Abstract
Introduction:
We evaluated whether Medicaid expansion (ME) was associated with improved two-year survival and time to treatment initiation (TTI) among patients with gastrointestinal (GI) cancer.
Methods:
GI cancer patients diagnosed 40–64 years were queried from the National Cancer Database. Those diagnosed from 2010–2012 were considered pre-expansion; those diagnosed from 2014–2016 were considered post-expansion. Cox models estimated hazard ratios (HR) and 95% confidence intervals (CI) for two-year overall survival. Generalized estimating equations (GEE) estimated odds ratios (OR) and 95%CI of TTI within 30- and 90-days. Multivariable Difference-in-Difference (DD) models were used to compare expansion/non-expansion cohorts pre/post-expansion, adjusting for patient, clinical, and hospital factors.
Results:
377,063 patients were included. No significant difference in two-year survival was demonstrated across ME and non-ME states overall or in site-based subgroup analysis. In stage-based subgroup analysis, two-year survival significantly improved among stage II cancer, with an 8% decreased hazard of death at two years (0.92;0.87–0.97). Those with stage IV had a 4% increased hazard of death at two years (1.04;1.01–1.07). Multivariable GEE models showed increased TTI within 30-days (1.12;1.09–1.16) and 90-days (1.22;1.17–1.27). Site-based subgroup analyses indicated increased likelihood of TTI within 30- and 90-days among colon, liver, pancreas, rectum, and stomach cancers, by 30-days for small intestinal cancer, and by 90-days for esophageal cancer. In subgroup analyses, all stages experienced improved odds of TTI within 30- and 90-days.
Conclusion:
ME was not associated with significant improvement in two-year survival for those with GI cancer. Although TTI increased after ME for both cohorts, the 30- and 90-day odds of TTI was higher for those from ME compared to non-ME states. Our findings add to growing evidence of associations with ME for those diagnosed with GI cancer.
Keywords: gastrointestinal neoplasms, Medicaid, survival, time-to-treatment
INTRODUCTION
The Patient Protection and Affordable Care Act (ACA) introduced the Medicaid expansion (ME) as a mechanism to increase the eligibility of Medicaid for low-income, non-elderly U.S. adults in expansion states.1 As of May 2023, 41 states and the District of Columbia have elected to expand Medicaid coverage.2 The expansion of Medicaid has had a significant impact on access to care and improvement in health outcomes for individuals who were previously uninsured in the United States.3
In cancer care, studies have shown that ME is associated with increased early-stage diagnosis, which is a critical factor in determining cancer outcomes.4–8 Individuals who are diagnosed with cancer at an early stage are more likely to receive treatment that is both effective and less invasive, resulting in improved mortality and lower costs.8,9 Furthermore, ME has been associated with reducing the time to treatment initiation (TTI) and increased overall survival for cancer patients.10 Gastrointestinal (GI) cancers, which account for nearly 20% of all cancers diagnosed, are particularly sensitive to timely diagnosis and treatment.11 In advanced stages, GI cancers are often asymptomatic and having health insurance is a strong predictor of timely diagnosis and cancer-directed treatment.10,11 However, despite the increasing number of studies on the early diagnosis of cancer, relatively few studies have examined the effect of ME on TTI and survival outcomes for patients with GI cancers.
Currently, evidence on the impact of ME on GI cancer outcomes is limited. One reason for this is that existing studies have aggregated a heterogeneous group of cancer types, making it difficult to isolate the effects of Medicaid expansion specifically on GI cancer outcomes.8,12 Additionally, the follow-up time in many studies is relatively short (typically one year or less), which may not be sufficient to fully understand the long-term impact of ME on GI cancer outcomes.11 This is an important consideration as the treatment of GI cancer often requires prolonged and complex care, and the effects of ME associated with access to care and outcomes may not be fully captured in short-term studies. Therefore, further research with longer follow-up periods, and studies that specifically focus on GI cancer outcomes, are needed to better understand the impact of ME on this population.
The aim of this study is to evaluate whether ME is associated with improved outcomes among patients diagnosed with GI cancer in states that adopted ME compared to those that did not. The primary outcome of this study is two-year survival for patients with GI cancers, and the secondary outcome is the receipt of cancer-directed TTI within 30 or 90 days.
MATERIALS AND METHODS
Data and study population
In this study, we utilized data from the National Cancer Database (NCDB), the largest patient-level cancer registry in the United States, which encompasses 70% of all new cancer diagnoses. The NCDB comprises data from over 30% of hospitals nationwide and includes data regarding patient characteristics, tumor type, stage, treatment received, and pertinent clinical data. The comprehensive nature of the NCDB allows for nuanced analysis of cancer outcomes and disparities, albeit not population-based.
We selected patients between the ages of 40 and 64 years old who were diagnosed with a new invasive gastrointestinal (GI) cancer between January 1, 2010 and December 31, 2012 (pre-expansion) and January 1, 2014 and December 31, 2016 (post-expansion). GI cancers included the following sites esophageal (C150-C159); stomach (C160-C169); small intestine (C170-C179); colon (C180-C189, C260); rectosigmoid junction (C199); rectum (C209); anus, anal canal, and anorectum (C210-C212, C218); liver (C220); and pancreas (C250-C259). Those diagnosed in 2013 were excluded to eliminate potential spillover effects of full implementation of ME beginning January 1, 2014.13–15 Patients were identified using the sequence number variable (coded 00 and 01) in the NCDB. Patients diagnosed prior to age 40 were not included because these data are not released at the patient-level due to privacy concerns. Patients who were diagnosed over the age of 64 years were excluded as they were not the intended target population for the ME policy change. Patients with missing information on state ME status, age, cancer stage, survival, those diagnosed with stage 0 GI cancer, and those with other types of GI cancer were excluded from the sample. The flowchart of the patient selection process can be found in Figure 1. This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.
Figure 1. Flow chart of patient selection.
Medicaid expansion status
The primary independent variable was ME status, which was determined using NCDB-provided data based on geographic location of patient’s residence by state and the state ME status at the time of diagnosis. The NCDB categorized states into four groups based on their ME status: 1) early expansion states (2010–2013), 2) January 2014 expansion states, 3) late expansion states (occurring after January 2015), and 4) non-expansion states. We dichotomized the states into two groups based on their status as of January 2014: ME states (early, January 2014, and late expansion states) and non-ME states (states that never expanded their Medicaid program). We defined the ME states as the intervention group and the non-ME states as the control group to follow a quasi-experimental research design.
Study outcomes
Our primary outcome was overall survival, defined as the number of months from the first cancer diagnosis to death. NCDB does not collect information on recurrence or cancer-specific survival. The secondary outcome was TTI, defined as the number of days from diagnosis to the earliest cancer-directed treatment of any type. In NCDB, time to treatment is coded in distinct variables delineating time to first surgery, most definitive surgery, radiotherapy, and systemic therapy (e.g., chemotherapy, immunotherapy, or hormonal therapy). From these, we derived a variable indicating the number of days from diagnosis to first cancer-directed treatment. As timeliness of care is an emerging indicator for quality of cancer care, we dichotomized TTI into two timepoints: within 30-days and within 90-days of diagnosis consistent with previous studies.16,17
Other covariates
Patient characteristics, clinical factors, and hospital or health system features were included as covariates. Patient characteristics were collected at diagnosis and/or first treatment and included age, sex, race/ethnicity, urban/metropolitan/rural residence, median household income, educational attainment, and insurance type. NCDB provides zip-code based data on urban, metropolitan, or rural status (defined by Rural Urban Continuum Code), median household income, and median educational attainment (according to the American Community Survey) for each patient.18 Clinical factors included year of diagnosis, primary site, stage at diagnosis (AJCC stage I-IV), comorbidity score, and receipt of palliative care. Because of competing risks of death for patients with GI cancer, examination of survival without adjustment for comorbidity may confound results. Therefore, we included Charlson/Deyo comorbidity scores to account for comorbidities present at diagnosis and/or treatment. Hospital or health system features included cancer program type (i.e., academic/research program, community cancer program, integrated network cancer program) and facility location (i.e., New England, Middle Atlantic, South Atlantic, East North Central, East South Central, West North Central, West South Central, Mountain, Pacific).
Statistical analysis
For all analyses, an alpha level of 0.05 was used to indicate significance. Statistical analyses were conducted using SAS v9.4 (SAS Institute Inc., Cary, NC, USA) and R 4.1.0. Chi-square tests compared sample characteristics between expansion and non-expansion groups. The differences in two-year survival and TTI before and after the 2014 expansion were compared between the ME and non-ME groups. Missing data were presented as a separate category, and none were excluded in the unadjusted analysis.
For the first outcome of survival, we used the Kaplan-Meier (K-M) method and log-rank test to determine the unadjusted association between the expansion period and two-year overall survival, stratified by ME status. Cancer site- and stage-specific K-M curves were generated to visualize and compare two-year survival between patients in ME and non-ME cohorts.
To evaluate if there were differences in outcomes between the two cohorts, we built the Difference-in-Differences (DD) estimator, which is an interaction term of year of diagnosis (pre-expansion: 2010–2012 or post-expansion: 2014–2016) and ME status. The significance of the DD estimator indicates that there are differences in the pre-and post-expansion time periods. We developed a basic Cox proportional hazards model by including ME status, year of diagnosis, and DD estimator to estimate the crude DD hazard ratio (HR). A HR of <1 indicates an improvement in two-year survival after ME compared with the pre-ME time period. We then performed a multivariable Cox proportional model adjusting for patient characteristics [age, sex, race/ethnicity, health insurance, residence (urban, metropolitan, or rural), annual household income, and educational attainment], clinical factors (cancer site, stage at diagnosis, Charlson-Deyo comorbidity score, receipt of palliative care), hospital or health system features (cancer program type, facility location) to estimate the adjusted DD HR.19,20 We assessed for multicollinearity using the variance inflation factor method. No multicollinearity was detected (all variance inflation factors <10), so we used the full model.21 We also performed unadjusted and adjusted subgroup Cox DD analyses by cancer site and stage at diagnosis. We then implemented DD analyses using crude univariable and adjusted multivariable linear probability models to assess the magnitude of percent change in the two time periods by state ME status.
For the second outcome, we used the DD estimator in unadjusted and adjusted generalized estimating equation models to assess associations with the proportion of TTI within 30- or 90-days. These models took a similar approach to the survival models by first estimating the crude odds ratios, followed by the adjusted odds ratios. These models treated the binary outcome (starting treatment within 30- or 90-days) as a continuous variable, with the mean estimates representing the proportion of patients who received TTI within the respective time period. The results were presented as adjusted DD odds ratios (OR) and their 95% confidence interval (CI). We tested for parallel trends in the study outcomes with visual inspection for similar trends in the outcomes in the pre-expansion periods (2010–2012; see Supplemental File Figure S-1) and falsification tests that treated the 2012–2013 as a hypothetical post-expansion period (see Supplemental File Tables S-1 and S-2). Overall survival outcomes showed no significant violation of parallel trends assumption, while TTI outcomes had limited violations. We also performed sensitivity analysis by excluding patients residing in the early expansion states or in the late expansion states; as there were no significant differences observed, we presented primary analysis results.
RESULTS
Sample characteristics
The final cohort included 377,063 patients diagnosed with GI cancer (see Table 1). Of the 229,921 from expansion states who were in the expansion cohort, 113,798 (49.5%) came from six early expansion states (expanded prior to 2014), 65,369 (28.4%) came from the 19 states expanding Medicaid in January 2014, and 50,754 (22.1%) came from the seven late expansions states (2015 or later). The remaining 147,142 (39%) people diagnosed with GI cancer were from non-expansion states that never expanded their Medicaid programs and were categorized in the non-expansion cohort. Colon (31.7%) represented the most common type of cancer among the study population, followed by pancreas (15.6%) and rectum (14.7%). Compared to those in the ME cohort, patients in the non-ME cohort were disproportionately non-Hispanic black race (12.1% vs. 19.6%), uninsured (11.6% vs. 4.5%), from an urban area (15.6% vs. 11.2%), live below the federal poverty line (26.3% vs. 14.7%), have less than high school education (27.8% vs. 19.4%), and seeking care at community cancer program (43.3% vs. 38.9%).
Table 1.
Characteristics of patients with gastrointestinal cancer by state Medicaid expansion status
| Overall Sample 377063 (100.0) |
Medicaid expansion status | ||
|---|---|---|---|
|
| |||
| Expansion 229921 (61.0) |
Non-expansion 147142 (39.0) |
||
|
| |||
| Patient characteristics | Sample size (percentage) | ||
|
| |||
| Age at diagnosis | |||
| 40–44 | 24015 (6.4) | 14400 (6.3) | 9615 (6.5) |
| 45–49 | 45181 (12.0) | 27319 (11.9) | 17862 (12.1) |
| 50–54 | 83056 (22.0) | 50580 (22) | 32476 (22.1) |
| 55–59 | 105869 (28.1) | 64415 (28) | 41454 (28.2) |
| 60–64 | 118942 (31.5) | 73207 (31.8) | 45735 (31.1) |
| Sex | |||
| Male | 229570 (60.9) | 140124 (60.9) | 89446 (60.8) |
| Female | 147493 (39.1) | 89797 (39.1) | 57696 (39.2) |
| Race/Ethnicity | |||
| Non-Hispanic white | 259108 (68.7) | 160958 (70.0) | 98150 (66.7) |
| Non-Hispanic black | 56781 (15.1) | 27912 (12.1) | 28869 (19.6) |
| Hispanic | 8362 (2.2) | 6394 (2.8) | 1968 (1.3) |
| Other | 52812 (14.0) | 34657 (15.1) | 18155 (12.3) |
| Insurance | |||
| Uninsured | 27232 (7.2) | 10232 (4.5) | 17000 (11.6) |
| Private | 242614 (64.3) | 151595 (65.9) | 91019 (61.9) |
| Medicaid | 54137 (14.4) | 37841 (16.5) | 16296 (11.1) |
| Medicare | 46105 (12.2) | 27304 (11.9) | 18801 (12.8) |
| Other government | 6975 (1.9) | 2949 (1.3) | 4026 (2.7) |
| Residence at diagnosis | |||
| Urban | 48721 (12.9) | 25730 (11.2) | 22991 (15.6) |
| Metropolitan | 311044 (82.5) | 194722 (84.7) | 116322 (79.1) |
| Rural | 6166 (1.6) | 2567 (1.1) | 3599 (2.5) |
| Unknown | 11132 (3) | 6902 (3.0) | 4230 (2.9) |
| Median income by zip code at diagnosis, $ | |||
| < $40,227 | 72571 (19.3) | 33899 (14.7) | 38672 (26.3) |
| $40,227 - $50,353 | 76195 (20.2) | 41093 (17.9) | 35102 (23.9) |
| $50,354 - $63,332 | 77441 (20.5) | 49196 (21.4) | 28245 (19.2) |
| >=$63,333 | 113314 (30.1) | 82720 (36.0) | 30594 (20.8) |
| Unknown | 37542 (10.0) | 23013 (10.0) | 14529 (9.9) |
| Less than high school education by zip code at diagnosis, % | |||
| >=17.6% | 85595 (22.7) | 44668 (19.4) | 40927 (27.8) |
| 10.9% - 17.5% | 91658 (24.3) | 52455 (22.8) | 39203 (26.6) |
| 6.3% - 10.8% | 90212 (23.9) | 59539 (25.9) | 30673 (20.9) |
| < 6.3% | 72746 (19.3) | 50683 (22.0) | 22063 (15.0) |
| Unknown | 36852 (9.8) | 22576 (9.8) | 14276 (9.7) |
|
| |||
| Clinical factors | |||
|
| |||
| Year of diagnosis | |||
| 2010–2012 (pre-expansion) | 173842 (46.1) | 105995 (46.1) | 67847 (46.1) |
| 2014–2016 (post-expansion) | 203221 (53.9) | 123926 (53.9) | 79295 (53.9) |
| Primary cancer site | |||
| Anus | 13266 (3.5) | 8247 (3.6) | 5019 (3.4) |
| Colon | 119555 (31.7) | 71677 (31.2) | 47878 (32.5) |
| Esophagus | 23876 (6.3) | 15002 (6.5) | 8874 (6.0) |
| Rectosigmoid junction | 17915 (4.8) | 11139 (4.8) | 6776 (4.6) |
| Liver | 44626 (11.8) | 26395 (11.5) | 18231 (12.4) |
| Pancreas | 58768 (15.6) | 35932 (15.6) | 22836 (15.5) |
| Rectum | 55593 (14.7) | 34370 (15.0) | 21223 (14.4) |
| Small intestine | 12825 (3.4) | 7703 (3.4) | 5122 (3.5) |
| Stomach | 30639 (8.1) | 19456 (8.5) | 11183 (7.6) |
| Stage at diagnosis | |||
| Stage I | 83954 (22.3) | 51991 (22.6) | 31963 (21.7) |
| Stage II | 85714 (22.7) | 51950 (22.6) | 33764 (23.0) |
| Stage III | 95803 (25.4) | 58254 (25.3) | 37549 (25.5) |
| Stage IV | 111592 (29.6) | 67726 (29.5) | 43866 (29.8) |
| Comorbidity score | |||
| 0 | 276915 (73.4) | 170942 (74.4) | 105973 (72.0) |
| 1 | 67422 (17.9) | 39685 (17.3) | 27737 (18.9) |
| 2 | 16256 (4.3) | 9643 (4.2) | 6613 (4.5) |
| 3 | 16470 (4.4) | 9651 (4.2) | 6819 (4.6) |
| Receipt of palliative care | |||
| No or unknown | 351171 (93.1) | 213769 (93.0) | 137402 (93.4) |
| Yes | 25892 (6.9) | 16152 (7.0) | 9740 (6.6) |
|
| |||
| Hospital or health system features | |||
|
| |||
| Type of cancer program | |||
| Community | 153220 (40.6) | 89529 (38.9) | 63691 (43.3) |
| Academic/research | 151803 (40.3) | 100325 (43.6) | 51478 (35.0) |
| Integrated network cancer program | 72040 (19.1) | 40067 (17.4) | 31973 (21.7) |
| Facility location | |||
| New England | 19814 (5.3) | 17762 (7.7) | 2052 (1.4) |
| Middle Atlantic | 57011 (15.1) | 56593 (24.6) | 418 (0.3) |
| South Atlantic | 81576 (21.6) | 13623 (5.9) | 67953 (46.2) |
| East North Central | 63163 (16.8) | 55655 (24.2) | 7508 (5.1) |
| East South Central | 27730 (7.4) | 7488 (3.3) | 20242 (13.8) |
| West North Central | 27951 (7.4) | 12852 (5.6) | 15099 (10.3) |
| West South Central | 37038 (9.8) | 7158 (3.1) | 29880 (20.3) |
| Mountain | 16262 (4.3) | 12479 (5.4) | 3783 (2.6) |
| Pacific | 46518 (12.3) | 46311 (20.1) | 207 (0.1) |
Insurance status changes
Upon further comparisons of the examined characteristics between the pre-and post-expansion periods, the percentage of uninsured patients decreased more in the ME cohort (from 6.3% to 2.9%) compared with that in the non-ME cohort (12.9% to 10.4%; see Table 2). There was a significant increase in patients covered by Medicaid in the ME cohort (from 13.9% to 18.6%).
Table 2.
Characteristics of patients with gastrointestinal cancer by state Medicaid expansion status, pre- and post-expansion periods
| Expansion states (N = 229,921) | Non-expansion states (N = 147,142) | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| Pre-ME (2010–2012) 105995 (28.1) | Post-ME (2014–2016) 123926 (32.9) | p value | Pre-ME (2010–2012) 67847 (18.0) | Post-ME (2014–2016) 79295 (21.0) | p value | |
|
| ||||||
| Patient characteristics | Sample size (percentage) | |||||
|
| ||||||
| Age | <0.001 | <0.001 | ||||
| 40–44 | 6857 (6.5) | 7543 (6.1) | 4496 (6.6) | 5119 (6.5) | ||
| 45–49 | 13174 (12.4) | 14145 (11.4) | 8726 (12.9) | 9136 (11.5) | ||
| 50–54 | 23970 (22.6) | 26610 (21.5) | 15289 (22.5) | 17187 (21.7) | ||
| 55–59 | 29446 (27.8) | 34969 (28.2) | 19069 (28.1) | 22385 (28.2) | ||
| Sex | 0.208 | 0.884 | ||||
| Male | 64745 (61.1) | 75379 (60.8) | 41257 (60.8) | 48189 (60.8) | ||
| Female | 41250 (38.9) | 48547 (39.2) | 26590 (39.2) | 31106 (39.2) | ||
| Race/Ethnicity | <0.001 | 0.016 | ||||
| Non-Hispanic White | 74120 (69.9) | 86838 (70.1) | 45221 (66.7) | 52929 (66.8) | ||
| Non-Hispanic Black | 12658 (11.9) | 15254 (12.3) | 13290 (19.6) | 15579 (19.7) | ||
| Hispanic | 2879 (2.7) | 3515 (2.8) | 978 (1.4) | 990 (1.3) | ||
| Other | 16338 (15.4) | 18319 (14.8) | 8358 (12.3) | 9797 (12.4) | ||
| Insurance | <0.001 | <0.001 | ||||
| Uninsured | 6660 (6.3) | 3572 (2.9) | 8731 (12.9) | 8269 (10.4) | ||
| Private | 70872 (66.9) | 80723 (65.1) | 41199 (60.7) | 49820 (62.8) | ||
| Medicaid | 14769 (13.9) | 23072 (18.6) | 7729 (11.4) | 8567 (10.8) | ||
| Medicare | 12327 (11.6) | 14977 (12.1) | 8313 (12.3) | 10488 (13.2) | ||
| Other government | 1367 (1.3) | 1582 (1.3) | 1875 (2.8) | 2151 (2.7) | ||
| Residence at diagnosis | 0.445 | 0.017 | ||||
| Urban | 11843 (11.2) | 13887 (11.2) | 10758 (15.9) | 12233 (15.4) | ||
| Metropolitan | 89765 (84.7) | 104957 (84.7) | 53462 (78.8) | 62860 (79.3) | ||
| Rural | 1215 (1.2) | 1352 (1.1) | 1710 (2.5) | 1889 (2.4) | ||
| Unknown | 3172 (3.0) | 3730 (3.0) | 1917 (2.8) | 2313 (2.9) | ||
| Median income by zip code at diagnosis, $ | 0.578 | 0.77 | ||||
| < $40,227 | 16014 (15.1) | 17885 (14.4) | 18308 (27.0) | 20364 (25.7) | ||
| $40,227 - $50,353 | 19478 (18.4) | 21615 (17.4) | 16512 (24.3) | 18590 (23.4) | ||
| $50,354 - $63,332 | 23183 (21.9) | 26013 (21.0) | 13286 (19.6) | 14959 (18.9) | ||
| >=$63,333 | 39293 (37.1) | 43427 (35.0) | 14476 (21.3) | 16118 (20.3) | ||
| Unknown | 8027 (7.6) | 14986 (12.1) | 5265 (7.8) | 9264 (11.7) | ||
| Less than high school education by zip code at diagnosis, % | 0.001 | 0.774 | ||||
| >=17.6% | 20804 (19.6) | 23864 (19.3) | 19278 (28.4) | 21649 (27.3) | ||
| 10.9% - 17.5% | 24814 (23.4) | 27641 (22.3) | 18514 (27.3) | 20689 (26.1) | ||
| 6.3% - 10.8% | 28344 (26.7) | 31195 (25.2) | 14434 (21.3) | 16239 (20.5) | ||
| < 6.3% | 24222 (22.9) | 26461 (21.4) | 10476 (15.4) | 11587 (14.6) | ||
| Unknown | 7811 (7.4) | 14765 (11.9) | 5145 (7.6) | 9131 (11.5) | ||
|
| ||||||
| Clinical factors | ||||||
|
| ||||||
| GI cancer site | <0.001 | <0.001 | ||||
| Anus | 3782 (3.6) | 4465 (3.6) | 2277 (3.4) | 2742 (3.5) | ||
| Colon | 33009 (31.1) | 38668 (31.2) | 22395 (33.0) | 25483 (32.1) | ||
| Esophagus | 7154 (6.8) | 7848 (6.3) | 4175 (6.2) | 4699 (5.9) | ||
| Rectosigmoid junction | 5341 (5.0) | 5798 (4.7) | 3191 (4.7) | 3585 (4.5) | ||
| Liver | 12227 (11.5) | 14168 (11.4) | 8173 (12.1) | 10058 (12.7) | ||
| Pancreas | 16230 (15.3) | 19702 (15.9) | 10425 (15.4) | 12411 (15.7) | ||
| Rectum | 15588 (14.7) | 18782 (15.2) | 9614 (14.2) | 11609 (14.6) | ||
| Small intestine | 3482 (3.3) | 4221 (3.4) | 2428 (3.6) | 2694 (3.4) | ||
| Stomach | 9182 (8.7) | 10274 (8.3) | 5169 (7.6) | 6014 (7.6) | ||
| Stage at diagnosis | 0.003 | <0.001 | ||||
| Stage I | 23671 (22.3) | 28320 (22.9) | 14901 (22.0) | 17062 (21.5) | ||
| Stage II | 24242 (22.9) | 27708 (22.4) | 15926 (23.5) | 17838 (22.5) | ||
| Stage III | 26918 (25.4) | 31336 (25.3) | 17359 (25.6) | 20190 (25.5) | ||
| Stage IV | 31164 (29.4) | 36562 (29.5) | 19661 (29.0) | 24205 (30.5) | ||
| Comorbidity score | <0.001 | <0.001 | ||||
| 0 | 79558 (75.1) | 91384 (73.7) | 48754 (71.9) | 57219 (72.2) | ||
| 1 | 18337 (17.3) | 21348 (17.2) | 13367 (19.7) | 14370 (18.1) | ||
| 2 | 4030 (3.8) | 5613 (4.5) | 2874 (4.2) | 3739 (4.7) | ||
| 3 | 4070 (3.8) | 5581 (4.5) | 2852 (4.2) | 3967 (5.0) | ||
| Palliative care receipt | <0.001 | <0.001 | ||||
| No or unknown | 99254 (93.6) | 114515 (92.4) | 63690 (93.9) | 73712 (93.0) | ||
| Yes | 6741 (6.4) | 9411 (7.6) | 4157 (6.1) | 5583 (7.0) | ||
| Mean time to treatment, days | 28.9 ± 43.4 | 30.9 ± 39.8 | <0.001 | 25.7 ± 38.1 | 28.8 ± 39.0 | <0.001 |
|
| ||||||
| Hospital or health system features | ||||||
|
| ||||||
| Type of cancer program | <0.001 | <0.001 | ||||
| Community | 40872 (38.6) | 48657 (39.3) | 29033 (42.8) | 34658 (43.7) | ||
| Academic/research | 46239 (43.6) | 54086 (43.6) | 23781 (35.1) | 27697 (34.9) | ||
| Integrated network cancer program | 18884 (17.8) | 21183 (17.1) | 15033 (22.2) | 16940 (21.4) | ||
| Facility location | <0.001 | 0.006 | ||||
| New England | 8227 (7.8) | 9535 (7.7) | 915 (1.4) | 1137 (1.4) | ||
| Middle Atlantic | 26369 (24.9) | 30224 (24.4) | 216 (0.3) | 202 (0.3) | ||
| South Atlantic | 6498 (6.1) | 7125 (5.8) | 31557 (46.5) | 36396 (45.9) | ||
| East North Central | 25895 (24.4) | 29760 (24) | 3432 (5.1) | 4076 (5.1) | ||
| East South Central | 3453 (3.3) | 4035 (3.3) | 9227 (13.6) | 11015 (13.9) | ||
| West North Central | 6148 (5.8) | 6704 (5.4) | 7060 (10.4) | 8039 (10.1) | ||
| West South Central | 3036 (2.9) | 4122 (3.3) | 13581 (20) | 16299 (20.6) | ||
| Mountain | 5613 (5.3) | 6866 (5.5) | 1757 (2.6) | 2026 (2.6) | ||
| Pacific | 20756 (19.6) | 25555 (20.6) | 102 (0.2) | 105 (0.1) | ||
ME, Medicaid Expansion; GI, gastrointestinal
Overall two-year survival
On K-M survival curves, we found significantly different two-year overall survival on the basis of ME status and time period (see Figure 2). Overall, GI cancer patients in the non-ME cohort had worse survival compared with those in the ME cohort. In the adjusted K-M curves depicting overall two-year survival by stage at diagnosis, we found that individuals from non-ME states post-expansion had the lowest survival for stage I or II GI cancer (see Figure 3). The difference was less pronounced in later stage cases (stages III and IV). Finally, in the adjusted K-M curves by cancer type, we see that individuals from non-ME states post-expansion had the lowest survival if diagnosed with GI cancer of the anus, colon, rectosigmoid junction, rectum, and small intestine (see Figure 4).
Figure 2. Adjusted Kaplan-Meier overall two-year survival by Medicaid expansion status, pre- and post-expansion.
Figure 3. Adjusted Kaplan-Meier overall two-year survival by stage at diagnosis and Medicaid expansion status, pre- and post-expansion.
Figure 4. Adjusted Kaplan-Meier overall two-year survival by GI cancer type and Medicaid expansion status, pre- and post-expansion.
The DD adjusted multivariable Cox regression model showed that no significant difference was found in the overall two-year survival between the two cohorts (DD HR, 1.00; 95% CI, 0.98–1.02) (see Table 3). Additionally, we did not observe any significant differences by cancer site in DD analysis. In the subgroup analysis by cancer stage, we found that two-year survival was significantly improved among patients diagnosed with stage II cancer in ME states, with an 8% decreased hazard of death at two years (DD HR, 0.92; 95% CI, 0.87–0.97). On the contrary, patients diagnosed with stage IV cancer had a 4% increased hazard of death at two years (DD HR, 1.04; 95% CI, 1.01–1.07). No significant differences were detected for those diagnosed with stage I or III GI cancer.
Table 3.
Change in two-year overall survival for all GI cancers, by GI cancer type, and by stage at diagnosis
| Expansion States | Non-Expansion States | Hazard Ratio | ||||||
|---|---|---|---|---|---|---|---|---|
| Pre-ME (%) | Post-ME (%) | Absolute Difference (%) | Pre-ME (%) | Post-ME (%) | Absolute Difference (%) | Crude Difference-in-Difference a | Adjusted Difference-in-Difference b | |
|
| ||||||||
| All GI cancers | 61.49 | 60.55 | −0.94 (−1.34, −0.54) | 59.52 | 57.48 | −2.04 (−2.55, −1.54) | 0.97 (0.95–0.99) | 1.00 (0.98–1.02) |
| By GI cancer type | ||||||||
| Anus | 81.72 | 78.09 | −3.63 (−5.36, −1.9) | 80.15 | 75.93 | −4.22 (−6.51, −1.93) | 0.99 (0.85–1.16) | 0.92 (0.79–1.08) |
| Colon | 77.20 | 74.26 | −2.94 (−3.57, −2.31) | 75.81 | 71.27 | −4.53 (−5.32, −3.74) | 0.95 (0.90–0.99) | 0.98 (0.93–1.03) |
| Esophagus | 38.84 | 38.25 | −0.59 (−2.15, 0.97) | 35.33 | 35.22 | −0.11 (−2.1, 1.88) | 1.00 (0.94–1.07) | 1.06 (0.99–1.14) |
| Rectosigmoid junction | 81.45 | 77.44 | −4.01 (−5.51, −2.52) | 78.10 | 74.62 | −3.48 (−5.5, −1.46) | 1.05 (0.92–1.20) | 1.06 (0.92–1.21) |
| Liver | 41.56 | 43.43 | 1.87 (0.67, 3.06) | 37.18 | 38.36 | 1.19 (−0.23, 2.6) | 0.98 (0.93–1.03) | 0.99 (0.94–1.04) |
| Pancreas | 25.58 | 30.21 | 4.63 (3.7, 5.55) | 23.96 | 27.64 | 3.68 (2.54, 4.81) | 0.98 (0.94–1.02) | 1.00 (0.96–1.04) |
| Rectum | 83.02 | 79.92 | −3.10 (−3.92, −2.28) | 80.62 | 76.98 | −3.64 (−4.74, −2.54) | 0.99 (0.92–1.07) | 1.01 (0.94–1.10) |
| Small intestine | 80.64 | 77.23 | −3.42 (−5.24, −1.59) | 80.23 | 75.61 | −4.62 (−6.89, −2.35) | 0.94 (0.81–1.09) | 0.91 (0.78–1.07) |
| Stomach | 48.89 | 48.41 | −0.49 (−1.89, 0.92) | 46.01 | 45.58 | −0.43 (−2.28, 1.42) | 1.00 (0.94–1.07) | 1.03 (0.96–1.10) |
| By stage | ||||||||
| Stage I-III | 75.14 | 73.65 | −1.49 (−1.92, −1.07) | 72.81 | 70.22 | −2.59 (−3.14, −2.04) | 0.96 (0.93–0.99) | 0.95 (0.92–0.98) |
| Stage I | 84.11 | 80.89 | −3.22 (−3.88, −2.57) | 80.95 | 77.10 | −3.85 (−4.74, −2.96) | 1.00 (0.94–1.06) | 0.98 (0.91–1.05) |
| Stage II | 72.87 | 71.75 | −1.13 (−1.9, −0.35) | 71.89 | 68.62 | −3.27 (−4.25, −2.3) | 0.92 (0.88–0.97) | 0.92 (0.87–0.97) |
| Stage III | 69.29 | 68.79 | −0.50 (−1.25, 0.25) | 66.68 | 65.84 | −0.84 (−1.8, 0.11) | 0.99 (0.95–1.04) | 0.97 (0.93–1.02) |
| Stage IV | 28.71 | 29.26 | 0.55 (−0.13, 1.24) | 26.94 | 28.47 | 1.53(0.69, 2.37) | 1.02 (1.00–1.05) | 1.04 (1.01–1.07) |
ME, Medicaid Expansion; GI, gastrointestinal
Crude Difference-in-Difference model adjusted for ME status, year of diagnosis, and Difference-in-Difference estimator to estimate the crude hazard ratio.
For all GI cancers, the adjusted Difference-in-Difference model adjusted for ME status, year of diagnosis, and Difference-in-Difference estimator and patient characteristics (age, sex, race/ethnicity, Charlson-Deyo comorbidity score, health insurance, residence, annual household income, educational attainment), clinical factors (primary site, stage at diagnosis, comorbidity score, and receipt of palliative care), and hospital or health system features (cancer program type, facility location) to estimate the adjusted hazard ratio. By GI cancer type, model was adjusted for all factors listed previously except GI cancer type. For cancer stage, the model was adjusted for all factors listed previously except cancer stage.
Time to treatment initiation
There are notable changes in TTI received within 30- and 90-days between the pre-expansion and post-expansion periods for ME and non-ME cohorts (see Table 4). We observed decreases in the absolute percentage of patients receiving early treatment within 30 days in both the ME and non-ME states. Results from the linear DD regression models showed a significant improvement in the proportion of patients who experienced TTI within 30-days and 90-days in ME cohort compared with non-ME cohort after 2014 for all cancer sites (30-day OR, 1.12; 95% CI, 1.09–1.16; 90-day OR, 1.22; 95% CI, 1.17–1.27). Although both cohorts experienced a decrease in the proportion of individuals receiving TTI within 30- or 90-days compared to before 2014, the ME cohort experienced an increase in the odds of receiving TTI within 30- and 90- days than those in the non-ME cohort.
Table 4.
Proportion of patients who received cancer-related treatment within 30 days and 90 days from diagnosis, overall, by GI cancer type, and by stage at diagnosis
| Expansion States | Non-Expansion | Odds Ratio | ||||||
|---|---|---|---|---|---|---|---|---|
| Pre-ME (%) | Post-ME (%) | Absolute Difference (%) | Pre-ME (%) | Post-ME (%) | Absolute Difference (%) | Crude Difference-in-Difference a | Adjusted Difference-in-Difference b | |
|
| ||||||||
| Treatment within 30 days | ||||||||
|
| ||||||||
| All GI cancers | 57.02 | 54.56 | −2.46 (−2.87, −2.05) | 61.88 | 57.00 | −4.89 (−5.39, −4.38) | 1.11 (1.08–1.14) | 1.12 (1.09–1.16) |
| By GI cancer type | ||||||||
| Anus | 53.60 | 47.44 | −6.16 (−8.32, −4.00) | 57.53 | 52.55 | −4.98 (−7.74, −2.22) | 0.96 (0.91–1.21) | 1.03 (0.83–1.13) |
| Colon | 77.58 | 75.04 | −2.55 (−3.17, −1.93) | 82.22 | 78.03 | −4.20 (−4.91, −3.48) | 1.13 (1.07–1.20) | 1.17 (1.10–1.24) |
| Esophagus | 42.49 | 41.11 | −1.39 (−2.97, 0.19) | 46.47 | 43.78 | −2.69 (−4.77, −0.62) | 1.05 (0.95–1.17) | 1.09 (0.97–1.23) |
| Rectosigmoid junction | 65.21 | 61.62 | −3.59 (−5.38, −1.80) | 70.6 | 65.05 | −5.56 (−7.78, −3.34) | 1.11 (0.97–1.26) | 1.09 (0.95–1.25) |
| Liver | 24.81 | 21.99 | −2.83 (−3.85, −1.80) | 27.48 | 22.54 | −4.94 (−6.21, −3.68) | 1.11 (1.02–1.21) | 1.16 (1.06–1.28) |
| Pancreas | 48.98 | 49.10 | 0.12 (−0.92, 1.16) | 51.93 | 50.05 | −1.88 (−3.18, −0.58) | 1.08 (1.01–1.16) | 1.11 (1.03–1.19) |
| Rectum | 53.78 | 49.46 | −4.32 (−5.38, −3.26) | 60.79 | 53.11 | −7.67 (−9.01, −6.34) | 1.15 (1.07–1.23) | 1.14 (1.06–1.23) |
| Small intestine | 67.09 | 64.94 | −2.15 (−4.27, −0.03) | 69.32 | 64.92 | −4.39 (−6.97, −1.82) | 1.11 (0.95–1.29) | 1.20 (1.02–1.42) |
| Stomach | 49.88 | 47.39 | −2.49 (−3.89, −1.08) | 55.77 | 51.38 | −4.39 (−6.25, −2.54) | 1.08 (0.98–1.19) | 1.11 (1.00–1.22) |
| By cancer stage | ||||||||
| Stage I-III | 59.04 | 55.55 | −3.49 (−3.97, −3.00) | 64.01 | 58.45 | −5.56 (−6.15, −4.97) | 1.10 (1.06–1.13) | 1.12 (1.08–1.16) |
| Stage I | 57.33 | 54.46 | −2.87 (−3.73, −2.02) | 60.87 | 55.52 | −5.35 (−6.44, −4.27) | 1.11 (1.05–1.17) | 1.12 (1.04–1.19) |
| Stage II | 59.38 | 55.78 | −3.60 (−4.45, −2.75) | 65.04 | 59.02 | −6.02 (−7.05, −4.98) | 1.11 (1.05–1.18) | 1.10 (1.03–1.17) |
| Stage III | 60.22 | 56.33 | −3.89 (−4.69, −3.09) | 65.76 | 60.42 | −5.34 (−6.31, −4.36) | 1.07 (1.02–1.13) | 1.14 (1.07–1.21) |
| Stage IV | 52.19 | 52.21 | 0.02 (−0.74, 0.77) | 56.68 | 53.70 | −2.98 (−3.91, −2.04) | 1.13 (1.08–1.18) | 1.15 (1.09–1.21) |
|
| ||||||||
| Treatment within 90 days | ||||||||
|
| ||||||||
| All GI cancers | 82.45 | 83.99 | 1.54 (1.23, 1.84) | 84.67 | 83.58 | −1.09 (−1.46, −0.72) | 1.21 (1.17–1.26) | 1.22 (1.17–1.27) |
| By GI cancer type | ||||||||
| Anus | 91.94 | 92.81 | 0.88 (−0.28, 2.03) | 93.94 | 93.36 | −0.58 (−1.93, 0.78) | 1.25 (0.94–1.65) | 1.33 (0.98–1.81) |
| Colon | 92.47 | 93.35 | 0.88 (0.50, 1.25) | 94.81 | 93.87 | −0.94 (−1.35, −0.52) | 1.36 (1.24–1.5) | 1.37 (1.23–1.52) |
| Esophagus | 82.01 | 84.35 | 2.34 (1.14, 3.54) | 83.98 | 83.68 | −0.3 (−1.83, 1.24) | 1.21 (1.05–1.39) | 1.22 (1.04–1.42) |
| Rectosigmoid junction | 89.78 | 91.72 | 1.94 (0.87, 3.02) | 92.29 | 92.75 | 0.46 (−0.80, 1.71) | 1.18 (0.95–1.47) | 1.24 (0.97–1.58) |
| Liver | 54.34 | 56.11 | 1.77 (0.57, 2.98) | 55.85 | 54.81 | −1.04 (−2.49, 0.41) | 1.12 (1.04–1.21) | 1.14 (1.05–1.24) |
| Pancreas | 71.81 | 74.55 | 2.74 (1.82, 3.66) | 74.05 | 73.94 | −0.11 (−1.25, 1.03) | 1.16 (1.07–1.25) | 1.16 (1.07–1.26) |
| Rectum | 89.68 | 90.86 | 1.19 (0.55, 1.82) | 91.95 | 90.44 | −1.51 (−2.27, −0.75) | 1.38 (1.23–1.56) | 1.36 (1.19–1.54) |
| Small intestine | 86.73 | 86.69 | −0.05 (−1.57, 1.48) | 88.18 | 86.27 | −1.91 (−3.74, −0.09) | 1.18 (0.96–1.46) | 1.20 (0.95–1.52) |
| Stomach | 80.96 | 83.13 | 2.17 (1.09, 3.25) | 84.31 | 83.54 | −0.77 (−2.14, 0.59) | 1.23 (1.08–1.39) | 1.27 (1.11–1.45) |
| By stage | ||||||||
| Stage I-III | 86.00 | 87.14 | 1.14 (0.81, 1.47) | 88.10 | 86.86 | −1.24 (−1.65, −0.84) | 1.24 (1.18–1.29) | 1.26 (1.20–1.34) |
| Stage I | 82.45 | 83.64 | 1.19 (0.54, 1.83) | 84.33 | 82.75 | −1.58 (−2.39, −0.76) | 1.22 (1.13–1.32) | 1.25 (1.14–1.36) |
| Stage II | 87.03 | 88.35 | 1.32 (0.75, 1.89) | 89.83 | 88.50 | −1.34 (−2.00, −0.67) | 1.3 (1.19–1.42) | 1.28 (1.16–1.41) |
| Stage III | 88.20 | 89.24 | 1.04 (0.53, 1.56) | 89.76 | 88.09 | −0.86 (−1.49, −0.24) | 1.21 (1.12–1.32) | 1.30 (1.18–1.43) |
| Stage IV | 73.92 | 76.45 | 2.52 (1.87, 3.18) | 76.26 | 76.11 | −0.15 (−0.95, 0.66) | 1.15 (1.09–1.22) | 1.17 (1.10–1.24) |
ME, Medicaid Expansion; GI, gastrointestinal
Crude Difference-in-Difference model adjusted for ME status, year of diagnosis, and Difference-in-Difference estimator to estimate the crude odds ratio.
For all GI cancers, the adjusted Difference-in-Difference model adjusted for ME status, year of diagnosis, and Difference-in-Difference estimator and patient characteristics (age, sex, race/ethnicity, Charlson-Deyo comorbidity score, health insurance, residence, annual household income, educational attainment), clinical factors (primary site, stage at diagnosis, comorbidity score, and receipt of palliative care), and hospital or health system features (cancer program type, facility location) to estimate the adjusted odds ratio. By GI cancer type, model was adjusted for all factors listed previously except GI cancer type. For cancer stage, the model was adjusted for all factors listed previously except cancer stage.
In the adjusted subgroup analysis by cancer site, we found significantly higher likelihoods of receiving TTI within 30-day among patients diagnosed with colon, liver, pancreas, rectum, small intestine, and stomach in the ME cohort compared with non-ME cohort after 2014. As for TTI received within 90-days, significant positive differences were found for those diagnosed with colon, esophagus, liver, pancreas, rectum, and stomach cancers. As for subgroup analysis by cancer stage, significantly higher likelihoods were found across all cancer stages in the ME cohort compared with non-ME cohort after 2014 for both TTI within 30- and 90-days.
DISCUSSION
In this study, we analyzed the National Cancer Database and found an association between ME and improved two-year survival in patients with early-stage GI cancer who resided in ME states. We also found that ME was associated with significantly improved TTI within 30- and 90-days for patients diagnosed with GI cancers in ME states compared to non-ME states after 2014. Our study confirms previous research on ME and provides further evidence of the positive impact of ME on key outcomes for those diagnosed with GI cancer.4–8,22,23 Our findings underscore the overall positive impact of Medicaid expansion on those diagnosed with GI cancers. Because TTI within 30- and 90-days were significantly improved for those from ME states diagnosed post-expansion, it is likely that these factors have programmatic or policy-based reasons for improvement in TTI. Resources at the local and state levels, individuals’ ability to access and use the care they need, and insurance differences and benefit coverage are all potential drivers of variation in TTI between expansion and non-expansion states.
Despite the positive impact of ME on TTI and early-stage two-year overall survival, we observed an overall decrease in two-year survival over time across all states, regardless of ME status. This suggests that ME may improve access to treatment but may not necessarily improve overall survival outcomes for patients with GI cancer. Health care providers and policymakers should keep this crucial aspect in mind as they strive to enhance patient outcomes and improve care accessibility for GI cancer patients. One possible explanation for these findings is that in states that adopted ME, there may have been an increase in the number of patients seeking screening (e.g., colonoscopy), but there may not have been a corresponding increase in the number of medical professionals available to provide those treatments in time. This could lead to delays in treatment and poorer overall outcomes for patients.24
We have also observed a decrease in the number of patients receiving early treatment (i.e., TTI within 30 days of diagnosis). This decrease could be attributed to insurance prior authorization that limits or delays access to treatment, while also adding to the growing administrative burden for health care providers. The impact of these authorizations on cancer care can be significant, resulting in treatment delays, increased stress anxiety, and out-of-pocket costs for patients, and reduced quality of care from care team providers.25 For example, if a patient requires chemotherapy, but their insurance company requires prior authorization, the approval process can take weeks or even months. This waiting period can be stressful and uncertain for the patient and their care team members, potentially leading to disease progression and reduced treatment effectiveness, ultimately negatively impacting quality of life and survival for these vulnerable patients. As such, it is crucial for insurance companies to consider the impact of prior authorization on cancer outcomes and streamline the process to minimize delays and stress for patients.
Additionally, we examined differences in survival outcomes across different types of GI cancer. Unlike prior studies, we did not observe an improvement in survival among patients diagnosed with colorectal cancer, potentially indicating a limitation in improving this type of cancer outcome, which may require more time post-implementation to see more significant improvements in survival due to improved access to care, screening, and engagement in the health system.26,27 Overall, our findings suggest that while ME can improve access to treatment for patients with GI cancer, it may not necessarily lead to improved survival outcomes. These findings highlight the need for further research and personalized, targeted interventions to improve survival for patients with GI cancer.
It should be noted that this study has limitations. Although we employed a rigorous quasi-experimental study design, this study was a retrospective, observational study. Thus, causal conclusions cannot be drawn regarding Medicaid expansion, overall two-year survival and 30- and 90-day TTI. Secondly, although the NCDB is an extensive database, it does not capture all potential confounding factors; for example, public use file of NCDB does not include information about specific patient comorbidities or social determinants of health, which likely impact TTI and overall two-year survival. Also, we are limited in the conclusions we can draw from the data collected on overall survival, as the NCDB does not delineate between cancer-specific mortality and mortality caused by other factors. This limits our ability to draw stronger cancer-specific conclusions about survival associated with ME. Moreover, the NCDB is a voluntary reporting system, which means that not all cancer cases are included in the database. This can affect the generalizability of the results of the study due to selection bias, in addition to unobserved confounders that are not measured given the nature of secondary data analysis. Although, the NCDB does cover 70% of Americans diagnosed with cancer. Additionally, GI cancers capture a heterogeneous group of malignancies, which different types of GI cancers having different treatment regimens and associated outcomes. Although we attempted to account for this by conducting subgroup analyses by GI cancer site, the results should be interpreted with caution as this may not reflect the experience of all people diagnosed with GI cancers following Medicaid expansion. Finally, we accounted for changes in the pre-expansion period by examining the parallel trends assumption. We found that there was not a significant violation of parallel trends observed in overall survival, but there were limited violations in TTI at 30- and 90-days, albeit minor (see Supplemental File Tables S-1 and S-2).
CONCLUSIONS
The expansion of Medicaid under the Affordable Care Act was not associated with a significant improvement in overall two-year survival for those diagnosed with GI cancer. However, early stage at diagnosis remains the mainstay leading to improved survival in ME states. Furthermore, although TTI increased after ME for both cohorts, the odds of accessing timely treatment at both 30- and 90-days was higher for those from ME compared to non-ME states. Future research should focus on identifying the drivers of decreased survival among late-stage GI cancer patients in ME states and investigate potential solutions. Additionally, future work should examine inequities in survival among specific subgroups (e.g., by race, ethnicity, insurance) to determine the presence of disparities. This information should be used to develop future interventions to target those subgroups who experience disparities using a personalized, tailored approach.
Supplementary Material
Synopsis for table of contents:
We evaluated whether Medicaid expansion (ME) was associated with improved two-year survival and time to treatment initiation (TTI) among those with gastrointestinal cancer. ME was not associated with improved survival; TTI increased after ME for both cohorts, yet, the odds of timely treatment were higher for those from ME versus non-ME states. The findings should be used to develop future interventions to target those subgroups of people diagnosed with gastrointestinal cancer who experience disparities using a personalized approach.
Acknowledgements
This study used data provided by the National Cancer Database (NCDB). The authors acknowledge the efforts of the American College of Surgeons, the NCDB, and cancer registrars in the creation of these data. We also want to thank the funders of this work, the American Gastroenterological Association and the National Institute on Aging, both of which made this work possible.
Funding
This study was partly supported by the American Gastroenterological Association Award in Digestive Health Disparities (AGA202022-21-05; PI: Hong). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Author Erin Mobley was supported by the National Institutes of Health, National Institute on Aging internal career development award at the University of Florida College of Medicine Jacksonville (grant number 5R33AG056540, JAX-ASCENT Junior Scholar Award).
Footnotes
Future meeting presentation: Preliminary results of this manuscript will be presented at the American College of Surgeons 109th Annual Clinical Congress, Scientific Forum, Boston, MA, October 2023 as an oral presentation.
SUBMISSION DETAILS
Author disclosure statement
The listed authors fulfill all authorship requirements as stated in the Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly work in Medical Journals and have no competing financial interests that exist.
Roles
EMM: conceptualization, investigation, funding acquisition, methodology, validation, visualization, writing - original draft, and writing - review and editing. GC: formal analysis, methodology, validation, visualization, investigation, writing - original draft, writing - review and editing. JX: investigation, writing - review and editing. LE: investigation, writing - review and editing. KP: investigation, writing - review and editing. MCD: investigation, writing - review and editing. ZTA: investigation, writing - review and editing. ASP: investigation, writing - review and editing. ZX: investigation, writing - review and editing. RS: investigation, writing - review and editing. SM: investigation, writing - review and editing. YRH: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing - original draft, and writing - review and editing.
REFERENCES
- 1.Patient Protection and Affordable Care Act of 2010. Public Law No: 111–148. [Google Scholar]
- 2.Kaiser Family Foundation. Status of State Action on the Medicaid Expansion Decision. Updated May 8. Accessed June 2, 2023. https://www.kff.org/health-reform/state-indicator/state-activity-around-expanding-medicaid-under-the-affordable-care-act/
- 3.Mazurenko Olena, Balio Casey P., Agarwal Rajender, Carroll Aaron E., Menachemi N. The Effects Of Medicaid Expansion Under The ACA: A Systematic Review. Health Affairs. 2018;37(6):944–950. doi: 10.1377/hlthaff.2017.1491 [DOI] [PubMed] [Google Scholar]
- 4.Mobley EM, Guerrier C, Tfirn I, et al. Impact of Medicaid Expansion on Stage at Diagnosis for US Adults with Pancreatic Cancer: A Population-Based Study. Journal of Racial and Ethnic Health Disparities. 2023/01/March 2023;doi: 10.1007/s40615-022-01459-4 [DOI] [PubMed] [Google Scholar]
- 5.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. Dec 1 2018;4(12):1713–1720. doi: 10.1001/jamaoncol.2018.3467 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kim U, Koroukian S, Statler A, Rose J. The effect of Medicaid expansion among adults from low-income communities on stage at diagnosis in those with screening-amenable cancers. Cancer. 2020;126(18):4209–4219. doi: 10.1002/cncr.32895 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.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. Journal of Clinical Oncology. 2017;35(35):3906–3915. [DOI] [PubMed] [Google Scholar]
- 8.Soni A, Simon K, Cawley J, Sabik L. Effect of Medicaid Expansions of 2014 on Overall and Early-Stage Cancer Diagnoses. American Journal of Public Health. 2018;108(2):216–218. doi: 10.2105/ajph.2017.304166 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.World Health Organization. Guide to cancer early diagnosis. World Health Organization; 2017. [Google Scholar]
- 10.Zavala VA, Bracci PM, Carethers JM, et al. Cancer health disparities in racial/ethnic minorities in the United States. British Journal of Cancer. 2021/01/January 2021;124(2):315–332. doi: 10.1038/s41416-020-01038-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hoehn RS, Rieser CJ, Phelos H, et al. Association Between Medicaid Expansion and Diagnosis and Management of Colon Cancer. Journal of the American College of Surgeons. 2020;232(2):146–156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.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 Network Open. 2020;3(2):e1921653–e1921653. doi: 10.1001/jamanetworkopen.2019.21653 [DOI] [PubMed] [Google Scholar]
- 13.Han X, Xiong KZ, Kramer MR, Jemal A. The Affordable Care Act and Cancer Stage at Diagnosis Among Young Adults. Journal of the National Cancer Institute. 2016;108(9):1–9. [DOI] [PubMed] [Google Scholar]
- 14.Hong YR, Holcomb D, Bhandari M, Larkin L. Affordable care act: comparison of healthcare indicators among different insurance beneficiaries with new coverage eligibility. BMC Health Services Research. 2016/April/04 2016;16(1):114. doi: 10.1186/s12913-016-1362-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sommers BD, Kenney GM, Epstein AM. New Evidence On The Affordable Care Act: Coverage Impacts Of Early Medicaid Expansions. Health Affairs. 2014;33(1):78–87. doi: 10.1377/hlthaff.2013.1087 [DOI] [PubMed] [Google Scholar]
- 16.Ho PJ, Cook AR, Binte Mohamed Ri NK, Liu J, Li J, Hartman M. Impact of delayed treatment in women diagnosed with breast cancer: A population-based study. Cancer Medicine. 2020;9(7):2435–2444. doi: 10.1002/cam4.2830 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nguyen D-D, Paciotti M, Marchese M, et al. Effect of Medicaid Expansion on Receipt of Definitive Treatment and Time to Treatment Initiation by Racial and Ethnic Minorities and at Minority-Serving Hospitals: A Patient-Level and Facility-Level Analysis of Breast, Colon, Lung, and Prostate Cancer. JCO Oncology Practice. 2021;17(5):e654–e665. doi: 10.1200/op.21.00010 [DOI] [PubMed] [Google Scholar]
- 18.Economic Research S. Urban Influence Codes: Documentation. United States Department of Agriculture, Economic Research Service. Updated November 27. Accessed June 2, 2023. https://www.ers.usda.gov/data-products/urban-influence-codes/documentation/ [Google Scholar]
- 19.Economic Research Service. Rural Urban Commuting Code: Documentation. Accessed May 27, 2019. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/documentation/
- 20.U.S. Census Bureau. Data from: 2013–2017 American Community Survey 5 Year Estimates. US Census Bureau. 2017. [Google Scholar]
- 21.Mansfield ER, Helms BP. Detecting Multicollinearity. The American Statistician. 1982;36(3):158–160. doi: 10.2307/2683167 [DOI] [Google Scholar]
- 22.Lam MB, Phelan J, Orav EJ, Jha AK, Keating NL. Medicaid Expansion and Mortality Among Patients With Breast, Lung, and Colorectal Cancer. JAMA Network Open. 2020;3(11):e2024366–e2024366. doi: 10.1001/jamanetworkopen.2020.24366 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Barnes JM, Johnson KJ, Adjei Boakye E, et al. Early Medicaid Expansion and Cancer Mortality. JNCI: Journal of the National Cancer Institute. 2021;113(12):1714–1722. doi: 10.1093/jnci/djab135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Miller S, Wherry LR. Health and Access to Care during the First 2 Years of the ACA Medicaid Expansions. New England Journal of Medicine. 2017;376(10):947–956. doi: 10.1056/NEJMsa1612890 [DOI] [PubMed] [Google Scholar]
- 25.Office of the Inspector General. Some Medicare Advantage Organization Denials of Prior Authorization Requests Raise Concerns About Beneficiary Access to Medically Necessary Care, OEI-09–18-00260. 2022. US Department of Health and Human Services, Office of the Inspector General. [Google Scholar]
- 26.Gan T, Sinner HF, Walling SC, et al. Impact of the Affordable Care Act on Colorectal Cancer Screening, Incidence, and Survival in Kentucky. Journal of the American College of Surgeons. 2019;228(4):342–353e1. doi: 10.1016/j.jamcollsurg.2018.12.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nathan NH, Bakhsheshian J, Ding L, Mack WJ, Attenello FJ. Evaluating Medicaid expansion benefits for patients with cancer: National Cancer Database analysis and systematic review. Journal of Cancer Policy. 2021/September/01/ 2021;29:100292. doi: 10.1016/j.jcpo.2021.100292 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.