A retrospective cohort study found the HOSPITAL score, a risk-assessment tool consisting of seven available clinical predictors, showed good discriminative ability and calibration for predicting the 30-day risk of potentially avoidable readmissions at nine not-for-profit hospitals.

Lead researcher Jacques D. Donze, MD, MSc, of Bern University Hospital in Switzerland, and colleagues published their results online in JAMA Internal Medicine on March 7.

They noted that 18 percent of Medicare patients are readmitted to hospitals within 30 days of discharge at a cost of more than $17 billion.

“The advantage of tools such as the HOSPITAL score is that we can use it to begin to answer the more relevant question of which patients are most likely to benefit from interventions designed to prevent readmissions,” they wrote. “Predicting potentially avoidable readmissions (as opposed to all readmissions) is an important step along this progression. This work is particularly important given recent incentives to reduce readmissions and several recent unsuccessful efforts to reduce them.”

In this study, the researchers analyzed 117,065 adults who spent at least one day in the hospital and were discharged alive between Jan. 1 and Dec. 31, 2011, at nine hospitals in the U.S., Canada, Israel and Switzerland. Patients were excluded if they transferred to another hospital or psychiatric hospital or who left against medical advice.

The researchers aimed to validate the HOSPITAL score, a prediction model for 30-day potentially avoidable readmissions that included the following clinical predictors: low hemoglobin level (less than 12 g/dL) at discharge; discharge from an oncology service; low sodium level (less than 135 mEq/L) at discharge; procedure during hospital stay; index admission type (urgent or emergent); number of hospital admissions during the previous year; and length of stay of five days or more.

They calculated a HOSPITAL score for each hospital discharge and categorized the risk for an admission to be followed by a 30-day potentially avoidable readmission into three groups: low risk (0 to 4 points); intermediate risk (5 to 6 points); and high risk (7 or more points).

At baseline, the mean age of patients was 60.8 years old. They had a median length of stay of four days.

Within 30 days after discharge, 15 percent of patients had a readmission and 9.7 percent had a potentially avoidable readmission. The researchers mentioned that patients with a potentially avoidable readmission were more likely to have an urgent or emergent index admission, be discharged from an oncology service, have a length of stay of more than five days, have more hospitalizations in the past year, to more often have a procedure, and more often to have a low hemoglobin or low sodium level at discharge.

The risk of 30-day potentially avoidable readmissions was 5.8 percent in the low risk group, 12.0 percent in the intermediate risk group and 22.8 percent in the high risk group. Of the patients, 62.4 percent of patients were categorized as low risk for potentially avoidable readmissions, 23.6 percent were considered intermediate risk and 14.0 percent were considered high risk.

The researchers noted that the HOSPITAL score had a C statistic of 0.72, which indicated good discrimination. The C statistic was 0.72 in U.S. hospitals, 0.78 in Canadian hospitals, 0.68 in Israeli hospitals and 0.68 in Swedish hospitals.

They mentioned a few limitations, including that the model only focuses on medical patients and may not be generalizable to surgical patients or other populations. They added that other predictors such as functional status and socioeconomic status may have improved the score and that 30 days may not be the best time period to judge preventable readmissions.

“The HOSPITAL score is easy to use and can be calculated before discharge, which makes it a practical tool for identification of patients at high risk for preventable readmission and the timely administration of high-intensity interventions designed to improve transitions of care,” they wrote.