SEARCH

SEARCH BY CITATION

Keywords:

  • health care reform;
  • insurance;
  • mammography;
  • breast cancer

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

BACKGROUND:

Massachusetts law requires all residents to maintain a minimum level of health insurance, and rates of uninsurance in that state decreased from 6.4% in 2006 to 1.9% in 2010. The authors of this report assessed whether health insurance expansion was associated with use of mammography and earlier stage at breast cancer diagnosis.

METHODS:

By using a prereform/postreform design with a concurrent control (California), mammography rates in the last year were assessed using the Behavioral Risk Factor Surveillance System survey and the diagnosis of stage I (vs II/III/IV) breast cancers based on cancer registry data among women ages 41 to 64. Propensity score analyses were used to compare California women who were most similar to women in Massachusetts with Massachusetts women.

RESULTS:

Among propensity-weighted cohorts, adjusted mammography rates in Massachusetts were 69.2% in 2006, 69.5% in 2008, and 69.0% in 2010. In California, the rates were 59% in 2006, 60.3% in 2008, and 56.2% in 2010 (P = .89 for interaction by state for 2010 vs 2006). Among propensity-weighted cohorts, adjusted rates of diagnosis with stage I cancers were 52.2% in 2006, 53.5% in 2007, and 52.4% in 2008 in Massachusetts versus 46.4% in 2006, 46.3% in 2007, and 45.7% in 2008 in California (P = .58 for interaction by state for 2010 vs 2006).

CONCLUSIONS:

Health insurance reform in Massachusetts was not associated with increased rates of mammography or earlier stage at diagnosis compared with California, possibly because of insurance and mammography rates that already were high. Additional research is needed to assess the impact of insurance expansions in other populations, especially those with higher uninsurance rates. Cancer 2013. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

In April 2006, Massachusetts enacted a health care reform law requiring all residents to carry a minimum level of health insurance.1 This law, which is enforced through the state tax return, is intended to increase health insurance coverage using a combination of government subsidies, Medicaid expansions, and health insurance market reforms. The new law also requires employers with >10 employees to make a “fair and reasonable” contribution toward the cost of health insurance or pay an annual assessment of up to $295 per employee. The Massachusetts health insurance expansion served as a model for the Patient Protection and Affordable Care Act.

Since passage of the law, rates of uninsurance in Massachusetts declined from about 6.4% in 2006 to 1.9% in 2010.2 Some early data from Massachusetts suggested that health insurance reform was associated with improvements in cost-related barriers to care but no improvements in access to a personal physician or health status.3 Other evidence has demonstrated sustained improvements in access to a usual source of care and receipt of preventive care as well as gains in the affordability of health care4, 5 and improvements in health status,6 although some patients still report challenges with access and affordability.4, 5

Women without insurance are less likely than the insured to receive preventive care, including mammograms, to seek care in a timely fashion, or to receive recommended treatment; and uninsured women are more likely to be diagnosed with breast cancers at a more advanced stage.7-14 Despite the association between lack of or inadequate health insurance and less mammography use, few data are available about whether expansion of health insurance can improve rates of mammography use or earlier breast cancer diagnosis. A recent study indicated that the expansion of Medicaid coverage to uninsured individuals in Oregon was associated with an increase in preventive care services, including mammograms.15 We assessed whether health insurance expansion in Massachusetts was associated with the use of mammography and earlier stage at breast cancer diagnosis compared with a control state, California, in which similar attempts at health insurance reform were unsuccessful.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Data

Mammography analysis

We used data from the Behavioral Risk Factor Surveillance System (BRFSS),16 which is a state-based health survey that collects information on preventive health practices, health behaviors, and health care access. State health departments, with assistance from the Centers for Disease Control and prevention, conduct cross-sectional random-digit-dial telephone surveys (conducted in English or Spanish) yearly of up to 1 adult per household contacted. The BRFSS Women's Health Module, which includes questions about mammography use, has been administered to women responders in even-numbered years since 2000. Response rates vary by state. In Massachusetts, the overall response rate was 35.9% in 2004, 38.6% in 2006, 48.2% in 2008, and 47.5% in 2010. In California, the overall response rate was 39.0% in 2004, 36.9% in 2006, 38.3% in 2008, and 42.7% in 2010. The response rate in other states ranged from 32.2% in New Jersey in 2004 to 66.5 in South Dakota in 2010.17

Stage at diagnosis analysis

We obtained data on all breast cancers diagnosed in Massachusetts and California from 2004 through 2008 from the Massachusetts Cancer Registry and the California Cancer Registry. Trained registry staff record patient demographics and stage at diagnosis for each incident cancer diagnosed in the state.

Women

Mammography analysis

We identified all women in Massachusetts and California ages 41 through 64 years who participated in BRFSS surveys during 2004, 2006, 2008, or 2010. We focused on women in these age groups who would be eligible for mammography (recommended for women ages 40-75 years) and who were likely to benefit from health insurance reform in Massachusetts (women aged <65 years, because most women aged ≥65 years were eligible for Medicare). We included 2004 survey data to have additional baseline data before the prereform 2006 survey data.

Stage at diagnosis analysis

We identified all women ages 41 through 64 years who were diagnosed with breast cancer during 2005 through 2008 (the most recent date for which data were available) in Massachusetts and California.

Variables for Mammography Analysis

Mammography

Women were asked, “Have you ever had a mammogram?” Women who responded yes were asked, “How long has it been since you had your last mammogram?” We characterized receipt of a mammogram in the last year for women who reported ever having a mammogram and that they had the mammogram within the last year. Because some guidelines recommend mammography for younger or average-risk women every 2 years, we also characterized receipt of a mammogram in the past 2 years.

Control variables

We characterized each woman's age, race/ethnicity, marital status, education, annual household income, employment, smoking status, and presence of diabetes, stroke, heart disease, and asthma based on patient report. Variables are categorized as in Table 1.

Table 1. Characteristics of the Mammography Cohorts, 2004 to 2010, Before and After Propensity Weighting
 Before Propensity Weighting, %After Propensity Weighting, %
CharacteristicMassachusetts Women (N = 16,847)California Women (N = 10,769)PaMassachusetts WomenCalifornia WomenPa
  • Abbreviations: SE, standard error.

  • a

    P values were calculated using weighted chi-square tests. All estimates apply the Behavioral Risk Factor Surveillance System survey weight. The postpropensity score weighting estimates also apply the propensity score weight, which was based on all variables listed in this table stratified by year.

Age: Mean±SE, y51.1±0.0851.3±0.09.0951.1±0.0851.1±0.09.99
Race  < .001  .20
 White89.379.3 89.388.3 
 African American3.76.5 3.73.8 
 Asian1.57.5 1.51.5 
 Other5.56.7 5.56.4 
Hispanic ethnicity  < .001  .87
 No94.367.8 94.394.3 
 Yes5.732.2 5.75.7 
Marital status  .005  .74
 Unmarried2932.4 2929.2 
 Married7167.6 7170.8 
Year of survey  .12  1.0
 200422.923.3 22.922.8 
 200624.125.2 24.124 
 20082525.4 2525 
 201027.926.1 27.928.1 
Education  < .001  .96
 High school only5.217.5 5.25.1 
 Some college46.245.2 46.246.1 
 College graduate48.336.3 48.348.6 
 Unknown0.21.1 0.20.2 
Annual household income, $US  < .001  .98
 <35,00019.334.7 19.319.4 
 35,000–75,0002724.7 2726.9 
 >75,00041.332.1 41.441.1 
 Unknown12.48.5 12.412.6 
Employed  < .001  .97
 No27.540.3 27.527.5 
 Yes72.559.7 72.572.5 
Comorbidities      
 Diabetes6.49.6< .0016.46.5.81
 Stroke1.21.8< .0011.21.2.89
 Heart disease1.11.2.411.11.1.81
 Asthma17.416.4.1517.417.6.74
Current smoker  < .001  .78
 No83.288 83.283 
 Yes16.812 16.817 
Variables for stage at diagnosis analysis

Stage at diagnosis was reported based on the American Joint Commission on Cancer Cancer Staging Manual, sixth edition. Registrars also documented patients' ages, race/ethnicity, marital status, place of birth, and diagnosis year. We linked patients' zip codes of residence to US Census data to obtain information about the proportion of individuals with a high school degree and the proportion living in poverty.

Analyses

Mammography

We compared mammography for women in Massachusetts with mammography for women in California who were most similar to the women in Massachusetts based on propensity score analysis. Comparisons assessed mammography rates over time in the 2 groups. Propensity score analyses involve comparing patients who have been matched, stratified, or weighted according to their propensity to be in a particular treatment group (in this analysis, the state of residence) to balance observed patient characteristics between treatment groups (as would occur in a randomized experiment) and obtain an estimate of the average treatment effect.18

To conduct the propensity score adjustment, we first used a logistic regression model to calculate the propensity of living in Massachusetts based on age, race (white, black, Asian, other), Hispanic ethnicity, marital status (married, unmarried/unknown), comorbidity (stroke, heart disease, asthma), employment (yes/no), current smoking (yes/no), education (less than high school, high school graduate/some college, college graduate), and income (<$35,000, $35,001-$75,000, >$75,000). Propensity score models were conducted for each survey year (2004, 2006, 2008, and 2010). We used the estimated regression coefficients and observed covariates to estimate the propensity for each woman to be living in Massachusetts (p). We used a standardized mortality ratio propensity score weight for each woman who had a score equal to 1 for Massachusetts women and p/(1 − p) (the propensity odds) for women in California.19, 20 This gives additional weight to California women who most resemble Massachusetts women, so that the weighted distribution of characteristics in the 2 cohorts is well balanced and equal to that of the original Massachusetts cohort. Standardized mortality ratio-weighted effects, thus, estimate the likelihood of mammography for a California resident if she were living in Massachusetts (eg, the treatment effect among the treated).

We compared the proportion of women reporting a mammogram in the last year over time according to state in the propensity-weighted cohort using logistic regression, testing the interaction of state by year and comparing mammography rates in 2008 and 2010 (postreform) with the rates in 2006 (prereform). We also used these models to calculate case mix-adjusted rates of mammography by year for each state. In sensitivity analyses, we assessed receipt of a mammogram in the past 2 years. In additional sensitivity analyses, we repeated analyses stratifying by race (white/nonwhite) and by income (<$35,000, $35-75,000, >$75,000). Finally, we repeated all analyses to compare changes in mammography in Massachusetts with the rest of the United States (rather than California only). We also repeated analyses including a variable for insurance in the propensity score model (insurance rates in 2004, 2006, 2008, and 2010 in Massachusetts were 92.8%, 93.5%, 96.7%, and 96.2%, respectively; in California, the respective rates were 85.6%, 84.8%, 86.1%, and 84.1%).

Stage at diagnosis

For the mammography analysis, we used propensity score adjustment to compare stage at diagnosis over time for women in Massachusetts and California. We first used a logistic regression model to calculate the propensity of living in Massachusetts based on age, race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, Asian, other), marital status (married, unmarried, unknown), birthplace (United States, foreign-born, unknown), the proportion of individuals in the zip code of residence with a high school degree, and the proportion of individuals in the zip code of residence living in poverty. We also included an indicator variable for the <3% of patients with unknown Census variables because of missing zip codes. Models were run for each diagnosis year (2005, 2006, 2007, and 2008). We used the estimated regression coefficients and observed covariates to estimate the propensity for each woman to be residing in Massachusetts (p). We used a standardized mortality ratio propensity score weight for each woman with a score equal to 1 for Massachusetts women and p/(1 − p) (the propensity odds) for women in California, as described above. We compared the proportion of women diagnosed with stage I cancers (vs stage II/III/IV cancers) by state in the propensity-weighted cohort using logistic regression, testing the interaction of state by year and comparing mammography rates in 2007 and 2008 (postreform) with the 2006 rates (prereform). We also used these models to calculate case mix-adjusted rates of stage I diagnosis by year in each state. In sensitivity analyses, we repeated analyses examining diagnosis with stage I/II disease versus stage III/IV disease and diagnosis with stage 0/I disease versus stage II/III/IV/unstaged disease.

Analyses were performed with SAS statistical software, version 9.2 (SAS Institute, Inc., Cary, NC) using weighted data to account for the complex survey design and to reflect the populations of the states studied, and standard errors were adjusted to account for the survey design. The study protocol was approved by the Harvard Medical School Committee on Human Studies, the Massachusetts Department of Public Health Institutional Review Board, and the California Committee for the Protection of Human Subjects.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Mammography

Characteristics of the study populations are included in Table 1. Before propensity score adjustment, women in Massachusetts were of similar age to women in California but were less likely to be of African American, Asian, or Hispanic race/ethnicity. Women in Massachusetts were more likely to be married, have a college education, have higher incomes, have health insurance, and smoke cigarettes. After propensity score weighting, characteristics of the populations were very similar and reflected characteristics of the Massachusetts cohort (Table 1). Over all study years, women in Massachusetts were more likely to have had a mammogram in the past year (69.7% vs 56.6%; P < .001) or in the past 2 years (82.6% vs 73.4%; P < .001).

When assessing changes in rates of mammography in the past year over time in Massachusetts versus California (using propensity-weighted data and additional adjustment with logistic regression), adjusted rates were highest in Massachusetts at the beginning of the study period and declined slowly over time (Fig. 1). In California, adjusted mammography rates increased slightly from 2004 to 2008 and then declined in 2010. Compared with 2006, the rates between Massachusetts and California did not differ in 2008 or 2010, suggesting no significant differences in trends in mammography use over time in the 2 states after health insurance reform in Massachusetts. The results were similar when we examined reports of mammography in the past 2 years (Fig. 1). Our findings also were similar when we examined the rest of the United States, rather than California alone, as a control population. When we stratified results by race/ethnicity and by income category, we identified no subgroups for which there was an association of mammography by state over time. All findings were similar when we added the insurance variable into the models (thus, insurance itself did not lead to more mammograms).2

thumbnail image

Figure 1. This chart illustrates the adjusted proportion of women reporting a mammogram in the last year over time in Massachusetts and California, 2004 to 2010. CI indicates confidence interval; BRFSS, Behavioral Risk Factor Surveillance System.

Download figure to PowerPoint

thumbnail image

Figure 2. This chart illustrates the adjusted proportion of women diagnosed with stage I (vs stage II, III, and IV) breast cancers over time in Massachusetts and California from 2005 to 2008.

Download figure to PowerPoint

Stage at Diagnosis

Among women who were diagnosed with breast cancer in Massachusetts and California, the women in Massachusetts were slightly younger, more often white, more often married, and more often born in the United States (Table 2). They also were more likely to live in areas with more high school graduates and fewer individuals living in poverty. After propensity score matching, these differences were attenuated substantially, although the matched cohort still differed slightly based on birthplace and the proportion of individuals living in poverty.

Table 2. Characteristics of Women Diagnosed With Breast Cancer in Massachusetts and California, 2005-2008, Before and After Propensity Weighting
 Before Propensity Weighting, %After Propensity Weighting, %
CharacteristicMassachusetts Women (N = 16,557)California Women (N = 67,938)PMassachusetts WomenCalifornia WomenPa
  • Abbreviations: SE, standard error.

  • a

    P values were calculated using unweighted (before) and weighted (after) chi-square tests for categorical variables and t tests for continuous variables. The estimates after-propensity score weighting apply the propensity score weight, which was based on all variables listed in Table 1 stratified by year of diagnosis.

Age Mean±SE, y52.0±0.0652.2±0.03.00552.0±0.0652.8±0.04.97
Race (%)  < .001  .61
 Non-Hispanic white87.960.9 87.987.5 
 Non-Hispanic black4.66.7 4.64.8 
 Hispanic417.8 44.1 
 Asian2.913.3 2.93 
 Other0.71.3 0.70.7 
Marital status at diagnosis  .03  .25
 Unmarried33.634.5 33.634.3 
 Married63.862.7 63.863.2 
 Unknown2.62.8 2.62.5 
Birthplace  < .001  .01
 United States41.133.1 41.140.2 
 Foreign-born9.318.5 9.310 
 Unknown49.648.4 49.649.8 
Diagnosis year  .69  1.00
 200523.824.2 23.823.8 
 200624.524.4 24.524.4 
 200725.425.3 25.425.4 
 200826.426.1 26.426.4 
Percentage in zip code with high school degree: Mean±SE86.5±079.8±0< .00186.5±0.0086.5±0.83
Percentage in zip code living in poverty: Mean±SE7.9±012±0< .0017.9±0.008.1±0< .001
Missing census information1.23< .0011.21.2.78

When assessing the proportion of women with breast cancer diagnosed with stage I disease over time (using propensity-weighted data and additional adjustment with logistic regression), adjusted rates in Massachusetts increased slightly in 2007 before declining slightly in 2008 (Fig 2). Adjusted rates of stage I cancers in California declined very slightly over the 4 year period. In models that assessed changes over time by state, there was no difference between 2007 (P = .37) or 2008 (P = .58) compared with 2006. Our findings were similar when we examined the proportion of stage I/II cancers versus III/IV cancers and the proportion of women diagnosed with stage 0/I cancers versus stage II/III/IV/unstaged cancers.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

We assessed self-reported mammography use and breast cancer stage at diagnosis over time in Massachusetts and California to assess whether health insurance expansion in Massachusetts was associated with increases in mammography use and earlier stage at diagnosis compared with a state in which health insurance expansion was considered but was not implemented. We observed higher rates of mammography use and earlier stage at breast cancer diagnosis in Massachusetts compared with California over all time periods. However, we did not observe statistically different trends by state to suggest that health insurance reform in Massachusetts was associated with increased mammography use or earlier stage at diagnosis.

Previous evidence suggests that women without insurance are less likely than other women to undergo mammograms,9, 10 and a recent study indicated that expansion of Medicaid coverage to uninsured individuals in Oregon was associated with an increase in preventive care services, including mammograms.15 In Oregon, uninsurance rates were 16.5% in 2008 when the Medicaid expansion occurred.21 However, our current study did not demonstrate a difference in mammography rates associated with health insurance expansion in Massachusetts compared with California, a state in which similar health insurance reform was proposed in 2007 but was not passed. There may be several reasons for this lack of effect. First, insurance rates already were quite high in Massachusetts before health insurance reform, with 94.6% of Massachusetts residents insured in 2004.2 Even if 25% of all uninsured Massachusetts residents in 2004 were women ages 41 to 64 years, and all of them had a mammogram after they obtained insurance, and use of mammography was unchanged for all other women, mammography rates would increase no more than 1.6 percentage points. In addition, before health insurance reform, low-income Massachusetts women without health insurance were eligible to obtain free breast and cervical cancer screening through the National Breast and Cervical Cancer Early Detection Program,22 and mammography rates in Massachusetts already were very high. Although mammography rates in the United States have declined since peaking in 2000,23 they have generally been higher in Massachusetts than in most other states.24

Second, access to health insurance does not automatically translate into receipt of mammography for eligible women; typically, women first must be seen by a primary care provider. One study indicated that, although cost-related barriers to care were reduced in Massachusetts after health insurance reform, there were no improvements in access to a personal physician,3 consistent with reports suggesting that health insurance reform in Massachusetts has been associated with a relative shortage of primary care providers25 or difficulty obtaining care because a provider was not accepting new patients.4, 5 Thus, it may take some time after a woman becomes insured before she receives a mammogram, because, typically, first, she would first need to establish primary care. Other recent data suggest that costs remain a barrier to care even among insured individuals.26 Third, despite gains in access to care after reform, some of the improvements observed in 2008 or 2009 were no longer evident in 2009 or 2010, possibly related to the economic climate and a decrease in employer-sponsored insurance because of unemployment.5, 6 Fourth, it is possible that California is not an ideal comparison state; however, the states are similar in other measures of health status, such as life expectancy at birth and obesity rates,27 and results of the mammography analysis were similar when we compared Massachusetts with the rest of the United States. Fifth, in most of the United States, mammography rates were declining in the early to middle 2000s.28 How underlying trends in mammography influenced our results is difficult to know, although the high rates in Massachusetts may have reflected a ceiling effect. In sensitivity analyses, our findings were similar when we used the rest of the United States as a control population. Finally, low response rates to the BRFSS survey and potential nonresponse bias may have limited our ability to detect a difference in mammography use.

Previous studies also have indicated that uninsured women are more likely to be diagnosed with breast cancers at more advanced stage.7, 11-13 Although 1 mechanism may be through increased use of mammography, insured women also may be more likely to go to the physician when they have symptoms and receive timely evaluation of their symptoms. They also may be more likely to follow through with additional diagnostic testing. We did not observe an association between Massachusetts insurance reform and breast cancer stage at diagnosis, possibly for many of the same reasons discussed above—notably, the very low uninsurance rate in Massachusetts before health reform, other safety net programs for breast cancer early detection, high baseline rates of mammography in Massachusetts, and potential challenges to accessing primary care.

Despite the lack of associations between health insurance expansion and increases in mammography and earlier stage at breast cancer diagnosis in Massachusetts compared with California, expanded access to care through insurance expansions probably will have important implications in other areas. It will be important to assess access to primary care providers and preventive care as health insurance expansion occurs throughout the United States, particularly for states in which the rates of uninsurance are high. Although recent guidelines recommend more discussion of the benefits and risks of mammography screening and informed patient decisions to screen, the benefits of mammography at least every 2 years for women ages 50 to 75 years are well accepted.29, 30 Understanding better how health insurance expansion influences mammography use and stage of breast cancer diagnosis in other populations is important and has implications for improving the prognosis for women who develop cancer.

Our findings should be interpreted in light of several limitations. First, we relied on self-reported information on the timing of mammography, which reportedly has moderate reliability and validity,31 although we have no reason to believe that reporting of mammography would vary by state. In addition, only patients with household telephones were contacted, and the response rates were relatively low and varied by state and year, although they generally increased over the study period. Next, our use of propensity score methods identified cohorts of patients who were similar based on observed characteristics; however, we cannot be certain whether unobserved characteristics may have influenced our results. Finally, we may have been underpowered to detect small differences; for example, with the number of respondents to the BRFSS surveys in each year, we estimate that we had sufficient power to detect a difference in the differences between rates for Massachusetts and California of approximately 5 percentage points.

In summary, the current investigation does not provide evidence that health insurance expansion in Massachusetts was associated with increases in mammography or earlier stage at breast cancer diagnosis. This may be partly because of already high insurance rates and mammography rates in Massachusetts and potential barriers to accessing primary care among newly insured women. Additional research is needed to assess the impact of insurance expansions in other populations, especially those with higher uninsurance rates.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

We thank Lawrence Zaborski for expert programming assistance.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

This work was supported by Research Scholar Grant RSGI-09-269-01-CPHPS from the American Cancer Society.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES
  • 1
    Mass Session Law §58. An Act Providing Access to Affordable, Quality, Accountable Health Care; 2006.
  • 2
    Division of Health Care Finance and Policy. Health care in Massachusetts: key indicators. May 2011 edition, released September 2011. Available at: http://www.mass.gov/eohhs/docs/dhcfp/r/pubs/11/2011-key-indicators-may.pdf. Accessed March 28, 2012.
  • 3
    Zhu J, Brawarsky P, Lipsitz S, Huskamp H, Haas JS. Massachusetts health reform and disparities in coverage, access and health status. J Gen Intern Med. 2010; 25: 1356-1362.
  • 4
    Long SK, Masi PB. Access and affordability: an update on health reform in Massachusetts, fall 2008. Health Aff (Millwood). 2009; 28: w578-w587.
  • 5
    Long SK, Stockley K. Sustaining health reform in a recession: an update on Massachusetts as of fall 2009. Health Aff (Millwood). 2010; 29: 1234-1241.
  • 6
    Long SK, Stockley K, Dahlen H. Massachusetts health reforms: uninsurance remains low, self-reported health status improves as state prepares to tackle costs. Health Aff (Millwood). 2012; 31: 444-451.
  • 7
    Ayanian JZ, Kohler BA, Abe T, Epstein AM. The relation between health insurance coverage and clinical outcomes among women with breast cancer. N Engl J Med. 1993; 329: 326-331.
  • 8
    Ayanian JZ, Weissman JS, Schneider EC, Ginsburg JA, Zaslavsky AM. Unmet health needs of uninsured adults in the United States. JAMA. 2000; 284: 2061-2069.
  • 9
    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.
  • 10
    Schootman M, Walker MS, Jeffe DB, Rohrer JE, Baker EA. Breast cancer screening and incidence in communities with a high proportion of uninsured. Am J Prev Med. 2007; 33: 379-386.
  • 11
    Halpern MT, Bian J, Ward EM, Schrag NM, Chen AY. Insurance status and stage of cancer at diagnosis among women with breast cancer. Cancer. 2007; 110: 403-411.
  • 12
    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.
  • 13
    Ward E, Halpern M, Schrag N, et al. Association of insurance with cancer care utilization and outcomes. CA Cancer J Clin. 2008; 58: 9-31.
  • 14
    Ward EM, Fedewa SA, Cokkinides V, Virgo K. The association of insurance and stage at diagnosis among patients aged 55 to 74 years in the National Cancer Database. Cancer J. 2010; 16: 614-621.
  • 15
    Finkelstein A, Taubman S, Wright B, et al. The Oregon health insurance experiment: evidence from the first year. NBER Working Paper Series. Available at: http://www.nber.org/papers/w17190.pdf?new_window=1. Accessed July 8, 2011.
  • 16
    Centers for Disease Control and Prevention (CDC). Behavioral Risk Factor Surveillance System Survey Data. Atlanta, GA: US Department of Health and Human Services, CDC; 2010.
  • 17
    Behavioral Risk Factor Surveillance System (BRFSS). BRFSS Annual Survey Data. Summary Data Quality Reports. Available at: http://www.cdc.gov/brfss/technical_infodata/quality.htm. Accessed April 3, 2012.
  • 18
    Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997; 127: 757-763.
  • 19
    Kurth T, Walker AM, Glynn RJ, et al. Results of multivariable logistic regression, propensity matching, propensity adjustment, and propensity-based weighting under conditions of nonuniform effect. Am J Epidemiol. 2006; 163: 262-270.
  • 20
    Sato T, Matsuyama Y. Marginal structural models as a tool for standardization. Epidemiology. 2003; 14: 680-686.
  • 21
    Office for Oregon Health Policy and Research, Oregon Health Authority. Oregon's Uninsured: Analysis of the 2008 American Community Survey. Salem, OR: Oregon Health Authority; 2010.
  • 22
    Liu MJ, Hawk H, Gershman ST, et al. The effects of a National Breast and Cervical Cancer Early Detection Program on social disparities in breast cancer diagnosis and treatment in Massachusetts. Cancer Causes Control. 2005; 16: 27-33.
  • 23
    Breen N, Gentleman JF, Schiller JS. Update on mammography trends: comparisons of rates in 2000, 2005, and 2008. Cancer. 2011; 117: 2209-2218.
  • 24
    Centers for Disease Control and Prevention (CDC). Vital signs: breast cancer screening among women aged 50–74 years—United States, 2008. MMWR Morb Mortal Wkly Rep. 2010; 59: 6-9.
  • 25
    Brown K. Massachusetts health insurance reform reveals doctor shortage. National Public Radio; November 30, 2008. Available at: http://www.npr.org/templates/story/story.php?storyId=97620520. Accessed March 2, 2012.
  • 26
    Radley DC, Schoen C. Geographic variation in access to care—the relationship with quality. N Engl J Med. 2012; 367: 3-6.
  • 27
    Kaiser Family Foundation. State Health Facts—Health Status. Available at: http://www.statehealthfacts.org/comparecat.jsp?cat=2&rgn=6&rgn= 23. 2012. Accessed June 14, 2012.
  • 28
    Miller JW, King JB, Ryerson AB, Eheman CR, White MC. Mammography use from 2000 to 2006: state-level trends with corresponding breast cancer incidence rates. AJR Am J Roentgenol. 2009; 192: 352-360.
  • 29
    Nelson HD, Tyne K, Naik A, et al. Screening for breast cancer: an update for the US Preventive Services Task Force. Ann Intern Med. 2009; 151: 727-737, W237-W242.
  • 30
    US Preventive Services Task Force. Screening for breast cancer: US Preventive Services Task Force recommendation statement. Ann Intern Med. 2009; 151: 716-726, W-236.
  • 31
    Nelson DE, Holtzman D, Bolen J, Stanwyck CA, Mack KA. Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS). Soz Praventivmed. 2001; 46(suppl 1): S3-S42.