A recent study examines hospital readmission risks to determine if they are affected by patient demographics or hospital quality.
Hospital readmission risks are high for certain diseases and can vary greatly from hospital to hospital. The causes for the readmissions can either be due to various patient factors or due to the events at the hospital during a previous admission. Research in the early 2000s focused on uncovering the unknown hospital events that may independently contribute to the increasing rates of readmission of patients. The significance of this data lies in the fact that readmissions that happen within a 30-day time frame may speak volumes on the effective delivery of healthcare at a particular hospital. A study was done on this topic in 2009 in the United States which contributed to the formation of the Affordable Care Act. After the implementation of the Act, readmission rates declined.
Since 2009, healthcare has changed partly as a result of an increase in the incidence of new diseases and multi-resistance forms of bacteria. Health professionals have therefore been questioning the adequacy of the risk adjustment in the previous act. As an extension of this inquiry, researchers wanted to find out the current hospital factors independent of patient factors that contributed to the risk of readmission for people under the care of the Centers of Medicare and Medicaid Service (MCS) health centers.
In order to conduct this study, they looked at a patient sample from July 2014 through June 2015. The centers used for collecting the data are hospitals that use Medicare and Medicaid insurance services and have records of patient readmissions in these hospitals. The records of these patients were divided into two random samples. One of the criteria for selecting the patients was to make sure that they were 65 years of age and above. In total, the study sample comprised of 37,508 patients who had two admissions for similar diagnoses at a total of 4,272 different hospitals. The two samples taken were further stratified into a hospital and principal diagnosis category.
In the first sample, the researchers used this data to calculate rates of unplanned readmission for any cause that patients may present with within 30 days of the previous admission. This data was calculated for each hospital. In turn, performance for these hospitals was calculated on a quarterly basis (every three months a year). The standard measure used to calculate the risk was the CMS readmission measure. The performance was judged by readmission rate and calculated as the ratio of predicted readmissions to the number of expected readmissions, multiplied by the nationally observed readmission rate. Lower readmission rates indicated better performance. The other half of the sample (principal diagnosis category) was used to compare patients who had been admitted to different hospitals for the same diagnosis. If the patient had more than one readmission in a calculated three-month period, only one of those readmissions was taken into the calculation.
The results of this study were recently published in the New England Journal of Medicine and demonstrated that the readmission rate was consistently higher among patients admitted to hospitals in low-performance hospitals, compared to high-performance hospitals for every quarter year analysis. In fact, the research shows that when patients from the low performing hospitals were readmitted to high-performance hospitals with the same complaint, there was a significant difference in rates of readmission within 30 days.
This outcome suggests that patients are affected by hospital quality independently from the care that patients give themselves and implies that measures to improve health delivery to patients should be looked into carefully. These results may reassure policymakers and hospital administrators that hospital readmission risks are indeed reliable markers of hospital performance.
Written by Dr. Apollina Sharma, MBBS, GradDip EXMD
Krumholz, Harlan M., et al. “Hospital-Readmission Risk—Isolating Hospital Effects from Patient Effects.” New England Journal of Medicine 377.11 (2017): 1055-1064.