Figure 1. A depiction of the relative importance of the various locations for the patients prior to admission to the surgical intensive care unit in the pre–step-down unit (SDU) and the post-SDU eras. OR indicates operating room; RR, recovery room; ED, emergency department; and IR, interventional radiology.
Figure 2. A depiction of the primary diagnoses for the patients admitted to the surgical intensive care unit only in the pre–step-down unit (SDU) era (PrICU), the post-SDU era (PoICU), and to the SDU (intermediate care unit) only in the post-SDU era. Possible diagnoses include gastrointestinal (GI) tract disease; vascular disease; urologic or gynecologic condition; neurologic or neurosurgical disease; sepsis; trauma; ear, nose, throat condition; transplant; cardiovascular; or patients admitted for monitoring after interventional radiology, and all other patients not included in one of these diagnoses (miscellaneous).
Eachempati SR, Hydo LJ, Barie PS. The Effect of an Intermediate Care Unit on the Demographics and Outcomes of a Surgical Intensive Care Unit Population. Arch Surg. 2004;139(3):315-319. doi:10.1001/archsurg.139.3.315
Many surgical intensive care units (SICUs) face bed shortages for acutely ill patients that may result from a large proportion of bed occupation by chronically ill patients. We hypothesized that the introduction of a new intermediate care or step-down unit (SDU) managed by surgically trained intensivists would allow the admission of more acutely ill patients while maintaining satisfactory outcomes.
Prospective retrospective comparison of SICU patient populations before and after the introduction of an SDU.
The SICU of NewYork-Presbyterian Hospital, New York Weill Cornell Center, a university hospital containing a level I trauma center.
Patients and Interventions
All patients in the SICU admitted from August 1, 1996, through June 30, 1997, were SICU patients prior to the introduction of the SDU. Patients admitted from August 1, 1997, through June 30, 1998, were SICU post-SDU patients, and SDU patients included those admitted to the SDU from August 1, 1997, through July 1, 1998.
Main Outcome Measures
For each of the 2 eras, patients were compared for age, sources of admission, Acute Physiology and Chronic Health Evaluation (APACHE) II and III scores, unit length of stay, and mortality. Other data collected included origin of admission, nature of admission, and diagnosis.
Six hundred sixty-six patients were admitted during the pre-SDU era, while a total of 1117 patients (619 SICU and 498 SDU patients) were admitted in the post-SDU era. After the introduction of the SDU, the mean (standard deviation) APACHE II scores of the SICU and SDU patients increased (14.2 vs 13.4, P = .04) without affecting mortality (6.0% in the post-SDU era vs 8.2% in the pre-SDU era, P = .07). The post-SDU era had a higher proportion of emergency admissions (42.3% vs 48.6%, P<.05) and interhospital transfers (7% vs 1%).
Opening an SDU resulted in a significant increase in the overall severity of the SICU population. Creation of an SDU managed by surgically trained intensivists may optimize the use of a hospital's resources, permit the expansion of emergency or tertiary care services, and improve outcomes for critically ill surgical patients.
The efficient use of critical care beds is a challenging endeavor for any hospital. Staffing these units to capacity with a full complement of physicians, nurses, and other personnel can be difficult. Additionally, the inpatient census of many hospitals is reflecting increased levels of acuity while the expense necessary to maintain these facilities is rising.1 Consequently, decreased bed availability may be most prominent in specialty intensive care units (ICUs) and cause transfers of critically ill patients from other areas of the hospital to be delayed or redirected to another ICU.
To decrease the strain on ICU beds, intermediate care or step-down units (SDUs) have been created in some hospitals.2 The functions of SDUs vary, but most accommodate patients who need either cardiac monitoring or more frequent nursing attention than allowed in the general ward.3- 6 Some SDUs perform long-term mechanical ventilatory treatment and some may even manage low-dose vasoactive drug infusions.7- 9 Nurses working in SDUs often have ICU experience or extra qualifications in critical care, such as the certified Critical Care Registered Nurse (CCRN) designation.
On July 11, 1997, our surgical ICU (SICU) moved to a new facility in a newly constructed portion of the hospital. Both units were managed by surgically trained intensivists and staffed by surgical residents, specialty nurses, and respiratory therapists. The former unit was a 9-bed unit in which surgical, trauma, and noncardiothoracic subspecialty patients were treated. The new unit was a 10-bed unit with an adjacent 4-bed intermediate care unit. Patients admitted to the new SICU area included those short-term treatment patients whose vital signs were monitored every 2 hours or less, those with special acute care issues such as pulmonary artery or intracranial monitoring, or those receiving acute ventilatory support. The other surgical patients requiring a monitored setting with less acute care issues could be placed in the step-down area. These patients might include those patients receiving chronic mechanical ventilation therapy who do not need frequent suctioning or those patients who are not receiving mechanical ventilation therapy but who do require more nursing care than offered by the general ward. Continuous electrocardiographic monitoring was available in both units and interpreted by trained critical care nursing personnel. The nursing ratios in the SICU and SDU areas were a maximum of 2:1 and 4:1, respectively. Twenty-four–hour bedside physicians were present on both units. The same critically care–trained nursing pool staffed both units.
On expansion of our facilities, we designed a prospective study intending to capture the effect of the new SDU on use and outcomes of the SICU patient population. We believed that in the pre-SDU era many urgent critical care transfers to our SICU were delayed or refused owing to lack of bed availability. We hypothesized that the extra beds created by the new SDU would allow us to treat more patients with high acuity in our own SICU. We also hypothesized that the demographics of our critically ill patients would not differ substantially from our prior patient mix because no new expansion of any particular surgical service was initiated during the time of the change.
Patients admitted between August 1, 1996, and July 1, 1998, who entered the SICU of NewYork-Presbyterian Hospital, New York Weill Cornell Center, a tertiary university center containing a level I trauma center, were enrolled in the study. The 4-bed SDU was opened adjacent to the SICU on July 11, 1997. All patients admitted to the SICU from August 1, 1996, through July 1, 1997, were considered pre-SDU (PrICU) patients. All SICU patients admitted from August 1, 1997, through July 1, 1998, were considered post-SDU (PoICU) patients. Step-down unit patients were those patients admitted to the SDU from August 1, 1997, through July 1, 1998. Both PoICU and SDU patients were the total number of SICU patients after the SDU inception. July 1997 was considered a transition month and was excluded from the analysis.
For data analysis, each patient counted as a single admission to prevent certain patients such as readmissions from inappropriately affecting the demographic or mortality data. Therefore, patients who spent time in more than 1 unit in the newer era or who were readmitted in the older era were only counted once. Readmissions were also analyzed separately. In the pre-SDU era, all patients transferred to the ward from the SICU were considered readmissions. In the post-SDU era, all patients transferred to the ward from the SICU or SDU and readmitted to either the SICU or SDU were considered readmissions.
Data were collected prospectively for all the patients admitted to the SICUs and SDUs for both eras. Data collected included age, sources of admission (operating room, emergency department, surgical or other hospital floor, recovery room, and interhospital transfer), Acute Physiology and Chronic Health Evaluation (APACHE) II and III scores, unit length of stay, and mortality. The APACHE III scores were not calculated for the SDU patients because its use has not been validated in a population with this level of acuity.10 Cumulative multiple organ dysfunction scores were also tabulated for each patient.
The primary diagnosis for the patients was also recorded. These diagnoses included gastrointestinal; vascular; neurologic or neurosurgical; sepsis; urologic or gynecologic; trauma; ear, nose, throat service; transplant; or cardiovascular conditions and patients who required monitoring after receiving an interventional radiological procedure. Destination after discharge from the SICU or SDU for each patient was also recorded and particular attention was paid to whether the patient went to the SDU, SICU, or surgical or other hospital floor after the stay in any of the designated units.
A separate analysis was performed for patients admitted emergently to the SICU in the pre-SDU vs the period post-SDU era. An emergent admission was defined as an admission from the emergency department or of a postoperative patient who was not listed on the published elective surgical schedule. Data collected for these patients included age, APACHE II and III scores, unit length of stay, and mortality.
Data are presented as mean (SE). Differences in coordinate data were analyzed by the χ2 and Fisher exact tests. Differences in continuous data were analyzed by univariate analysis of variance. Nonparametric tests (Kolmogorov-Smirnov test) were used to test differences in unit length of stay. Statistically significant differences were accepted for P<.05. Data gathered in this study were approved for review by our institutional review board as part of an ongoing quality assurance project evaluating outcomes in our SICU.
Total admission figures to the SICU before and after the creation of the SDU are given in Table 1. Because of the increased number of beds, the number of patients admitted after the SDU increased to 1117 from 666. However, the severity of critical illness by mean (SE) APACHE II scores of the entire patient cohort also increased in the post-SDU era (13.4 [0.3]-14.2 [0.3], P = .04). The mean age and unit length of stay did not change between the 2 eras. The mortality trended to a lower value (6.0% vs 8.2%, P = .07) in the post-SDU era despite the increased acuity.
Other demographic data between the 2 eras were also compared. The proportion of patients admitted postoperatively decreased in the post-SDU era, whereas the proportion of patients admitted emergently increased. Patients admitted to the SICU after the introduction of the SDU were more likely to be admitted emergently (P<.001), have higher APACHE III scores, and have a longer unit length of stay than the previous cohort of SICU patients.
Other specific characteristics of the patient groups were also examined. A depiction of the source of SICU patients pre- and post-SDU as differentiated by emergency room, surgical or other hospital floor, another hospital, interventional radiology, operating room, or recovery room is shown in Figure 1. The primary diagnoses for all the patients are shown in Figure 2. Patients with neurologic illness, sepsis, and trauma increased in proportion to the SICU, whereas patients with gastrointestinal tract illness and patients for cardiac monitoring increased in the SDU compared with their previous proportions.
Several other patient group analyses were performed. In the post-SDU era, the APACHE II and III scores were notably higher for the PoICU patients compared with the PrICU patients. Another analysis was performed for patients admitted only in the post-SDU era (Table 2). As expected, APACHE II scores and mortality rates for SDU patients were significantly lower than those for ICU patients in the same period.
The demographic data of patients emergently admitted are listed in Table 3. Of the total proportion of admissions, 42.5% were admitted emergently pre-SDU and 64.9% (P<.001) were admitted emergently post-SDU. The origin of the patients was further analyzed. One hundred ninety-five patients entered the SDU from the ICU and 25 patients entered the ICU from the SDU. Fifty-two patients required more than 1 admission to the SICU during the in-hospital stay and 13 patients required more than 1 admission to the SDU. In the pre-SDU era, 79 (6.3%) of the total 666 patients were readmitted to the SICU or SDU. In the post-SDU era, 111 (4.3%) of the total 1117 patients were readmitted to the SICU or SDU.
We also studied the cumulative multiple organ dysfunction score for the different cohorts of patients. In the pre-SDU era, the mean (SE) multiple organ dysfunction score was 2.4 (0.17) for the 666 PrICU patients. In the post-SDU era the multiple organ dysfunction score was 2.6 (0.18) for the 619 PoICU patients. These 2 scores were not statistically significantly different (P = .37).
Although studies have shown that perioperative care by trained intensivists in specialized critical care settings may produce better outcomes, many hospitals have decreased the number of critical care beds.11- 17 The major reasons for these declines include staffing issues and cost. The job-related burnout rate for critical care nurses may be much higher than other areas in the hospital because ICUs may demand the most labor-intensive responsibilities from their nurses.18 Additionally, staffing ICUs to capacity may be more expensive than other hospital units for a variety of reasons including the low nurse-patient ratios.19- 22
Having all the critical care beds filled in a particular hospital affects both inpatients of that hospital and certain patients outside the hospital. From an inpatient standpoint, inpatient transfers are delayed or refused and acutely ill patients from the ward may be delayed in finding a bed in the ICU. When a patient transfer to a specialty ICU is delayed, the patient may be put in another critical care unit less well equipped to provide the particular specialization of care needed by that patient. However, patients outside the institution may be affected as well as the fact that the emergency department may be forced to go on ambulance and trauma diversion if all critical care beds in a facility are occupied.
Unlike many institutions, we were afforded the opportunity to study the effects of expansion on our SICU population. We found the provision of an SDU affected the population of the SICU in several discernible ways. Notably, the overall level of acuity managed in our new SICU actually increased after the addition of the new SDU. Logically, adding a lower acuity unit would bring some healthier patients to the SDU and decrease the overall acuity of the patient mix.23 One explanation of this finding is that perhaps the extra beds facilitated the admission of more emergently ill patients, many of whom might have been transferred to other units in the pre-SDU era. Our data support this explanation as our proportion of emergency admissions increased with the expansion.
In our study the overall acuity increased by the mortality of the most severly ill patients (the PrICU and PoICU patients) did not differ between the 2 eras. This observation paralleled the finding that the multiple organ dysfunction scores did not differ between the 2 eras. Therefore, we were able to treat graver overall pathologic conditions of greater severity without the patients developing increased multiple organ failure. This finding may relate to a multiplicity of factors including more timely bed availability, clustering the most severely ill patients, as well as improvements in other critical care management strategies. Perhaps with more critical care beds, we were able to admit patients earlier in their critical illness without delay and proceed immediately with supportive care to diminish the development of subsequent multiple organ failure.24- 26 We can defend this explanation with certainty since we did not compare waiting times for ICU admission or interventions performed by the primary services between the 2 eras.
Another theory is that more beds may allow patients with impending organ failure manifested by tachypnea, hypotension, or oliguria to be treated earlier and more aggressively. Other authors27,28 have suggested that holding ill patients in the ICU before they are fully ready for transfer may limit the development of new organ dysfunction by subsequent readmission or an adverse cardiopulmonary event. With more critical care bed availability, we have been less pressured to transfer out "borderline" ward patients.
Our study also revealed other important features of the changing demographics of critically ill surgical patients. More recently, emergently ill patients have not only become a higher proportion of the SICU population, but they have also been a population with an increased acuity. Consequently, our SICU has seen an increased number of patients admitted from the general ward, the emergency department, and other institutions. Many of these patients may not have been able to be transferred to the SICU in the pre-SDU era in which fewer beds were available. Concurrently, routine postoperative SICU admissions have decreased. This finding may be partially because of the decreased number of open vascular procedures performed at our institution as well as the emergence of minimally invasive surgery.
Certain limitations of our study must be acknowledged. We did not capture the number of ICU or intermediate care admissions refused because of bed unavailability during the study period. Similarly, information regarding bed occupancy was not meaningful since maximum bed capacity was not a constant factor but dependent on other factors such as nurse staffing and the acuity of the patient's condition. We were also unable to stratify detailed cost data for individual patients as they sometimes were present at times during their hospital stay in both units. Also, while we used APACHE II and III scoring as our method of illness severity, these modalities may not accurately describe the pathologic condition of these surgical patients. Finally, although our outcomes improved after the introduction of the SDU, we cannot conclude that the availability of the new unit itself improved outcomes.
Nevertheless, our study did show that an ICU managed by surgically trained, critical care personnel could maintain its outcome standards despite increased acuity and volume. The increase in the number of total critical care beds may have facilitated more immediate care to be provided to acutely ill patients as well as a closer follow-up of care to these patients after the period of highest acuity. Despite the increased overall acuity, these satisfactory patient outcomes seemed to justify the expansion of our SICU to include an SDU. Individual hospitals should assess whether the addition of an SDU staffed by surgically trained intensivists may improve the outcomes of critically ill surgical patients at their institution.
Corresponding author and reprints: Soumitra R. Eachempati, MD, 525 E 68th St, Payson 718 A, NewYork-Presbyterian Hospital, New York, NY 10021 (e-mail: email@example.com).
Accepted for publication October 29, 2003.