Abstract
Background
Medicaid expansion under the Affordable Care Act extends eligibility for participating states and has been associated with improved outcomes by facilitating access to care. Delayed initiation of adjuvant chemotherapy is associated with worse outcomes among patients with early-stage breast cancer (BC). The impact of Medicaid expansion in narrowing delays by race and ethnicity has not been studied, to our knowledge.
Methods
This was a population-based study using the National Cancer Database. Patients diagnosed with primary early-stage BC between 2007 and 2017 residing in states that underwent Medicaid expansion in January 2014 were included. Time to chemotherapy initiation and proportion of patients experiencing chemotherapy delays (>60 days) were evaluated using difference-in-difference and Cox proportional hazards models in preexpansion and postexpansion periods according to race and ethnicity.
Results
A total 100 643 patients were included (63 313 preexpansion and 37 330 postexpansion). After Medicaid expansion, the proportion of patients experiencing chemotherapy initiation delay decreased from 23.4% to 19.4%. The absolute decrease was 3.2, 5.3, 6.4, and 4.8 percentage points (ppt) for Black, Hispanic, White, and Other patients. Compared with White patients, statistically significant adjusted difference-in-differences were observed for Black (−2.1 ppt, 95% confidence interval [CI] = −3.7% to −0.5%) and Hispanic patients (−3.2 ppt, 95% CI = −5.6% to −0.9%). Statistically significant reductions in time to chemotherapy between expansion periods were observed among White patients (adjusted hazard ratio = .11, 95% CI = 1.09 to 1.12) and those belonging to racialized groups (adjusted hazard ratio = 1.14, 95% CI = 1.11 to 1.17).
Conclusions
Among patients with early-stage BC, Medicaid expansion was associated with a reduction in racial disparities by decreasing the gap in the proportion of Black and Hispanic patients experiencing delays in adjuvant chemotherapy initiation.
The Patient Protection and Affordable Care Act (ACA) extended Medicaid eligibility to all individuals with income equal to or less than 138% of the Federal Poverty Level. Medicaid was created to increase access to care for low-income individuals in response to the lack of universal health care in the United States. Since its implementation, Medicaid expansion has been associated with a decreased percentage of uninsured patients, increased access, and improved outcomes (1-5).
Despite improvements in cancer survival, disparities in outcomes persist among racialized groups and among patients of low socioeconomic status. Delays in adjuvant chemotherapy initiation are associated with worse survival outcomes among patients with early-stage breast cancer (BC) (6-9). Due to discriminatory policies and practices that limit access to health insurance coverage and access to care, individuals from communities targeted for marginalization, such as Black and Hispanic patients, are more likely to experience delays in receipt of chemotherapy (10). Limited access to health care and lack of insurance coverage are also known to affect staging and overall survival (11-13). These observations emphasize the importance of timely chemotherapy initiation as a strategy to improve patient outcomes, especially for underserved populations and communities targeted for marginalization. The ACA expansion of Medicaid eligibility can potentially address cancer disparities. A study evaluating the association of Medicaid expansion with timely treatment among patients with metastatic tumors identified in the Flatiron Health database reported a statistically significant decrease in disparities, demonstrated by a greater benefit among Black patients compared with White patients (14).
In this study, we examined the association between Medicaid expansion and adjuvant chemotherapy initiation delays among patients with early-stage BC according to race and ethnicity. We hypothesize that Medicaid expansion decreased the proportion of patients experiencing chemotherapy delays, with a greater reduction among patients belonging to racialized groups, thereby narrowing disparities.
Methods
Data were obtained from the National Cancer Database (NCDB), a database sourced from hospital registries jointly sponsored by the American Cancer Society and the American College of Surgeons. The NCDB collects data from more than 1500 Commission on Cancer (CoC)-accredited facilities, representing approximately 70% of newly diagnosed cancer cases nationwide across more than 34 million historical records (15). The CoC’s NCDB and the hospitals participating in the CoC NCDB are the source of the deidentified data used herein.
We conducted a retrospective cross-sectional study comparing time to adjuvant chemotherapy initiation among patients residing in states that underwent Medicaid expansion in January 2014 (KY, NV, CO, OR, NM, WV, AR, RI, AZ, MD, MA, ND, OH, IA, IL, VT, HI, NY, DE). These states were selected based on a uniform date of Medicaid expansion. We included women diagnosed with their first primary invasive BC (International Classification of Disease for Oncology [0-3] codes C50.0-C.50.9) between 2007 and 2017. Patient age was limited to 40-64 years because NCDB suppresses information for patients aged 39 years or younger, and those aged 65 years and older are eligible for Medicare benefits. Patients with early-stage I-III BC who had surgery and adjuvant chemotherapy (within 6 months after surgery) were included. We excluded those who received neoadjuvant chemotherapy and those who died within 90 days after surgery. The final cohort included 100 643 patients (Supplementary Table 1, available online).
The primary exposure variable was race and ethnicity, as reported in the NCDB. Patients were grouped as Hispanic (any race), Non-Hispanic Black (Black), and Non-Hispanic White (White). Patients identified as Asian American or Pacific Islander and American Indian or Alaska Native were evaluated separately and grouped together with those of unknown race and ethnicity labeled as Non-Hispanic Other (Other). The primary intervention was the implementation of Medicaid expansion in January 2014. The preexpansion period was defined as 2007-2013 and the postexpansion period as 2014-2016. Our primary outcome was adjuvant chemotherapy initiation delay, defined as more than 60 days from surgery and the first dose of chemotherapy (6,9).
Variables were compared between the preexpansion or postexpansion periods via χ2 tests. Patient sociodemographic and clinical characteristics evaluated included age; Charlson-Deyo comorbidity index (16-18); pathological stage; BC subtype, including hormone receptor-positive (estrogen receptor [ER] or progesterone receptor [PR]-positive and HER2 receptor–negative or unknown), HER2-positive (regardless of ER and PR), and triple-negative (ER, PR, and HER2-negative); surgery (breast-conserving surgery, mastectomy); primary insurance status (private, Medicaid, Medicare, other governmental insurance, and uninsured); hospital transfer (yes/no); and distance to facility. Area-level variables included geographic region, residential area (metropolitan, rural, urban, unknown), ZIP code–based education status measured by percentiles of adults without a high school diploma, and median household income. Facility-level variables included facility type (community cancer program, comprehensive community cancer program, academic, integrated network) and facility BC volume tertile. Covariates with missing values were grouped as a separate “unknown” category.
Our primary analysis compared changes in the proportions of patients experiencing chemotherapy initiation delays between the preexpansion and postexpansion periods according to race and ethnicity. The primary independent variable included preexpansion and postexpansion, race and ethnicity, and an interaction between the 2 variables. We calculated difference-in-difference (DID) estimates using multivariable linear regression models. Delayed adjuvant chemotherapy initiation was a binary outcome for DID estimates, which we then treated as a continuous variable in the linear regression model. Mean DID estimates represent the proportions of patients who experienced a delay in chemotherapy initiation.
All covariates were initially included in the model, retaining variables based on both clinical and statistical significance (P < .05). The final model included age, comorbidity, stage, BC subtype, surgery, hospital transfer, distance to facility, residence area, region, facility type, and facility volume. We excluded insurance and area-level income from the final model because they are qualifying factors for Medicaid eligibility and are therefore related to the exposure variable. The parallel trends assumption for DID method was evaluated in the preexpansion period by conducting a test in which we included a linear year-by-race interaction on the delayed chemotherapy outcome. Results are presented as adjusted DID percentage points (ppt) with 95% confidence intervals (CIs) between White (reference category) and Black, Hispanic, and Other race and ethnicities in the preexpansion and postexpansion periods. A negative DID estimate indicates a reduction in chemotherapy delay.
We used Cox proportional hazards regression models to compare the change in time to chemotherapy (TTC) initiation between preexpansion and postexpansion periods according to race and ethnicity. Proportional hazard assumption was tested. To examine whether Medicaid expansion was associated with racial and ethnic groups differently, interaction between preexpansion and postexpansion and race and ethnicity was tested. Covariates were included based on clinical and statistical significance (P < .05). We expressed results as adjusted hazard ratios (aHR) and DID with 95% confidence intervals. An adjusted hazard ratio >1 indicated shorter TTC of postexpansion vs preexpansion, and an adjusted DID hazard ratio >1 indicated racial disparity reduction compared with White race.
Subgroup analyses were performed among patients in the lowest income quartile and those with Medicaid insurance. Sensitivity analyses were performed excluding patients diagnosed in 2013 to avoid the spillover effect of the Medicaid expansion policy (19).
To further test our hypothesis that Medicaid expansion was associated with decreased racial disparity and assess the effect on temporal trends, we performed a falsification analysis in a cohort of patients residing in states that have not undergone Medicaid expansion as of November 2021 (TN, NC, ID, GA, FL, MO, AL, MS, KS, TX, WI, UT, SC, SD, VA, OK, NE, WY, ME), referred to herein as nonexpansion states. Except for residing in nonexpansion states, other selection criteria remained the same as for the analysis cohort. In addition, we implemented a triple-differences (difference-in-difference-in-difference) method (20) based on a propensity score weighted cohort (21) that combined data from both expansion states and nonexpansion states (Supplementary Methods, available online).
Statistical significance was set at P < .05. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA) and R version 4.0.0. The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline. We obtained approval to conduct this study from our institutional review board.
Results
Study population
A total of 100 643 patients were included in our main analyses. Of those, 63 313 (62.9%) were diagnosed in the preexpansion and 37 330 (37.1%) in the postexpansion period. The median age at surgery was 53 years (range = 40-64 years). Overall, 4.2% of the patients were Asian American or Pacific Islander, 0.3% American Indian or Alaska Native, 11.4% Black, 4.9% Hispanic, 77.4% White, and 1.5% Unknown. Table 1 compares cohort characteristics between patients diagnosed in the preexpansion and postexpansion periods. Medicaid expansion was associated with an increase of early-stage disease presentation (stage I: 48.2% preexpansion vs 50.9% postexpansion), Medicaid insurance coverage (9.2% preexpansion vs 12.1% postexpansion), and care sought in academic facilities (38.4% preexpansion vs 43.7% postexpansion).
Table 1.
Characteristics of patients with stage I-III BC identified in the National Cancer Database (NCDB) between 2007 and 2017 residing in a state that underwent Medicaid expansion in January 2014, according to expansion perioda
| All patients | Preexpansion | Postexpansion | P b | |
|---|---|---|---|---|
| (N = 100 643) | (n = 63 313) | (n = 37 330) | ||
| No. (%) | No. (%) | No. (%) | ||
| Age at surgery, y | ||||
| 40-49 | 33 269 (33.1) | 21 713 (34.3) | 11 556 (31.0) | <.001 |
| 50-59 | 45 336 (45.0) | 28 322 (44.7) | 17 014 (45.6) | |
| 60-64 | 22 038 (21.9) | 13 278 (21.0) | 8760 (23.5) | |
| Race and ethnicity | ||||
| American Indian and Alaska Native NH | 269 (0.3) | 166 (0.3) | 103 (0.3) | <.001 |
| Asian American NH | 3212 (3.2) | 1866 (2.9) | 1346 (3.6) | |
| Black | 11 805 (11.7) | 7060 (11.2) | 4745 (12.7) | |
| Hispanic | 4923 (4.9) | 2789 (4.4) | 2134 (5.7) | |
| Pacific Islander NH | 1028 (1.0) | 618 (1.0) | 410 (1.1) | |
| Other/Unknown | 1493 (1.5) | 904 (1.4) | 589 (1.6) | |
| White | 77 913 (77.4) | 49 910 (78.8) | 28 003 (75.0) | |
| AJCC stage | ||||
| I | 49 512 (49.2) | 30 522 (48.2) | 18 990 (50.9) | <.001 |
| II | 43 008 (42.7) | 27 622 (43.6) | 15 386 (41.2) | |
| III | 8123 (8.1) | 5169 (8.2) | 2954 (7.9) | |
| Comorbidity score | ||||
| 0 | 87 029 (86.5) | 54 961 (86.8) | 32 068 (85.9) | <.001 |
| 1 | 11 279 (11.2) | 7059 (11.1) | 4220 (11.3) | |
| 2+ | 2335 (2.3) | 1293 (2.0) | 1042 (2.8) | |
| Subtype | ||||
| Hormonal receptor positive | 43 569 (43.3) | 22 948 (36.2) | 20 621 (55.2) | <.001 |
| HER2 positive | 17 734 (17.6) | 8930 (14.1) | 8804 (23.6) | |
| Triple negative | 15 063 (15.0) | 8171 (12.9) | 6892 (18.5) | |
| Unknown | 24 277 (24.1) | 23 264 (36.7) | 1013 (2.7) | |
| Surgery type | ||||
| Lumpectomy | 51 291 (51.0) | 32 053 (50.6) | 19 238 (51.5) | .005 |
| Mastectomy | 49 352 (49.0) | 31 260 (49.4) | 18 092 (48.5) | |
| Primary insurance | ||||
| Private | 80 483 (80.0) | 51 247 (80.9) | 29 236 (78.3) | <.001 |
| Medicaid | 10 365 (10.3) | 5835 (9.2) | 4530 (12.1) | |
| Medicare | 5817 (5.8) | 3488 (5.5) | 2329 (6.2) | |
| Other government insurance | 1022 (1.0) | 616 (1.0) | 406 (1.1) | |
| Uninsured | 1832 (1.8) | 1375 (2.2) | 457 (1.2) | |
| Unknown | 1124 (1.1) | 752 (1.2) | 372 (1.0) | |
| Transfer of care | ||||
| No | 56 364 (56.0) | 36 450 (57.6) | 19 914 (53.3) | <0.001 |
| Yes | 44 279 (44.0) | 26 863 (42.4) | 17 416 (46.7) | |
| Distance from patient to hospital | ||||
| 0-30 miles | 77 789 (77.3) | 50 205 (79.3) | 27 584 (73.9) | <.001 |
| 31-50 miles | 11 686 (11.6) | 7094 (11.2) | 4592 (12.3) | |
| >50 miles | 11 168 (11.1) | 6014 (9.5) | 5154 (13.8) | |
| Patients treated in >1 CoC facility | ||||
| No | 77 011 (76.5) | 48 183 (76.1) | 28 828 (77.2) | <.001 |
| Yes | 23 632 (23.5) | 15 130 (23.9) | 8502 (22.8) | |
| Percent nonhigh school graduates | ||||
| First quartile | 12 544 (12.5) | 7207 (11.4) | 5337 (14.3) | <.001 |
| Second quartile | 20 164 (20.0) | 12 650 (20.0) | 7514 (20.1) | |
| Third quartile | 27 324 (27.1) | 18 238 (28.8) | 9086 (24.3) | |
| Fourth quartile (most educated) | 29 385 (29.2) | 19 168 (30.3) | 10 217 (27.4) | |
| Unknown | 11 226 (11.2) | 6050 (9.6) | 5176 (13.9) | |
| Median household income | ||||
| First quartile | 11 703 (11.6) | 7107 (11.2) | 4596 (12.3) | <.001 |
| Second quartile | 15 921 (15.8) | 10 376 (16.4) | 5545 (14.9) | |
| Third quartile | 21 498 (21.4) | 14 313 (22.6) | 7185 (19.2) | |
| Fourth quartile (wealthiest) | 40 262 (40.0) | 25 445 (40.2) | 14 817 (39.7) | |
| Unknown | 11 259 (11.2) | 6072 (9.6) | 5187 (13.9) | |
| Residence areac | ||||
| Metro | 85 036 (84.5) | 53 329 (84.2) | 31 707 (84.9) | <.001 |
| Urban | 11 646 (11.6) | 7361 (11.6) | 4285 (11.5) | |
| Rural | 1329 (1.3) | 796 (1.3) | 533 (1.4) | |
| Unknown | 2632 (2.6) | 1827 (2.9) | 805 (2.2) | |
| Region | ||||
| South | 18 734 (18.6) | 11 877 (18.8) | 6857 (18.4) | <.001 |
| Northeast | 28 079 (27.9) | 16 529 (26.1) | 11 550 (30.9) | |
| Midwest | 38 341 (38.1) | 25 100 (39.6) | 13 241 (35.5) | |
| West | 15 489 (15.4) | 9807 (15.5) | 5682 (15.2) | |
| Facility type | ||||
| Community cancer program | 9827 (9.8) | 6360 (10.0) | 3467 (9.3) | <.001 |
| Comprehensive CCP | 39 661 (39.4) | 25 907 (40.9) | 13 754 (36.8) | |
| Academic | 40 424 (40.2) | 24 122 (38.1) | 16 302 (43.7) | |
| Integrated network | 10 731 (10.7) | 6924 (10.9) | 3807 (10.2) | |
| Facility volume tertile | ||||
| Low | 9983 (9.9) | 6560 (10.4) | 3423 (9.2) | <.001 |
| Medium | 24 538 (24.4) | 15 884 (25.1) | 8654 (23.2) | |
| High | 66 122 (65.7) | 40 869 (64.6) | 25 253 (67.6) | |
Participants include patients aged 40 years to 64 years in the National Cancer Database from January 1, 2007, to December 31, 2017. Preexpansion included patients diagnosed in 2007-2013, and postexpansion included patients diagnosed in 2014-2017. AJCC = American Joint Committee on Cancer; CCP = community cancer program; CoC = Commission on Cancer; NH = non-Hispanic.
P values were derived from Chi-square test.
Estimated by matching the state/county Federal Information Processing Standards code of the patient at diagnosis to 2013 data published by the US Department of Agriculture Economic Research Service. Metropolitan counties are defined as having a population size of the metropolitan area greater than 250 000. Urban counties are defined as nonmetropolitan with a population size of at least 2500. Rural counties have a population of fewer than 2500.
Chemotherapy initiation delay
Table 2 presents the adjusted rates of chemotherapy initiation delay according to race and ethnicity in the prexpansion and postexpansion periods with DID estimates. Overall, 21.9% of women experienced a chemotherapy initiation delay, declining from 23.4% in 2010-2013 to 19.4% in 2014-2017 (P < .001, χ2 test). Among White patients, chemotherapy delay rates decreased from 22.2% preexpansion to 19.0% postexpansion. Among Black patients, chemotherapy initiation delay decreased from 33.2% to 27.9%, resulting in an adjusted DID of −2.1 ppt (95% CI = −3.7 to −0.5 ppt, P = .01). Similarly, Hispanic patients experienced a reduction in delay from 30.8% to 24.5%, resulting in an adjusted DID of −3.2 ppt (95% CI = −5.6 ppt to −0.9 ppt, P = .008). Despite a small size (n = 2484), Asian American or Pacific Islander patients had a reduction in delay from 25.2% to 20.4%, with an adjusted DID of −2.3 ppt (95% CI −4.9 ppt to 0.2 ppt, P = .07) with marginally statistical significance. The exclusion of patients diagnosed in 2013 yielded similar results with a statistically significant adjusted DID for Asian American or Pacific Islander patients (Supplementary Table 2, available online).
Table 2.
Adjusted rates of delayed chemotherapy after Medicaid expansion among patients with breast cancer who resided in the January 2014 expansion states according to race and ethnicitya
| Race | Preexpansion 2007-2013 |
Postexpansion 2014-2017 |
Adjusted DID, percentage points (95% CI)b | P | ||||
|---|---|---|---|---|---|---|---|---|
| No. | Adjusted rate of delayed chemo >60 days | Adjusted difference from White participants, percentage points | No. | Adjusted rate of delayed chemo >60 days | Adjusted difference from White participants, percentage points | |||
| White | 49 910 | 22.2 | 0.0 | 28 003 | 19.0 | 0.0 | Ref | Ref |
| AAPI | 2484 | 25.2 | 3.0 | 1758 | 19.7 | 0.7 | −2.3 (−4.9 to 0.2) | .07 |
| AIAN | 166 | 25.9 | 3.8 | 103 | 15.7 | −3.3 | −7.1 (−17.1 to 3) | .17 |
| Black | 7060 | 33.2 | 11.0 | 4745 | 27.9 | 8.9 | −2.1 (−3.7 to −0.5) | .01 |
| Hispanic | 2789 | 30.8 | 8.6 | 2134 | 24.5 | 5.5 | −3.2 (−5.6 to −0.9) | .008 |
| Other (AAPI, AIAN, unknown) | 3554 | 25.0 | 2.8 | 2448 | 20.4 | 1.3 | −1.6 (−3.8 to 0.6) | .14 |
| Other/unknown | 904 | 24.1 | 2.0 | 589 | 23.0 | 4.0 | 2.0 (−2.2 to 6.3) | .35 |
Participants include patients aged 40 years to 64 years in the National Cancer Database from January 1, 2007, to December 31, 2017, residing in a state undergoing Medicaid expansion in the January 2014. Preexpansion included patients diagnosed in 2007-2013, and postexpansion included patients diagnosed in 2014-2017. AAPI = Asian American or Pacific Islander; AIAN = American Indian or Alaska Native; CI = confidence interval; DID = difference-in-difference.
Indicates the regression coefficient on an interaction term between Medicaid expansion status and race, adjusted for patient age, comorbidity, stage, subtype, surgery, hospital transfer, treated at >1 Commission on Cancer hospitals, region, facility type, and facility volume.
The median TTC initiation was 42 days (interquartile range = 31-57 days) for the entire cohort. For White patients, the preexpansion and postexpansion median TTC was 42 and 41 days. For patients of racialized groups (all categories combined), it was 46 and 43 days, respectively. In the adjusted Cox proportional hazards models assessing TTC according to expansion period (Table 3), we observed a statistically significant reduction in the TTC initiation in the postexpansion period among White patients (aHR = 1.11, 95% CI = 1.09 to 1.12, P <0 .001) and among patients of racialized groups (aHR = 1.14, 95% CI = 1.11 to 1.17, P <0 .001), with a greater reduction in TTC among racialized groups (DID aHR = 1.03, 95% CI = 1.00 to 1.06, P = .05) compared with White patients.
Table-3.
Multivariable Cox proportional hazard model for time to chemotherapy by time period (preexpansion, postexpansion) according to race and ethnicity
| Cohorts | No. | Postexpansion vs preexpansion, adjusted HR (95% CI) | P | Adjusted DID, HR (95% CI)a | P |
|---|---|---|---|---|---|
| White | 77 913 | 1.11 (1.09 to 1.12) | <.001 | Ref | Ref |
| Racialized group | 22 730 | 1.14 (1.11 to 1.17) | <.001 | 1.03 (1.00 to 1.06) | .05 |
Refers to ratio of preexpansion to postexpansion hazard ratio among patients belonging to a racialized group compared with preexpansion to postexpansion hazard ratio among patients with White race. Ratios greater than 1 indicate improvement (shorter time from surgery to chemotherapy) in a racialized group using White as a reference. Model adjustment included age, comorbidity, stage, subtype, surgery, treated at >1 CoC facility, hospital transfer, region, residence area, area-level education quartiles, facility type and facility volume. DID = difference-in-difference; HR = hazard ratio.
Subgroup analyses
Among patients residing in the lowest income quartile (n = 11 703), there was a similar pattern of decrease in the proportion of patients experiencing delays among all racialized groups, with reductions of 1.4, 5.7, 9.4, 7.7, 23.9, and 6.0 ppt among White, Black, Hispanic, Asian American or Pacific Islander, American Indian or Alaska Native, and Unknown patients, respectively. Using White patients as the reference category, American Indian or Alaska Native (n = 78) patients experienced the greatest reduction in chemotherapy delay, followed by Black and Hispanic patients. Despite the small sample size, patients identifying as American Indian or Alaska Native had an adjusted DID of −21.5 ppt (95% CI −42.9% to 0%, P = .05). Black patients had an adjusted DID of −4.3 ppt (95% CI −7.9% to −0.6%, P = .02), and Hispanic patients had an adjusted DID of −8.0 ppt (95% CI −13.7% to −2.2%, P = .007) (Table 4).
Table 4.
Adjusted rates of delayed chemotherapy after Medicaid expansion according to race and ethnicity among those who resided in the lowest income quartilea
| Race | Preexpansion 2007-2013 |
Postexpansion 2014-2017 |
Adjusted DID, percentage points (95% CI)b | P | ||||
|---|---|---|---|---|---|---|---|---|
| No. | Adjusted rate of delayed chemo >60 days | Adjusted difference from White participants, percentage points | No. | Adjusted rate of delayed chemo >60 days | Adjusted difference from White participants, percentage points | |||
| Patients who resided in lowest income quartile (N = 11 703) | ||||||||
| White | 4133 | 25.2 | 0.0 | 2540 | 23.8 | 0.0 | Ref | Ref |
| AAPI | 110 | 28.4 | 3.2 | 109 | 20.7 | −3.0 | −6.2 (−18.0 to 5.6) | .30 |
| AIAN | 55 | 26.4 | 1.2 | 23 | 3.5 | −20.3 | −21.5 (−42.9 to 0) | .05 |
| Black | 2165 | 33.0 | 7.8 | 1384 | 27.3 | 3.5 | −4.3 (−7.9 to −0.6) | .02 |
| Hispanic | 576 | 34.6 | 9.4 | 481 | 25.2 | 1.4 | −8.0 (−13.7 to −2.2) | .007 |
| Other/unknown | 87 | 27.3 | 2.1 | 65 | 21.3 | −2.5 | −4.6 (−18.8 to 9.6) | .53 |
Participants include patients aged 40 to 64 years identified in the National Cancer Database from January 1, 2007, to December 31, 2017 and residing in a state that underwent Medicaid expansion in January 2014. Preexpansion included patients diagnosed in 2007-2013, and postexpansion included patients diagnosed in 2014-2017. AAPI = Asian American or Pacific Islander; AIAN = American Indian or Alaska Native; CI = confidence interval; DID = difference-in-difference.
For patients who resided in the lowest income quartile, the adjusted DID indicates the regression coefficient on an interaction term between Medicaid expansion status and race, adjusted for patient age, comorbidity, stage, subtype, surgery, hospital transfer, residence area, and facility type and facility volume tertiles.
Among Medicaid beneficiaries (n = 10 365), delays in chemotherapy initiation decreased among all groups, but magnitudes varied (Supplementary Table 2, available online). Compared with White patients, American Indian or Alaska Native (n = 56) and Black patients had the greatest reductions in delay. American Indian or Alaska Native patients had an adjusted DID of −23.4 ppt (95% CI = −46.9% to 0.1%, P = .051) compared with White patients, and Black patients experienced improvements in delay resulting an adjusted DID of −7.0 ppt (95% CI = −11.5% to −2.6%, P = .002).
In the analyses among (n = 120 309) patients residing in nonexpansion states, we observed similar trends; however, they were of smaller magnitude, and no statistically significant decrease in racial disparity was observed (Supplementary Table 3, available online). The triple-differences sensitivity analysis suggested that Medicaid expansion was associated with a statistically significant racial disparity reduction of −2.5 ppt (95% CI = −4.0% to −0.9%, P = .002) in the rate of delayed chemo between expansion and nonexpansion states (Supplementary Table 4, available online).
Discussion
In this study, we observed a statistically significant reduction in TTC after Medicaid expansion among women with early-stage BC residing in states that expanded Medicaid eligibility in January 2014. Furthermore, we show that patients belonging to racialized groups had worse access and greater disparities to timely care pre-ACA that were reduced after Medicaid expansion. Medicaid expansion was associated with an increased proportion of women presenting with early-stage disease and receiving care at academic facilities. Although our findings corroborate and extend those of previous studies (22-25), we offer unique insight into the positive association of Medicaid expansion in reducing racial disparities among BC patients by reporting reduction in the rates of chemotherapy delays.
Time to treatment initiation has consistently been found to affect cancer outcomes, with a clear association between TTC and worse overall and BC-specific survival (8,26). Discriminatory policies and practices create barriers in access to health care, leading to longer delays and worse survival outcomes among communities targeted for marginalization. In our study, we defined delay to adjuvant initiation as more than 60 days, which we find relevant among underserved populations who experience delays at all stages of the cancer care continuum (9).
Racialized groups experience worse BC outcomes. Inequitable access to care among communities targeted for marginalization is well-documented in the literature. In the United States, Black patients have lower stage-specific survival compared with White patients (27) and are 42% more likely to die from BC than White women (28). The survival disparity is even larger for American Indian or Alaska Native patients, who have a long history of discriminatory policies and barriers in access to health care (29), resulting a 51% higher risk of cancer death than White patients (27). Although differences of BC subtypes and high-penetrance genes across racial and ethnic populations could explain some of the disparity (30-32), the impact of social determinants of health and differences in cancer care delivery due to modifiable factors, such as health care access among communities targeted by marginalization, play a major role (33). For instance, Black and Hispanic patients are less likely to undergo timely surgery after BC diagnosis compared with White patients (34-37), more likely to present with more advanced stages (38), experience greater delays in treatment initiation (35), and less likely to receive guideline-concordant care (39). Because all these factors are closely related to access to health care and insurance coverage (11-13), understanding the association of Medicaid expansion on timely access to care, particularly by race and ethnicity, is critical.
Han et al. (40) examined the relationship of dependent coverage expansion under the ACA with time to BC treatment among young women. Investigators observed a non-statistically significant decrease in uninsured rates among women aged 19-25 years that did not translate into a decrease in chemotherapy delays. In an NCDB study evaluating the association of Medicaid expansion with insurance status, stage at diagnosis, and timely treatment among newly diagnosed patients with invasive breast, colon, or non-small cell lung cancer (22), no changes in timeliness to treatment among patients residing in states that expanded Medicaid were observed. Conversely, in a similar analysis among patients with metastatic cancers identified in the Flatiron Health database, a decrease in Black–White racial disparities in receiving timely first-line systemic treatment within 30 days was reported (14). Recently, a study comparing preexpansion and postexpansion 2-year survival among newly diagnosed patients from 42 states’ population-based cancer registries reported that states expanding Medicaid eligibility had a greater increase in survival compared with nonexpansion states (41). Authors also showed that improvements in survival were greater among Black patients and those living in rural areas (41), suggesting that Medicaid expansion is associated with narrowing disparities by race and rurality. We reported similar results after observing that the racial disparity of 2-year mortality among patients with de novo stage IV BC narrowed due to Medicaid expansion (1).
In this study, we observed a decrease in TTC after Medicaid expansion, with decreases of greater magnitude among patients belonging to racialized groups. Compared with White patients, Black or Hispanic patients experienced greater benefits from Medicaid expansion, resulting in reduced racial and ethnic disparities. Althouh the small sample size warrants careful interpretation, patients identified as American Indian or Alaska Native experienced the largest reduction in chemotherapy initiation delay. All racial and ethnic groups similarly experienced these improvements among all subgroup analyses. Our findings emphasize the importance of addressing disparities and highlight the positive effect of implementing policies to improve health equity.
Using a Cox model, the median TTC decreased for all patients between the preexpansion and postexpansion periods. Furthermore, the magnitude of benefit was greater among patients belonging to racialized groups, consistent with our DID analysis. We conducted falsification confirmatory analyses among patients residing in nonexpansion states to control for temporal change and observed a slight decrease in chemotherapy initiation delays that was not associated with statistically significant reduction in racial disparities, supporting the observation that the improvements in shortening TTC are likely associated with Medicaid expansion.
In this study, we evaluated BC disparities between expansion periods among states that expanded Medicaid eligibility. We selected this approach to limit potential biases from differences between states that chose to expand Medicaid eligibility and those that did not. Our findings are novel and relevant because they explain how policy-level interventions aimed at improving health insurance coverage can diminish barriers to care and improve health equity measures. Populations targeted for marginalization had worse access to timely care preexpansion but saw statistically significant improvements in BC outcomes with increased access to care. We believe that the newfound reductions in TTC initiation among patients with early-stage BC, in conjunction with increased Medicaid insurance coverage and presentation of earlier-stage disease, suggest that Medicaid expansion may have a positive association with BC outcomes, with the potential to reduce many persisting disparities.
To our knowledge, this is the first study to report on the association of Medicaid expansion with TTC among a large group of patients with stage I-III BC. Our study design limits the ability to infer causality; nevertheless, our results show an association between Medicaid expansion and reductions in racial disparity in chemotherapy initiation delays using both a binary DID approach and time-to-event analysis. We acknowledge that Hispanic ethnicity accounted for only 4.9% of our cohort, lower than what has been seen in other cohorts (6), because the NCDB does not link data to the North American Association of Central Cancer Registries Hispanic and Asian/Pacific Islander Identification Algorithm (42). The lower proportion of patients belonging to underrepresented minorities and moderate sample size may have limited DID estimates from achieving statistical significance. Furthermore, the proportion of patients of Hispanic origin residing in states that underwent expansion is lower than that among states not undergoing expansion. Although a spillover effect cannot be excluded, a sensitivity analysis excluding patients diagnosed in 2013 supports our findings. Future studies exploring further racial breakdown, broader age ranges including younger patients, and longer follow-ups are necessary to expand our findings’ generalizability and determine the potential survival association of this policy.
Implementation of Medicaid expansion reduced racial disparities in TTC between White patients and those belonging to racialized groups with early-stage BC. Our results highlight the importance of increasing access to health care by implementing policies aimed at improving access to care for all and eliminating barriers to care. Our findings suggest that cancer outcomes could be improved if other states expand Medicaid.
Supplementary Material
Acknowledgements
The CoC’s NCDB and the hospitals participating in the CoC NCDB are the source of the deidentified data used; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors.
The funder did not play a role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.
Findings from this manuscript were presented, in part, at the San Antonio Breast Cancer Symposium in 2020.
Contributor Information
Mariana Chavez-MacGregor, Department of Health Services Research, Division of Cancer Prevention and Population Sciences, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; Department of Breast Cancer Oncology, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
Xiudong Lei, Department of Health Services Research, Division of Cancer Prevention and Population Sciences, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
Catalina Malinowski, Department of Health Services Research, Division of Cancer Prevention and Population Sciences, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
Hui Zhao, Department of Health Services Research, Division of Cancer Prevention and Population Sciences, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
Ya-Chen Shih, Department of Health Services Research, Division of Cancer Prevention and Population Sciences, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
Sharon H Giordano, Department of Health Services Research, Division of Cancer Prevention and Population Sciences, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; Department of Breast Cancer Oncology, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
Data availability
The primary dataset (National Cancer Database) is available publicly through the American College of Surgeons (https://www.facs.org/quality-programs/cancer/ncdb). The datasets generated and/or analyzed during this study are available from the corresponding author on reasonable request.
Author contributions
Mariana Chavez-MacGregor, MD, MSc (formal analysis; funding acquisition; investigation; methodology; project administration; resources; supervision; validation; writing—original draft; writing—review and editing); Xiudong Lei, PhD (data curation; formal analysis; methodology; writing—original draft; writing—review and editing); Catalina Malinowski, MPH (conceptualization; investigation; project administration; writing—original draft; writing—review and editing); Hui Zhao, PhD (supervision; writing—review and editing); Ya-Chen Shih, PhD (conceptualization; writing—review and editing); Sharon H. Giordano, MD, MPH (conceptualization; funding acquisition; supervision; writing—review and editing).
Funding
The study was supported by the Susan G. Komen Foundation (SAC150061, SAC220221); the National Cancer Institute at the National Institutes of Health (P30 CA016672); The Breast Cancer Research Foundation (BCRF-22-190) and Conquer Cancer, the ASCO Foundation.
Conflicts of interest
Authors have no relevant conflicts of interest to report at this time.
References
- 1. Malinowski C, Lei X, Zhao H, Giordano SH, Chavez-MacGregor M.. Association of Medicaid expansion with mortality disparity by race and ethnicity among patients with de novo stage IV breast cancer. JAMA Oncol. 2022;8(6):863-870. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Emerson MA, Golightly YM, Aiello AE, et al. Breast cancer treatment delays by socioeconomic and health care access latent classes in Black and White women. Cancer. 2020;126(22):4957-4966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Miller BC, Bowers JM, Payne JB, Moyer A.. Barriers to mammography screening among racial and ethnic minority women. Soc Sci Med. 2019;239:112494. [DOI] [PubMed] [Google Scholar]
- 4. Farias AJ, Wu W-H, Du XL.. Racial differences in long-term adjuvant endocrine therapy adherence and mortality among Medicaid-insured breast cancer patients in Texas: findings from TCR-Medicaid linked data. BMC Cancer. 2018;18(1):1-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Parikh-Patel A, Morris CR, Kizer KW.. Disparities in quality of cancer care: the role of health insurance and population demographics. Medicine. 2017;96(50):e9125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Smith-Graziani D, Lei X, Giordano SH, Zhao H, Karuturi M, Chavez-MacGregor M.. Delayed initiation of adjuvant chemotherapy in older women with breast cancer. Cancer Med. 2020;9(19):6961-6971. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. de Melo Gagliato D, Lei X, Giordano SH, et al. Impact of delayed neoadjuvant systemic chemotherapy on overall survival among patients with breast cancer. Oncologist. 2020;25(9):749-757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Chavez-MacGregor M, Clarke CA, Lichtensztajn DY, Giordano SH.. Delayed initiation of adjuvant chemotherapy among patients with breast cancer. JAMA Oncol. 2016;2(3):322-329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. de Melo Gagliato D, Gonzalez-Angulo AM, Lei X, et al. Clinical impact of delaying initiation of adjuvant chemotherapy in patients with breast cancer. J Clin Oncol. 2014;32(8):735-744. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Chavez‐MacGregor M, Unger JM, Moseley A, Ramsey SD, Hershman DL.. Survival by Hispanic ethnicity among patients with cancer participating in SWOG clinical trials. Cancer. 2018;124(8):1760-1769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Sommers BD, Gawande AA, Baicker K.. Health insurance coverage and health—what the recent evidence tells us. N Engl J Med. 2017;377(6):586-593. [DOI] [PubMed] [Google Scholar]
- 12. Halpern MT, Brawley OW.. Insurance status, health equity, and the cancer care continuum. Cancer. 2016;122(20):3106-3109. [DOI] [PubMed] [Google Scholar]
- 13. Ward E, Halpern M, Schrag N, et al. Association of insurance with cancer care utilization and outcomes. CA Cancer J Clin. 2008;58(1):9-31. [DOI] [PubMed] [Google Scholar]
- 14. Adamson BJ, Cohen AB, Gross CP, et al. ACA Medicaid expansion association with racial disparity reductions in timely cancer treatment. Am J Managed Care. 2021;27(7):274-281. [DOI] [PubMed] [Google Scholar]
- 15. Boffa DJ, Rosen JE, Mallin K, et al. Using the National Cancer Database for outcomes research: a review. JAMA Oncol. 2017;3(12):1722-1728. [DOI] [PubMed] [Google Scholar]
- 16. Sundararajan V, Henderson T, Perry C, Muggivan A, Quan H, Ghali WA.. New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality. J Clin Epidemiol. 2004;57(12):1288-1294. [DOI] [PubMed] [Google Scholar]
- 17. Deyo RA, Cherkin DC, Ciol MA.. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613-619. [DOI] [PubMed] [Google Scholar]
- 18. Charlson ME, Pompei P, Ales KL, MacKenzie CR.. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383. [DOI] [PubMed] [Google Scholar]
- 19. Dimick JB, Ryan AM.. Methods for evaluating changes in health care policy: the difference-in-differences approach. JAMA. 2014;312(22):2401-2402. [DOI] [PubMed] [Google Scholar]
- 20. Stimpson JP, Pintor JK, McKenna RM, Park S, Wilson FA.. Association of Medicaid expansion with health insurance coverage among persons with a disability. JAMA Netw Open. 2019;2(7):e197136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Stuart EA, Huskamp HA, Duckworth K, et al. Using propensity scores in difference-in-differences models to estimate the effects of a policy change. Health Serv Outcomes Res Methodol. 2014;14(4):166-182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Takvorian SU, Oganisian A, Mamtani R, et al. Association of Medicaid expansion under the Affordable Care Act with insurance status, cancer stage, and timely treatment among patients with breast, colon, and lung cancer. JAMA Netw Open. 2020;3(2):e1921653. [DOI] [PubMed] [Google Scholar]
- 23. Soni A, Simon K, Cawley J, Sabik L.. Effect of Medicaid expansions of 2014 on overall and early-stage cancer diagnoses. Am J Public Health. 2018;108(2):216-218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. 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 Oncol. 2018;4(12):1713-1720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. 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(35):3906-3915. [DOI] [PubMed] [Google Scholar]
- 26. Bleicher RJ. Timing and delays in breast cancer evaluation and treatment. Ann Surg Oncol. 2018;25(10):2829-2838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Siegel RL, Miller KD, Fuchs HE, Jemal A.. Cancer statistics, 2021. CA A Cancer J Clin. 2021;71(1):7-33. [DOI] [PubMed] [Google Scholar]
- 28. Yedjou CG, Sims JN, Miele L, et al. Health and racial disparity in breast cancer. Breast Cancer Metastasis Drug Resistance. 2019;1152:31-49. doi: 10.1007/978-3-030-20301-6_3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Cromer KJ, Wofford L, Wyant DK.. Barriers to healthcare access facing American Indian and Alaska Natives in rural America. J Community Health Nurs. 2019;36(4):165-187. [DOI] [PubMed] [Google Scholar]
- 30. Pan J-W, Zabidi MMA, Ng P-S, et al. The molecular landscape of Asian breast cancers reveals clinically relevant population-specific differences. Nat Commun. 2020;11(1):1-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Serrano-Gómez SJ, Fejerman L, Zabaleta J.. Breast cancer in Latinas: a focus on intrinsic subtypes distribution. Cancer Epidemiol Prev Biomarkers. 2018;27(1):3-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA. 2006;295(21):2492-2502. [DOI] [PubMed] [Google Scholar]
- 33. Eaglehouse YL, Georg MW, Shriver CD, Zhu K.. Racial differences in time to breast cancer surgery and overall survival in the US military health system. JAMA Surg. 2019;154(3):e185113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Navarro S, Yang Y, Ochoa CY, et al. Asian ethnic subgroup disparities in delays of surgical treatment for breast cancer. JNCI Cancer Spectrum. 2022;6(1):pkab089. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Reeder‐Hayes KE, Mayer SE, Olshan AF, et al. Race and delays in breast cancer treatment across the care continuum in the Carolina Breast Cancer Study. Cancer. 2019;125(22):3985-3992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. George P, Chandwani S, Gabel M, et al. Diagnosis and surgical delays in African American and White women with early-stage breast cancer. J Womens Health (Larchmt). 2015;24(3):209-217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Smith EC, Ziogas A, Anton-Culver H.. Delay in surgical treatment and survival after breast cancer diagnosis in young women by race/ethnicity. JAMA Surg. 2013;148(6):516-523. [DOI] [PubMed] [Google Scholar]
- 38. Nahleh Z, Otoukesh S, Mirshahidi HR, et al. Disparities in breast cancer: a multi‐institutional comparative analysis focusing on American Hispanics. Cancer Med. 2018;7(6):2710-2717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Hirko KA, Rocque G, Reasor E, et al. The impact of race and ethnicity in breast cancer—disparities and implications for precision oncology. BMC Med. 2022;20(1):1-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Han X, Zhao J, Ruddy KJ, Lin CC, Sineshaw HM, Jemal A.. The impact of dependent coverage expansion under the Affordable Care Act on time to breast cancer treatment among young women. PLoS One. 2018;13(6):e0198771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Han X, Zhao J, Yabroff KR, Johnson CJ, Jemal A.. Association between Medicaid expansion under the Affordable Care Act and survival among newly diagnosed cancer patients. J Natl Cancer Inst. 2022;114(8):1176-1185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Mallin K, Browner A, Palis B, et al. Incident cases captured in the National Cancer Database compared with those in US population based central cancer registries in 2012–2014. Ann Surg Oncol. 2019;26(6):1604-1612. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The primary dataset (National Cancer Database) is available publicly through the American College of Surgeons (https://www.facs.org/quality-programs/cancer/ncdb). The datasets generated and/or analyzed during this study are available from the corresponding author on reasonable request.