ABSTRACT & COMMENTARY
Disparities in Hospital Transfers: Not What You Would Expect
By Deborah J. DeWaay, MD, FACP
Assistant Professor, Medical University of South Carolina, Charleston, SC
Dr. DeWaay reports no financial relationships in this field of study.
SYNOPSIS: Administrative data extracted from the Nationwide Inpatient Sample indicates that uninsured patients and women were less likely to be transferred from one acute hospital to another.
SOURCE: Hanmer J, Lu X, Rosenthal G, Cram P. Insurance Status and the Transfer of Hospitalized Patients: An Observational Study. Annals of Internal Medicine. 2013; 160: 81-90.
In 1986, the Emergency Medical Treatment and Active Labor Act (EMTALA) was enacted in order to assure that all patients, irrespective of sex, ethnicity, race or ability to pay were given emergent medical care. The act mandates that all patients are examined, treated and stabilized until they are stable for discharge, even if they cannot pay. There have been several studies in the past showing that once a patient is stabilized that hospitals may transfer patients to other hospitals for non-medical reasons. The authors of this article used the 2010 data from the Nationwide Inpatient Sample (NIS) to examine the relationship between inter-hospital transfer and insurance status for patients who have already been admitted. They hypothesized that hospitals would transfer their underinsured patients to other acute care hospitals at a greater rate in order to lessen their financial burden.
The 2010 NIS data set contains discharge data for about 8 million hospitalizations from 1051 hospitals. The authors purchased this data from the Agency for Healthcare Research and Quality (AHRQ). The hospitals included in this study are only acute care hospitals. Patients who were admitted under observation are not included in the data set. The authors analyzed discharges as opposed to individual patients because there are no unique identifiers in the database which could be used to track patients across different hospital admissions. The authors selected all discharges for patients 18-64, excluding those 65 and older because they presumably had Medicare. The following discharge records were also excluded: those of patients whose primary payer was listed as "other" (as opposed to Medicare, Medicaid, private or uninsured), those of patients who died, and those of patients who left against medical advice. The authors selected 5 common diagnoses — biliary tract disease, septicemia, skin infections, pneumonia, and chest pain — to study further. Then the authors linked this data to the AHRQ's hospital weights file in order to analyze data such as hospital ownership and financial status.
The authors used the chi-square test for all categorical variables and t-tests for continuous variables to compare the patients transferred to those that were not transferred in the following areas: demographic characteristics, presence of key comorbid conditions, average comorbid conditions, and insurance coverage. They also used weighted analyses to look at the odds of a patient transfer to another acute care hospital. They controlled for patient demographics, comorbid conditions, and hospital characteristics. The authors did multiple sensitivity analyses to make sure their data were robust and held true under a variety of assumptions.
Across all diagnoses, women and the uninsured were less likely to be transferred than men and those with private insurance. For example, the odds ratio for women to be transferred with chest pain was 0.67 (0.58-0.76; P < 0.001). The uninsured were significantly less likely to be transferred if they had biliary tract disease, chest pain, septicemia and skin infections. The uninsured with pneumonia were less likely to be transferred, but the odds ratio was not statistically significant. Patient transfer also was different depending upon which type of hospital the patient was originally admitted. Non-teaching hospitals transferred more frequently than teaching hospitals.
The authors were surprised to find their hypothesis was incorrect; the uninsured were less likely to be transferred than those with private insurance. Although these data only reveal associations and do not prove causation, the authors argue that the receiving hospitals may not be accepting patients without insurance. It is not clear why women would be less likely to be transferred. Further research is necessary to explain this difference. The consequence of this difference is also unclear. Patients who are not transferred may receive substandard care because they are not transferred to facilities with better technology or subspecialty care. Or, patients who are transferred may receive tests and procedures that are not necessary, therefore they are exposed to extra risk with transfer. The major limitation to this study is that it used administrative data, so the exact reason for each transfer is unknown. In addition, the clinical outcomes for patients are unclear.
The finding that women and the uninsured are less likely to be transferred is an interesting finding because it illustrates that there are differences in treatment in these groups. However, the consequences of this difference are unclear. Importantly, the fact that administrative data was used is a major limitation since the reason for transfer in each case is unknown. Further research is necessary to see if these differences remain when the reason for transfer is known. An interesting follow-up question would be how Medicare patients with different supplemental plans are treated versus patients with private insurance. In addition, this study will need to be redone after the full effects of the Affordable Care Act are in place.