Introduction

In January, 2014, the Patient Protection and Affordable Care Act [commonly known as the Affordable Care Act (ACA)] included a Medicaid expansion component which extended access to health insurance coverage to over 20 million previously uninsured individuals, with a resultant 21% increase in Medicaid enrollment [1, 2]. The primary driver of this sharp increase was expansion of Medicaid eligibility to adults under 65 years of age earning up to 138% of the federal poverty level. However, a 2012 Supreme court ruling allowed states to opt out of Medicaid expansion. Currently 32 states and the District of Columbia have elected to expand Medicaid, whereas 19 states have not [1]. Following Medicaid expansion in California, there was an estimated 45% reduction in the uninsured population, with 1.3 million individuals gaining Medicaid coverage in January, 2014 and an additional 2.1 million individuals gaining Medicaid coverage by December, 2014 [3, 4].

There is considerable controversy regarding the impact of Medicaid expansion on health care costs, hospital readmission rates and quality of care [1, 5–7]. There are few data evaluating the impact of Medicaid expansion on Emergency Department (ED) utilization. Assessment of ED utilization after Medicaid expansion is important, because the ED is a suboptimal environment for delivery of non-urgent care. ED visits for primary care treatable conditions are costly, lead to ED overcrowding, and result in fragmented care. Furthermore, in the ED environment, care is often delivered without knowledge of a patient’s comprehensive medical history and comorbidities. Prior to full Medicaid expansion, annual Emergency Department (ED) visit rates in the United States were increasing more than would be expected from population growth,[8] with a rising proportion of Medicaid or uninsured patients [9, 10]. One aim of the ACA was to facilitate better access to primary care. Providing Medicaid coverage to the uninsured as a part of Medicaid expansion was an important component of this plan, however, Medicaid patients are among the highest utilizers of the ED [8]. Determining the impact of Medicaid expansion on ED utilization and hospitalizations is therefore a matter of great interest.

In this study we evaluate all ED visits and the top ED visit diagnoses, the percentage of patients admitted to hospitals from the ED, all hospitalizations and top hospitalization diagnoses over the five-year period leading up to Medicaid expansion and the one year following Medicaid expansion, and compare results between a Medicaid expansion state (California) and a non-expansion state (Florida).

Methods

Results

Demographics

Population demographics over the study period are depicted in Table 1. The overall US population was 306,771,529 in 2009 and 318,907,401 in 2014 (3.96% increase). The population of California was 36,961,229 in 2009 and 38,792,291 in 2014 (4.95% increased) and the population of Florida was 18,652,644 in 2009 and 19,905,569 in 2014 (6.72% increased) (Table 1). California and Florida represented 12.05% and 6.08% of the US population in 2009 and 12.16% and 6.24% of the US population in 2014, respectively. There are significant differences in age, race and ethnicity between the two states. Florida has a higher proportion of people aged 65 years of age and older. In 2009, 17.19% of Florida’s population was aged ≥ 65 years of age and older, compared to 11.46% of the California population and 12.92% of the US population. An increased proportion of all populations were 65 and older in 2014: 19.05% in Florida, 12.85% in California, and 14.49% in the US. In 2009 and 2014, Hispanics comprised 37.2% and 38.6% of California’s total population, compared with 22.2% and 24.1% of Florida’s population and 16.1% and 17.4% of the US population, respectively. In 2009 and 2014, Black individuals comprised 16.8% and 17.3% of the Florida population, compared with only 7.3% and 7.2% of the California population and 13.5% and 13.9% of the US population, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Total population, racial/ethnic.

https://doi.org/10.1371/journal.pone.0182346.t001

ED utilization

Prior to Medicaid expansion, Medicaid patient ED visits in California increased from 8.4 per 100 residents in 2009 to 9.4 per 100 residents in 2013 (11.9% increase over 5 years). Over the same period of time, Medicaid patient ED visits in Florida increased from 10.7 to 14.1 per 100 residents (31.8% increase over 5 years, S1 Table, Fig 1B). The increase in Medicaid ED visits in California in 2014 was significantly greater than predicted based on linear regression analysis of 2009–2013 data (Fig 2A), with the actual rates being 32.2% higher than predicted (p<0.0001). In contrast, Florida Medicaid ED visit rates in 2014 closely followed those predicted from 2009–2013 data (Fig 2B). After Medicaid expansion in 2014, California Medicaid patient ED visits increased to 12.5 per 100 residents—a 33% increase in the rate of Medicaid patient ED visits (S1 Table, Fig 1A) over a single year. From 2013 to 2014, a 25% decrease in self-pay patient visits was also observed in California (S1 Table, Fig 1A). In contrast, in Florida, only a 5.7% increase in Medicaid patient ED visits was noted between 2013 and 2014, with a 5.1% decrease in rate of self-pay patient visits and an 11.7% increase in the rate of private insurance patient visits in 2014 relative to 2013 (S1 Table, Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Emergency department utilization in California (A) and Florida (B) between 2009 and 2014 by payer, expressed as annual visits per 100 state residents. Hospital utilization in California (C) and Florida (D) between 2009 and 2014 by payer, expressed as annual hospitalizations per 100 state residents.

See S1 Table for corresponding data tables.

https://doi.org/10.1371/journal.pone.0182346.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

(A) Predicted vs. Actual Emergency Department visits in Medicaid enrollees in California (per 100 state residents). In 2014, the actual rate was 32.2% higher than predicted based on 2009–2013 data (p<0.0001). (B) Florida Medicaid ED visit rates did not differ from those predicted based on 2009–2013 data.

https://doi.org/10.1371/journal.pone.0182346.g002

The total number of ED visits in California increased from 31.6 per 100 residents in 2009 to 33.1 per 100 residents in 2013 and further increased to 34.5 per 100 residents in 2014. Overall ED visits increased by 4.2% in California from 2013 to 2014, compared to an average increase of around 2% over the previous 3 years. (S1 Table, Fig 1A and 1B).

Hospitalizations

In the five years preceding Medicaid expansion, Medicaid hospitalizations in California decreased from 2.8 per 100 residents in 2009 to 2.6 per 100 residents in 2013 (Fig 3A and 3B, S1 Table). From 2013 to 2014, overall hospitalizations remained stable (9.9 per 100 in 2013 and 9.8 per 100 in 2014), but California hospitalization trends by payer demonstrated a 15.4% increase in the rate of inpatient stays for Medicaid patients and a 25% decline in the rate of self-pay hospitalizations (Fig 3A and 3B, S1 Table). Hospitalizations in Florida increased from 2.7 to 2.9 per 100 residents from 2009–2013 and remained stable at 2.9 per 100 residents from 2013 to 2014. Trends for other payers, including Medicare, private insurance and self-pay remained stable from 2013–2014 in Florida (S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Rate of hospitalization California (A) and Florida (B) between 2009 and 2014 by payer, expressed hospitalizations per 100 emergency department visits.

See S2 Table for corresponding data tables.

https://doi.org/10.1371/journal.pone.0182346.g003

Discussion

Emergency departments are often viewed as a safety net for medically underserved patients with limited access to primary care. Adults with Medicaid insurance and patients without health insurance are high ED utilizers [8]. It was hoped that full implementation of the ACA in January, 2014 would provide insurance coverage to the uninsured (primarily through Medicaid expansion) and thus allow improved access to primary care physicians, thereby decreasing the demand for ED services, at least for primary care treatable conditions. However, many hypothesized that, despite the expansion of insurance coverage with the ACA, inadequate access to primary care for the newly insured (attributable to higher demand for care) combined with a reduction in financial barriers to using the ED, might actually increase demand for ED services [15, 16]. Although few data are available to support this hypothesis, an initial survey of over 2,000 ED physicians conducted by the American College of Emergency Physicians indicated that 75% believe that ED visits have increased since implementation of the ACA [17]. In 2008, Oregon's Medicaid program underwent what has been considered a randomized controlled trial of Medicaid expansion [18, 19]. While individuals with new Medicaid coverage in Oregon reported superior access to primary care, overall healthcare spending increased by approximately $1000 per patient annually, and Medicaid patients were found to use the ED 40% more frequently than other adult populations [18, 19]. Other studies have found that Medicaid patients, rather than the uninsured, account for the largest proportion of patients using EDs for non-urgent conditions [20]. These non-urgent ED visits are rarely preceded by a visit to a primary care provider [20]. On the other hand, a study of the 2010 ACA dependent coverage expansion provision, which allowed adults under 26 to retain insurance coverage under their parent’s insurance plan, noted a 0.5% decrease in the number of ED visits per 1,000 people in young adults aged 19 to 25 years, compared to 26–31 year-old adults unaffected by the provision [21].

We compared ED utilization in two states, one that implemented Medicaid expansion (California) and the other that opted out of this provision (Florida), using data from five years before and one year after Medicaid expansion to analyze trends over time. In the five years prior to Medicaid expansion, Medicaid patient ED visits in California increased by 11.9%, potentially associated with implementation of California’s Low Income Health Plan, a Medicaid-like plan which was implemented on a rolling county-by-county basis from November, 2010 to December, 2013 and ended in January, 2014 at which time enrollees in this plan were rolled into California’s Medicaid expansion program. As a frame of reference, following Medicaid expansion in California, 3.4 million individuals gained Medicaid coverage in 2014, compared with just 624,115 individuals who transitioned from the Low Income Health Plan to Medicaid in January, 2014 [3, 4].

In the five years prior to Medicaid expansion, Medicaid patient ED visits in Florida increased by 31.8%. Our data indicate small changes in overall ED visits with Medicaid expansion, but demonstrate a dramatic shift in the payer mix for these ED visits. Medicaid expansion resulted in a 33% increase in Medicaid patient ED visits and a corresponding sharp decline (25%) in self-pay patient visits in California. Over the same time period, in Florida, a non-expansion state, there was a small (5.7%) overall increase in Medicaid patient visits, correlating with the small increase in Medicaid enrollment from 2013 to 2014,[22] similar to the trend in preceding years. These shifts in the California payer profile are consistent with the expansion of Medicaid and the resultant coverage of previously uninsured and self-pay individuals as well as those previously enrolled in the Low Income Health Plan. The stability of overall ED visits may be attributable to a payer shift between two high ED utilization populations as uninsured individuals obtain Medicaid insurance. Overall hospitalization rates are similarly stable after Medicaid expansion in California, with a slight increase in total hospitalization rates in Florida over the same time period. Trends in payer mix for hospitalizations mirror those for ED utilization, but the overall increase in Medicaid patient hospitalizations in California after Medicaid expansion is lower, with a corresponding decline in self-pay and private insurance hospitalizations.

We show that the percentage of patients hospitalized after ED visits has been decreasing steadily over the past 5 years in California and Florida. An overall 5.4% decrease in the rate of ED visits resulting in hospitalization was noted in California after Medicaid expansion, while a smaller 2.1% decrease in ED visits resulting in hospitalizations was evident in Florida over the same time period. The stably higher percentage of patients hospitalized after ED visits in Florida is probably a consequence of its older population, many of whom may have multiple co-morbidities. Presenting diagnoses for ED visits did not significantly change over time (S3 Table). Prior to Medicaid expansion, there was speculation that ED visits for non-emergent conditions might increase significantly due to the relative lack of primary care access, given the sudden inundation of the healthcare system with a large number of newly insured Medicaid and other insured patients. If such a trend for increasingly non-emergent presentations to the ED had been seen, it might indicate that prior insurance coverage was inadequate and that previously uninsured individuals who now had insurance were seeking medical attention in the ED for lower acuity symptoms. Scrutiny of the top 10 conditions prompting ED visits (S3 Table) reveals no change in the top 10 conditions for which patients present to the ED after Medicaid expansion. However, the five-year trend of same-institution hospitalizations in California is notable for a more dramatic decline in the rate of admission from the ED after Medicaid expansion than would be projected, based on the rate of admission decrease over the 5 years preceding Medicaid expansion. This greater than expected decline in admission rates suggests that ED use for lower acuity conditions that do not fall within the list of top 10 ED diagnoses may have increased following Medicaid expansion.

The Medicaid population evaluated here includes both Medicaid fee-for-service (FFS) and Medicaid Managed Care patients. It would be of interest to understand the independent trends of these two Medicaid populations and how they may differ. Medicaid Managed Care involves a centralized care setting with managed care principles to minimize cost and optimize access and quality of care [23]. ED visits for patients with Medicaid Managed Care may then be more amenable to incentives to enhance primary care management of non-urgent conditions and decrease ED visits in the longer term [23]. However, it is also possible that barriers to accessing medical care under Medicaid Managed Care may result in an increase in ED visits and preventable hospitalizations [24].

A limitation of this study is that although we evaluated trends over five years before Medicaid expansion, we had only a single year’s data following Medicaid expansion. However, this transition year is particularly worthy of study, as it may help identify and highlight trends and bottlenecks that must be addressed to optimize future achievement of the stated goals of the ACA. Medicaid patients often encounter difficulty in finding primary care physicians who will accept their coverage, as Medicaid reimbursement rates are lower than Medicare reimbursement rates [25]. Medicaid pays only approximately 56% and Medicare 80%, of what private insurers typically pay. In anticipation of this issue, Medicaid reimbursement rates were federally funded to match Medicare reimbursement rates for a defined set of primary care-related services over 2013 and 2014, in an effort to enhance access to primary care for the new and growing number of Medicaid enrollees [26]. This ACA initiative improved primary care access as measured by appointment availability for Medicaid patients,[27] which might explain the overall stability of ED visit rates in California after Medicaid expansion. The reimbursement increase was transient and expired at the end of 2014 in most states including California [26]. It will be of great interest to determine the impact of withdrawal of this rate increase and the estimated 42.8% reduction in Medicaid reimbursement on primary care access and ED visits over subsequent years [28]. Whereas primary care physicians may continue to see Medicaid patients who have established care with their practices even though reimbursements decline, it is also possible that access to primary care may again diminish with a consequent increase in ED visits for Medicaid patients. The impact of withdrawal of the reimbursement rate increase can be evaluated as some states opted to use state funds to continue to support the Medicaid reimbursement increase in 2015, whereas several did not [29].

Studying the impact of ACA on health care utilization is challenging, because the process is not a randomized, controlled study. Other cost analyses have noted that slowing of the increase in health care costs may be attributable to the recent economic recession rather than Medicaid expansion [6]. Expensive therapeutic advances in health care may further modulate health care costs [6]. These variables prevent direct attribution of any changes in health care cost to Medicaid expansion and also, to a lesser extent, impacts our analysis of ED utilization and hospitalizations, as health care utilization may be modulated by economic downturns. Our analysis of the six-year trend in ED and hospital utilization provides longitudinal support and a context for interpreting changes that followed Medicaid expansion.

In summary, our study has demonstrated a significant shift in payer for ED visits and hospitalizations after Medicaid expansion in California. There were notably fewer self-pay and significantly more Medicaid patient admissions and ED visits after Medicaid expansion in California, with relatively stable overall ED and hospital utilization. This is consistent with a shift in the burden of reimbursement from patients and hospitals to the government. The stability of ED utilization and analyses of top diagnoses indicate that individuals with newly obtained Medicaid insurance coverage utilize the ED for conditions of similar seriousness as those who are self-pay or uninsured. If financial barriers had previously prevented uninsured patients with significant illness from seeking medical care, one would expect a dramatic increase in ED visits and hospitalizations when these individuals obtained Medicaid coverage. We found no evidence of such a trend, indicating either that health care access was sufficient prior to Medicaid expansion or that the temporary increase in Medicaid reimbursement rates have buffered the impact of Medicaid expansion on ED utilization.

Supporting information