| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
RESEARCH AND PRACTICE |
Cathy J. Bradley is with the Department of Medicine, and Charles W. Given is with the Department of Family Practice, Institute for Health Care Studies, Michigan State University, East Lansing, Mich. Caralee Roberts is with the Roberts Research Associates, East Lansing, Mich.
Correspondence: Requests for reprints should be sent to Cathy J. Bradley, PhD, Associate Professor, Michigan State University, Department of Medicine, B413 Clinical Center, East Lansing, MI 48824 (e-mail: cathy.bradley{at}ht.msu.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Objectives. We compared cervical cancer incidence, stage at diagnosis, and survival in Medicaid-insured and non–Medicaid-insured populations.
Methods. We stratified the sample by age and used ordered logistic regression to predict stage at diagnosis and used Cox proportional hazards regression to predict survival.
Results. Medicaid insured nearly one quarter of women diagnosed with cervical cancer. The likelihood of late-stage disease was greatest for women who enrolled in Medicaid after diagnosis. Women younger than 65 years who enrolled in Medicaid after diagnosis were more likely to die from cervical cancer than were women who were not insured by Medicaid (hazard ratio=2.40, 95% confidence interval=1.49, 3.86).
Conclusions. Our study underscores the importance of cervical cancer screening programs targeted at low-income women.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Since 1991, the Michigan Department of Community Health has administered a comprehensive breast and cervical cancer control program (BCCCP) funded by the US Centers for Disease Control and Prevention. The overarching goal of the BCCCP is to reduce disparate cancer outcomes among low-income women by providing screening services free of charge to women whose incomes are at or below 250% of the federal poverty level.19 Insured women may also be screened as part of the BCCCP, provided that they are not enrolled in Medicare Part B or a managed care or health maintenance organization and that they are underinsured for screening services and meet income eligibility requirements. As of 2001, any woman who is diagnosed with breast or cervical cancer in Michigan is automatically enrolled in Medicaid, which will pay for all treatment costs until her physician indicates that she is cancer free.19 Before 2001, women diagnosed through the BCCCP were enrolled in Medicaid only if they met disability or Aid to Families with Dependent Children eligibility criteria as well as asset and income requirements.
Using linked statewide Cancer Registry and Medicaid enrollment databases, we compared cervical cancer incidence rates, cancer stage at detection, and chances of survival between women insured by Medicaid and women not insured by Medicaid. Our study identified a low-income, insured population and distinguished between women who enrolled in Medicaid after being diagnosed with cancer and women who were diagnosed with cancer while enrolled in Medicaid. Women enrolled in Medicaid before their diagnosis had an opportunity to benefit from covered health care services that may have resulted in early-stage cancer detection and treatment. By contrast, women enrolling in Medicaid after diagnosis may have been previously uninsured or underinsured, with limited access to health care. Our analysis provides information about the population of women served by the BCCCP who may subsequently enroll in Medicaid. This inquiry is particularly relevant now that many states are considering curtailing Medicaid services to resolve the budgetary crises they are facing.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Michigan Department of Community Health analysts matched incident cervical cancer cases to the 1996 and 1997 Medicaid enrollment file. Patients who matched on all variables, on all variables except address, or on name (first and last), date of birth, gender, and either the first 5 digits or the last 4 digits of the Social Security Number were considered valid matches.9 This process identified 1125 women who were insured by Medicaid. The Medicaid enrollment file contained patients’ complete historical enrollment information extending before the 1996–1997 study period.
Because Medicaid is provided only to individuals who are medically indigent, Medicaid insures the lowest socioeconomic stratum of the population—one that is associated with complex medical conditions and low use of preventive services.20 Many low-income persons who do not meet the categorical or financial qualifications for Medicaid, as well as uninsured individuals, were present in the control population. Although imperfect, our method correctly identified an important segment of low-income, insured women for whom Medicaid policy changes and programs such as the BCCCP can greatly influence health outcomes.
We stratified the sample by patients younger than 65 years and patients aged 65 years and older. This distinction is important, because the 2 groups of women differ systematically in health status and alternative sources for health insurance. To enroll in Medicaid, recipients must either qualify under the Aid to Families with Dependent Children program or have a disabling condition expected to last 1 or more years. The majority of women younger than 65 years qualified for Medicaid under the Aid to Families with Dependent Children program (66%); therefore, many of these patients were of childbearing age and were likely to be reasonably healthy. Medicaid enrollment for younger women varied from month to month, depending on recipients’ family status, age of dependent children, and assets and income. In contrast, Medicare insures nearly every US citizen 65 years and older and may be associated with greater access to health care providers and continuous coverage relative to Medicaid insurance alone. In addition, women 65 years and older who were dually eligible for Medicaid and Medicare were likely to be disabled; 95% of the women who were 65 years or older qualified for Medicaid because of a disability.
We distinguished between patients who enrolled in Medicaid after they were diagnosed with cancer, who were enrolled in Medicaid before diagnosis, and who were non–Medicaid-insured. If a patient was enrolled in Medicaid during the same or later month and year as the month and year of diagnosis, we coded this individual as "enrolled after diagnosis." If the patient had been enrolled in Medicaid for 1 or more months before the date of cancer diagnosis, we considered the patient "enrolled at diagnosis." Patients were "non-Medicaid" if they were not enrolled in Medicaid at any time before or during the study period. Enrollment in the same month of diagnosis, in many cases, indicates that once the beneficiary was determined to meet enrollment criteria, which likely occurred some months following a cancer diagnosis, Medicaid enrollment was made retroactive to the date of diagnosis.
To estimate cervical cancer incidence in the Medicaid and non-Medicaid samples, we aggregated female Medicaid enrollees and all women residing in Michigan into 5-year age groups. The number of cancer cases was then calculated for female Medicaid enrollees and all women residing in Michigan for 1996 and 1997 and divided by the total number of female Medicaid enrollees and the total female population, respectively, by age group, for 1996 and 1997. We multiplied the resulting quotients by 1000 to obtain an incidence rate per 1000 women for each age group. In this analysis, we could not subtract Medicaid enrollees from the general population; thus, differences observed in cervical cancer incidence between the 2 groups underestimated the true difference in incidence. Although the actual difference in incidence is likely to be greater than we estimated, our estimates are informative because these incidence rates help to demonstrate the relative cancer burden in the 2 populations.
We used ordered logistic regression to analyze cervical cancer stage at diagnosis (in situ, local, regional, and distant). Patients whose cancers were not staged were excluded from the analysis (n = 86, women younger than 65 years; and n = 24, women aged 65 years and older). Using predicted probabilities computed from the coefficients derived in the ordered logistic regression, we estimated the likelihood of cancer detection at each stage for non–Medicaid-insured women, women insured by Medicaid at the time of diagnosis, and women enrolled in Medicaid after diagnosis. We used a multivariate Cox proportional hazards model to estimate the risk of death from cervical cancer during the study period. We also estimated the risk of death from all-cause mortality (results are not shown but are available on request). We tested the proportional hazard assumption for individual variables and performed a global test of proportionality with the Schoenfeld residuals.21 Age and race/ethnicity (White, African American, and Other) were controlled in all regression models, and early-stage diagnosis (defined as either in situ or local cancer) was controlled in the Cox model. All statistical analyses were performed with Stata version 7.0 (Stata Corp, College Station, Tex).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
shows that compared with the incidence in Michigan’s general population, the age-adjusted incidence of cervical cancer among Medicaid-insured women was much higher, particularly among women in younger age groups. The highest incidence for both Medicaid (2.43 per 1000 women) and the general population (1.98 per 1000 women) was in women aged 25–29 years. The age group with the largest difference in incidence was women 50–54 years old (1.21 per 1000 Medicaid-insured women vs 0.32 per 1000 women in the general population). Overall, the cervical cancer incidence rate was higher in the Medicaid population in every age group up to 80 years of age. Because women insured by Medicaid were included in the general population of women, the true difference in incidence rates is higher than that reported here.
|
shows the descriptive statistics for the cervical cancer sample. During 1996 and 1997, 5049 cases of cervical cancer were diagnosed in Michigan, and Medicaid insured 22% of these women (n = 1125). For women younger than 65 years, the non–Medicaid- and Medicaid-insured samples were comparable in age (mean: 36.6 years and 35 years, respectively). Although both samples had a high percentage of White women (71% and 57%, respectively), only 10% of non–Medicaid-insured women were African American, whereas 28% of the Medicaid-insured women were African American. In both samples, most cancers were detected at the in situ or local stage (
90%), although there was a higher percentage of regional, distant, and invasive/ unknown cancers in the Medicaid sample (10%) relative to the non–Medicaid-insured sample (5%). As would be expected given the more severe stage of disease, a higher percentage of women in the Medicaid sample died relative to women in the non-Medicaid sample (6% vs 2%). In both samples, the vast majority of deaths were caused by cervical cancer. In the Medicaid sample, approximately one quarter of the women were enrolled in Medicaid after the detection of cervical cancer.
|
, column 4). The average age of the women was 74 years. Nearly 34% of the older women insured by Medicaid were African American. Later-stage cancers and deaths were far more common in women aged 65 years and older relative to women younger than 65 years. Although 71% of older, non–Medicaid-insured women received diagnoses of cervical cancer at in situ or local stage, only 55% of older, Medicaid-insured women received diagnoses at an early stage. Approximately 18% and 29% of non–Medicaid-insured women and Medicaid-insured women, respectively, died during the study period, and as found with the younger women, the majority of these deaths were related to cervical cancer. On the basis of univariate analysis alone, Medicaid insured a substantial proportion of the total cervical cancer cases, with one quarter of Medicaid recipients enrolled in Medicaid after diagnosis. In addition, women who were insured by Medicaid were more likely to have late-stage disease when diagnosed and were more likely to die compared with women who were not insured by Medicaid. The descriptive analysis also revealed a heightened vulnerability for late-stage cervical cancer among women aged 65 years and older relative to younger women. However, these older women represented fewer than 7% of the total cervical cancer population.
Cervical Cancer Detection
In an ordered logistic regression predicting cancer stage at diagnosis, age, "other or unknown" race/ethnicity, and enrollment in Medicaid both before and after diagnosis were associated with later-stage disease for both women younger than 65 years and women aged 65 years and older (Table 2). Table 3 shows the estimated predicted probabilities by cancer stage for women who were not insured by Medicaid, enrolled in Medicaid before a cervical cancer diagnosis, and enrolled in Medicaid after a cervical cancer diagnosis. We used the mean values for age and race/ethnicity in these estimations. A consistent pattern emerged from the data: Women who were not insured by Medicaid received diagnoses at early stages, whereas women who were enrolled in Medicaid at the time of diagnosis had a slightly lower chance of early-stage detection, and women enrolled in Medicaid after diagnosis were much less likely to have an early-stage cancer at detection. This pattern held for both younger and older women (Table 3). In younger women, for example, the probability of receiving a cervical cancer diagnosis in situ was 92% for those who were not insured by Medicaid, 88% for those who were enrolled in Medicaid at the time of diagnosis, and 73% for those who enrolled in Medicaid after diagnosis. The most dramatic differences in the predicted cancer stage probabilities were observed in older women with in situ cancer, for whom the predicted probabilities were 54%, 36%, and 19% for the non–Medicaid-enrolled, Medicaid-enrolled at the time of diagnosis, and Medicaid-enrolled after diagnosis categories, respectively.
|
|
shows the results of a Cox proportional hazards model predicting death from cervical cancer after control for age, early-stage disease (in situ or local), race/ethnicity, and Medicaid enrollment status. Age and cancer stage were associated with survival, whereas race/ethnicity was not. For women younger than 65 years, the hazard ratio for age was 1.06 (95% confidence interval [CI] = 1.04, 1.09). The hazard ratio for women enrolled in Medicaid at the time of diagnosis was 1.77 (95% CI = 1.02, 3.07), and the hazard ratio for women enrolled in Medicaid after diagnosis was 2.40 (95% CI = 1.49, 3.86). This result is striking, given that cancer stage was controlled in the model. Age was positively associated with an increased risk of death for women aged 65 years and older. Medicaid enrollment, however, was not associated with an increased risk of death in older women. This was partly a result of the collinearity between Medicaid enrollment and cancer stage at diagnosis (Table 3). When early-stage cancer at diagnosis was removed from the model, the hazard ratio for women enrolled in Medicaid after diagnosis was 3.20 (95% CI = 1.25, 8.17; results not shown), and the hazard ratio for women enrolled in Medicaid at the time of diagnosis was not statistically significant (P < .05).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Our findings with regard to women enrolled in Medicaid after diagnosis suggest that expanding Medicaid insurance to include uninsured women is one way to reduce, although not eliminate, disparities in cervical cancer. The economic challenge of providing health care through the Medicaid program has been widely discussed.24 Nevertheless, our study indicates that reducing coverage may have adverse economic consequences, because low-income women are likely to enroll in Medicaid after a cancer has progressed and is more expensive to treat. In 1996 and 1997, the Michigan Medicaid program spent nearly $2.5 million dollars for the treatment of confirmed cervical cancer cases. A question to examine in future research is whether fewer women with late-stage disease (which is less expensive to treat) enrolled in Medicaid after 2001, when Medicaid eligibility became automatic for women diagnosed through the BCCCP.
Our study demonstrates that there are 2 distinct populations of women with cervical cancer—those younger than 65 years and those aged 65 years and older—who require different strategies for cervical cancer screening and treatment. Promotion of more intensive screening among elderly women, particularly if they are nursing home residents, presents a unique set of challenges. Residents in long-term care facilities might not benefit from improvements in cervical cancer detection and treatment. These women may have comorbid conditions (e.g., dementia, reduced physical functioning) that negatively influence a physician’s willingness to recommend screening or treatment. The average age of women in the sample who were older than 65 years was 74 years. Nonetheless, the utility of increased cervical cancer screening and treatment for elderly women should be investigated. However, Fahs found that triennial screening reduced mortality by 74% at a cost of $7345 per year of life saved for women 65 years and older.25 Other studies have shown that a strategy of combined Pap smears and human papillomavirus testing every 2 years up to age 100 years can be implemented at $70347 per quality-adjusted life-year saved.26 Such investigations assist providers and policymakers in understanding the benefits of intensive screening efforts relative to their costs.
A limitation of our study is that we could not identify uninsured individuals within our sample. We could identify only those who were insured by Medicaid, and thus, the younger, non–Medicaid-insured group contained individuals who were uninsured. We hypothesized, but could not test, that the outcomes (e.g., stage at detection, mortality) of uninsured patients are similar to those of patients insured by Medicaid; thus the presence of uninsured women in the non-Medicaid group underestimates the true difference between women who are Medicaid insured and women who are privately insured. Although it is likely that women who enrolled in Medicaid after diagnosis were previously uninsured, we cannot be certain that this is the case—we only know they were not insured by Medicaid. In addition, the findings from our study of a single state may not be generalizable to other geographical areas.
Our research suggests that several measures can be taken to reduce disparities in cervical cancer. Medicaid insurance, for example—although it was not comparable to other forms of insurance—improved stage at diagnosis and survival for women who were enrolled at the time of diagnosis. Because disparities in diagnosis and survival were observed in the sample of women enrolled in Medicaid at the time of diagnosis, further investigation regarding the quality of care provided to women insured by Medicaid is warranted. As the medical community considers how best to address disparities in health care, it must face the financial and programmatic commitment required to establish regular patterns of care with vulnerable populations and to ensure that the care delivered is of high quality. Some of the costs to provide the needed care to low-income women may be offset by a reduction in future Medicaid enrollment after the disease has advanced to a stage in which it is expensive to treat and recovery is less likely.
|
Acknowledgments |
|---|
Human Participant Protection
This study was approved by Michigan State University’s institutional review board.
|
Footnotes |
|---|
Accepted for publication December 6, 2003.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
2. Shinagawa SM. The excess burden of breast carcinoma in minority and medically underserved communities. Cancer. 2000;88(5 suppl):1217–1223.[ISI][Medline]
3. Lannin DR, Mathews HF, Mitchell J, et al. Influence of socioeconomic and cultural factors on racial differences in late-stage presentation of breast cancer. JAMA. 1998;279:1801–1807.
4. Figueroa JB, Breen N. Significance of underclass residence on the stage of breast or cervical cancer diagnosis. Am Econ Rev. 1995;85:112–116.[ISI][Medline]
5. Greenwald HP, Borgatta EF, McCorkle R, et al. Explaining reduced cancer survival among the disadvantaged. Milbank Q. 1996;74:215–238.[ISI][Medline]
6. Roetzheim RG, Pal N, Tennant C, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst. 1999;91:1409–1415.
7. Haynes MA, Smedley BD, eds. The Unequal Burden of Cancer: An Assessment of NIH Research and Programs for Ethnic Minorities and the Medically Underserved. Washington, DC: National Academy Press; 1999.
8. Eley JW, Hill H, Chen VW, et al. Racial differences in survival from breast cancer. Results of the National Cancer Institute Black/White Cancer Survival Study. JAMA. 1994;272:947–954.[Abstract]
9. Bradley CJ, Given CW, Roberts C. Disparities in cancer diagnosis and survival. Cancer. 2001;91:178–188.[Medline]
10. National Cancer Institute. The Nation’s Investment in Cancer Research: A Budget Proposal for Fiscal Year 2001. Bethesda, Md: National Institutes of Health; 2001. Publication 02–4373.
11. National Institutes of Health Consensus Conference on Cervical Cancer. Bethesda, Maryland, April 1–3, 1996. J Natl Cancer Inst Monogr. 1996;(21):1–148.
12. Mandelblatt J, Andrews H, Kerner J, et al. Determinants of late stage diagnosis of breast cancer and cervical cancer: the impact of age, race, social class, and hospital type. Am J Public Health. 1991;81:646–649.
13. Mundt AJ, Connell PP, Campbell T. Race and clinical outcomes in patients with carcinoma of the uterine cervix treated with radiation therapy. Gynecol Oncol. 1998;71:151–158.[ISI][Medline]
14. Grisby PW, Hall-Daniels L, Baker S, et al. Comparison of clinical outcomes in black and white women treated with radiotherapy for cervical carcinoma. Gynecol Oncol. 2000;79:357–361.[ISI][Medline]
15. Lundin FE, Christopherson WM, Mendez WM, et al. Morbidity from cervical cancer: effects of cervical cytology and socioeconomic status. J Natl Cancer Inst. 1965;35:1015–1025.
16. Howell EA, Chen YT, Concato J. Differences in cervical cancer mortality among black and white women. Obstet Gynecol. 1999;94:509–515.
17. Jones WB, Shingleton HM, Russell A, Chmiel JS, et al. Patterns of care for invasive cervical cancer. Cancer. 1995;76:1934–1947.[ISI][Medline]
18. Serur E, Fruchter RG, Maiman M, et al. Age, substance abuse, and survival of patients with cervical carcinoma. Cancer. 1995;75:2530–2538.[ISI][Medline]
19. Michigan BCCCP Web page. Available at: http://www.michigancancer.org/michiganbcccp.htm. Accessed October 24, 2003.
20. DeVoe JE, Fryer GE, Phillips R, Green L. Receipt of preventive care among adults: insurance status and usual source of care. Am J Public Health. 2003;93:786–791.
21. Grambsch P, Therneau T. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81:515–526.
22. Weiss-Jibaja ML, Volk RJ, Kingery P, et al. Tailored messages for breast and cervical cancer screening of low-income and minority women using medical records data. Patient Educ Couns. 2003;50:123–132.[ISI][Medline]
23. Bradley CJ, Given CW, Roberts C. Race, socioeconomic status, and breast cancer treatment and survival. J Natl Cancer Inst. 2002;94:493–496.
24. Boyd DJ. The bursting state fiscal budget and state Medicaid budgets. Health Aff. 2003;22:46–61.
25. Fahs MC, Mandelblatt J, Schechter C, et al. Cost effectiveness of cervical cancer screening for the elderly. Ann Intern Med. 1992;117:520–527.
26. Mandelblatt JS, Lawrence WF, Womack SM, et al. Benefits and costs of using HPV testing to screen for cervical cancer. JAMA. 2002;287:2372–2381.
This article has been cited by other articles:
|
|
 |
|
  L. R. Shugarman, M. E.S. Sorbero, H. Tian, A. K. Jain, and J. S. Ashwood An Exploration of Urban and Rural Differences in Lung Cancer Survival Among Medicare Beneficiaries Am J Public Health, July 1, 2008; 98(7): 1280 - 1287. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. N. Westin, D. Bustillos, J. B. Gano, M. M. Fields, A. L. Coker, C. C. Sun, and L. M. Ramondetta Social Factors Affecting Treatment of Cervical Cancer: Ethical Issues and Policy Implications Obstet. Gynecol., March 1, 2008; 111(3): 747 - 751. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Bradley, C. W. Given, Z. Luo, C. Roberts, G. Copeland, and B. A. Virnig Medicaid, Medicare, and the Michigan Tumor Registry: A Linkage Strategy Med Decis Making, August 1, 2007; 27(4): 352 - 363. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
