9 billionfor first-listed narrowly defined chronicangina and an upper boundary of between$33.0 billion for first-listed CAD to $74.8 billionfor any-listed CAD. It is proposed thatthe actual direct cost of chronic angina lieswithin this range.

Acknowledgments

The data analysis and writing of this manuscriptwas supported by CV Therapeutics,Inc. The authors thank Roman Alper of IMSfor assistance with analysis of medicationuse data.

Nurse Pract.

1. Beattie S. Management of chronic stable angina. 1999;24:49-61.

B Med Bull.

2. Davies SW. Clinical presentation and diagnosis ofcoronary artery disease: stable angina. 2001;59:17-27.

PDNT.

3. Hatchett R. Coronary heart disease: 2. The assessment,diagnosis and management of stable angina. 2001;97:39-42.

4. Gibbons RJ, Abrams J, Chatterjee K, et al. ACC/AHA2002 guideline update for the management of patientswith chronic stable angina: a report of the AmericanCollege of Cardiology/American Heart Association TaskForce on Practice Guidelines (Committee to Update the1999 Guidelines for the Management of Patients withChronic Stable Angina). 2002. Available at:www.acc.org/clinical/guidelines/stable/stable.pdf.Accessed May 13, 2004.

Health Technol Assess.

5. Sculpher MJ, Petticrew M, Kelland JL, et al. Resourceallocation for chronic stable angina: a systematicreview of effectiveness, costs and cost-effectiveness ofalternative interventions. 1998;2:1-178.

Am J Cardiol.

6. Russell MW, Huse DM, Drowns S, Hamel EC, Hartz SC.Direct medical costs of coronary artery disease in theUnited States. 1998;81:1110-1115.

Am JCardiol.

7. Wittels EH, Hay JW, Gotto AM. Medical costs ofcoronary artery disease in the United States. 1990;65:432-440.

Med Care.

8. Eisenstein EL, Shaw LK, Anstrom KJ, et al. Assessingthe clinical and economic burden of coronary arterydisease: 1986-1998. 2001;39:824-835.

Health Aff.

9. Cohen JW, Krauss NA. Spending and service useamong people with the fifteen most costly medicalconditions, 1997. 2003;22:129-138.

Med Care.

10. Hodgson TA, Cohen AJ. Medical care expendituresfor selected circulatory diseases. Opportunities forreducing national health expenditures. 1999;37:994-1012.

Ann Intern Med.

11. Finkler SA. The distinction between costs andcharges. 1982;96:102-109.

Am J Cardiol.

12. Pepine CJ, Abrams J, Marks RG, et al.Characteristics of a contemporary population withangina pectoris. 1994;74:226-231.

Theory and Methods of EconomicEvaluation of Health Care.

13. Johannesson M. Dordrecht, Netherlands:Kluwer Academic Publishers; 1996.

Principlesof Pharmacoeconomics.

14. Larson LN. Cost determination and analysis. In:Bootman JL, Townsend RJ, McGhan WF, eds. Cincinnati, Ohio: HarveyWhitney Books Co; 1991:38-49.

Meta-analysis, Decision Analysis, andCost-effectiveness Analysis.

15. Petitti DB. New York, NY: OxfordUniversity Press; 1994:175-177.

16. Centers for Medicare and Medicaid Services. 100%MEDPAR Inpatient Hospital Fiscal Year 2000. Obtainedfrom http://www.cms.hhs.gov/statistics/medpar/shtstay00drg.pdf. Accessed January 8, 2003.

17. Miller ME, Welch WP. Analysis of Hospital MedicalStaff Volume Performance Standards: Technical Report.Washington, DC: The Urban Institute; 1993.

Program Memorandum Intermediaries/Carriers.

18. HCFA. Transmittal AB-00-88. September 18, 2000.

Program Memorandum Intermediaries/Carriers.

19. HCFA. Transmittal AB-00-131. December 22, 2000.

Federal Register.

20. Office of the Inspector General; Medicare ProgramProspective Payment System for Hospital OutpatientServices; Final Rule. April 7, 2000(65)68.

21. Procedural Cross Coder. Salt Lake City, Utah:Ingenix, Inc; 2002.

CPT 2000.

22. American Medical Association. Chicago,Ill: American Medical Association; 2000.

Federal Register.

23. Medicare Program; Prospective Payment System forHome Health Agencies; Final Rule. July 3, 2000(65)128.

Program Memorandum Intermediaries.