In the United States, there are approximately 22,000 new diagnoses of ovarian cancer and 14,000 disease-related deaths annually. Although accounting for only 3–5% of all cancers in women, ovarian cancer is the leading cause of death from gynecologic malignancy and the fifth most common cause of cancer death in women.1 In the absence of a screening test for this malignancy, the majority of women are diagnosed at late stages and succumb to the disease within 5 years of diagnosis.2 Early diagnosis and adherence to national treatment guidelines (eg, timely receipt of surgery and chemotherapy) are independent predictors of improved disease-specific survival, irrespective of race–ethnicity and socioeconomic status. However, a substantial proportion of women in the United States may experience a delay in diagnosis and treatment or never receive the recommended treatment at all.3 Thus, interventions that may improve early diagnosis and access to life-saving cancer therapies are critical to improving outcomes in ovarian cancer.
Subtle symptoms, such as abdominal pain and bloating, are often present for months to years before diagnosis.4,5 Initial presentation to primary care with such nonspecific symptoms eventually leads to a workup and detection.6 As with other health conditions, lack of insurance or underinsurance may delay care-seeking and ovarian cancer diagnosis.7–9
The 2010 Affordable Care Act (ACA) expanded insurance access in the United States and led to 20 million Americans gaining insurance.10 Multiple studies have found increased access to a primary care doctor and usual source of care associated with the insurance gains under the ACA.10 In addition to the ACA's provisions expanding insurance coverage—the dependent coverage mandate, Medicaid expansion, health insurance marketplaces—multiple components of the ACA may have improved access to cancer care, including treatment of gynecologic malignancy.11,12 For example, the ACA's essential health benefits reduced cost-sharing for primary care visits, and removal of pre-existing condition exclusions allowed more individuals to purchase insurance. Several studies have suggested improvements in screening and early-stage diagnosis of nongynecologic cancers under the ACA.13–16
Our objective was to estimate how implementation of the ACA might be associated with stage at diagnosis and receipt of timely treatment in women with ovarian cancer. We also examined whether the ACA was associated with racial, insurance, and other socioeconomic disparities in ovarian cancer care.
This retrospective cohort study was granted exempt status by the Johns Hopkins institutional review board, because it was determined not to be human subjects research.
We used a difference-in-differences study design to compare trends in diagnoses and treatment of women with ovarian cancer before and after the ACA. The intervention group was women aged 21–64 years old, that is, women affected by the ACA. The comparison group was women aged 65 years and older, that is, women eligible for Medicare throughout the study period and unlikely to be affected by the ACA. Our study period included 6 years pre-ACA (2004–2009) and 5 years post-ACA (2011–2015).
The difference-in-differences approach compares trends over time in the intervention and comparison. It approximates a randomized controlled trial by controls for background trends (ie, the Great Recession) and baseline differences between the intervention and comparison groups.17 There were no major changes in federal or state health policy other than the ACA during the study period.
We used hospital-reported data from the National Cancer Database, a program of the American Cancer Society and the American College of Surgeons' Commission on Cancer. The National Cancer database is a hospital-based registry that collects standardized patient data from more than 1,500 cancer programs. It includes more than 70% of new cancer diagnoses in the United States.18
We identified women with new diagnoses of ovarian cancer during the study period. We included all histologies. We excluded women who did not have a tissue-based diagnosis or who had a delay of more than 365 days from cancer diagnosis to treatment. We compared outcomes prereform (2004–2009) and postreform (2011–2015). We then analyzed outcomes at two separate periods postreform: 1) early post-ACA (2011–2013), during which the dependent coverage mandate and expanded coverage of preventive care for private and public insurance were implemented; and 2) late post-ACA (2014–2015), after which Medicaid expansion, health insurance marketplaces, coverage for pre-existing conditions, and the individual mandate were implemented.10
We classified a woman as insured if she was recorded as having either public or private insurance at the time of initial diagnosis or treatment. We classified a woman as privately insured if her primary insurance was recorded as “private insurance or managed care.” We defined a woman as publicly-insured if her primary insurance was Medicaid, Medicare, or another government plan. We excluded cases missing insurance status from analyses, which were less than 3% of cases in each survey year.
Using the American Journal on Cancer Commission staging and pathologic data in the National Cancer Database, we defined early-stage diagnosis as stage I–II and late-stage diagnosis as stage III–IV. There were no changes in ovarian cancer staging during the study period.
We chose receipt of treatment within 30 days of diagnosis as an outcome measure because of prior studies showing improvements in survival with treatment initiated within this time interval.7,19,20 We excluded cases with treatment delays of more than 365 days.
We analyzed outcomes for women overall and then by type of insurance (public, private), race (white, nonwhite), and income (low-income, not low-income). We adjusted analyses for socio-demographic, clinical, and hospital-level factors. Patient factors included were race (white, black, other), insurance type (privately-insured, publicly-insured, or uninsured), living in a rural area (metropolitan area greater than, equal to, or less than 50,000 people), area-level household income, and area-level education attainment. In the National Cancer Database, household income is defined as the median income in the patient's zip code at the time of diagnosis from the 2008–2012 American Community surveys and divided into quartiles (less than $38,000, $38,000–47,999, $48,000–62,999, $63,000 or more). We defined low-income as the lowest quartile of income (less than $38,000). Education attainment is defined by proportion of individuals who did not graduate high school in the patient's zip code at the time for diagnosis from the 2008–2012 American Community surveys and divided into quartiles (21% or more, 13–20.9%, 7–12.9%, less than 7%). Although the National Cancer Database lacks individual-level data on income and education, the zip code-based variables appear to be highly correlated with patient income and education levels.21,22
Clinical factors were distance traveled for treatment and patient comorbidities. We divided distanced traveled for treatment into quartiles: 0–4.9 miles, 5–11.4 miles, 11.5–29.6, and 29.7 or more miles. We used the Charlson Comorbidity Index score to classify patients with no, one, or two or more comorbidities. Hospital factors included were treatment at an academic institution and census region. We excluded cases missing any of the above patient, clinical, or hospital factors from our analysis (n=19,212 cases).
We assessed for baseline differences in demographic characteristics between the intervention and comparison groups using χ2 tests. To assess pre–post trends, we subtracted the proportion with each outcome pre-ACA from the proportion post-ACA for the intervention and comparison groups. We used a multivariate linear regression model adjusted for patient, clinical, and hospital covariates to examine significance of pre–post trends. We used a multivariate linear regression model adjusted for patient, clinical, and hospital covariates to estimate difference-in-differences. A positive difference-in-differences suggests that the ACA improved the outcome of interest; this may occur when the trend over time in the intervention and control group is positive–positive, negative–negative, positive–negative, or no change–negative because of greater differences over time in the intervention compared with the comparison group. A difference-in-differences of zero suggests that there was no relationship between the ACA and the outcome of interest.
We conducted three sensitivity analyses. First, we looked at outcomes among women with serous epithelial ovarian cancer only. Second, we compared differences-in-differences for 2011–2013 (early post-ACA) to 2014–2015 (late post-ACA). Third, we added age clustering to our difference-in-differences model.
We considered P=.05 to be significant. We had 80% power to detect pre–post differences of 1% at P=.05. Analyses were conducted with Stata 11.
A total of 39,999 ovarian cancer cases prereform and 36,564 postreform were identified for women aged 21–64 years. There were 31,290 cases prereform and 29,807 postreform for women aged 65 years and older. Women aged 21–64 years were more likely to be uninsured and have higher income and education levels pre-ACA than women aged 65 and older (Table 1).
Uninsurance did not change significantly for women aged 21–64 years post-ACA (P for trend=.48) or for the comparison group of women aged 65 and older post-ACA (P for trend=.08). In the overall difference-in-differences model, the ACA was associated with a nonsignificant decrease in uninsurance (difference-in-differences 0.1%, 95% CI −0.3 to 0.5) (Table 2). In the early post-ACA model, the ACA was associated with a significant increase in uninsurance as insurance increased in the Medicare-eligible population (difference-in-differences −1.3%, 95% CI −0.1 to 1.8). In the late post-ACA model, the ACA was associated with a significant decrease in uninsurance as insurance rates increased in women aged 21–64 years old and stayed the same in women aged 65 and older (difference-in-differences −2.2%, 95% CI −2.7 to −1.6).
Early-stage diagnosis increased for women aged 21–64 years post-ACA (P for trend <.001) and for the comparison group of women aged 65 and older post-ACA (P for trend <.001) (Fig. 1). In the overall difference-in-differences model, the ACA was associated with a significant greater increase in early-stage diagnosis for women aged 21–64 compared with women aged 65 years and older (difference-in-differences 1.4%, 95% CI 0.4–2.4). In the early post-ACA model, the ACA was associated with a significant increase in early-stage diagnosis (difference-in-differences 1.7%, 95% CI 0.6–2.9). In the late post-ACA model, the ACA was associated with a nonsignificant increase in early-stage diagnosis (difference-in-differences 0.9%, 95% CI −0.4 to 2.3).
Receipt of treatment within 30 days of diagnosis decreased slightly for women aged 21–64 years post-ACA (92.8–91.1%, P for trend <.001) and decreased for the comparison group of women aged 65 and older post-ACA (88.4–84.5%, P for trend <.001) (Fig. 2). In the overall difference-in-differences model, the ACA was associated with a significant increase in receipt of treatment within 30 days of diagnosis for women aged 21–64 compared with women aged 65 years and older (difference-in-differences 2.3%, 95% CI 1.7–3.0). The ACA was associated with a significant increase in receipt of treatment within 30 days of diagnosis in both the early and late post-ACA models (difference-in-differences 2.0%, 95% CI 1.3–2.7 and difference-in-differences 2.9%, 95% CI 2.0–3.7).
Among publicly-insured women, the ACA was associated with a significant increase in early-stage diagnosis (difference-in-differences 2.7%, 95% CI 1.0–4.5) and in receipt of treatment within 30 days of diagnosis (difference-in-differences 2.5%, 95% CI 1.2–3.8) (Table 3). For privately insured women, the ACA was associated with a nonsignificant increase in early-stage diagnosis (difference-in-differences 1.1%, 95% CI −1.3 to 3.6) and a significant increase in receipt of treatment within 30 days of diagnosis (difference-in-differences 1.4%, 95% CI 0.2–2.7).
For nonwhite women, the ACA was associated with a significant increase in receipt of treatment within 30 days of diagnosis (difference-in-differences 4.2%, 95% CI 2.0–6.3) (Table 4). The ACA was not associated with a significant change in early-stage diagnosis for nonwhite women (difference-in-differences 2.5%, 95% CI −0.1 to 5.6). In the late post-ACA period, the ACA was associated with a significant decrease in uninsurance for nonwhite women (difference-in-differences −3.4, 95% CI −1.5 to −5.3). For white women, the ACA was associated with significant increases in early-stage diagnosis (difference-in-differences 1.3%, 95% CI 0.2–2.3) and receipt of treatment within 30 days of diagnosis (difference-in-differences 2.1, 95% CI 1.4–2.8). For white women, the ACA was associated with significant decrease in uninsurance in the late post-ACA period (difference-in-differences −2.0, 95% CI −1.5 to −2.5).
For low-income women, the ACA was associated with a significant increase in receipt of treatment within 30 days of diagnosis (difference-in-differences 3.3, 95% CI 1.6–5.0) (Table 5). For higher-income women, the ACA was associated with significant increases in early-stage at diagnosis and treatment within 30 days of diagnosis (difference-in-differences 1.5, 95% CI 0.4–2.6 and 2.2, 95% CI 1.5–2.9).
For women with serous epithelial ovarian cancer, the ACA was associated with significant increase in treatment within 30 days of diagnosis, including for nonwhite and low-income women (Appendix 1, available online at http://links.lww.com/AOG/B674). In sensitivity analyses, similar increases in early-stage diagnosis and receipt of treatment within 30 days of diagnosis were seen when adjusted for age clustering (Appendices 2 and 3, available online at http://links.lww.com/AOG/B674). Appendix 4 (available online at http://links.lww.com/AOG/B674) shows our full regression model, and Appendix 5 (available online at http://links.lww.com/AOG/B674) shows women with ovarian cancer per year.
Women with ovarian cancer were more likely to be diagnosed at an early-stage and receive treatment within 30 days of diagnosis after implementation of the ACA. We estimate that after the ACA, approximately 100 more women aged 21–64 were diagnosed at early-stage and 70 more received timely treatment each year. Publicly-insured women experienced even larger gains in early-stage diagnosis and timely receipt of treatment. Nonwhite and low-income women also experienced improvements in insurance rates and timely treatment, especially after Medicaid expansion.
One of the study's strengths is the analysis at multiple time points corresponding to the implementation of different ACA provisions. Other studies found increases of 0.4–2.0% in early-stage diagnosis largely attributed to the ACA's Medicaid expansion.14,16 We found improvements in early-stage diagnosis and timely treatment early post-ACA period (2011–2013); the only income-based insurance expansion during this time was early Medicaid expansion in six states. During this time, the ACA's dependent coverage mandate and expanded coverage of preventive health services were implemented (eg, elimination of cost-sharing for wellness visits for privately-insured women and increased coverage of wellness visits and physician reimbursement for publicly insured women). Our previous research on young women with gynecologic cancer similarly found improvements in early-stage diagnosis under these provisions of the ACA.11 Lower cost-sharing may prompt women to have a lower threshold to seek care for the subtle symptoms of ovarian cancer. Additionally, the ACA's essential benefits provisions removed cost-sharing for genetic testing, expanding access to ovarian cancer susceptibility testing (eg, BRCA 1-2), which may have led to more women at high risk of ovarian cancer seeking care.23
Greater access to health care providers is associated with earlier stage of diagnosis and better survival outcomes across multiple cancer subtypes.24 We found improvements in insurance and further gains in receipt of timely treatment in the late post-ACA period (2014–2015), especially for publicly insured and low-income women. By the end of 2014, Medicaid expansion occurred in 26 states and the District of Columbia, and health insurance marketplaces were available nationwide.12,25 We also saw improvements in insurance rates and timely treatment for racial–ethnic minorities under the ACA, although disparities remain.
Our study has limitations, including those inherent to a difference-in-differences study design. Histology of ovarian cancer differs by age with high-grade serous ovarian cancer more common in older women and low-grade serous ovarian cancer more common in younger women. Nonetheless, we found significant improvement in time to treatment and insurance status in our subgroup analysis of women with serous epithelial ovarian cancer. The ACA introduced free annual wellness examinations for Medicare recipients in 2011, which may have biased results toward the null, given the potential for increased access to care for the control group with this provision.26 The National Cancer Database does not include state-level data, so we were unable to assess the direct effects of Medicaid expansion, including early expansion in several large states (California, Colorado, Connecticut, Minnesota, New Jersey, and the District of Columbia). Although the National Cancer Database is the largest cancer database in the United States and includes more than 70% of new cancer diagnoses, it excludes cancer programs not accredited by the Commission on Cancer. Most Marketplace plans implemented under the ACA (95%) include one or more Commission on Cancer accredited hospitals, but some women with marketplace plans—and thus insurance gains related to the ACA—may not be in our dataset.27 We plan on looking at long-term outcomes, including survival, once further follow-up data are available.
As earlier stage at diagnosis and prompt treatment are major determinants of survival, these gains under the ACA may result in long-term benefits for women with ovarian cancer. Scaling back components of the ACA could jeopardize these advances in women's health.
1. Torre LA, Trabert B, DeSantis CE, Miller KD, Samimi G, Runowicz CD, et al. Ovarian cancer statistics, 2018. CA Cancer J Clin 2018;68:284–96.
2. Sopik V, Iqbal J, Rosen B, Narod SA. Why have ovarian cancer mortality rates declined? Part I. Incidence. Gynecol Oncol 2015;138:741–9.
3. Bristow RE, Chang J, Ziogas A, Campos B, Chavez LR, Anton-Culver H. Sociodemographic disparities in advanced ovarian cancer survival and adherence to treatment guidelines. Obstet Gynecol 2015;125:833–42.
4. Goff BA, Mandel L, Muntz HG, Melancon CH. Ovarian carcinoma diagnosis. Cancer 2000;89:2068–75.
5. Goff BA, Mandel LS, Melancon CH, Muntz HG. Frequency of symptoms of ovarian cancer in women presenting to primary care clinics. JAMA 2004;291:2705–12.
6. Devlin SM, Diehr PH, Andersen MR, Goff BA, Tyree PT, Lafferty WE. Identification of ovarian cancer symptoms in health insurance claims data. J Womens Heal 2010;19:381–9.
7. Seagle BLL, Butler SK, Strohl AE, Nieves-Neira W, Shahabi S. Chemotherapy delay after primary debulking surgery for ovarian cancer. Gynecol Oncol 2017;144:260–5.
8. Hinchcliff E, Melamed A, Bregar A, Diver E, Clemmer J, Del Carmen M, et al. Factors associated with delivery of neoadjuvant chemotherapy in women with advanced stage ovarian cancer. Gynecol Oncol 2018;148:168–73.
9. Bercow AS, Chen L, Chatterjee S, Tergas AI, Hou JY, Burke WM, et al. Cost of care for the initial management of ovarian cancer. Obstet Gynecol 2017;130:1269–75.
10. Kominski GF, Nonzee NJ, Sorensen A. The Affordable Care Act's impacts on access to insurance and health care for low-income populations. Annu Rev Public Health 2017;38:489–505.
11. Smith AJB, Fader AN. Effects of the Affordable Care Act on young women with gynecologic cancers. Obstet Gynecol 2018;131:966–76.
12. Frean M, Gruber J, Sommers BD. Premium subsidies, the mandate, and Medicaid expansion: coverage effects of the Affordable Care Act. J Health Econ 2017;53:72–86.
13. White-Means S, Osmani A. Affordable Care Act and disparities in health services utilization among ethnic minority breast cancer survivors: evidence from longitudinal medical expenditure panel surveys 2008–2015. Int J Environ Res Public Health 2018;15:1860.
14. Han X, Yabroff KR, Ward E, Brawley OW, Jemal A. Comparison of insurance status and diagnosis stage among patients with newly diagnosed cancer before vs after implementation of the patient protection and Affordable Care Act. JAMA Oncol2018;4:1713–20.
15. Eguia E, Cobb AN, Kothari AN, Molefe A, Afshar M, Aranha GV, et al. Impact of the Affordable Care Act (ACA) Medicaid expansion on cancer admissions and surgeries. Ann Surg 2018;268:584–90.
16. Jemal A, Lin CC, Davidoff AJ, Han X. Changes in insurance coverage and stage at diagnosis among nonelderly patients with cancer after the Affordable Care Act. J Clin Oncol 2017;35:3906–15.
17. Dimick JB, Ryan AM. Methods for evaluating changes in health care policy: the difference-in-differences approach. JAMA 2014;312:2401–2.
18. Boffa DJ, Rosen JE, Mallin K, Loomis A, Gay G, Palis B, et al. Using the National Cancer Database for outcomes research. JAMA Oncol 2017;3:1722–8.
19. Tewari KS, Java JJ, Eskander RN, Monk BJ, Burger RA. Early initiation of chemotherapy following complete resection of advanced ovarian cancer associated with improved survival: NRG Oncology/Gynecologic Oncology Group study. Ann Oncol 2016;27:114–21.
20. Hofstetter G, Concin N, Braicu I, Chekerov R, Sehouli J, Cadron I, et al. The time interval from surgery to start of chemotherapy significantly impacts prognosis in patients with advanced serous ovarian carcinoma—analysis of patient data in the prospective OVCAD study. Gynecol Oncol 2013;131:15–20.
21. Narla NP, Pardo-Crespo MR, Beebe TJ, Sloan J, Yawn B, Williams AR, et al. Concordance between individual vs. area-level socioeconomic measures in an urban setting. J Health Care Poor Underserved 2015;26:1157–72.
22. Pardo-Crespo MR, Narla NP, Williams AR, Beebe TJ, Sloan J, Yawn BP, et al. Comparison of individual-level versus area-level socioeconomic measures in assessing health outcomes of children in Olmsted County, Minnesota. J Epidemiol Community Health 2013;67:305–10.
23. Han X, Jemal A. Recent patterns in genetic testing for breast and ovarian cancer risk in the U.S. Am J Prev Med 2017;53:504–7.
24. Ananthakrishnan AN, Hoffmann RG, Saeian K. Higher physician density is associated with lower incidence of late-stage colorectal cancer. J Gen Intern Med 2010;25:1164–71.
25. Soni A, Sabik LM, Simon K, Sommers BD. Changes in insurance coverage among cancer patients under the Affordable Care Act. JAMA Oncol 2018;4:122–4.
26. Chung S, Lesser LI, Lauderdale DS, Johns NE, Palaniappan LP, Luft HS. Medicare annual preventive care visits: use increased among fee-for-service patients, but many do not participate. Health Aff 2015;34:11–20.
27. Kehl KL, Liao KP, Krause TM, Giordano SH. Access to accredited cancer hospitals within federal exchange plans under the Affordable Care Act. J Clin Oncol 2017;35:645–51.