View all items in this collection

National Health Spending Trends

An RWJF Collection

View all items in this collection

RWJF Project Tracks Impact of Reform on Hospital Utilization

    • July 8, 2014
Kathy Hempstead Headshot

Katherine Hempstead, PhD, MA, director and senior program officer, leads RWJF's work on health insurance coverage.

To monitor the impact of health reform on hospital utilization, the Robert Wood Johnson Foundation has embarked on an important surveillance project, working in collaboration with 24 state hospital associations. The RWJF Hospital ACA Monitoring Project, or HAMP, collects data each quarter on all inpatient admissions and emergency department (ED) visits by payer. HAMP also collects some information on a subset of diagnoses and procedures that are believed to be sensitive to insurance status.

Clearly, there are a variety of ways in which health reform may affect hospital utilization. Conventional wisdom might suggest that coverage expansion will result in fewer preventable hospitalizations, and less use of the ED for ambulatory care sensitive conditions. However, we saw in the Oregon Medicaid experiment that increased Medicaid enrollment resulted in more ED use. Another possibility is that increased demand for primary care may overwhelm the ambulatory care system, resulting in increased use of the hospital for primary care treatable conditions, even among those who were previously insured.

The potential financial impact of health reform on hospitals is also unclear. While the reduction in uncompensated care is clearly a plus, there may be significant increases in utilization by patients who have payers that reimburse at relatively low rates. Further, there is a possibility of increased bad debt from patients with Marketplace plans, which require significant cost-sharing. Additionally, there are reductions in DSH payments and other simultaneous changes in Medicare payments.

The HAMP effort is designed to shed light on some of the effects of health reform on hospitals and provide extremely timely data to researchers, policymakers and hospital leaders. Seventeen state hospital associations submitted information from individual hospitals, while seven others submitted state-level data. There are approximately 1,700 hospitals included in this data set, which is roughly one-third of all hospitals in the country. The participating states and the number of individual hospital submissions are shown in Table 1.

The individual diagnoses and procedures being collected include three reasons for inpatient hospitalization that are considered to be preventable, and have been defined as such by the federal Agency for Health Research and Quality (AHRQ). These include short-term complications of diabetes, hypertension and urinary tract infection. Additionally, HAMP is monitoring admissions for knee replacement—an inpatient procedure that may be sensitive to insurance status. It is also monitoring ED visits for three specific diagnoses that are considered primary care treatable or at least ambulatory care sensitive: upper respiratory infection, urinary tract infection and headache. More details about these specific conditions are shown in Table 2.

The types of hospitals that participate in this project are shown in Table 3. While about 65 percent are acute care hospitals, there are a number of critical access hospitals, particularly in Western states and rural areas. The distribution of participating hospitals is shown in Table 4.  

Baseline Data Offers Trends to Watch

Data from 2013 from participating states show great variation in the number of inpatient admissions and ED visits, as shown in Tables 5 and 6. Tables 7 and 8 show the payer mix in the inpatient and ED setting. There are clearly differences between the two. While only about 6 percent of inpatients are reported as being “self-pay”, about 20 percent of ED visits are attributable to the uninsured. The state variation in the percentage of admissions and visits which fall into the “self-pay” category are significant. This range can be seen in Tables 9, 10 and 11. It is clear that certain states have a relatively high share of uncompensated care. For example, 30 percent of South Carolina’s ED visits are in the “self-pay” category, as compared to about 11 percent of those in Nebraska. Inpatient admissions range from about 2 percent self-pay in Minnesota to about 12 percent in Wyoming. Table 11 makes clear that states with a high percentage of self-payers in the inpatient setting tend to also have a high share of self-pay patients in the ED. Tables 12-14 show similar patterns for Medicaid.

One thing that is clear from these tables is that some of the participating states that have expanded Medicaid were exposed to relatively little uncompensated care in 2013. This is the case for Minnesota, Michigan, Connecticut and New York. However there are also a number of expanding states (New Jersey, Nevada, Colorado and Kentucky) that in 2013 had a significant amount of self-pay utilization both in the inpatient and ED settings. Depending on the degree of eligibility and take-up among these uninsured patients, these states may experience a fairly significant change in utilization patterns upon expansion. Early reports from national hospital chains suggest increases in Medicaid utilization and decreases in uncompensated care in expanding states—and no change in Medicaid and increase in uncompensated care in non-expanding states. Similarly, data released by the Colorado Hospital Association showed similar trends in Medicaid and uncompensated care as a percent of charges in Q1 2014 in expanding versus non-expanding states.

Future posts will provide more information about payer mix and utilization by state and variation within states and for more specific diagnoses and procedures. Data for Q1 2014 are expected by the end of the summer.