Rampant premature discharge to wards may be behind high ICU readmissions

Many readmission rates may not be high at all but normal

Many critical care units are using a medical standard that is too low when deciding to transfer patients out of the ICU. By setting the bar so low, physicians are almost ensuring that the patient will have to return to the ICU, often sicker and with more serious medical complications.

However, researchers aren’t so certain this is the case. It is only one of several theories being offered to explain a nagging problem of readmissions in critical care that is steadily growing worse.

For decades, nursing administrators have wondered about the phenomenon that causes patients to return to the ICU within 72 hours following an otherwise medically appropriate discharge to another inpatient unit of the hospital. Some readmissions have occurred on the same day of a transfer that was deemed medically correct in all respects.

Several theories have been offered to help explain the reasons for high readmission rates. Pressures leading to early discharge, managed care, and even the limited value of critical care as a medical intervention have been cited as reasons why some patients inevitably return to the ICU.

But in fact, no one is certain whether any single set of factors is responsible for ICU readmissions. "There are no absolutes regarding the reasons for high readmissions," says Derek C. Angus, MD, MPH, associate professor of anesthesia, critical care and medicine at the University of Pittsburgh School of Medicine.

Some authorities even argue that in many cases readmission rates aren’t actually high, but are a function of a particular hospital’s effect on certain patients much like the unemployment rate: There will always be a certain unemployment rate no matter how healthy the economy is.

According to this view, a readmission rate perceived as high at one hospital could be the norm at another. Adding confusion, a dearth of national benchmarks or statistics on readmission norms ultimately makes rational assessments about high rates compared to low ones impossible to determine, according to some critical care experts.

Angus is among them. A certain level of patient recidivism in ICUs appears to be a norm, he observes.

A certain rate will be a constant at hospitals no matter what you do. But it’s difficult to gauge systematically when the rate becomes too high, although everyone knows it intuitively. "Like obscenity, which evades an accurate definition, you know it when you see it," Angus says.

Patient returned to ICU in one hour

In some cases, the situation is undeniable. One veteran nurse recalls a critically-ill surgical ICU patient who was transferred to a medical unit of her hospital at 3 p.m., only to return an hour later because he was too ill to remain on a general med-surg floor.

Minutes before his transfer out of the ICU, nurses had declared him stable. He was not on titration, seemed conscious, and was vaguely alert. His blood pressure had not changed for several hours; nurses had taken him off suctioning the previous night by order of the physician on duty.

But, by some indications, the patient wasn’t considered sufficiently ready for transfer. Instead, his condition was "borderline" at best, says Kathleen T. Bergin, RN, a surgical nurse practitioner at MetroHealth Medical Center in Cleveland, who was there at the time.

The patient could have gone either way. He could have remained on the general medical floor and continued to do well. Or, his condition could have deteriorated as is what actually occurred as a result of the transfer; his readmission would have been a foregone conclusion, Bergin states.

Until recently, medical researchers did not really question factors driving recidivism rates in critical care. Today, overriding concerns about cost and hospital financial results have fueled a demand for reductions in patient utilization and lengths of hospital stays. But, statistical facts that can serve to point the way to achieving these results in critical care have lagged behind the enthusiasm to create standards and best practices, Angus notes.

"The Society for Critical Care Medicine’s Quality Indicators Committee ranked ICU readmissions within 48 hours as the top indicator for judging ICU quality. Yet aside from a few single-center studies, there has been little systematic evaluation of ICU readmission rates," states Angus in a Critical Care Medicine editorial.1

"What’s driving the need for hospitals to perform better? Is it money? Is it an interest in quality? There isn’t enough data about readmission rates across any spectrum to help achieve either of these ends. How can we intelligently speak about the readmission phenomenon?" Angus asks.

Hampering the effort is the fact that hospitals tend to stop collecting data on patients once they leave an ICU. No one really knows statistically how patients react outside the unit, which may be responsible for the readmission, Angus notes. Whatever in-hospital data is available is usually site-specific. Therefore, conclusions regarding what may cause ICU readmissions don’t necessarily apply to other hospitals either in a region or across the country.

Saying that a hospital’s readmission rate is far too high may not have any basis in fact because it becomes impossible to compare two or more institutions. Furthermore, the factors responsible for the readmissions at each facility may be so different that there still would be no basis for comparisons, according to some researchers.

The result is that "questions regarding readmission rate and its correlation with ICU quality remain unexplored," according to Angus.1

Study compares readmission causes

Some researchers are making an effort nonetheless. Physicians at the Critical Care Research Network of the London Health Sciences Centre in Ontario, Canada, studied patients discharged from ICUs at six different hospitals to determine the causes that led to their readmissions.

The retrospective study also investigated the clinical features and outcomes for the cases that were readmitted.2

During the study period, which ranged between Jan. 1, 1995 and Feb. 29, 1996, researchers studied the outcomes of 236 patients who were discharged from the ICU but later readmitted. The study found that patients with gastrointestinal and neurologic disorders faced the greatest risk of an ICU readmission.

Cardiovascular and respiratory diseases were reported as the major reasons for the readmission.

Of the readmitted cases, nearly half, or 45%, had a worsening or recurrence of the initial disease; 39% experienced new complications; and 14% required additional surgery.

Aspiration or bacterial pneumonia comprised 60% of the new respiratory complications in patients admitted at a group of community hospitals. In comparison, 30.2% of readmitted cases were diagnosed with respiratory arrest; 25.6% with pulmonary edema; and 16.3% with pneumonia. These were patients treated at a group of teaching hospitals.

Stark differences between hospitals

In fact, the research revealed stark differences in outcomes between community and teaching hospitals that participated in the study. Three ICUs from two teaching hospitals and four ICUs from four community hospitals participated.

For example, the number of patients who needed at least one ICU readmission during the same hospital stay was higher at the teaching hospitals than the community hospitals. Readmission within 24 hours of initial ICU discharge occurred in 26% of teaching hospital cases compared with 30% in community hospital patients. (For additional comparisons between the two hospital types, see bar graphs, pp. 2-3.)

Researchers didn’t attempt to explain the statistical differences between the two hospital types. But Claudio M. Martin, MD, MSc, one of the project lead investigators, noted that inherent factors that normally distinguish teaching hospitals from non-teaching ones might account for the differences.

Higher 24-hour staff-to-patient coverage and the prevalence of additional residency education resources, including intermediate and other step-down unit facilities at teaching institutions, compared with community hospitals may have contributed to a lower readmission rate. It also may have been a factor in the number of cases that needed more than one readmission at the teaching facilities during the same stay. But these reasons are more speculation than fact, Martin told Critical Care Management.

Yet, one of the surprising trends encountered by investigators was that readmitted patients at either type of hospital faced a high risk of hospital death. Readmitted patients who did not survive were older and more frequently readmitted for new complications than those surviving the hospitalization.

Patient bumping plays a role

Does this fact mean that premature ICU discharges may be at the core of the problem? Angus and Martin hesitate to draw any conclusions. Complications that lead to an ICU readmission in, say, a post operative cardiac patient "may be more reflective of surgical technique than of post-operative ICU care," Angus asserts.

Similarly, a readmission may be due to sub-par care on a general medical ward and may not stem from a premature ICU discharge, Angus adds.

In his study, Martin leaves the point moot: "Further studies are required to determine if patients at risk for readmission can be identified early to improve the outcome."2

Bergin of MetroHealth is convinced that the continuing trend toward premature ICU discharges is having a detrimental affect on patient outcomes.

"The process of bumping patients into general medical wards too soon is at the heart of the problem," she contends. Innovations such as step-down and subacute care units have helped at many facilities.

Nurses are still seeing far too many patients transferred, only to have them return sicker and in need of higher clinical resources. "If all this is being done to save money, it certainly isn’t meeting those goals," Bergin says.


1. Angus DC. Grappling with intensive care unit quality — does the readmission rate tell us anything? Crit Care Med 1998; 26:1,779-1,780.

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.