Association of State-Level Medicaid Expansion With Treatment of Patients With Higher-Risk Prostate Cancer

Key Points Question How is Medicaid expansion associated with the insurance status and treatment patterns among patients with higher-risk prostate cancer? Findings In this cohort study that included 15 332 men newly diagnosed with higher-risk prostate cancer, Medicaid coverage increased after expansion but was not associated with higher treatment rates at the population level after expansion or compared with uninsured patients. Meaning Although there was no difference in higher-risk prostate cancer treatment rates after Medicaid expansion in this study, it is important to continue addressing disparities in at-risk populations, given that insurance coverage alone is not sufficient to improve cancer care.


Introduction
Amid major advances in medicine and technology, inequity in access to health care services persists.
The Patient Protection and Affordable Care Act (ACA) increased health insurance coverage through multiple mechanisms, including permitting states to expand Medicaid eligibility at their discretion, with some states expanding prior to 2014 and some opting not to adopt expansion. ACA implementation took effect on January 1, 2014, and allowed states to expand coverage for childless adults younger than 65 years whose income was 133% of the federal poverty level or lower. To date, more than 20 million previously uninsured individuals have gained coverage, with the proportion of Americans without health insurance reaching a historic low. 1,2 Newly covered individuals disproportionately belonged to minority racial/ethnic groups, were older, had poorer health, and were more likely to delay care due to cost. [3][4][5] Racial disparities in all-cause mortality were reduced following increased insurance coverage. 1,4,6 After ACA implementation, there were decreases in the proportion of uninsured patients among lower-income residents in expansion states 5,6 along with increased insurance coverage and health care utilization for low-income adults. 7 Increased insurance coverage has the potential to improve access to cancer care and possibly reduce persistent disparities in cancer survival based on race/ethnicity 8 and socioeconomic status. 9 These disparities are driven partly by a lack of health insurance, which affects receipt of cancer screening, 10 resulting in higher rates of advanced malignant neoplasms. [11][12][13][14][15][16] Medicaid expansion may also help at-risk men who otherwise would not undergo prostate cancer screening. 17 Acknowledging that improved screening may risk possible overtreatment of indolent disease, contemporary practice emphasizes shared decision-making and often includes secondary screening tools and further imaging to minimize potential harms. The ability to successfully diagnose and definitively treat prostate cancer in younger and/or healthy men has considerable implications due to excellent survival rates. However, even with improved access to care disparities in timely and appropriate treatment may result in differences in survival outcomes. An analysis of Surveillance, Epidemiology, and End Results (SEER; National Cancer Institute) data from 1992 to 2009 found evidence of a substantial difference in quality of surgical care for prostate cancer in Black patients, with not just longer treatment delay but also lower likelihood of receiving radical prostatectomy within 3 months of diagnosis compared with non-Hispanic White men. 18 A similar study found that men younger than 65 years with nonpalpable prostate cancer were more likely to receive conservative management if they were Black and had no insurance-or even state Medicaid coverage-than if they had private insurance during times of economic hardship. 19 We hypothesized that increasing access to care through Medicaid expansion would be associated with increased prostate cancer treatment at a population level for men diagnosed with higher-risk prostate tumors (Gleason grade group [GG] 3-5; GG 2 with prostate-specific antigen (PSA) level Ն10 ng/mL [to convert to micrograms per liter, multiply by 1.0]; or GG 1 with PSA level >20 ng/mL) that would result in definitive treatment. 20 To test this, we used cancer registry data to compare patterns of prostate cancer treatment in states that did and did not expand Medicaid coverage. Demonstration of the downstream consequences of insurance expansion on prostate cancer treatment patterns would have important policy implications in terms of identifying mechanisms to decrease barriers to potentially life-saving interventions for men with prostate cancer.

Data Set
We used SEER data from January 2010 through December 2016. 21 The SEER program reports information on cancer incidence and mortality and provides patient-level data on selected demographic and tumor-specific factors from 18 cancer registries, covering approximately 28% of the US population. 21 Per the Common Rule, this study was deemed exempt from institutional review board oversight because it relied on publicly available deidentified data. Reporting of this study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Study Population
The analytic cohort was created as displayed in eFigure 1 in the Supplement. We identified men aged 50 to 64 years newly diagnosed with prostate cancer during the 7-year study period, which extended from January 2010 to December 2016. We limited our analysis to SEER registries with statewide coverage and excluded patients residing in states that partially expanded Medicaid coverage prior to 2010 (ie, California and Connecticut). We excluded nonhistologically confirmed cases and cases from autopsy or death certificate reporting. We limited the analysis to patients with higher-risk tumors that are more appropriate for definitive cancer-directed treatment (ie, any biopsy with GG 3-5; GG 2 with PSA level Ն10 ng/mL; or GG 1 with PSA level >20 ng/mL).

Exposures and Outcomes
The primary exposure was state Medicaid expansion status. On January 1, 2014, 5 states with SEER data expanded Medicaid coverage (ie, Hawaii, Iowa, New Mexico, Kentucky, and New Jersey) and 3 did not (ie, Georgia, Louisiana, and Utah). We assessed whether Medicaid expansion was associated with a change in the proportion of patients with Medicaid coverage and lack of insurance. Our primary outcome of interest was treatment with radical prostatectomy or radiation therapy

Statistical Analysis
We performed bivariate analyses to evaluate unadjusted associations between residence in a Medicaid expansion state (vs nonexpansion state) and receipt of treatment with various demographic and clinical patient characteristics. Multivariable logistic regression models were fit to examine the association between factors of interest with Medicaid coverage, lack of insurance, and the receipt of treatment. Models assessing insurance coverage adjusted for age (continuous), race/ ethnicity (categorical), and marital status (categorical) and included an interaction term for expansion status and time before or after expansion. Models assessing receipt of treatment included those covariates plus insurance status. Models were adjusted for clustering at the state level with a robust variance estimator. The variance inflation factor was assessed for all models without multicollinearity observed between the variables. The results of the logistic regression analyses were expressed as adjusted odds ratios (aORs) and corresponding 95% CIs. Estimated probabilities with 95% CIs were generated using postestimation commands.
Trends over time were assessed using an interrupted time-series regression model with a multiple group comparison (eAppendix in the Supplement). 22 The interruption point was assumed to coincide with Medicaid expansion on January 1, 2014. The main analysis compared trends in states with and without Medicaid expansion. Newey-West SEs were used to address autocorrelation and

JAMA Network Open | Urology
State-Level Medicaid Expansion and Treatment of Patients With Higher-Risk Prostate Cancer possible heteroscedasticity. From these models, the slope over time for each group was estimated before and after the intervention, and the difference-in-differences of slopes before and after the intervention was then calculated. Finally, the slopes of changes in treatment receipt were plotted for easier visualization of results.
All analyses were performed using Stata version 16.1 (StataCorp). Data were analyzed between August and December 2019. A 2-sided P < .05 was considered statistically significant for all analyses.
Patients with missing data (accounting for approximately 10% of the total patients) were excluded to minimize bias that can be associated with imputation.

Results
The sociodemographic and clinical characteristics of the study cohort are outlined in Table 1 The association between patient factors and insurance type is displayed in Table 2. The associations between race, insurance status, expansion status, and treatment receipt are depicted in Table 3. Compared with White men, non-Hispanic Black men (aOR, 0.65; 95% CI, 0.52-0.81) and Hispanic/Latino men (aOR, 0.71; 95% CI, 0.54-0.94) were less likely to undergo prostatectomy or radiation therapy. Men with private insurance or Medicare were more likely to receive treatment than men without insurance (aOR, 1.52; 95% CI, 1.25-1.86).

Discussion
Our analysis produced 3 major findings. First, we showed that patients with prostate cancer who had clinically significant tumors and resided in states that expanded Medicaid eligibility after January 2014 were significantly more likely to have Medicaid coverage after expansion and that similar changes in Medicaid coverage were not seen in nonexpansion states. Second, despite increases in Medicaid coverage at the patient level, state-level expansion of Medicaid coverage was not associated with any significant short-term change in treatment trends among men with clinically   It is imperative to continue dedicating efforts to improve outcomes and survival for patients who may be the most marginalized-the uninsured. Expansion of insurance coverage will not accomplish this alone. Furthermore, efforts to improve prostate cancer care for those most at risk of adverse outcomes should also focus on Medicaid patients. Among a group of uninsured patients with prostate cancer previously managed by a disease-specific, state-funded treatment program, men who left the program after gaining comprehensive insurance coverage had inferior physical health compared with those who stayed in the program. 35 Although some improvement is required for patients with prostate cancer, Medicaid expansion has considerable benefits at a population level with regards to financial security, chronic disease care, and overall well-being. 10 Ongoing policy

Limitations
This study should be interpreted in the context of its limitations. First, the treatment status for prostate cancer patients in SEER is not 100% sensitive, particularly for those receiving radiation therapy. However, we did not observe major differences in types of treatments received based on insurance status, so we should not anticipate a major bias in our results related to this fact. Second, because the proportion of patients with Medicaid or no insurance was relatively small, we may have not been able to observe subtle changes in treatment rates due to insufficient sample size.
Furthermore, due to the geographic basis of the SEER cancer registry, we are unable to account for patients who change residence out of areas captured during the time frame of the study.

Conclusions
In this cohort study, expansion of Medicaid coverage at the state level in 2014 was associated with increased Medicaid coverage and a lower proportion of uninsured men among patients with clinically significant prostate tumors. However, Medicaid expansion was not associated with changes in treatment receipt in this population.