In Massachusetts, Potentially Avoidable Emergency Department Use Persists 5 Years After Health Reform

This blog was originally published on The Medical Care Blog. 

The right care, for the right person, in the right place, at the right time.

That’s a reasonable goal of a high-quality health system. In the US, however, many people go to the emergency department (ED) when a primary care provider (PCP), or perhaps an urgent or express care center, would be better. In a paper published in the Feburary 2019 issue of Medical Care, our team found that this reliance on the ED was more common for people with public health insurance (in this case, mostly MassHealth, the Massachusetts Medicaid program) compared with private insurance enrollees.

In theory, health insurance improves financial access to primary care, which should, in turn, reduce inappropriate ED use. In Massachusetts, one motivator for the health reform efforts that began in 2006 with the nation’s first universal health coverage program–a model for the Affordable Care Act–was the high rate of expensive ED visits deemed potentially avoidable (estimated at about half). Now, nearly everyone in Massachusetts has some form of health insurance, but there’s still a lot of potentially avoidable ED use.

So why do people use the ED when a primary care or urgent care provider might be better?

In surveys with non-urgent users of the ED, the answers might surprise you:

• In a hospital-based survey in Colorado, about half of non-urgent ED users said they went there because their PCP referred them there.
• In a Massachusetts survey, about half of ED users said they couldn’t get a timely appointment with their PCP.
• In a national survey, 55% of people who went to an ED thought that only hospital-based care could help them, 43% thought their problem was too serious for primary care, and almost 20% were referred by their PCP.
• In a survey of adult Medicaid enrollees presenting at an urban, academic medical center ED for low-acuity conditions, nearly half expressed a preference for the ED over their PCP. Their reasons included: perceived need for diagnostic equipment, such as X-rays; mistrust of their PCP; communication difficulties with their PCP; and feeling that their pain was too bad to wait for a PCP appointment.

Identifying and reducing non-urgent ED use has been a major challenge.

Health care systems, health insurers, and policymakers have been trying to reduce these non-urgent ED visits for decades, both in the US and elsewhere, without much success.  Even for highly trained and experienced ED doctors and nurses, it’s not always easy to tell what constitutes an emergency condition. In fact, research has shown that professionals disagree with each other about half the timeon how to classify the seriousness of an ED patient’s condition.

Also, the diagnosis code that ends up being recorded on a person’s hospital bill after triage and evaluation may look like a non-urgent complaint — even in cases where the patient experienced severe symptoms. Presenting complaints don’t always correspond with discharge diagnoses.  Any efforts to use claims and administrative data to evaluate the severity of a person’s condition are bound to be wrong some of the time. That’s why we believe that such ED visits should continue to be reimbursed by payors, and why the algorithm we used in our study to evaluate these types of ED visits should not be thought of as determining the “appropriateness” of any individual visit.

Looking more deeply at the factors that influence non-urgent ED visits.

With those caveats, our new paper in Medical Care examined characteristics of insured Massachusetts residents that were associated with non-urgent ED use and compared such use for people under age 65 with public (Medicaid) versus private insurance. We studied more than 2.2 million individuals in 2011-12; about 40% had public insurance and the rest had private insurance.

We refer to this type of ED use as primary care sensitive (PCS), because we think it could be reduced by increasing the supply, accessibility, or quality of primary care. In our definition, PCS ED visits fall into three categories:

1. Ambulatory care sensitive conditions (ACSCs) – for example, uncontrolled diabetes.  The term “ambulatory” here refers to care outside of a hospital setting (most often primary care).
2. Cases in which care was not truly needed on urgent basis, with urgent defined as within 12 hours (for example, a sore throat).
3. Cases in which care was probably needed within 12 hours, but the care could have been provided in a less-acute setting (such as cellulitis care, which can be delivered in a primary or urgent care setting).

We found that PCS ED use was significantly higher among the publicly insured: on an unadjusted basis, more than 4 times higher. After adjusting for a range of potential confounders, such as the vastly different morbidity burdens in the two groups, public insurance in 2011 was associated with a PCS ED visit rate ratio of 2.53 (95% CL: 2.49–2.56) when compared to private insurance.

We also found that people with public insurance were less likely to have any PCP visits than people with private insurance. More than 80% of people in the private group had at least one primary care encounter, compared to only 70% of the public group. The public group also had fewer visits to their PCP of record, even though nearly all of them had an officially designated PCP.

So, why else might people with public insurance go to the ED instead of their PCP?

Even with insurance, many barriers to accessing primary care remain:

• In 2006, 69% of Massachusetts internists were accepting new patients [PDF]; in 2008, only 52% were. According to insurers’ submissions to the state’s health care claims database (MA APCD), during the time of our study, only 15% of PCPs were accepting new patients.
• In Massachusetts in 2005 (pre-reform), the average wait time for a new patient to get an appointment to see an internist was 17 days; in 2008, after reform, it was 31 days. And that’s for all patients – the situation is worse for MassHealth enrollees, since PCPs are less likely to accept MassHealth than other kinds of doctors.
• Even though state legislation increased reimbursements for providers seeing MassHealth beneficiaries in 2006, the proportion of PCPs who accepted MassHealth declined from 2006 to 2012. From 2008 to 2012, which includes our study period, Medicaid fees for primary care services decreased by about 4% in Massachusetts, with the ratio of Medicaid to Medicare fees [PDF] decreasing from 0.78 to 0.68 over the same period.
• Massachusetts also faces the same kinds of primary care clinician shortages as the rest of the US. About 30% of Massachusetts physicians practice primary care, compared to about 70% in other developed countries and in the US 50 years ago.

Beginning to solve what seems like an intractable problem.

Clearly, with low reimbursement rates and an already inadequate supply of PCPs, increases in insurance coverage were not enough to improve access to primary care. It may also be unrealistic to expect reductions in ED use given certain financial incentives in our health care system, such as fee-for-service and procedure-based reimbursement systems, and the unique profitability of ED admissions for hospitals.

According to a 2014 systematic review of non-ED-based interventions to reduce ED overuse, the most successful approaches have been managed care, patient education, and patient financial incentives. State-level “personal responsibility” efforts have largely failed and carry serious risks of reducing appropriate ED use along with less-desirable use. What hasn’t been tried much — yet — is to give PCPs a reason to reduce their patients’ use of the ED, such as a potential bonus if they achieve reductions in ED use rates relative to what would be expected given their case mix. We explored this idea in a 2017 paper.

Some value-based payment reform efforts have targeted ED use as a quality measure, but they generally measure total ED use rather than the subset of ED visits that a PCP could reasonably be expected to affect – in other words, PCS ED use. We suggest that future ED-based quality measures would benefit from a more targeted approach.

In the meantime, PCPs can use some simple office management techniques to help their patients avoid the ED and get the right care at the right time and in the right place.