Predicting ICU Readmission After Cardiac Surgery

Abstract & Commentary

Synopsis: Patients most likely to be readmitted to the ICU were those with the longest time in the ICU prior to discharge. The most common reason for readmission was pulmonary problems.

Source: Cohn WE, et al. Chest 1999:116:688-692.

To evaluate factors that increase the likelihood of intensive care unit (ICU) readmission, Cohn and colleagues surveyed 2388 consecutive patients who underwent cardiac surgery at their institution from 1994-1997. Patients were excluded from analysis if they died during surgery (n = 15), in the ICU (n = 87), or when available data were incomplete (n = 58), yielding a final sample of 2228 patients. Of these patients, 128 were readmitted to the ICU. One investigator who identified the single most prominent reason for ICU readmission reviewed the discharge summary of each patient readmitted to the ICU.

Over the four-year period, ICU stay decreased (P = 0.048), but ICU readmission as a percentage of discharges increased from 3.9% (1994) to 8.4% (1997) (P = 0.005). When patients readmitted to the ICU were divided into four equal groups based on their initial ICU length of stay (LOS; 1 = shortest; 4 = longest), the incidence of readmission was 3.9%, 5.2%, 4.7%, and 9.2%, respectively. Patients who required readmission had a longer (P = 0.008) mean initial ICU LOS as compared with those who did not require readmission. "Long stay" patients also had a longer secondary stay compared with "short stay" patients.

Preoperative factors associated with a greater likelihood of ICU readmission included a history of congestive heart failure (P = 0.0001), lower left ventricular ejection fraction (P = 0.0080), female gender (P = 0.019) and > three comorbidities. Postoperative and intraoperative factors associated with increased likelihood of ICU readmission included a greater weight gain during ICU stay (P = 0.038), longer initial ventilator time (P = 0.038), and homologous blood use in the operating room (P = 0.045). Three factors accounted for 75% of all readmissions: respiratory problems (40.6%), dysrhythmias (21.1%), and the need to return to the operative room (13.3%). The percentage of patients readmitted for pulmonary problems rose from 1.5% (1994) and 0.9% (1995) to 3.2 % (1996) and 3.0% (1997) (P= 0001).

Comment by Leslie A. Hoffman, PhD, RN

This study was prompted by a review on the part of the cardiac surgical team that indicated an apparent increase in the return of patients to the ICU after an initial discharge. This observation raised concerns that the decrease in post-cardiac surgery ICU LOS prompted by managed care had adversely affected patient outcomes. However, the central finding of the study was that patients most likely to return to the ICU were those with the longest stay, not those with the shortest stay.

Patients in this sample seemed typical of those individuals who require cardiac surgery: mean age was 65 ± 11.8 years, 32.4% were female, and 73.8% underwent coronary artery bypass grafting. However, no further information was given regarding preoperative health status. Patients were managed by one group of cardiac surgeons, with transfer to a surgical ICU staffed by surgical and anesthesiologist intensivists. After their ICU stay, patients were transferred to a dedicated rehabilitation unit with telemetry. ICU discharge was protocol-based, not time-based, and there was no evidence that unit capacity caused premature discharge.

Although many predictive factors identified in this study are not amenable to intervention, there were some notable exceptions. Patients who required ICU readmission were likely to have a greater mean weight gain while in the ICU (10.3 ± 4.7 kg vs 8.5 ± 5.6 kg), and longer time on mechanical ventilation. In almost half the cases, the reason for ICU readmission was a pulmonary problem which developed on the clinical unit. No information was given regarding stability of the nurse:patient ratios or respiratory care staffing during the four-year data collection interval. One of the many consequences of managed care has often been a decrease in the number of nurses and respiratory therapists available to manage the care of acutely ill patients, combined with an increase in patient severity of illness. We have substantially decreased ICU LOS for most cardiac surgery patients without adverse outcomes.

As Cohn et al note, the next challenge is to devise management strategies targeted at patients at high risk for ICU readmission that might prevent the need for readmission or, at the least, decrease the overall length of ICU stay. One method of accomplishing this goal is analysis of the extensive data in databases to determine causes of the problem. Another method involves testing innovative solutions. Kirby and Durbin (Respir Care 1996:41:903) tested the impact of a dedicated respiratory therapy assessment team for non-ICU patients that visited all patients prior to and following ICU discharge. The team was successful in reducing mortality of readmitted patients and the proportion of patients readmitted for respiratory failure. Broader testing of this and other creative solutions is needed to help patients at highest risk bridge the critical interval between ICU discharge and recovery.

Cardiac surgery patients most likely to be readmitted to the ICU had:

a. the shortest initial ICU stays.

b. the longest initial ICU stays.

c. intraoperative complications.

d. varying lengths of time in the ICU.