Proactive Palliative Care in the Medical ICU Affects Length of Stay for High-Risk Patients

Abstract & Commentary

By James E. McFeely, MD, Medical Director Critical Care Units, Alta Bates Summit Medical Center, Berkeley, CA

Dr. McFeely Reports no financial relationship to this field of study.

Synopsis: Using a simple clinical screening tool, patients considered at high risk for death in this closed medical ICU received a basic palliative care consultation. Those with unmet needs received a full consultation with ongoing intervention from the palliative care team. This process shortened ICU length of stay without affecting mortality rates or discharge disposition.

Source: Norton SA, et al. Crit Care Med. 2007;35(6):1530-1535.

Use of specially trained palliative care teams in the ICU has increased in frequency in recent years. Consultation by these teams tends to occur very late in patients' length of stay. Few studies have been performed evaluating outcomes, such as length of stay. This study from the University of Rochester reports the results of prospective evaluation of a performance improvement project in which the goal was to intervene earlier in the patient's ICU course for those high risk patients identified as potentially needing palliative care intervention.

The study was conducted in a 17-bed medical ICU in an academic tertiary care hospital. All patients were screened within 72 hours of admission to identify those with a high adverse burden/benefit ratio and a high risk of death. Patients had to meet at least one of several inclusion criteria, including admission to the ICU following a current hospital stay of greater than a week, age greater than 80 years and the presence of two or more co-morbid diagnoses, diagnosis of stage IV malignancy, status post cardiac arrest, or diagnosis of an intracerebral hemorrhage requiring mechanical ventilation.

Previously palliative care consultations were available through request of the attending physician and occurred relatively infrequently and late in the patient's hospital course. After development of the screening tool, all patients were screened within 72 hours of admission, and anyone who met the inclusion criteria received either a basic or complete palliative care consult. The extent of the palliative care consult was determined after discussions between the palliative care and MICU management team.

During the 4 month long pre-intervention period, an average of 3.8 patients per week screened positive using the High Risk Assessment Tool, but only 5 patients were referred for palliative care consultation using the usual referral process. On average, this consult occurred on day 14 of the ICU stay. During the intervention phase, an average of 4 patients screened positive per week and all were evaluated by the palliative care team.

Baseline demographics of the pre- and post-groups were similar. The total hospital death rates of the two groups were also similar (54% usual care, 59% proactive palliative care). The medical ICU death rate was also unchanged (38% vs 37%). The ICU length of stay for patients dying in the ICU was 5.7 days in the intervention group and 14 days in the usual care group. The ICU length of stay for patients dying elsewhere in the hospital was 9 days in the intervention group and 16 in the usual care group. For those patients who died in the hospital but not in the ICU, there was no difference in their overall hospital length of stay.


This study, while having some methodologic flaws, provides further common-sense evidence supporting the value of palliative care early in the course of a high risk patient in the ICU. The study hospital was fortunate enough to have an on-site trained palliative care team. They developed a clinically relevant and easily implementable screening tool and had the resources available to see all these patients within an average of 1.7 days. This aggressive schedule of intervening resulted in a reduction of the ICU length of stay by an average of 8 days while having no net effect on overall mortality rates. This lack of effect on overall mortality suggests that a palliative care consultation did not hasten or otherwise impact patients' death, but rather assisted with appropriate resource utilization for patients who were going to die anyway.

For those hospitals with a palliative care service, the study provides further motivation for using the service early in high-risk patients. For those without a palliative care service, it provides further compelling information that may be used with hospital administrators in order to develop such a program.