Medical Care blog

Comments from the scientific community regarding Medical Care.

Friday, January 2, 2015

Emergency Department Use in Massachusetts for Low-Income Adults with Subsidized Health Insurance
by Lisa M. Lines, PhD, MPH

Emergency department (ED) use has been increasing in the US for several decades, and some estimate that about half of all outpatient ED visits are potentially avoidable (also referred to as primary-care sensitive, or PCS). ED visits are expensive and may signify issues with access to, and quality of, care. Thus, reducing ED use is frequently a goal of clinical and policy interventions.

In Massachusetts, where health insurance reforms that began in 2006 were a model for the Affordable Care Act, evidence for the effects of reform on ED use has been mixed. The state’s health statistics agency, the Center for Healthcare Information and Analysis (CHIA), released a report in 2012 that showed about a 6% increase in both overall and PCS ED visits between 2006 and 2010, although the growth was slower than in previous years. In fact, from 2009 to 2010, the volume of total visits declined by 0.3%, and PCS visits declined by 0.6%. These trends may be partially explained by the recession, which decreased levels of healthcare utilization across the country, but policymakers hope that the expansion of health insurance and delivery system reforms such as patient-centered medical homes might also be responsible. At least 6 other peer-reviewed studies have reported decreases in ED use since reforms were implemented, but 1 reported no change in ED use and 2 reported increases.

A recent Medical Care article by Lee and colleagues analyzes ED use trends in a particular subpopulation in Massachusetts: low-income residents who obtained subsidized insurance coverage through Commonwealth Care (CommCare). The authors used public health insurance data from 2004-2008 matched with ED claims for all adults who enrolled in CommCare by September 30, 2009. They then examined overall ED use in the pre- and post-enrollment periods for each person, comparing individuals’ own use of the ED before and after enrollment in CommCare using conditional logistic regression models. The authors estimated separate models for those who were previously enrolled in MassHealth (Medicaid), Health Safety Net (HSN, formerly the Uncompensated Care Pool), or were not publicly insured. The main analysis focused on 112,146 adults who were enrolled in CommCare for at least 3 months and made at least one ED visit.

The authors report that the overall odds of visiting the ED after CommCare enrollment decreased by about 4%. However, the results varied substantially depending on a person’s prior coverage. The HSN group’s odds of visiting the ED decreased by 18%, and the MassHealth (Medicaid) group’s odds were 7% less. However, among those who were not publicly insured prior to CommCare enrollment, a person’s odds of visiting the ED increased by an average of 12%. Because of data limitations, the authors did not know whether those who were not publicly insured before CommCare enrollment had private insurance or were uninsured during the entire pre-enrollment period, although all of them were uninsured at the time of enrollment and none had been covered by public insurance during the previous 2 years.

The “pre-post” study design, in which each person serves as their own control, is subject to numerous potential biases, including regression to the mean, time period bias, and asymmetrical lengths of exposure to the “intervention” (in this case, enrollment in CommCare). In addition, the design cannot control for individual changes in health behaviors or health status that may have triggered changes in ED use behavior.

Nevertheless, it is interesting that the authors observed a large spike in ED utilization in the 3 months pre- and post-CommCare enrollment, especially in those who were previously not publicly insured. The authors surmise that this spike was caused by individuals becoming ill and subsequently enrolling in CommCare.

An alternative explanation is that newly insured individuals, who may have had substantial unmet care needs before enrollment, deliberately sought care in the ED (rather than in less-acute settings). In Massachusetts, as in the rest of the country, individuals with public insurance often delay obtaining primary and specialty care because of cost barriers as well as Medicaid’s low reimbursement rates (which limits the pool of available providers). Although Massachusetts policymakers increased reimbursements for Medicaid beneficiaries beginning in 2006, the number of PCPs who accepted Medicaid still declined from 2006 to 2009, according to a state report from 2010. During the same period, more people sought care as a result of the expansion in insurance coverage, while fewer PCPs were accepting new patients. Moreover, low-income adults may also be less likely to have jobs that provide paid sick leave, meaning they may be more likely to seek care during non-business hours. In our imperfect healthcare system, even “avoidable” ED visits may be hard to prevent.

Overall, the results of this study suggest that the effects of obtaining subsidized health insurance on ED utilization among low-income adults will vary substantially based on prior coverage. More research is needed to further understand the effects of subsidized insurance coverage in the context of other predictors, such as health status and access to care.

Lisa Lines is a health services researcher at RTI International and the co-editor of the Medical Care Blog.