Predicting Avoidable 30-Day Readmissions

Article Info :

By: Jacques Donzé, MD,

 

References: 

Donzé J, Aujesky D, Williams D, Schnipper JL. Potentially avoidable 30-day hospital readmissions in medical patients: derivation and validation of a prediction model. JAMA Intern Med.

 

Predicting Avoidable 30-Day Readmissions

 

For medical patients, a simple prediction model appears to identify the risk of potentially avoidable early readmissions before discharge. The model can help clinicians seek out patients who may need more intensive transitional care interventions.

Throughout the United States, readmission rates are increasingly being used for benchmarking across hospitals. Some hospital readmissions may be avoidable, which in turn has led to the levying of financial penalties on hospitals with high risk-adjusted rates. Recent studies have estimated that the 30-day readmission rate for Medicare beneficiaries is almost 20%, and these occurrences cost the U.S. healthcare system as much as $17 billion annually.

In June 2009, CMS began publishing 30-day readmission data for select medical diseases, resulting in hospital readmissions becoming an important metric for measuring the quality of patient care. The changing regulations issued by CMS means that hospital reimbursements can be reduced based on an adjustment factor determined by a hospital’s expected and observed 30-day readmission rates. These changes have also raised the bar for decreasing unnecessary surgical readmissions. In addition to the financial implications, unplanned hospital readmissions further limit hospital resources. For each patient readmitted, there is an opportunity lost to treat another patient who needs care.

 

New Prediction Model for 30-Day Readmission In JAMA Internal Medicine, Dr. Donzé and colleagues had a study published that derived and validated a prediction model for potentially avoidable 30-day hospital readmissions in medical patients. The model used administrative and clinical data that was readily available prior to discharge. “Our purpose was to help clinicians target transitional care interventions most efficiently,” Dr. Donzé says. “The goal was to develop a score to predict potentially avoidable readmissions. In other words, we wanted to predict which patients may be most likely to benefit from intensive interventions.” The HOSPITAL score is able to indicate readmission risk before a patient is discharged. This allows clinicians to target a timely transitional care intervention. In their retrospective analysis, Dr. Donzé and colleagues analyzed potentially avoidable 30-day readmissions at three hospitals using a validated computerized algorithm. Several factors were identified as being independently associated with the risk of potentially avoidable readmission.

These factors were then used to build a prediction score—referred to as the HOSPITAL score—for early readmission: hemoglobin at discharge; discharge from an oncology service; sodium level at discharge, procedure during the index admission; index type of admission; number of admissions during the last 12 months; and length of stay (Table 1). “The strength of the HOSPITAL score is its simplicity,” says Dr. Donzé.

 “Physicians can easily review these seven variables at a patient’s bedside prior to discharge. The more of these risk factors a patient has, the greater the risk of readmission. If a patient is deemed high risk for readmission, a return trip to the hospital could be prevented by offering additional interventions, such as a home visit from nurses, patient coaching, or a pharmacist consultation.” Validating Results on the HOSPITAL Score Using the HOSPITAL score, Dr. Donzé and colleagues stratified the risk of potentially avoidable readmission into three categories: low, intermediate, and high. Low-risk patients, who scored between 0 and 4 points, had a 5.2% estimated risk of potentially avoidable readmission.

The observed proportion in the derivation set was 5.4%. High-risk patients, who scored 7 or higher on the HOSPITAL score, had an estimated probability of potentially avoidable readmission of 18.3% and an observed probability of 18.7% (Table 2). When compared with other prediction models, Dr. Donzé says the HOSPITAL score had fair discriminatory power and good calibration in identifying the risk of potentially avoidable readmission.

“The HOSPITAL score was developed to identify readmissions that might be prevented and are therefore potentially actionable,” he says. “It can be used for all medical patients regardless of their primary cause of admission. Perhaps most important is the fact that the HOSPITAL score is able to indicate readmission risk before a patient is discharged. This allows clinicians to target a timely transitional care intervention.” The HOSPITAL score can be a valuable tool in the national effort to reduce healthcare costs and improve the quality of care, according to Dr. Donzé. “Identifying patients who at least have the potential to benefit from more intensive transitional interventions is an important first step in reducing hospital readmissions. All patients should receive high-quality transitional care that meets certain standards. We’re not recommending that low- and average-risk patients be deprived of effective transitional interventions. That said, certain interventions that have been shown to be successful are resource intensive. One way to make the best use of limited resources is to reserve them for those who are most likely to benefit from them.” Validating Readmissions Prediction Model Although the HOSPITAL model is promising, no prediction model will be a perfect indicator of preventable readmissions. Because the model was created and validated at one hospital, a multicenter international validation of the model is currently underway. “Predicting potentially avoidable readmissions is only a proxy for identifying who might benefit from specific interventions,” says Dr. Donzé. “Studies of interventions that target this patient group are needed to definitively prove the usefulness of the HOSPITAL score.

 

Patients at Highest Risks for Readmission Based on analysis of ICD-9 coding data, Dr. Sweeney and colleagues found that gastrointestinal complications carried a high risk of readmission (27.6%), while surgical infections reached 22.1%. These top two reasons accounted for nearly half of all readmissions, according to study findings. Complex gastrointestinal procedures, such as pancreatectomy, colectomy, and liver resection, likely had higher complication rates because of the complexity of these surgeries. The top surgical complications identified in the study were wound infections, pulmonary complications, and urinary tract infections (UTIs). Patients with postoperative sepsis or UTIs were about five times more likely to be readmitted than patients without these complications. Postoperative wound infections and postoperative pulmonary complications were associated with a 3.5-fold increase in readmission rates. Additionally, several comorbidities substantially affected readmission risk, most notably cancer, open wounds, and dyspnea. Patients who were immunosuppressed, poor wound healers, and with baseline pulmonary disease were also vulnerable to complications that raise the likelihood of postoperative readmission.