Photo Credit: SDI Productions

Emergency department physician admission rates do not appear to affect the likelihood of their patients dying, according to recently published research.

Patients of emergency department (ED) physicians who admit patients at higher rates do not appear to be less likely to die, results of a large cross-sectional study in JAMA Internal Medicine suggest.

“Physicians in the emergency department greatly differ in how often they hospitalize the patients they see,” study author Dan P. Ly, MD, PhD, MPP, of the David Geffen School of Medicine at UCLA, tells Physician’s Weekly (PW). “Some physicians are almost twice as likely as others working in the same ED to admit the patients they see with similar prior health status. Patients who see these high-admitting physicians also receive more testing, but they are not less likely to die than patients seeing lower-admitting physicians.”

According to the study authors, the wide variation in ED physicians’ propensities to admit their patients to the hospital had been documented, but little was known about whether and how these differences affected patient outcomes.

Veterans Affairs EHR Data Analysis

Dr. Ly and his coauthor, Stephen Coussens, PhD, Abett, Inc., reviewed nationwide Veterans Affairs electronic health record data from 2,098 physicians and 2,137,681 patient visits for chest pain, shortness of breath, or abdominal pain in 105 EDs from 2011 through 2019. The mean patient age was 63 years, and 90.2% of patients were male.

The researchers compared the admission rates of physicians practicing within the same ED and analyzed the relationships between the physicians’ admission rates and their patients’ inpatient stays under 24 hours and 30-day mortality rates.

They found that:

• Physicians’ adjusted admission rates varied widely within the same ED despite the lack of association between adjusted admission rates and patients’ prior health status, as measured by their pre-ED-visit Elixhauser Comorbidity Index score.
• The mean admission rate was 41.2%. However, patients treated by ED physicians with admission propensities at the 90^th percentile were almost twice as likely to be admitted compared to patients treated by ED physicians with propensities at the 10^th percentile (for example, admission rates of 58.2% versus 39.1%, respectively, for patients with shortness of breath, a difference of 19.1 percentage points).
• Patients admitted by ED physicians in the 90^th percentile were more likely to be discharged within 24 hours compared to patients with physicians in the 10^th percentile (for example, discharge rates of 31.0% versus 24.8%, respectively, for patients with chest pain, a difference of 6.2 percentage points).
• Overall, about 2.5% of patients died within 30 days. Patients admitted by ED physicians in the 90^th percentile were not less likely than those admitted by ED physicians in the 10^th percentile to die up to one year later.

Dr. Ly discussed the study and its implications for ED patient admissions with PW.

PW: Why was using EHR data the optimal approach for your study?

Dr. Ly: By using EHR data, our study was able to account for variables important to consider, such as time of ED arrival and certain areas within the ED that might house patients who are extremely sick or very stable. It’s important to account for such factors when studying physicians in the ED to make apples-to-apples comparisons.

Did the results surprise you?

Yes. Although there might have been a slight health benefit to being more likely to be hospitalized, I was surprised that we saw no mortality benefit, at least in our study.

How may the findings potentially affect admission policies and patient outcomes?

We want to be clear that our study does not support indiscriminately pushing high-admitting physicians to reduce their admission rates. However, by better understanding both low-admitting physicians and high-admitting physicians, we may learn how to safely discharge patients from the ED.

What questions remain unanswered?

Further understanding of why high-admitting physicians hospitalize so many of their patients and how to safely reduce such hospitalizations are important areas of future study. It’s important to understand that physicians do differ in their decisions. Unpacking how and why they differ provides a significant opportunity to learn and improve the care we deliver to patients.

An Editorial Perspective

In an editorial accompanying the published study, Jerard Z. Kneifati-Hayek, MD, MS, and Michael A. Incze, MD, MSEd, cautioned that unnecessary hospital admissions “expose patients to added risk (eg, hospital-acquired infection), accumulate high costs, and often encompass inconsequential testing … all without added patient benefit.”

“Innovative payment models and quality improvement programs that incentivize standardized admission decision-making for common conditions are needed to reduce low-value care in the ED,” they advised.