Osteoporosis International

, Volume 17, Issue 2, pp 252–258

Fragility fracture-related direct medical costs in the first year following a nonvertebral fracture in a managed care setting

Authors

  • Robert L. Ohsfeldt
    • College of Public HealthUniversity of Iowa
    • Iowa City VA Medical CenterDepartment of Health Management and Policy
    • Proctor and Gamble Pharmaceuticals Inc.
  • Richard L. Sheer
    • Proctor and Gamble Pharmaceuticals Inc.
Original Article

DOI: 10.1007/s00198-005-1993-2

Cite this article as:
Ohsfeldt, R.L., Borisov, N.N. & Sheer, R.L. Osteoporos Int (2006) 17: 252. doi:10.1007/s00198-005-1993-2
  • 114 Views

Abstract

The objective of this study was to estimate the fracture-related direct medical costs during the first year following a fragility nonvertebral fracture in a managed care setting. This was a retrospective cohort study conducted among patients (aged 45+ years) with a primary diagnosis for a fragility nonvertebral fracture between July 1, 2000, and December 31, 2000, using MarketScan, an integrated administrative, medical, and pharmacy claims database. All patients had 6 months of observation prior to their fracture and 12 months following a nonvertebral fracture. Fracture-related direct medical costs were evaluated in the 12-month period following fracture diagnosis using 2003 Medicare fee schedule payments. The costs per fracture per year (PFPY) for specific nonvertebral fracture sites were determined, as well as costs by type of care (i.e., outpatient, inpatient, and other). A total of 4,477 women and men fulfilled the inclusion criteria. The sample was comprised of 73% women and the mean age was 70 years. The most prevalent nonvertebral fracture sites were wrist/forearm (37%), hip (25%), and humerus (15%). Mean total costs per patient per year were highest for fractures of the hip ($26,856), femur ($14,805), tibia ($10,224), and pelvis ($10,198). On average, 84% of the annual fracture-related costs were inpatient; 3% were outpatient, and 13% were long-term care and other costs. In a patient population aged 45+ years, the first month following a nonvertebral fracture has a major impact on medical care costs. The most costly nonvertebral fracture sites were hip, femur, and tibia fractures.

Keywords

Claims database analysisFracture-related health care costsNonvertebral fractureRetrospective cohort study

Introduction

Nonvertebral fractures are a major cause of morbidity in the aging population [1, 2,3]. Hip fractures are the most serious and contribute most to the health care burden of osteoporosis [1,2]. The costs of specific types of nonvertebral fractures have been estimated in several studies [4,5]. However, none has provided cost estimates from a large managed care organization’s (MCO) perspective by level of care, nor do they report cost estimates on a monthly basis. The purpose of this study was to fill these gaps by estimating the fracture-related medical costs to a managed care organization (MCO) in the first year following a fragility nonvertebral fracture.

Methods

A retrospective cohort study was conducted using Medstat’s MarketScan integrated administrative, medical, and pharmacy claims data. The MarketScan Commercial Claims and Encounters Research Database contains administrative medical claims for over 7 million employees and dependents covered by the health benefit programs of large United States employers. These claims data were collected from approximately 100 different insurance companies, Blue Cross Blue Shield plans, and third-party administrators. These data represent the medical experience of insured employees and their dependents for active employees, early retirees, COBRA continues and Medicare-eligible retirees with employer-provided Medicare Supplemental plan. No Medicaid or Workers Compensation data are included.

A dataset of administrative and medical records was obtained for patients aged 45 years or older with a primary diagnosis for a fragility nonvertebral fracture between July 1, 2000, and December 31, 2000 (6 months). All patients had 6 months of observation prior to their fracture and 12 months following a nonvertebral fracture. The fracture was identified using the fracture-related International Classification of Diseases, 9th Edition, Clinical Modification diagnosis codes (see Appendix for codes). Only closed nonvertebral fractures were examined in this study to ensure selection of fragility fractures. Open-wound fractures were assumed to be traumatic and were not included.

Patients were included in the study (Fig. 1) if they had: (1) continuous drug and medical enrollment throughout the study period; (2) no fracture diagnosis at the same fracture site in the 6-month period prior to the index diagnosis (to exclude re-fractures); (3) more than one fracture-related outpatient visit during the 12-month follow-up period (to reduce the likelihood of including costs related to a visit for a previous fracture); and (4) an emergency room visit or inpatient hospital stay for a nonvertebral fracture that would usually require acute care (i.e., hip, femur, tibia, humerus).
Fig. 1

Flowchart of patients’ selection criteria and process

The fracture-related direct medical costs were evaluated by applying 2003 Medicare fee schedule payments [6] to specific nonvertebral fracture diagnoses and procedures collected from MarketScan data. The costs were estimated on an annual basis (per fracture per year [PFPY]) and stratified by nonvertebral fracture sites including clavicle, femur, hip, humerus, pelvis, tibia/fibula, and wrist/forearm. The costs were also evaluated by type of care: (1) outpatient cost defined as the cost of fracture-related physician services including outpatient facility claims; (2) inpatient costs defined as the cost of fracture-related hospital inpatient care, ER use, hospital outpatient care, and ambulatory surgical care; (3) long-term care (LTC) costs including fracture-related costs of long-term care hospital, skilled nursing facilities (SNF), home health care, hospice, and (4) other costs including claims from independent laboratories and from unknown/unlisted facilities.

Prospective payments for inpatient and long-term care were derived using fracture-specific diagnostic-related groups (DRGs). Prospective payments for outpatient care were based on the Physician’s Current Procedural Terminology, 4th edition, and the Healthcare Common Procedure Coding System National Codes. Where no billable procedure was available, average costs for the fracture site, bill type (professional versus facility), and level of care were applied to that record to derive expected payment. These estimated costs were used in lieu of zero values. LTC costs were calculated using DRG-based Medicare prospective payment rates. For SNF, hospice, and home health care, a per diem amount was multiplied by the patient’s length of stay. For LTC hospital, a DRG-based fixed dollar amount was applied. All DRG-based costs included physician services, facility, and pharmacy claims.

Results

A total of 4,477 women and men aged 45 and older fulfilled the inclusion criteria. The sample comprised 73% women, and the mean age was 70 years. Sixty-four percent of the sample was patients aged 65 and older. The most prevalent nonvertebral fracture sites were wrist/forearm, hip, and humerus, (37%, 25%, and 15%, respectively). Patients with a hip fracture were the oldest cohort, with a mean age of 79 years. The youngest cohort were patients with a tibia/fibula fracture (mean age 63 years).

In the first year after fracture, the most expensive nonvertebral fracture sites were hip, femur, tibia, and pelvis with mean total cost per patient per year of $26,856, $14,805, $10,224, and $10,198, respectively (Table 1). For hip fracture, median total cost per patient per year was $17,012, and 25% of hip fracture patients had costs of $27,554 or more during the 12-month follow-up period. Cost distributions across patients with nonvertebral fractures at other sites in general were similarly skewed.
Table 1

Nonvertebral fracture-related direct medical costa during the first year after fracture

Fracture site

Fractures

Cost PFPY*

25%

Median

75%

First month cost per fracture

n (%)

US dollars

(%)

Clavicle

146

(3)

1,373

353

558

974

1,090

(79)

Femur

252

(6)

14,805

1,285

8,611

17,625

11,766

(79)

Hip

1,125

(25)

26,856

6,876

17,012

27,554

22,815

(85)

Humerus

663

(15)

5,567

591

1,051

4,519

4,483

(81)

Pelvis

299

(7)

10,198

581

3,948

12,080

8,991

(88)

Tibia/fbula

340

(8)

10,224

618

1,972

10,435

8,864

(87)

Wrist/forearm

1,652

(37)

2,688

548

986

1,941

2,123

(79)

Overall

4,477

(100)

10,900

2,187

5,711

10,932

9,152

(84)

aIn 2003 US dollars; *PFPY per fracture per year

Overall, most fracture-related medical costs (84%) occurred in the first month after fracture, approximately 6% of annual costs occurred in the second month and approximately 3% occurred in the third month [data not shown]. Thus, on average 93% of total fracture costs for the 12 months following nonvertebral fracture occurred within the first 3 months.

Comparing patient populations aged 45–64 and 65+ (Table 2), the most prevalent nonvertebral fracture site for the younger group was wrist/forearm (52%), while for the older group the most prevalent nonvertebral fracture site was hip (35%). The overall mean fracture-related cost among patients aged 65 and older was $12,382 per fracture per year, compared to an overall mean fracture-related cost among the younger patient group of $8,260 per fracture per year (33% lower). Although the overall mean cost per fracture per year among patients aged 45–64 years was lower than in a patient population aged 65 and older, the costs by specific fracture site generally were higher for patients aged 45–64. Thus, the higher overall mean fracture cost among the older cohort is attributable to a higher incidence of fractures at more costly sites.
Table 2

Nonvertebral fracture-related direct medical costa distribution by age groups

Fracture site

Age 45–64

Age 65 and older

Fractures

Cost PFPY*

25%

Median

75%

Fractures

Cost

PFPY

25%

Median

75%

n (%)

US dollars

n (%)

US dollars

Clavicle

82

(5)

1,626

364

618

1,262

64

(2)

1,047

289

516

881

Femur

59

(4)

22,784

1,899

9,405

39,981

193

(7)

12,366

1,219

8,347

15,413

Hip

119

(7)

33,198

2,087

15,567

46,822

1,006

(35)

26,106

7,377

17,080

26,666

Humerus

219

(14)

6,432

713

1,226

4,133

444

(15)

5,141

547

971

4,574

Pelvis

76

(5)

16,702

529

2,680

16,470

223

(8)

7,982

644

4,116

11,246

Tibia/fibula

214

(13)

11,743

623

1,972

11,167

126

(4)

7,644

601

1,966

9,034

Wrist/forearm

840

(52)

3,180

586

1,091

2,462

812

(28)

2,179

509

892

1,614

Overall

1,609

(100)

8,260

604

1,305

4,318

2,868

(100)

12,382

705

2,261

15,175

aIn 2003 US dollars; *PFPY per fracture per year

Comparing nonvertebral fracture costs between males and females (Table 3), the overall mean fracture-related cost among female patients was $11,013 per fracture per year, which is similar to the overall mean fracture-related cost for male patients ($10,593 per fracture per year). However, there were some differences in fracture costs at specific fracture sites. The most expensive fracture site was hip for both groups; however, females incurred 20% higher hip fracture cost than males ($28,110 vs $23,423). In contrast, the mean cost for a fracture of the pelvis among male patients was 50% higher than for female patients ($13,507 v $8,990).
Table 3

Nonvertebral fracture-related direct medical costa distribution by gender

Fracture site

Females

Males

Fractures

Cost

PFPY*

25%

Median

75%

Fractures

Cost PFPY

25%

Median

75%

n (%)

US dollars

n (%)

US dollars

Clavicle

75

(2)

1,511

340

541

930

71

(6)

1,226

353

615

1,153

Femur

174

(5)

14,698

1,294

7,638

17,135

78

(7)

15,045

1,219

10,017

18,758

Hip

824

(25)

28,110

7,928

17,745

28,589

301

(25)

23,423

3,410

14,168

23,623

Humerus

520

(16)

5,672

572

1,034

4,784

143

(12)

5,188

672

1,053

2,686

Pelvis

219

(7)

8,990

562

3,779

11,679

80

(7)

13,507

616

4,755

13,412

Tibia/fibula

225

(7)

10,284

604

1,529

11,265

115

(10)

10,107

664

2,668

8,931

Wrist/forearm

1,245

(38)

2,472

538

976

1,811

407

(34)

3,349

569

1,042

2,363

Overall

3,282

(100)

11,013

659

1,679

11,724

1,195

(100)

10,593

674

1,895

10,956

aIn 2003 US dollars; * PFPY per fracture per year

The majority of fractures (91%) required initial inpatient care (51% required an initial hospitalization and 40% of fractures required an initial emergency room or urgent care visit). However, 8% of all fractures were treated exclusively in a physician’s office (19% of wrist/forearm, 16% of clavicle, and 11% of pelvis fractures, data not shown).

Overall, hip and femur were the most resource-intensive fractures (Table 4). Ninety-seven percent of hip and 92% of femur fractures required a hospital stay with an average length of 10 days and 11 days, respectively. Fifty-six percent of hip and 27% of femur fractures were discharged to long-term care that required 9 days and 11 days of average stay, respectively.
Table 4

Nonvertebral fracture-related resource utilization during the first year after fracture

Fracture site

Fractures

Hospital stay

Physician services

LTCb

Patients

Patients

Stay

Mean LOSa

Inpatient

Outpatient

Patients

Mean LOS

n

n (% of Fractures)

n

(SD)

n per fracture

n per fracture

n (% of fractures)

(SD)

Clavicle

146

31

(21)

33

5

(5.5)

611

4

555

4

6

(4)

4

(3)

Femur

252

232

(92)

247

11

(13)

1,798

7

554

2

67

(27)

11

(16)

Hip

1,125

1,094

(97)

1,480

10

(16)

9,100

8

1,687

1.5

630

(56)

9

(20)

Humerus

663

257

(39)

285

7.5

(12)

3,667

5.5

3,522

5

77

(12)

8

(13)

Pelvis

299

209

(70)

257

8

(11)

1,736

6

506

2

85

(28)

8

(14)

Tibia/fibula

340

192

(56)

263

7

(11)

1,906

6

1,359

4

74

(22)

9

(15)

Wrist/forearm

1,652

285

(17)

303

6

(9)

6,230

4

7,401

4.5

36

(2)

9

(11)

Overall

4,477

2,300

(51)

3,171

8.8

(14)

25,048

6

15,584

3

975

(22)

8.8

(18)

aLOS length of stay, bLTC long-term care, such as, skilled nursing facility, home health care, or hospice

In terms of the distribution of nonvertebral fracture-related direct medical cost by level of care, overall, 84% of annual fracture-related costs were for inpatient care, 3% for outpatient care, and 9% for long-term care costs (Table 5). Hip and femur fractures by far were the most expensive nonvertebral fracture sites in an inpatient setting ($23,266 and $11,575 per fracture per year, respectively) and in a long-term care facility ($2,950 and $1,572 per fracture per year, respectively). However, in an outpatient setting, the most expensive nonvertebral fractures were wrist, humerus and tibia fractures ($467, $420, and $392 per fracture per year, respectively).
Table 5

Nonvertebral fracture-related direct medical costa distribution by place of service during the first year after fracture

Fracture site

Inpatientb

Outpatientc

LTCd

Othere

Total cost

US dollars (% of total)

Clavicle

935

(68)

311

(23)

5

(0)

123

(9)

1,373

Femur

11,575

(78)

138

(1)

1,572

(11)

1,521

(10)

14,805

Hip

23,266

(87)

96

(0)

2,950

(11)

544

(2)

26,856

Humerus

4,572

(82)

420

(8)

328

(6)

248

(4)

5,567

Pelvis

8,720

(86)

152

(1)

963

(9)

364

(4)

10,198

Tibia/fibula

8,431

(82)

392

(4)

657

(6)

744

(7)

10,224

Wrist/forearm

1,885

(70)

467

(17)

138

(5)

197

(7)

2,688

Overall

9,124

(84)

316

(3)

1,044

(9)

416

(4)

10,900

aIn 2003 US dollars

bInpatient defined as hospital stay, emergency room, or urgent care. Prospective payments for inpatient care were derived using fracture-specific DRGs that included fracture-specific inpatient physician services, facility claims and pharmacy claims

cOutpatient defined as outpatient physician services including facility claims. Prospective payments for physician services were derived using fracture-specific procedure codes

dSuch as SNF, home health care, or hospice. LTC costs were calculated using DRG-based Medicare prospective payment rates. For SNF, hospice, and home health care, a per diem amount was multiplied by the patient’s length of stay. For LTC hospital, a DRG-based fixed dollar amount was applied

eOther costs include claims from independent laboratory and from unknown/unlisted facilities

Discussion

Several past studies have attempted to estimate the direct medical costs of nonvertebral osteoporotic fractures. Two fundamentally different approaches have been used to estimate fracture-related costs using insurance claims or administrative data. The first method, used in the present study, employs information from diagnosis or procedure codes to define the resource utilization related to fracture and the associated costs. The second method relies on an indirect attribution of costs to fractures using a case-control approach. Post-fracture medical care costs among patients with incident fractures (cases) are compared with the direct medical costs among matched individuals without fracture (controls), usually matched by age, gender and comorbidity factors. The difference in direct medical costs between cases and controls (incremental costs) is an estimate of the direct medical costs specifically attributable to the incident fracture, rather than the background clinical conditions of individuals with incident fractures.

In most cases, one would expect estimates of fracture costs using the incremental cost estimation method to be lower than estimates produced using more direct measurement of fracture-related costs. For example, coincidental treatment for comorbid conditions during an episode of care (e.g., inpatient stay) may be assigned to fracture-related costs if a primary or secondary diagnosis of fracture in the insurance claim is used to label costs for the episode as fracture-related costs. Presumably, the incremental cost estimation method would purge the cost impact of coincidental treatment for comorbid conditions from the estimate of fracture costs.

The estimated mean cost per hip fracture over 1 year reported in Table 1 was $26,856, which compares with previously reported estimates of $29,300 and $20,800–$23, 900 (expressed in 2003-equivalent dollars) for studies using fracture-related cost estimation methods [7, 8]. These estimates are significantly higher than a median incremental cost estimate of $13,600 (in 2003 dollars) reported by Gabriel and colleagues [4], though their median cost estimate is not dramatically lower than the median cost estimate reported in Table 1 ($17,012).

In practice, however, the incremental cost method can over-adjust fracture costs estimates given unobserved differences between cases and controls, coupled with the large variances in costs. For example, the estimated mean cost per wrist/forearm fracture reported in Table 1 ($2,688) is reasonably similar to the estimated median incremental cost per fracture “like a wrist” fracture of $2,400 (in 2003 dollars) reported in Gabriel et al., and is substantially greater than the median cost estimate reported in Table 1 ($986).

Overall, the estimates of the direct medical costs associated with fractures by fracture site over the first year following fracture generally are consistent with previous estimates published in the literature. However, prior studies have not attempted to quantify the timing of fracture-related costs following an incident fracture. The results reported here indicate that most facture-related costs over 1 year post-fracture occur within 1 month of facture.

Conclusion

In a patient population aged 45+ years, patients who experienced a nonvertebral fragility fracture had mean costs of about $11,000 over 12 months post-fracture, with approximately 84% of the costs occurring during the first month. Although the most expensive nonvertebral fracture site was the hip, costs associated with fractures of the femur, tibia and pelvis also were substantial. Thus, hip and other nonvertebral fractures can have a major impact on managed care costs, particularly among managed care organizations providing coverage for Medicare or retiree populations.

Acknowledgements

This study was funded by Procter & Gamble Pharmaceuticals, Mason, OH, and Aventis Pharmaceuticals, Bridgewater, NJ

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2005