Are early transfers, discharges to blame for high readmissions?

Quest for a normal’ readmission rate can be elusive

Is there a "normal" readmission rate for the ICU? After complaining about the high cost of critical care, ICU professionals have not fully determined whether there is (or should be) a universal average in intensive-care readmissions.

With a debate over cost still raging, the solution to why so many patients are being readmitted during the same hospitalization and what to do to curb their numbers seems to elude the experts, according to individuals who have studied the problem.

The finger of blame seems to point to the current trend of early transfers and discharges to less intensive, less costly levels of acute care. But is early discharge from the ICU really to blame, some experts ask?

Uncertainty about early transfers

Veteran managers defend early transfers and assert that hospitals are getting better at case management and fostering greater collaboration between ICUs and general medical-surgical floors. The goal has been to ensure both appropriate and timely patient transfers.

Nevertheless, patients are returning to the ICU in worrisome numbers. Studies show that a large percentage of those patients are returning sicker, and in many cases with new medical problems.

They also "require stays in the ICU that are longer than if they had not been transferred in the first place," says researcher Charles G. Durbin Jr., MD, associate professor of anesthesiology at the Virginia Health Sciences Center in Charlottesville, who has studied ICU readmission rates.

Within 48 to 72 hours, between 30% and 50% of all readmitted cases present in the ICU with new complications, compared with 30% of patients who were not previously transferred to other units, he says.

Medical ICUs (MICUs), according to Durbin, generally see a higher rate of complications, compared to surgical ICUs (SICUs). And cardiac surgery cases have a better track record than general surgery patients.

The chief reason, according to Durbin, is that most SICU patients tend to be in better general health and have fewer predisposing complications than MICU patients. Otherwise, they would not have made good surgery candidates, Durbin adds.

While some studies suggest that better respiratory assessment and other planned interventions can decrease mortality and morbidity in readmitted patients, the numbers have not been overwhelming thus far.1

Meanwhile, ICU care continues to be increasingly expensive. More than 20% of every dollar spent on hospital care goes to treating critical care cases, according to figures from the U.S. Office of Technology Assessment in Washington, DC.

That amount has been increasing rapidly, experts say. Meanwhile, the "positive effects of critical care treatment on patient outcomes have been difficult to gauge," according to Durbin.

Absence of universal benchmarks

The answer, experts agree, isn’t longer ICU stays. But most intensivists assert that something has to be done to slow the rate of readmissions.

But determining a "normal" or a typical readmission rate for any unit isn’t so simple, muses epidemiologist Liddy M. Chen, MB, MSc, a biostatistician who holds a medical degree in China and is employed by PPD Development, a pharmaceutical research firm in Raleigh, NC.

Chen studied ICU readmission cases as a researcher at the Critical Care Research Network of the London Health Sciences Centre in Ontario, Canada.

Reducing readmissions by making comparisons to a predetermined standard or absolute rate won’t work because very likely there is no absolute, universal rate in existence, Chen says.

Nevertheless, nurses can achieve a consistent unit-specific baseline rate over time based on internal environmental factors for each unit. Here’s how:

Analyze early transfer trends.

Certainly, early transfers play a role in high morbidity and mortality among readmitted cases, Chen says. Although many departments have strong case management and discharge planning protocols, hospitals still face enormous pressures to shorten lengths of stay.

Analyzing whether the balance between judicious, appropriate discharges and economic imperatives does not inadvertently favor one at the expense of another may help determine whether you are transferring patients too early, says Chen.

Assess the quality of external support.

At the same time, concentrating on correcting potentially flawed discharge planning protocols can mask other factors leading to high readmission rates.

The problem actually may lie in the poor quality or lack of sufficient patient support services in a stepdown or receiving medical-surgical floor, some studies show.2

For example, the nurse-to-patient ratio or the effectiveness of the respiratory therapy staff on those floors may be inferior or inadequate to support your patients adequately, Chen notes.

If such is the case, managers and discharge planners should consider modifying their discharge policies to account for the potential risk associated with a patient transfer.

Consider the unit’s case mix.

A problem associated with benchmarking lies in drawing faulty comparisons between different sets of patient acuity and diagnosis factors between the benchmark hospital and one’s own facility, studies show.

Readmissions skewed by case mix

While striving for reduced readmission rates may be a worthy goal, in reality, readmission levels vary markedly by diagnosis. They also vary by the patients’ levels of medical need, acuity differences, and the overall case mix within a department, Chen says.

For example, in one study of readmission rates among teaching hospitals, cardiovascular conditions led among a group of four leading diagnostic causes of ICU readmissions.

The same study showed that among disease categories, the leading causes of readmission, barring new or original complications, were renal disorders, gastrointestinal and neurologic problems, and sepsis.

But more than 50% of patients readmitted to the ICU suffered from a cardiovascular condition as the primary diagnosis. (The other three causes were respiratory, gastrointestinal, and neurologic. The graphs on p. 4 illustrate those and other comparisons.)2

Furthermore in the same study, respiratory problems increased significantly (by about 40 percentage points) in readmitted patients, compared to those originally admitted with other problems.

The level and sophistication of technology can also influence a department’s readmission rates, complicating the definition of a "normal" rate for any unit, says Chen. "Under this scenario, things are always changing and are different every day," she says.

Still, there are ways to determine what for a given unit may be a normal, or typical, readmission rate. What you may be aiming for is a statistical average, or mean. But you also may be looking for the disease category with the highest frequency of readmissions in a sample size.

Several suggestions follow:

Make intelligent comparisons.

By keeping the above-cited variables constant, Chen says, a department can compare itself with another in setting benchmarks based on retrospective discharge and diagnostic data. The data can be shared especially when a facility belongs to a large system or group of affiliated providers.

Large urban hospitals are more suitable for comparisons due to the breadth and variety of their disease categories. But that tenet doesn’t always hold true, Chen says, due to other factors that can make two urban providers quite dissimilar.

Another caveat: Teaching, community, or rural hospitals should be compared to facilities of the same type but not with each other, regardless of any similarities in their patient mix, according to Durbin’s study.

Stratify the findings.

When looking for a normal readmission rate, focus the search on narrowly defined diagnosis groups, such as all 45-year-old male patients with perforated gastrointestinal tracts.

Avoid drawing conclusions regarding the unit’s overall readmission rate. The unit’s overall rate will not necessarily mean anything because it says virtually nothing about specific cases and disease categories, which are the key to what drives readmissions, says Chen.

Focus your search.

It is often easier to calculate an average or median, which is the middle rate between the highest and lowest range when you compare similar diagnoses and disease groups.

But ensure that the sample size is substantial. Three or four years of retrospective data should be sufficient, says Sue Chapman, RN, a clinical coordinator at York County Hospital in Ontario, who helped compile data for Chen’s study.

It is fairly simple to extrapolate those figures from a patient database. But it means you will be able only to speak of a normal readmission rate for a specific diagnosis or disease category and not the whole unit, Chapman says.

Account for high-risk cases.

Determining a "normal" readmission rate for high-risk patients can be useful in setting policies to lower the number of high-risk cases who return to the unit.

When conditions in the unit change, clinicians can use the data to help track the reasons that readmitted patients are developing, such high-risk complications outside the department, says Chen.

Was it that certain patients were transferred too soon? Or were existing clinical resources throughout the hospital inadequate to address the needs of the readmitted patients at the time?

Allow for discharge planning changes.

Finally, changes in discharge planning policies and procedures may account for wide variations in subsequent readmissions. Track any changes in such policies during the three years of reviewing the patient data, Chapman advises. They could offer a key to reducing the readmission rate to what in the process may be deemed normal for the department.

Ultimately, the only prediction that can be made with certainty is that a unit will have a minimum readmission rate of some kind. Even that may change over time.

And due to changing factors in the unit, the minimum will likely seldom be achieved. Instead, the usual rate will probably be much higher and subject to further change, Chen says.

Even while appropriate lengths of ICU stays can help achieve better outcomes and fewer medical complications, a nagging question will remain as to which patients are most likely to benefit most from such policies. As yet, that isn’t known.

"Information about this group of patients is required to improve clinical decision-making," Chen says. "Unfortunately, only a few studies in the United States have investigated these patients and the reasons for their readmissions."


1. Kirby EG, Durbin Jr. CG. Establishment of a respiratory assessment team is associated with decreased mortality in patients readmitted to the ICU. Resp Care 1996; 41:903-907.

2. Chen LM, Martin CM, Keenan SP, et al. Patients readmitted to the intensive care unit during the same hospitalization: Clinical features and outcomes. Crit Care Med 1998; 26:1,834-1,841.