Direct medical costs of complications of diabetes in the United States: estimates for event-year and annual state costs (USD 2012)

Abstract

Objective:

To estimate the direct medical costs associated with managing complications, hypoglycemia episodes, and infections associated with type 2 diabetes expressed in 2012 United States dollars (USD).

Methods:

Direct data analysis and microcosting were used to estimate the costs for an event leading to either a hospital admission or outpatient care, and the post-acute care associated with managing macrovascular and microvascular complications, hypoglycemia episodes, and infections. Data were obtained from many sources, including inpatient and emergency department databases, national physician and laboratory fee schedules, government reports, and literature. Event-year costs reflect the resource use during an acute care episode (initial management in an inpatient or outpatient setting) and any subsequent care provided in the first year. State costs reflect annual resource use required beyond the first year for the ongoing management of complications and other conditions. Costs were assessed from the perspective of a comprehensive US healthcare payer and expressed in 2012 USD.

Results:

Event-year costs (and state costs) for macrovascular complications were as follows: myocardial infarction $56,445 ($1904); ischemic stroke $42,119 ($15,541); congestive heart failure $23,758 ($1904); ischemic heart disease $21,406 ($1904); and transient ischemic attack $7388 ($179). For two microvascular complications the event-year and state costs were assumed the same: $71,714 for end stage renal disease, and $2862 blindness. The event-year cost was $9041 for lower extremity amputations, and $2147 for diabetic foot ulcers. Costs were also determined for managing hypoglycemic episodes: $176–$16,478 (depending on treatment required), and infections: vulvovaginal candidiasis $111, lower urinary tract infection $105.

Conclusions:

This study, which provides up-to-date cost estimates per patient, found that managing macrovascular and microvascular complications results in substantial costs to the healthcare system. This study facilitates conduct of other research studies such as modeling the management of diabetes and estimating the economic burden associated with complications.

Keywords::

• Diabetes
• Cost
• Economic

Introduction

Approximately 22 million Americans are diagnosed with diabetes, and the prevalence is projected to increase in the coming decades1,2. The economic implications are substantial, as the average annual medical costs of caring for individuals with diabetes are ∼2.3-fold higher than other patients1. The American Diabetes Association (ADA) estimated the total cost of diagnosed diabetes to be $245 billion in 2012 ($176 billion in direct medical costs and $69 billion in reduced productivity)1. Inpatient care accounted for 43% of the medical expenditures attributed to diabetes1, and about half of the direct medical costs are believed to be associated with management of the complications attributed to diabetes1.

Recent publications have not included a comprehensive set of cost estimates for complications that are readily translated into patient-level cost inputs for an economic model. O’Brien et al.3,4 last addressed this by publishing estimates of the annual cost per patient for 15 complications and published updated estimates in 2000 United States dollars (USD). These two comprehensive studies derived estimates by combining data from inpatient and emergency department (ED) care databases, national fee schedules, government reports, as well as the published literature. While database studies have since been published in this field, no single US database can be used to generate all these costs estimates.

The current study follows the same approach as O’Brien et al.3,4 by conducting analyses of databases and combining data from many sources to obtain current estimates of the direct medical costs per patient of managing macrovascular and microvascular complications, hypoglycemia episodes, and infections associated with some treatments, and expresses these cost estimates in 2012 USD. The objective was to derive the event-year costs and, when appropriate, the subsequent ongoing care costs associated with continued management of the complications and other conditions.

Methods

A summary of the sources used to derive the event and state estimates for macrovascular and microvascular complications, hypoglycemia episodes, and infections is provided in Table 1. Data were obtained from inpatient and ED database analyses, as well as from national physician and laboratory fee schedules, government reports, and published literature. The event-year cost per patient refers to costs associated with management of an acute care episode (initial management in an inpatient, emergency department, or outpatient setting) and any subsequent care provided within the first year following the acute episode. The state cost per patient reflects annual resource use required beyond the first year for the ongoing management of the given health state and applies while that particular health state is present for the remainder of the patient’s life. The costs were assessed from the perspective of a comprehensive US healthcare payer.

Table 1. Summary of methods used to develop event-year and state costs.

Complication	Event year			State
	Database	Microcosting	Publication	Publication	Microcosting
Cardiovascular
Acute Myocardial Infarction	•		•	•
Congestive Heart Failure	•		•	•
Ischemic Heart Disease	•		•	•
Cerebrovascular
Ischemic Stroke	•	•	•	•	•
Transient Ischemic Attack	•*	•	•	•	•
Nephropathy
End-stage Renal Disease			•	•
Gross Proteinuria		•
Microalbuminuria		•
Retinopathy
Blindness			•	•
Macular Edema		•			•
Proliferative Retinopathy		•			•
Neuropathy
Lower Extremity Amputation			•
Diabetic Foot Ulcer			•
Symptomatic Neuropathy		•	•
Peripheral Vascular Disease		•	•	•	•
Hypoglycemia
Episode Requiring Hospitalization/ED Visit	•*	•
Episode Requiring Glucagon Injection		•
Infections
Vulvovaginal Candidiasis		•
Lower Urinary Tract Infection		•

*Acute care estimates derived from ED and inpatient database, other complications used inpatient database only; Publication = government publications or published literature.

ED, emergency department.

Estimates were derived by following these three steps:

• Step 1. Identification of relevant healthcare services;

• Step 2. Quantification of the frequency of use and proportion of users for each healthcare service in the profile; and

• Step 3. Valuation of the resources used.

A specific resource use profile was developed for each complication and condition, and then the costs for each component were determined by direct data analysis, or microcosting (Table 1). The acute care cost estimate was derived by applying the mean cost per inpatient stay or ED visit (including physician fees) or outpatient care to the proportion of cases managed in each setting. For some profiles initial inpatient care was assumed for all cases (myocardial infarction, congestive heart failure, ischemic heart disease, and ischemic stroke), whereas for others either outpatient management only (e.g., for proteinuria) or a combination of the two (e.g., for transient ischemia attack or hypoglycemia) were considered. The acute care cost of a transient ischemic attack was a weighted average of hospitalization and management in the ED5, and the event-year cost also includes the post-acute follow-up care costs (Table 2 provides details). For some short-term conditions (i.e., proteinuria, infections, hypoglycemia) and complications (diabetic foot ulcers, lower extremity amputations), only event costs were developed. Costs were derived for hypoglycemic events that led to a glucagon injection and physician visit, an ED visit, or a hospitalization. The event costs derived for episodes leading to an emergency department visit or an inpatient stay were also assumed to lead to follow-up care from a primary care physician (details are provided in Table 2). Literature searches were conducted to confirm the profiles as well as the annual management and post-acute care cost estimates.

Table 2. Microcosting: resource use profiles and cost estimate derived per patient (reported in 2012 USD).

Complication	Physician visit CPT	Test/procedure HCPCS	Drugs	Total cost
Transient ischemic attack*	99203, 99213	NA	Aspirin	$179
Gross proteinuria	99214	81000	NA	$109
Microalbuminuria	99213	82043	NA	$79
Proliferative retinopathy	92014, 99213	92250, 67228	NA	$614
Macular edema	92012, 99213	92235, 67210, 67028	Triamcinolone acetonide	$798
Symptomatic neuropathy	99214	95861	Pregabalin	$878
Peripheral vascular disease	NA	NA	Aspirin, atorvastatin	$126
Hypoglycemia, requiring hospitalization/ED visit*	99203, 99213	82947	NA	$181
Hypoglycemia, requiring glucagon injection	99213	82947	Glucagon	$176
Vulvovaginal candidiasis	99214	NA	Clotrimazole	$111
Lower urinary tract infection	99214	NA	Co-trimoxazole	$105

*For transient ischemic attack and hypoglycemia episodes requiring an ED visit or hospitalization, these profiles were combined with database estimates to derive the event-year costs.

CPT, Current Procedural Terminology; ED, emergency department; HCPCS, Healthcare Common Procedure Coding System; NA, not applicable.

All event and state costs are expressed in 2012 US dollars. Where 2012 values were not available, older estimates were inflated using the medical care component of the US Consumer Price Index reported by the Bureau of Labor Statistics6.

Data sources

The key sources for developing cost estimates were as follows:

• Inpatient stay: Hospital discharge data (2010) from 1147 hospitals in six states (California, Florida, Massachusetts, Maryland, New Jersey, and Washington)7–12. These databases contain merged discharge-level demographic, clinical, and economic data for all hospital discharges within a given year for patients of all ages and covered by all payers (e.g., Medicare or private).

• ED visit: Emergency department data (2010) from 67 emergency departments in Massachusetts acute care hospitals and satellite emergency facilities13. Data elements include patient demographics, clinical characteristics, services provided, and charges.

• Outpatient visits, tests, and procedures: Medicare national fee schedules in Centers for Medicare and Medicaid Services 2012 Payment Files (Physician Fee Schedule, Hospital Outpatient Prospective Payment System, Clinical Laboratory Fee Schedule)14–16.

• Outpatient medications: Wholesale Acquisition Cost (WAC)17. In cases where more than one option was available, the lower cost treatment was included in the estimate.

Inpatient stay and ED visit

Analyses conducted on the inpatient hospital and ED databases used the following inclusion criteria to identify cases:

• Adults (≥18 years old);

• Secondary discharge diagnosis for diabetes (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] diagnosis codes 250.x0, 250.x2); and

• Principal discharge diagnosis for the selected complication (ICD-9 CM codes for each complication are provided in Table 3).

Table 3. Diabetes complications: inpatient acute care (mean cost per stay) and emergency department (cost per visit) (reported in 2012 USD).

Complication	Principal diagnosis ICD-9-CM* codes	Cases	Cost per stay† Mean	Stay, days Mean
Inpatient
Acute Myocardial Infarction	410.x0 410.x1	41,523	$31,482	5.1
Congestive Heart Failure	428.x	92,917	$16,960	4.9
Ischemic Heart Disease	411.1 411.81 413.x	3573	$11,210	2.2
Ischemic Stroke	433.x1 434.x1	34,190	$18,835	5.2
Transient Ischemic Attack	435.x	14,275	$11,321	2.6
Hypoglycemia Episode	249.8 250.80 250.82	9723	$16,297	5.6
Emergency Department			Cost per visit‡ Mean	Visit, hours Mean
Mean
Transient Ischemic Attack	435.x	106	$2574	5.0
Hypoglycemia Episode	249.8 250.80 250.82 251.2	232	$1129	3.5

*ICD-9-CM codes (International Classification of Diseases, Ninth Revision, Clinical Modification) refer to principal diagnoses; all patients also have a secondary diagnosis code for diabetes (ICD-9-CM 250.x0, 250.x2).

†Includes physician fees ($2067 per stay).

‡Includes physician fees ($279 per visit).

Costs per inpatient stay and per ED visit were derived by applying cost-to-charge ratios (CCRs), the addition of an average physician payment, and the inflation of the cost components to 2012 USD. Average hospital charges per stay for each state were adjusted to costs using CCRs obtained from the Healthcare Cost and Utilization Project (HCUP) CCR files (California =0.288, Florida = 0.287, Massachusetts = 0.544, Maryland = 0.767, New Jersey = 0.231, Washington =0.624)18. The physician fees are not included in the charge variable of the inpatient or ED databases, so average Medicare physician payments were assumed to represent the average costs per stay or per visit and were added as a cost component in the mean cost per stay or emergency department visit19.

Outpatient and post-acute care

Outpatient care costs were based on the profiles provided in Table 2; these were developed after review of recent publications on management, guidelines, and clinical advice5,20,21. The unit costs for the pertinent physician visit, diagnostic and therapeutic procedures, and laboratory tests were each extracted from the relevant Medicare fee schedules for the specified Current Procedural Terminology (CPT) or Healthcare Common Procedure Coding System (HCPCS) codes, or WAC for medications14–16. Only hypoglycemic episodes where patients used formal healthcare services were considered, and these were defined in this study as an episode leading to a physician visit (and receiving a glucagon injection), an ED visit, or an inpatient stay.

The post-acute care cost for stroke includes a weighted average cost derived using the hospital discharge disposition of survivors (i.e., discharged to home with healthcare = 16%, inpatient rehabilitation unit = 13%, skilled nursing facility = 17%, intermediate care facility = 2%, or long-term care = 11%). Patients admitted from a similar facility were not included in the estimates developed. The appropriate Medicare program payment was applied for home healthcare, skilled nursing facility, and long-term care, and the pertinent physician fees for each care setting were extracted from the Medicare and Medicaid statistical supplement19. As the intermediate care facility payment varies by state, we used the average per diem covered by Medicaid in California as our estimate22.

Literature searches were conducted to support derivation of annual management and post-acute care cost estimates for selected complications (see Table 1). Published estimates were identified for the incremental direct medical costs in the first year following the acute event for macrovascular complications (myocardial infarction, congestive heart failure, ischemic heart disease, and ischemic stroke)23,24 that included re-admissions. Annual state costs for cardiovascular complications were obtained from a Medical Expenditure Panel Survey report on heart condition-associated office visits and medications25. The annual state cost for a stroke was estimated from published studies and a weighted average cost for intermediate care facility and long-term care.

Event-year annual management costs associated with lower extremity amputations and diabetic foot ulcers, plus state costs related to blindness and end stage renal disease (ESRD) annual management were each derived from estimates presented in published reports26–28. The ESRD cost estimate was a weighted annual cost of the Medicare expenditure per person treated with hemodialysis, peritoneal dialysis, and transplant27. The excess medical costs related to blindness were extracted from a report by Prevent Blindness America, an organization dedicated to fighting blindness and saving sight28. The Agency for Healthcare Research and Quality (AHRQ) released a report on a study of the economic burden of diabetic foot ulcers and lower extremity amputations among Medicare beneficiaries26.

Results

The event and annual state costs developed for each complication and other condition are provided in Table 4. The acute care costs estimated by conducting specific analyses of the inpatient and ED databases are presented in Table 3, and the cost estimates derived by micro-costing in Table 2.

Table 4. Diabetes complications: event-year and state cost estimates per patient (reported in 2012 USD).

Complication	Event-year cost*	State, annualized cost†
Cardiovascular
Acute Myocardial Infarction	$56,445	$1904
Congestive Heart Failure	$23,758	$1904
Ischemic Heart Disease	$21,406	$1904
Cerebrovascular
Ischemic Stroke	$42,119	$15,540
Transient Ischemic Attack	$7388	$179
Nephropathy
End-stage Renal Disease	$71,714	$71,714
Gross Proteinuria	$109	NA
Microalbuminuria	$79	NA
Retinopathy
Blindness	$2862	$2862
Macular Edema	$798	$70
Proliferative Retinopathy	$614	$70
Neuropathy
Lower Extremity Amputation	$9041	NA
Diabetic Foot Ulcer	$2147	NA
Symptomatic Neuropathy	$878	$1101
Peripheral Vascular Disease	$126	$126
Hypoglycemia
Episode Requiring Hospitalization	$16,478	NA
Episode Requiring ED visit	$1311	NA
Episode Requiring Glucagon Injection	$176	NA
Infections
Vulvovaginal Candidiasis	$111	NA
Lower Urinary Tract Infection	$105	NA

*Event costs are associated with management of the acute episode and any subsequent care in the first year.

†State costs reflect resource use beyond the first year for the ongoing management of the complication.

ED, emergency department; NA, not applicable.

Macrovascular complications

Acute myocardial infarction, congestive heart failure, ischemic heart disease, and ischemic stroke event-year costs were derived by assuming that all cases initially receive inpatient care (acute care costs are provided in Table 3), which accounted for the majority of the costs accrued in the year of the event. The post-acute care components included inpatient care, outpatient visits, and medication costs attributed to the specified condition accrued in the year of the initial hospital stay23,24. The state cost included office visits and medications associated with the specified conditions25.

The event-year post-acute care costs for stroke combined the inpatient costs of $18,835 (Table 3), the post-acute medical care in the event year derived from published studies ($9024), and sub-acute care cost component ($14,260), and this yielded the event-year cost of $42,119. The state cost was estimated as $15,541 and includes medical care derived from published studies ($9024) and a weighted average cost for intermediate care facility and long-term care ($6517).

The acute care cost of a transient ischemic attack was $7209, a weighted average of hospitalization (53% of cases) and management in the ED (Table 3)5. The transient ischemic attack event-year cost ($7388) also includes the post-acute follow-up care costs ($179), listed in Table 2.

Microvascular complications

Estimates for ESRD, blindness, lower extremity amputations, and diabetic foot ulcers were derived (Table 4). The ESRD cost estimate was a weighted annual cost of the Medicare expenditure per person treated with hemodialysis ($82,295), peritoneal dialysis ($61,588), and transplant ($29,983)27 in 2009 USD and then inflated to 2012 USD (Table 4). The excess medical costs related to blindness in 2004 USD ($2157) were extracted28, and inflated to 2012 USD (Table 4). The annual reimbursement per Medicare beneficiary for the services related to managing these two complications in 2008 was extracted (diabetic foot ulcers $1900 and lower extremity amputations $8000), and inflated to 2012 USD (Table 4)26.

Hypoglycemic episodes

Costs were derived for events that led to a glucagon injection and physician visit, an ED visit, or a hospitalization (Table 4). The event costs derived for episodes leading to an emergency department visit ($1129) or an inpatient stay ($16,297) (Table 3) were also assumed to lead to follow-up care from a primary care physician ($181) (details of this cost component are provided in Table 2). Hypoglycemic episodes where patients did not use formal healthcare services were not considered.

Discussion

This study developed estimates of the annual cost per patient for a number of diabetes complications in 2012 USD, providing an update for the cost estimates for complications developed in 2000 USD more than a decade ago by O’Brien et al.4 and including some additional common comorbidities (congestive heart failure, peripheral vascular disease) and treatment-related infections (lower urinary tract infection and vulvovaginal candidiasis). It might be expected that medical inflation would lead to the cost increase of managing these complications and other conditions over the past decade, which was the case for most except for amputations, diabetic foot ulcers, and blindness. In part this is because some fees (e.g., Medicare outpatient fee schedules) have not kept pace with the increases expected from applying medical inflation rates29.

For some complications, there have been clinical management changes that have affected the costs. For example, the amputation and diabetic foot ulcer costs were derived from a recent database analysis reflecting current management practices that allow some patients undergoing certain minor procedures to avoid hospital stays26. However, when patients were hospitalized, the cost estimate per stay remained high for these complications (lower extremity amputation $32,611 and diabetic foot ulcer $15,713). Current management of retinopathy has changed with the availability of intra-vitreal injection of anti-Vascular Endothelial Growth Factor, which may now be used alone or in combination with a retinal laser to enhance the retinopathy treatment effect and lower the risk of blindness. The direct medical care estimates for blindness are low compared with other serious complications included in this study, as the large indirect costs on patients and their families were not captured.

After taking into consideration the likely impact of medical inflation on the costs published by O’Brien et al.4 in 2000 USD, the changes in current management patterns do appear to have increased costs by at least 15% for cardiovascular disease, transient ischemia attacks, and ESRD. As cardiovascular disease is one of the most prevalent comorbidities among patients with diabetes and has been shown in many studies to impact healthcare utilization, this trend will be increasing the annual expenditure by payers, and as so many patients are elderly, this especially affects Medicare30–32. However, the ICD-9 CM codes used to identify ischemic heart disease cases were changed, and this may explain some cost differences observed. The new analysis includes intermediate coronary syndrome and acute coronary occlusion without myocardial infarction (ICD-9 CM codes 411.1 and 411.81) to align with the United Kingdom Prospective Diabetes Study Outcomes Model (UKPDS 68) codes for ischemic heart disease33. This modification was made because the risk functions published by the UKPDS 68 are often used to predict long-term complication rates in models. A high proportion of medical costs for diabetes are attributable to nephropathy progressing to ESRD, which is managed with dialysis or kidney transplantation30. The increase in ESRD costs observed here is consistent with the trend over the past 3 years evident from the data published for total Medicare expenditure for hemodialysis, peritoneal dialysis, and transplants, which appears to be increasing by ∼7% each year34. One study estimated that almost half (46%) of Medicare beneficiaries with type 2 diabetes had multiple co-morbid conditions, and the high costs associated with managing many of these complications concurrently will be contributing to the overall management costs1,30–32.

The estimates presented here provide a comprehensive picture of the direct medical costs attributable to complications, hypoglycemia, and infections associated with diabetes and its management. However, the estimates presented are reflective of healthcare costs, and do not capture, for example, lost productivity, or care provided by an unpaid caregiver, and so are an under-estimate of the true economic implications1. These estimates were derived by using a combination of direct data analysis (e.g., hospital care, emergency visits) and micro-costing based on typical management profiles (e.g., post-acute care), and rely on multiple sources (national physician and laboratory fee schedules, government reports, and published literature). This approach was adopted as no single database study can generate all these cost estimates; however, as a result, there are several limitations worth mentioning. Cost estimates available in other published studies cannot be readily compared with these estimates; for example, some of the other studies include annual cost estimates for both newly diagnosed and established cases, report annual costs, do not separate out complication-specific costs, or consider a specific payer perspective (e.g., managed care)1,30,32. In addition, when the data available reflected the costs from previous years, these costs were inflated and reported in 2012 USD. For example, whilst the inpatient hospital databases for each state report charges were incurred during 2010, some published costs were inflated over 5 years23. In addition, although the acute care costs were obtained from database analyses which selected patients with diabetes and the comorbidities of interest or hypoglycemia, many post-acute or outpatient care cost estimates were developed by applying unit costs to a typical management profile; as a result, development of these profiles required some assumptions. For example, management of macrovascular complications such as a myocardial infarction yields high costs in the year of the event and, while inpatient care data analyses conducted account for the majority of this cost estimate, some of the post-acute care and state cost estimates were from studies that did not select patients with diabetes. The ischemic stroke event-year costs in the current study capture inpatient care and use of sub-acute care facilities, so the post-acute care estimate required combining many sources to estimate costs for each of these discharge locations by applying various published fee schedules and government reports.

Conclusions

This study provides updated estimates of the direct medical costs associated with the treatment of various diabetes complications, hypoglycemia, and infections. Managing complications results in substantial costs to the healthcare system, as macrovascular complications yield high inpatient care costs in the event year, and long-term management of complications such as ESRD lead to high ongoing annual costs. Expressing each of the costs as event-year and annualized state costs per patient allows the estimates to be readily used for economic modeling. In addition, this format facilitates the conduct of other studies such as analyses modeling the management of diabetes and estimating the economic burden associated with complications.

Transparency

Declaration of funding

This study was sponsored by Janssen Scientific Affairs, LLC, Raritan, NJ, USA.

Declaration of financial/other relationships

AW and PA are employed by Evidera, a company that received funding from Janssen to conduct this study. LV and SM are employees of Janssen.

Acknowledgments

No assistance in the preparation of this article is to be declared.