|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||
Background Policies that increase patients' share of health care expenses decrease the use of discretionary health services but also may reduce the use of important preventive care such as mammography.
Methods We reviewed coverage for mammography within 174 Medicare managed-care plans from 2001 through 2004. Among 550,082 individual-level observations for 366,475 women between the ages of 65 and 69 years, we compared rates of biennial breast-cancer screening in plans requiring cost sharing for mammography with screening rates in plans with full coverage. We also performed a longitudinal analysis of screening rates in plans that changed from full coverage to cost sharing for mammography as compared with rates in matched control plans that did not institute cost sharing.
Results The number of plans with cost sharing for mammography, which we defined as requiring a copayment of more than $10 or coinsurance of more than 10% for screening mammography, increased from 3 in 2001 (representing 0.5% of women) to 21 in 2004 (11.4% of women). Biennial screening rates were 8.3 percentage points lower in cost-sharing plans than in plans with full coverage, a difference that persisted in adjusted analyses (P<0.001). The effect of cost sharing was magnified among women residing in areas of lower income or educational levels (P<0.001 for each interaction). Screening rates decreased by 5.5 percentage points in plans that instituted cost sharing and increased by 3.4 percentage points in matched control plans that retained full coverage (P<0.001 for the adjusted analysis).
Conclusions Relatively small copayments were associated with significantly lower mammography rates among women who should undergo screening mammography according to accepted clinical guidelines. For effective preventive services such as mammography, exempting elderly adults from cost sharing may be warranted.
Payers and insurers have increased cost-sharing requirements in recent years.4 Since 2001, the average deductible in employer-based plans has increased by 60%, and the most common copayment for an office visit has doubled.5 A growing number of employers offer "consumer-driven" health plans that require enrollees to exceed a high deductible before receiving coverage for health benefits.5 For elderly Americans, cost sharing is a prominent feature of the Medicare program, including its recently enacted prescription-drug benefit.
Persons who are subject to cost sharing use fewer health services and have lower total health care spending than do persons with full coverage.6,7,8,9,10,11 Cost sharing through deductibles and copayments may result in more prudent use of medical care and avoidance of low-value services for which the expected benefit is outweighed by a patient's financial contribution. Such requirements, however, also may reduce the use of appropriate care,12 especially for vulnerable groups such as low-income patients or those who are chronically ill. Therefore, it is crucial to understand the effect of copayments on the use of specific preventive services that are known to improve health, particularly among elderly adults, who may have greater adverse health effects from forgoing medical care and may be more sensitive to price, given their typically fixed incomes.13 If patients consider preventive care to be discretionary or if they simply cannot afford out-of-pocket payments, then even modest cost sharing may lead patients to reduce their use of effective preventive services, with potentially negative effects on health.
Breast cancer is the second leading cause of cancer-related death among women and most commonly occurs among older women.14 Because regular screening can reduce morbidity from breast cancer, clinical guidelines developed by the U.S. Preventive Services Task Force and the American Cancer Society recommend regular screening mammography for women over 40 years of age.15,16 Although rates of breast-cancer screening increased substantially during the 1990s, two recent studies reported a decline in the use of mammography from 2000 to 2005.17,18 Our study assessed the extent to which elderly women forgo this recommended preventive service when faced with cost sharing.
Methods
Data Sources and Study Population
We obtained data from the Medicare Health Plan Employer Data and Information Set (HEDIS) from 2001 through 2004 through the Centers for Medicare and Medicaid Services (CMS), which contained 24,468,528 observations for enrollees in Medicare managed-care plans. Each observation included an enrollee's health plan and variables designating adherence to HEDIS quality indicators. Information about data collection, measure specifications, and CMS-sponsored audits has been published previously.19,20 Brown University's Human Research Protections Office and the CMS Privacy Board approved the study protocol; informed consent from the patients was not required.
We matched 97% of the observations in the HEDIS data set to the Medicare enrollment file to determine enrollees' demographic characteristics. We excluded observations from enrollees who were male, who were not between the ages of 65 and 69 years, or who had died during the measurement year.
To obtain information on health plan characteristics, we linked data from 2002 through 2004 (2,189,983 observations from 178 health plans) to the Interstudy Competitive Edge database.21 For 17 health plans that could not be matched to this database, we ascertained these characteristics by contacting the plans directly. We excluded 4 plans for which we were unable to obtain this information, yielding 2,143,566 observations from 174 plans. Our final study sample included 550,082 of these observations that were assessed for the HEDIS mammography measure, representing 366,475 enrollees.
We obtained monthly information on health-plan benefits from 2001 through 2004 from the CMS, including each plan's required copayment or coinsurance for a screening mammogram. ZIP Code–level data on income and education were obtained from the 2000 U.S. Census.
Variables
Our dependent variable was a screening mammogram during the previous 2 years (yes or no). This variable was assessed annually, with more than one observation for enrollees who were included in the sample in multiple years. The primary independent variable was enrollment in a Medicare plan with cost sharing for screening mammography (yes or no). We designated plans requiring copayments of more than $10 or coinsurance of more than 10% for screening mammography as cost-sharing plans. Three plans with copayments of $10 or less or coinsurance of 10% or less for mammography were grouped with full-coverage plans. Assuming an average price of $100 to $150 for a screening mammogram,22,23,24 we calculated that 10% coinsurance represents an out-of-pocket cost of $10 to $15. For three plans that provided a range of copayments, we designated the plan's required copayment as the median value. Other covariates included year, race, and the proportions of women who were 65 years of age or older in the patient's ZIP Code who had an income below the federal poverty level or who attended college. Plan-level covariates were Census region, model type (staff or group model, independent practice association, network, or mixed model), average yearly Medicare enrollment, the age of the plan in years, and tax status (for-profit or not-for-profit).
Statistical Analysis
We assessed the characteristics of the women who were eligible for screening mammography and rates of biennial screening in full-coverage and cost-sharing plans. We constructed linear regression models with generalized estimating equations that adjusted for race, ZIP Code–level income and education, Census region, number of Medicare enrollees in the plan, age of the plan, tax status, model type, measurement year, clustering of observations within plans, and repeated measurements among enrollees. We determined whether the effect of cost sharing varied according to income, education, and race by assessing two-way interactions of these variables with cost sharing. For these cross-sectional analyses, the cost-sharing status of a health plan was defined according to the plan's benefit policy in January of the measurement year.
We identified seven health plans that instituted mammography copayments of more than $10 or coinsurance of more than 10% during the 2003 calendar year. For each of these plans, we randomly selected two matched control plans with continuous Medicare participation from 2002 through 2004 that did not institute copayments. Plans were matched on the basis of Census region, tax status, and model type. We compared the change in rates of mammography among the seven cost-sharing plans with the corresponding change among the 14 matched control plans after adjusting for income, education, and race. There were no plans that eliminated mammography cost sharing during the study period.
All analyses were performed with the use of SAS statistical software (version 9.1), and the results are reported with either two-tailed P values or 95% confidence intervals.
Results
Characteristics of Patients and Plans
As compared with enrollees in plans with full coverage, enrollees in plans with cost sharing for mammography were more likely to be black and living in ZIP Code areas characterized by lower income and education levels (Table 1). Enrollees in cost-sharing plans were also more likely to be located in the Northeast and South and to be enrolled in non–staff-model or non–group-model plans with for-profit tax status, smaller numbers of Medicare enrollees, and fewer years of operation.
|
Screening Rates
Rates of breast-cancer screening were 77.5% in plans with full coverage and 69.2% in plans with cost sharing. Annual differences in screening rates between cost-sharing plans and full-coverage plans ranged from 8 to 11 percentage points (Figure 1).
|
|
|
|
Cost sharing at the point of service encourages patients to consider costs before seeking medical care and relies on their ability to distinguish between essential and nonessential services. In this nationally representative study of Medicare managed-care enrollees, a copayment of more than $10 or coinsurance of more than 10% was associated with significantly reduced rates of mammography among women who should receive regular screening mammography, according to accepted clinical guidelines. The association of copayments with mammography rates was strong, persisted after adjustment for individual and health-plan covariates, and was consistent both in cross-sectional analyses of all Medicare health plans and in longitudinal analyses of matched health plans that did or did not add copayment requirements over time.
Cost-sharing policies disproportionately affected vulnerable population groups. Black enrollees and those from communities with lower income and educational levels were more likely to enroll in plans requiring copayments for mammography. Moreover, the effect of cost sharing was magnified for enrollees living in communities with lower levels of education and income. However, although the magnitude of cost-sharing effects differed according to socioeconomic factors, we observed a significant negative effect on mammography rates for every demographic group we analyzed.
The prevalence of cost sharing for mammography increased substantially during the years of our study. These trends mirror those described for commercial insurance plans.5 Managed-care organizations have traditionally used provider-focused mechanisms such as capitation, utilization review, and regulated networks to contain costs while providing a generous benefit structure with full coverage of preventive care. In recent years, however, increasing numbers of managed-care plans, including those that participate in Medicare, have instituted patient-directed mechanisms, such as greater cost sharing at the point of service, to reduce costs.25,26
Why would modest copayments discourage the use of a service with substantial acceptance among patients and clinicians? First, as compared with insured adults under the age of 65 years, adults 65 years of age or older often have increased out-of-pocket health care expenses associated with chronic medical conditions.27,28 Therefore, even a small copayment may represent a substantial financial deterrent to seeking routine medical care, including mammography.29,30 Second, patients may be particularly sensitive to the price of preventive services because such offerings may be viewed as less essential than health care for symptomatic conditions. For example, studies of the effect of cost sharing on emergency care have shown that copayments did not delay use of emergency care for persons with acute coronary syndromes or reduce such care for patients with severe conditions.9,31 In contrast, in the RAND Health Insurance Experiment, cost sharing decreased the use of preventive care and resulted in adverse outcomes in visual acuity and blood-pressure control in a low-income subgroup.32,33,34 A direct comparison of the effect of cost sharing on mammography with the effect on other services for Medicare enrollees was beyond the scope of this study and should be further evaluated.
The design of benefits for Medicare enrollees is an area of considerable policy debate. The goal of such policies is to ensure that Medicare enrollees have access to appropriate health services and are protected from financial risk while providing incentives to curb the inappropriate use of services. For its first two decades, Medicare did not provide coverage for cancer screening. Because insurance coverage for effective preventive services can promote their use, Medicare instituted coverage for mammography, Papanicolaou tests, fecal occult-blood tests, and colonoscopy during the 1990s.35 However, the Medicare fee-for-service program still requires a copayment for most preventive services, and we found that an increasing number of Medicare managed-care plans also have implemented such copayments for screening mammography.
The economic rationale for maintaining copayments is that first-dollar coverage of health services leads to overuse by shielding enrollees from the costs of care and providing an incentive to use services of marginal value or to use them excessively — a problem known as moral hazard.36 However, first-dollar coverage for mammography would be unlikely to foster this problem for several reasons. First, for most preventive services covered by Medicare, including screening for breast cancer, colon cancer, and osteoporosis, the appropriate frequency and population for these services have been established in clinical guidelines based on extensive evidence. Indeed, the benefit policies we reviewed typically limited coverage to one annual screening mammogram. In addition, preventive care is generally underused. Only three quarters of the eligible women in our study underwent biennial mammography. Rates of breast-cancer screening are even lower in the Medicare fee-for-service program: only 63% of women between the ages of 65 and 69 years underwent biennial mammography in 2003 or 2004.37 Finally, mammography is a service of high value to women in this age group.38 The cost-effectiveness of such screening has been demonstrated in several studies and compares favorably with that of many other medical services covered by Medicare.38,39,40,41
Cost sharing for mammography yields two striking policy contradictions. First, in an effort to monitor and improve clinical performance, the CMS has required that health plans report mammography rates in their enrolled populations publicly. Yet the increasing prevalence of copayments in these same plans could reduce mammography use, despite any quality-improvement efforts to increase such use. Second, Medicare provides coverage for its enrollees for expensive treatment of advanced breast cancer, such as mastectomy and chemotherapy. Yet coinsurance could reduce screening and thereby contribute to greater costs arising from delayed diagnoses of disease. Therefore, a key policy implication of our study is that cost sharing should be tailored to the underlying clinical and societal value of the service in question.12 For certain preventive services of demonstrated benefit, full coverage will promote appropriately greater use.
Our study has a number of strengths. First, we studied women 65 years of age or older, a group that was excluded from the RAND Health Insurance Experiment and other well-known studies of cost sharing.6,8,9 Our study analysis was based on data from 174 health plans operating in 38 states rather than data from a single plan or region. Other studies have assessed the association of supplemental insurance coverage with the use of preventive services.42,43 However, supplemental coverage is only an indirect estimate of the degree of cost sharing. Our study reviewed specific benefit policies, thereby allowing greater precision in the assessment of cost sharing. Instead of relying solely on a cross-sectional analysis, we also conducted a longitudinal analysis focusing on plans that introduced cost sharing for mammography, as compared with a matched control group of plans that maintained full coverage. We found a similar magnitude of effect with the two analyses. Finally, assessments of both copayment policies and screening mammography were ascertained through administrative data, which are typically more accurate than patients' reports.44,45
The most important limitation of our study is that enrollees were not randomly assigned to cost-sharing plans. Although we adjusted for several individual and health-plan characteristics that could explain the relationship between cost sharing and the use of mammography, selection effects or residual confounding may explain some part of our results. For example, women who are more inclined to seek mammography may enroll in plans without cost sharing, or plans without cost sharing may have other mechanisms to promote mammography. We did not have detailed clinical information to determine the effect of copayments on women who are at increased risk for breast cancer or to estimate excess late-stage diagnoses or breast-cancer deaths in cost-sharing plans. We could not analyze differential effects of specific copayment amounts, since there were limited numbers of plans in some cost-sharing ranges. The study used ZIP Code–level measures of income and education, which may underestimate the effects of individual-level socioeconomic measures.46 We were unable to identify which enrollees in our study were eligible for Medicaid coverage of Medicare managed-care copayments. Finally, we did not have information regarding physician recommendations or referrals for mammography, which are important determinants of screening.47
Ensuring that all eligible women undergo appropriate breast-cancer screening is an important public health goal. We found that a rising proportion of women in the Medicare managed-care program are subject to copayments for screening mammography and that such copayments are associated with significantly reduced rates of screening. For cost-effective preventive services such as mammography, exempting elderly beneficiaries from cost sharing may increase rates of appropriate use.
Supported by a grant (5T32-HS-000020-20) from the Agency for Healthcare Research and Quality.
Dr. Ayanian reports receiving consulting fees from RTI International and DxCG. No other potential conflict of interest relevant to this article was reported.
We thank Vincent Mor for his helpful comments on a previous version of the manuscript; Kim Elmo, Judy Giles, and Michael McCann of the CMS for their assistance with data acquisition; and the Research Data Assistance Center of the University of Minnesota for advice on obtaining Medicare data.
Source Information
From the Department of Community Health, Warren Alpert Medical School of Brown University (A.N.T., W.R.), and the Target Research Enhancement Program, Providence Veterans Affairs Medical Center (A.N.T.) — both in Providence, RI; and the Division of General Medicine, Brigham and Women's Hospital; the Department of Health Care Policy, Harvard Medical School; and the Department of Health Policy and Management, Harvard School of Public Health — all in Boston (J.Z.A.).
Address reprint requests to Dr. Trivedi at the Department of Community Health, Warren Alpert Medical School of Brown University, Box G-S121, Providence, RI 02912, or at amal_trivedi{at}brown.edu.
References
40 years -- United States, 2000-2005. MMWR Morb Mortal Wkly Rep 2007;56:49-51. [Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Effect of Cost Sharing on Screening Mammography
Dhalla I. A., Kiran T., Golden H. E., Trivedi A. N., Rakowski W., Ayanian J. Z.
Extract |
Full Text |
PDF
N Engl J Med 2008;
358:2411-2412, May 29, 2008.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
