Medicaid Prescription Formulary Restrictions and Arthritis Treatment Costs

[Guest guest]

Medicaid Prescription Formulary Restrictions and Arthritis Treatment Costs

Posted on: Thursday, 10 July 2008, 06:00 CDT

By , Tricia J Stahl-Moncada,

Objectives. We used the Arizona Medicaid program as a model to examine the

consequences of the relative restrictiveness of nonsteroidal

anti-inflammatory drug (NSAID)-preferred drug lists on health care use and

costs for Medicaid enrollees with arthritis.

Methods. In a retrospective, cross-sectional study of Medicaid enrollees

with rheumatoid arthritis or osteoarthritis, we used data from the Arizona

HealthQuery database and generalized linear regression models to estimate

the effect of the restrictiveness of formularies on the association between

number of NSAID drugs covered and the number of emergency department visits,

ambulatory physician visits, hospital stays, and total health expenditures.

Results. For plans with NSAID formularies that were more restrictive,

enrollees with rheumatoid arthritis experienced 22% fewer ambulatory visits

and 29% more hospitalizations, and enrollees with osteoarthritis experienced

38% fewer ambulatory visits and 52% more hospitalizations. These plans spent

an additional $935 for medical care and prescription drugs annually per

enrollee with rheumatoid arthritis.

Conclusions. Formularies that are more restrictive significantly change the

patterns of health care and prescription drug use and may have unintended

consequences in terms of more frequent and, for those with rheumatoid

arthritis, more expensive medical care.

(Am J Public Health. 2008;98:1300-1305. doi:10.2105/ AJPH.2007.118133)

Arthritis and other rheumatic diseases are the leading cause of disability

in the United States. Rheumatoid arthritis affects 3 million adults, and

osteoarthritis affects over 21 million adults.1 Arthritis is the eighth most

costly medical condition.2 Spending on hospitalizations, ambulatory visits,

and prescription drugs is twice as high for persons with arthritis than for

those with other chronic conditions and more than 8 times higher than for

those with no chronic conditions.3

Rheumatoid arthritis is an autoimmune inflammatory disease that targets the

joints. Aggressive treatment with pharmaceutical drugs can slow the

progression of joint degeneration and help control symptoms. Nonsteroidal

antiinflammatory drugs (NSAIDs) and corticosteroids are often prescribed for

pain management of rheumatoid arthritis. Traditionally, disease-modifying

antirheumatic drugs (DMARDs) have been prescribed to slow rheumatoid

arthritis's progression. Since 1998, biological response modifiers (BRMs),

drugs that stimulate the body's response to infection and disease, have been

used as an alternative to DMARDs to treat rheumatoid arthritis.

Osteoarthritis is a degenerative, rather than autoimmune, joint disease.

Treatment for osteoarthritis is limited and includes the use of NSAIDs,

analgesics, and topical creams to alleviate symptoms including joint

swelling and pain. Although rheumatoid arthritis and osteoarthritis are two

different diseases affecting the joints, NSAIDs are commonly used for pain

management in both diseases. In 1998, the Food and Drug Administration (FDA)

approved the first cyclooxygenase-2 (COX-2) inhibitor, celecoxib, a subclass

of NSAIDs, to help reduce pain and inflammation of arthritis while reducing

gastrointestinal complications associated with older NSAIDs.4 Since their

introduction, NSAIDs have remained a mainstay for pain management.

Medicaid is an important source of health insurance for persons with

arthritis. All state Medicaid programs include prescription drug benefits,

even though states are not required to do so. Medicaid spending associated

with prescription drugs doubled (from 5.6% to 12%5) between 1992 and 2002.

State Medicaid programs have used prescription drug formularies and

preferred drug lists to restrict access to more-expensive prescription

medications and control rising prescription drug costs. Decisions regarding

which drugs to include on a preferred drug list are based on the health

plan's assessment of relative clinical benefit within a therapeutic class

and judgment about the value to the state on the basis of total cost.6 As of

2003, 29 state Medicaid fee-for-service programs had obtained legislative

approval for a preferred drug list or were in the process of implementing

such a list with expanded prior authorization.6

Although the purpose of formulary restrictions is to reduce prescription

drug expenditures, they may have some unintended consequences. Studies of

the consequences of preferred drug lists and prior authorization

requirements for other drugs, such as statins and hypertensive medications,

have shown that these mechanisms may encourage the substitution of lower

cost alternatives that may not be therapeutic equivalents or may increase

nonadherence, causing adverse events that ultimately increase medical care

costs.7-10 Research has shown that prior authorization requirements for

NSAIDs and, specifically, COX-2 inhibitors have been successful in reducing

NSAID use. Fischer et al.11 found that a prior-authorization requirement for

COX-2 inhibitors in state Medicaid programs reduced NSAID use by 15%.

Smalley et al.12 found that NSAID prescriptions decreased by 19% after the

Tennessee Medicaid program implemented an NSAID prior-authorization program.

The short- and long-term effects of prior-authorization requirements for

arthritis medications on health outcomes, however, are not yet clear. One

study examined the health impact of NSAID prior- authorization requirements,

but results were limited to an 8-week follow-up period after implementing an

NSAID brand name prior- authorization requirement, thus providing a very

short-term picture of the consequences.13

Preferred drug lists often limit access to more expensive drugs or drugs

that have entered the market more recently with lessproven clinical

effectiveness as a way to control prescription drug spending. Lichtenberg14

found that the average vintage (i.e., FDA approval year) of drugs used by

Medicaid enrollees relative to non- Medicaid enrollees has increased in

recent years and that the average age of pain medications, including

valdecoxib, celecoxib, and rofecoxib, prescribed for Medicaid enrollees

compared with non- Medicaid enrollees was more than 1.2 years older for

Medicaid enrollees. Although this study provides some evidence that Medicaid

preferred drug lists have had more restrictions on recent-vintage drugs, the

consequences for health outcomes is not known, because the relative clinical

effectiveness of individual drugs was not taken into consideration.

Preferred drug lists are a mechanism to reduce prescription drug spending,

but increases in medical care may offset prescription drug savings. Most

studies have not evaluated potential substitution effects (i.e., greater use

of nonrestricted drugs or other nondrug health services) in a

well-controlled manner.15 It remains an important but unanswered question

whether formularies that are more restrictive reduce or increase spending on

medical care and prescription drugs.

We used the Arizona Medicaid program as a model to understand the impact of

the relative restrictiveness of NSAID-preferred drug lists on health care

use and costs for Medicaid enrollees with arthritis. Arizona has a long and

well-established history of managed care, with a number of health plans,

each of which is responsible for designing and managing its own preferred

drug list. A single state with multiple Medicaid health plans is ideal for

examining the effect of preferred drug list comprehensiveness on arthritis

health care use.

METHODS

The Arizona Medicaid program, Arizona Health Care Cost Containment System

(AHCCCS), covers adults 18 years and older with a household income less than

or equal to 100% of the federal poverty level ($1533 per month for a family

of 4 in 2003).16 The AHCCCS covered approximately 639000 children, 643700

nondisabled adults 18 years and older and 110800 persons who were blind or

disabled in 2004.17 AHCCCS included 13 managed care plans and a small

fee-for- service plan.18 Each of the health plans established and managed

its own prescription formulary, and the fee-for-service formulary was

established and managed by a pharmacy benefit management company.

Our study followed a cross-sectional design to examine the association

between the comprehensiveness of NSAID prescription drug formularies and

health care use for Medicaid beneficiaries with arthritis. The health care

outcomes included ambulatory physician visits, emergency department visits,

hospitalizations, and medical costs in the calendar year 2003. Although the

sample included dually eligible Medicare-Medicaid beneficiaries, Medicare

did not offer a prescription discount card or other prescription drug

benefits during this period. The sample was limited to individuals 18 years

and older who were continuously enrolled in a Medicaid-managed care plan for

the entire year. We excluded 1 Arizona health plan because it did not

provide coverage for the entire calendar year, and the fee-for-service plan

because the number of enrollees with arthritis was too small. As a result,

we included 7 acute health care plans and 5 long-term care plans.

AHCCCS data were obtained from the Arizona HealthQuery (AZHQ) database

provided by Arizona State University. AZHQ is an integrated database of

medical records from public and private data partners in Maricopa County and

is a joint project of Arizona State University and St Luke's Health

Initiatives. AZHQ includes claims data for services covered by AHCCCS and

Medicare as well as detailed data for fee-for-service and managed Medicaid

enrollees. International Classification of Diseases Clinical Modification

(ICD-9-CM)19 diagnosis codes were used to identify enrollees with a primary

or secondary diagnosis of rheumatoid arthritis (ICD-9-CM code 714.0) or

osteoarthritis (ICD-9-CM codes 715.00-715.98) on the basis of the ICD-9-CM

diagnosis codes associated with physician office visit, outpatient hospital

visit, emergency department visit, and hospitalization claims in the AZHQ.

To our knowledge, this was the first study to examine the relative

restrictiveness of a preferred drug list, so no theoretical foundation for

defining each prescription drug formulary as restrictive or lenient was

available. We classified each health plan's formulary as either lenient or

restrictive on the basis of the number of NSAIDs included. The threshold

between lenient and restrictive formularies was empirically determined by

examining the distribution of drugs covered across the 12 health plans.

Lenient formularies included those with 12 or more unique drug choices,

whereas restrictive formularies included those with fewer than 12 unique

drugs choices. We used a sensitivity analysis to test whether the results

were sensitive to the 12-drug threshold for a lenient formulary. We reran

the regression models for a series of different quantity thresholds defining

a lenient formulary.

We were specifically interested in health care use associated with

osteoarthritis and rheumatoid arthritis and stratified the models by type of

condition. We measured use by the number of ambulatory visits, including

physician, clinic, and other outpatient visits; emergency department visits;

and inpatient hospital stays with a primary or secondary ICD-9-CM diagnosis

code of arthritis. Independent variables in the models included formulary

restrictiveness (lenient and restrictive), whether the enrollee was in an

acute or longterm care health plan, age (18-44 years, 45-54 years, 55-64

years, 65-74 years, or 75 years and older), race/ ethnicity (White,

non-Hispanic; Hispanic; Black; or other), geographic origin as a measure of

access to care (proportion of the population residing in urban areas with 50

000 or more persons), and whether the enrollee had both rheumatoid arthritis

and osteoarthritis or only 1 disease. An interaction term for the type of

NSAID formulary (restrictive or lenient) and presence of both osteoarthritis

and rheumatoid arthritis was included to control for potential differences

in the incremental impact of formulary restrictiveness for those with both

diseases.

A second model estimated total health care expenditures to quantify whether

formulary restrictiveness was associated with medical care and prescription

drug spending. Health care expenditures included the total amount paid by

Medicare and Medicaid for ambulatory and emergency department visits,

hospital stays, and prescription drugs used to treat arthritis. The cost of

BRMs administered in a physician's office or hospital outpatient setting

were captured in the ambulatory care costs, whereas the cost of self-

administered BRMs were captured in the prescription drug costs.

To account for the discrete nature of our data and skewed distribution, we

used multivariate generalized linear models with Poisson distributions and

log link functions to estimate the number of ambulatory physician visits,

emergency department visits, and hospitalizations for persons with

osteoarthritis and rheumatoid arthritis separately. Because our sample

included all individuals with at least 1 service (ambulatory or emergency

department visit or hospitalization) with either osteoarthritis or

rheumatoid arthritis coded as a primary or secondary diagnosis code, nearly

all individuals had 1 or more payments recorded in the AZHQ. We estimated

total health care expenditures with a generalized linear model with a gamma

distribution and log link function.20

RESULTS

Of the 12 health plans included in the study, 2 acute plans and 3 long-term

care plans had NSAID formularies that we classified as lenient, representing

69% of enrollees with rheumatoid arthritis and 74% of enrollees with

osteoarthritis. We classified the remaining plans as restrictive. Table 1

presents the enrollee characteristics for persons with rheumatoid arthritis

and osteoarthritis. Of the 6918 enrollees, 14% had rheumatoid arthritis

only, 78% had osteoarthritis only, and 7% had both rheumatoid arthritis and

osteoarthritis. For those with rheumatoid arthritis enrolled in a health

plan with a lenient formulary, average health care expenditures were $2089

compared with $2548 for those enrolled in a health plan with a restrictive

formulary. For enrollees with osteoarthritis in a health plan with a lenient

formulary, average health care expenditures were $1077 compared with $1098

for enrollees with a restrictive formulary.

Rheumatoid Arthritis

Table 2 summarizes the parameter estimates from the use and expenditure

models for enrollees with rheumatoid arthritis. Restrictive NSAID

formularies were associated with 22% fewer ambulatory visits (b=-0.251; 1 -

exp[-0.251]=-22.2%), translating into 0.77 fewer ambulatory visits per

enrollee than per enrollees with lenient formularies. A restrictive NSAID

formulary was associated with 29% more hospitalizations (0.25 more hospital

stays) per enrollee but no difference in emergency department use.

The models also tested for differences in use between those with rheumatoid

arthritis only and those with rheumatoid arthritis and osteoarthritis.

Having both osteoarthritis and rheumatoid arthritis was associated with 38%

more ambulatory visits, 34% more hospital stays, and 61% more emergency

department visits compared with individuals with rheumatoid arthritis only.

The interaction between restrictive NSAID formularies and having both

osteoarthritis and rheumatoid arthritis was insignificant.

Table 2 also presents results for health care spending. Enrollment in a

health plan with a restrictive NSAID formulary was associated with 45%

higher spending for medical and prescription drugs, translating into $935

more for those having a restrictive NSAID formulary, after demographic and

health differences were controlled. Having both rheumatoid arthritis and

osteoarthritis was associated with 45% higher health care spending than

having rheumatoid arthritis only. For those with both rheumatoid arthritis

and osteoarthritis, having a restrictive NSAID formulary was associated with

42% lower expenditures than having a lenient formulary.

Osteoarthritis

Table 3 presents the results from the generalized linear regression models

for enrollees with osteoarthritis. Restrictive NSAID formularies were

associated with 38% fewer ambulatory visits related to the disease,

translating into 0.71 fewer ambulatory visits per enrollee compared with

those with formularies that are lenient. A restrictive NSAID formulary was

associated with 52% more hospitalizations (0.26 more hospital stays) per

enrollee but no difference in emergency department use.

Enrollees with both rheumatoid arthritis and osteoarthritis used

significantly more services than did those with osteoarthritis only,

including nearly 2.5 times more ambulatory visits and hospitalizations and

over 3 times more emergency department visits. For those with rheumatoid

arthritis and osteoarthritis, restrictive NSAID formularies were associated

with fewer hospitalizations compared with those with formularies that are

lenient.

Overall, total health care expenditures were similar between enrollees with

restrictive and those with lenient NSAID formularies. For those with

osteoarthritis and rheumatoid arthritis, health care expenditures were 86%

higher than expenditures incurred by enrollees with osteoarthritis only.

Sensitivity Analysis

To test the robustness of our results to the inclusion of enrollees with

either a primary or a secondary diagnosis of rheumatoid arthritis or

osteoarthritis, we reestimated our models for enrollees with a primary

diagnosis of 1 of the 2 conditions. The results were remarkably similar to

the full set of results (Table 4). Restrictive NSAID formularies were

associated with 16% fewer ambulatory visits and 35% more hospitalizations

for those with rheumatoid arthritis, translating into 61% higher medical and

prescription drug expenditures or $1448. For those with osteoarthritis,

enrollees with restrictive formularies had 23% more ambulatory visits and

71% more hospitalizations, but there was no difference in expenditures.

DISCUSSION

We examined prescription formulary characteristics to assess whether the

relative leniency of a Medicaid health plan's formulary related to NSAID

coverage was associated with different patterns of health care use. Our

findings indicate that restrictions on access to NSAIDs shifted care from

the ambulatory to the inpatient setting for both rheumatoid arthritis and

osteoarthritis. Differences were somewhat larger in percentage terms for

those with osteoarthritis than for enrollees with rheumatoid arthritis, but

the absolute differences in use were similar, because enrollees with

rheumatoid arthritis used more medical care overall. After factoring in

prescription drug expenditures, total expenditures were more than one third

higher for enrollees with rheumatoid arthritis in health plans with

more-restrictive formularies. For the 456 enrollees with rheumatoid

arthritis in health plans with restrictive formularies, the additional

spending associated with restrictive NSAID formularies translated into a net

increase of $426000 in costs for the health plans. Although the formulary

restrictiveness did not affect total spending for those with osteoarthritis,

it was associated with more hospitalizations, suggesting that lower

ambulatory spending may have offset the increase in spending for hospital

care. The medical use and expenditure results taken together suggest that

restrictive formularies significantly change the patterns of health care and

prescription drug use and may have unintended consequences in terms of more

frequent and, for those with rheumatoid arthritis, more expensive medical

care. Although employer-sponsored health plans can control costs by using a

combination of the prescription drugs covered by a formulary and enrollee

cost sharing (i.e., deductibles and copayments), the primary tool available

to control costs for Medicaid plans is the formulary or preferred drug list.

Our review of the prescription formularies showed that the newest and most

expensive arthritis medications available on the market were often not

included even on the most lenient Medicaid health plan formularies; thus the

formularies that were more lenient had somewhat less expensive drugs and did

not cover the gamut of arthritis drugs as do many private health plans.14

From a clinical perspective, physicians may be more concerned with the

availability of newer drugs than with the number of drugs covered. Although

Lichtenberg's results suggest that the average vintage of NSAIDs covered by

the Medicaid health plans has increased in recent years relative to the

vintage of those covered by private plans, higher copayments for these drugs

may have lessened their use.

Our findings suggest the importance of considering the effect on medical

care use when a health plan determines the comprehensiveness of its

prescription formulary, particularly for drugs used to manage chronic

conditions. The Arizona Medicaid program relies on primary care gatekeeper

physicians to provide routine medical care and make specialty care

referrals. The homogeneity in the lack of prescription drug cost sharing

strengthens these results.

The next step in disentangling the relation between formulary

restrictiveness and health care use is to understand whether differences in

use are directly related to NSAID formulary coverage or to more-general

characteristics of health plans in which the comprehensiveness of the NSAID

formulary is a proxy for broader institutional factors. It may be, for

example, that health plans with restrictive coverage of NSAIDs have less

comprehensive primary care physician networks or primary care physicians

with less capacity to treat Medicaid enrollees or who are located in areas

that are less accessible to enrollees. Although we could not quantify the

extent to which each of the health plans used case management and disease

management to help enrollees handle their arthritis, differences in use

across health plans may have been the result of different case and disease

management programs that were correlated with the number of NSAIDs covered

by their respective formularies.

The comprehensiveness of the NSAID formulary may also be a proxy for the

extent to which the formulary covers other drugs used to treat rheumatoid

arthritis, such as more-expensive DMARDs and BRMs. If this is the case, we

would expect enrollees with restrictive NSAID formularies to have lower

prescription drug costs overall because of more restrictions on obtaining

DMARDs and biological response modifiers. Our results, however, do not

support this explanation, and prescription drug costs were nearly $300

higher for those with restrictive NSAID formularies. Future work should

delve into the restrictiveness of the formularies for BRMs and DMARDs and

examine how the relation between the restrictions on prescribing each of the

3 groups of drugs affects health care use and expenditures for people with

rheumatoid arthritis.

Ultimately, interventions to reduce unnecessary hospitalizations and

emergency department care depend on the underlying mechanism driving these

differences. Although we found a strong relation between NSAID formulary

restrictiveness and hospitalization rates, our data did not allow us to

pinpoint the medical reasons for these differences. Future studies should

explore whether enrollees with restrictive formularies have higher rates of

joint replacement, for example, to better understand these differences.

Limitations

Despite the strengths of this study in providing a benchmark for the impact

of prescription formulary restrictions on arthritis treatment in a

well-established managed care environment, this study had 3 main

limitations. The AHCCCS-managed care plans each manage their own

prescription formulary; some plans were unable to provide the 2003

formulary, and consequently we used the 2005 formulary. A comparison of the

available 2003 formularies with the 2005 formularies showed minimal

differences in the number and type of covered drugs for the DMARDs and

NSAIDs. Second, we counted the number of drugs covered by the health plan

formulary or preferred drug list, rather than comparing the specific drugs

covered. The logical next step in understanding the effect of formulary

restrictions is to examine the extent to which restrictions on specific

drugs are related to health care use and outcomes. A final limitation stems

from changes in the available prescription drugs used to treat arthritis.

Although several of the highest cost arthritis drugs, such as Vioxx, have

been removed from the market, new and expensive DMARDs have been introduced

since 2003. Prior authorization may have covered moreexpensive DMARD drugs,

but this is not the same as being automatically covered or included on the

formulary.

Conclusions

Despite these limitations, our analysis extends beyond previous research

that has examined the impact of preferred drug lists by exploring

differences in medical care. We have provided a methodology to examine the

effect of formulary restrictions on health care use for chronic conditions

more generally. It is critical that policymakers and health plan

administrators understand how to most efficiently and cost effectively

manage chronic conditions, particularly those that rely on maintenance drugs

for treatment, while preserving the highest level of health status.

Although restrictions on access to NSAIDs are a powerful incentive to reduce

use, increases in emergency department visits and hospitalizations may

offset prescription drug cost savings. Our findings demonstrate the

importance of weighing the costs and benefits associated with restrictions

on formularies used to treat chronic conditions such as arthritis.

References

1. Centers for Disease Control and Prevention. State prevalence of

self-reported doctor-diagnosed arthritis and arthritis- attributable

activity limitation-United States, 2003. MMWR Morb Mortal Weekly Rep. 2005;

55:477-480.

2. Thorpe KE, Florence CS, Joski P. Which medical conditions account for the

rise in health care spending? Health Aff. 2004;W4:437-445.

3. Yelin E. Medical care expenditures and earnings losses of persons with

arthritis and other rheumatic conditions in the United States in 1997.

Arthritis Rheum. 2004;50:2317-2326.

4. US Food and Drug Administration. Patient Information Sheet: Celecoxib

(Marketed as Celebrex) 2005. Available at: http://

www.fda.gov/cder/drug/infopage/ celebrex/celebrex-ptsk.htm. Accessed January

31, 2008.

5. Cunningham PJ. Medicaid cost containment and access to prescription

drugs. Health Aff. 2005;24: 780-789.

6. Owens MK. State Medicaid program issues: preferred drug lists. Reston,

VA: National Pharmaceutical Council; 2003.

7. J, Axelsen K, Tang S. Medicaid prescription drug access

restrictions: exploring the effect on patient persistence with hypertension

medications. Am J Manag Care. 2005;11:SP27-SP34.

8. Abdelgawad T, Egbuonu- L. Preferred drug lists and Medicaid

prescriptions. Pharmacoeconomics. 2006;24(suppl 3):55-63.

9. Ridley DB, Axelsen KJ. Impact of Medicaid preferred drug lists on

therapeutic adherence. Pharmacoeconomics. 2006;24(suppl 3):65- 78.

10. Goldman DP, Joyce GF, Zheng Y. Prescription drug cost sharing:

associations with medication and medical utilization and spending and

health. JAMA. 2007;298:61-69.

11. Fischer MA, Schneeweiss S, Avorn J, DH. Medicaid

prior-authorization programs and the use of cyclooxygenase-2 inhibitors. N

Engl J Med. 2004;351: 2187-2194.

12. Smalley WE, MR, Fought RL, Sullivan L, Ray WA. Effect of a

prior-authorization requirement on the use of nonsteroidal anti-

inflammatory drugs by Medicaid patients. N Engl J Med. 1995;332:1612- 1617.

13. Momani AA, Madhaven SS, Nau DP. Impact of NSAIDs prior authorization

policy on patients' quality of life. Ann Pharmacother. 2002;36:1681-1691.

14. Lichtenberg FR. The effect of access restrictions on the vintage of

drugs used by Medicaid enrollees. Am J Manag Care. 2005;11:SP7-SP13.

15. Walser BL, Ross-Degnan D, Soumerai SB. Do open formularies increase

access to clinically useful drugs? Health Aff. 1996;15:95- 109.

16. US Dept of Health and Human Services. Annual update of the poverty

guidelines. Fed Regist. 2003;68: 6456-6458.

17. Kaiser Commission on Medicaid and the Uninsured. State Medicaid Fact

Sheets: Arizona and United States. Available at: http:/

/www.statehealthfacts.org/ mfs.jsp?rgn=4 & rgn=1 & x=9 & y=8. Accessed December 4,

2007.

18. Arizona Health Care Cost Containment System. AHCCCS Eligibility and

Enrollment Reports-December 2003. Available at:

http://www.ahcccs.state.az.us/ Statistics/AHCCCSpopulation/2003/ Dec.

Accessed December 20, 2006.

19. International Classification of Diseases, Ninth Revision, Clinical

Modification. Hyattsville, MD: National Center for Health Statistics; 1980.

DHHS publication PHS 80-1260.

20. Diehr P, Yanez D, Ash A, Hornbrook M, Lin DY. Methods for analyzing

health care utilization and costs. Annu Rev Public Health. 1999;20:125-144.

Tricia J. , PhD, and Stahl-Moncada, MS

About the Authors

Tricia J. is with the Center for Health Management and Policy

Research and the Department of Health Systems Management, Rush University,

Chicago, IL. At the time of writing, Stahl-Moncada was with Health

Systems Management, Rush University, Chicago. Requests for reprints should

be sent to Tricia J. , Director, Center for Health Management and

Policy Research, Department of Health Systems Management, Rush University,

1700 West Van Buren St, TOB Suite 126B, Chicago, IL 60612 (e-mail:

tricia_j_johnson@...).

Contributors

T. J. and S. Stahl-Moncada originated the study and drafted and

revised the article. T. J. acquired the data and conducted the data

analysis and interpretation. S. Stahl-Moncada assisted with the study design

and data analysis.

Acknowledgments

This research was supported by a grant from the Flinn Foundation (grant

25061542) and the Arizona State University Office of the Vice President for

Research and Economic Affairs.

We wish to thank , Amy Bartels, Goldberg, Diane

, Karl Matuszewski, Odwazny, Swisher, Shane, and

Del Swan for helpful comments. We thank Miwa Edge and Wade Bannister for

data management expertise.

Human Participant Protection

This study was reviewed and approved as an exempt study by the Rush

University Medical Center institutional review board.

Copyright American Public Health Association Jul 2008

© 2008 American Journal of Public Health. Provided by ProQuest Information

and Learning. All rights Reserved.

Source: American Journal of Public Health