Executives at hospitals getting socked with readmissions penalties up to 3% of their Medicare reimbursement may dream of a formula that accurately predicts which patients will return within 30 days, and which of those they can prevent.
But a team from Brigham and Women's Hospital in Boston, collaborating with a team in Bern, Switzerland, is closer to making that dream come true for patients with medical illnesses, (but not for surgical patients).
Through a simple risk score, they say they can identify roughly one-fourth of a hospital's patient population with the highest likelihood of being readmitted, and then within that group the18% whose readmissions were potentially avoidable, for whom more expensive, intensive efforts might be worth the money.
Among 9,212 adult discharges from July 1, 2009 to June 30, 2010, and 2,398 readmissions, the researchers were able to prospectively identify 879, or 8.5% of all discharges, who were at high risk of a potentially avoidable readmission.
Jacques Donzé, MD, lead author of the report published in JAMA Internal Medicine, and colleagues wrote, "This easy-to-use model enables physicians to prospectively identify approximately 27% of the patients as high risk of having a potentially avoidable readmission." They added that it "would allow targeting intensive transitional care interventions to patients who might benefit the most."
What's more, it allows hospital teams to identify these patients at or before discharge, before they left the hospital, and before their health status was jeopardized again.
The formula they designed weighs seven factors:
- Hemoglobin levels at discharge
- Whether the patient was discharged from an Oncology service
- Sodium level at discharge
- The Procedure performed during the admission
- The Type of index admission,
- The frequency of Admissions in the prior 12 months
- The Length of the patient's hospital stay
Lined up, the algorithm's components spell HOSPITAL, which makes it easy to remember and score, Donzé said.
In response to questions, co-author Jeffrey L. Schnipper, MD, also of Brigham and Women's, in an e-mail, elaborated on how the equation would work for a hospital where 22.3% of 1,000 patients were readmitted within 30 days, or 223 patients.
Out of those 1,000 patients, those whose HOSPITAL scores were in the top 25% would be identified. From that group, 18% would have a potentially avoidable readmissions.
If "intensive transition interventions" were applied to those 250 patients, and the interventions were 100% effective, "then you would prevent 18% x 250 or 45 readmissions," reducing the number of readmissions from 223 per 1,000 to 178, "a 4.5% absolute reduction in the readmission rate."
Practically speaking, he added, those interventions would not be 100% effective [it would be] more like 50%. "So it may be more like half that, a 2.3% absolute reduction, i.e. from 22.3% to 20%. But this is still substantial. If we could do this nationally, it would save society about $2 billion."
The team excluded unavoidable readmissions, for example a patient originally treated for heart failure, but who was readmitted because of newly developed conditions not related to known diseases during the index admission. They also excluded planned readmissions, for example a cancer patient's scheduled chemotherapy treatment.
"To our knowledge, no previous studies included the number of procedures performed during the index stay in their models," and hemoglobin level as a way to measure anemia has only been used in one other study, they said.
In particular, Donzé wrote, number of prior hospitalizations and their length of stay of the index admission "were important predictors," because they represent a way to measure a patient's illness severity.
It's unclear how translatable such success in a Boston hospital might be, since Massachusetts health reform laws gave hospitals in that state a head start incentive to reduce readmissions through penalties imposed on Medicaid readmissions too.
Schnipper suggested that hospitals should consider implementing the following interventions:
- A patient coaching intervention, similar to the Coleman Care Transitions Intervention.
- Inpatient and post-discharge pharmacist interventions, including medication reconciliation, patient counseling at discharge, and post-discharge phone follow-up.
- An inpatient "discharge advocate," similar to Project RED.
He emphasized that all patients should receive high-quality discharge processes, "but effective and potentially expensive interventions like those mentioned above should be reserved for those patients most likely to benefit."
In an invited commentary in the same issue, Eric Marks, MD, of the Department of Medicine of the Uniformed Services University of the Health Sciences in Bethesda, MD, suggested that the Brigham and Women's score deals "primarily with severity of illness, which is "commonly identified as a major contributor to the risk of subsequent admissions, but specifics as to how these risks can be modified or contained continues to be an issue."
The other problem is what to do about it, with "documented adherence" to well-known interventions failing to consistently reduce readmission frequency.
Even the penalty, now up to 3% of a hospital's Medicare base DRG for three conditions, has failed to prompt a significant reduction in readmission rates so far, although some progress has been seen.
"This lack of significant improvement in readmission rates suggests that there is a much more complex relationship between risk factors, interventions, and intended outcomes," Marks wrote.