Rapid Response Team Calls and Unplanned Transfers to the Pediatric Intensive Care Unit in a Pediatric Hospital

This study was a retrospective cohort study with deidentified data. After the study was approved by the appropriate institutional review board, data were collected from databases compiled at Miami Children’s Hospital, a single tertiary care, freestanding 289-bed facility, for the period July 2007 through September 2012. The databases included local RRT logs and the VPS database (a clinical database dedicated to standardized data sharing and benchmarking among PICUs). The hospital has a neonatal ICU, a PICU, and a pediatric cardiac ICU. The hospital is a level 1 pediatric trauma center.

The RRT was instituted in 2007 and consists of the PICU charge nurse, a respiratory therapist, and a third-year pediatric resident. Backup team members include a PICU fellow and a PICU attending physician. Rapid response calls can be initiated by health care providers, hospital employees, and patients’ family members. Each rapid response call is evaluated by the RRT, and disposition is based on the team’s clinical judgment. Criteria for RRT activation include worsening of respiratory or cardiovascular status, change in mental status, and a health care provider’s concern about a child’s condition. An RRT call is not required before every unplanned transfer. If a patient’s health care team on an acute care unit decides that a child needs a higher level of care, then the patient is transferred to the PICU after discussion with the PICU health care team without activation of the RRT team. The hospital does not have intermediate care or a high-dependency unit. All unplanned transfers to a higher level of care are to one of the ICUs (pediatric, cardiac, or neonatal).

With the initiation of the RRT, a database was created in which demographic information, time and date of call, and disposition were recorded. The PICU charge nurse was responsible for completing the database immediately after the evaluation of a rapid-response patient. Data collected included time of rapid response and disposition of the patient after the rapid response and demographic information on the child. Demographic information included age, weight, and sex. In addition, data on unplanned transfers to the PICU during the same period as the study period were retrieved from the local VPS database and were compared with data on all other PICU admissions.

The rate of RRT calls during the day shift and the night shift and differences in disposition rate after RRT calls between day and night shifts were analyzed by using χ² analysis. For disposition, binary data (transfer to PICU or not) were used. The differences in mean length of stay and scores on the Pediatric Index of Mortality 2 risk of mortality between unplanned transfers and other PICU admissions were analyzed by using the Mann Whitney test, and differences in mortality rate were analyzed by using χ² analysis. The value P < .05 was considered significant.

During the study period, 542 rapid responses were initiated. Of these, 321 (59.2%) were called during the daytime (7 am–7 pm). During the responses, 323 children (59.6%) were transferred to the PICU, 164 (30.3%) remained in an acute care unit, 19 (3.5%) required resuscitation and eventually were transferred to the PICU, 5 (0.9%) were transferred to the emergency department, and 31 (5.7%) had an unknown disposition. More children (P = .048) were transferred to the PICU after RRT calls during the daytime (66% of daytime calls) than during the nighttime (59% of nighttime calls). During the same period, the total number of unplanned transfers to the PICU from the acute care units was 1313. Of these, 323 children were transferred after RRT activation. Age, sex, scores on the Pediatric Index of Mortality 2 risk of mortality, length of stay, and mortality rate did not differ according to the time of transfer among all unplanned transfers from acute care units to the PICU. The mortality rate among children transferred from acute care units (3.8%) to the PICU was significantly higher (P < .001) than the rate among other PICU admissions (1.4%).

We found that more rapid responses were called during the daytime shift than during the night shift and that the mortality rate among patients transferred to the PICU from the acute care units was greater than that of PICU admissions from all other sources, such as the emergency department or operating room. Most patients who required an RRT call were transferred to a higher level of care. Of the total unplanned transfers from an acute care unit to the PICU, only 25% occurred after an RRT call. In contrast, in a study of adults, Jäderling et al⁹  reported that 51% of unplanned transfers occurred after an RRT call. At Miami Children’s Hospital, an RRT call is not required before every transfer to the PICU. If the health care team on an acute care unit decides that a child needs a higher level of care, then the patient is transferred to the PICU after discussion with the PICU health care team without activation of the RRT team. When activated, the RRT stabilizes the patient’s condition, determines if a child requires a higher level of care, and expedites transfer. In our study, the rate of transfer to a higher level of care after RTT activation was higher than that reported from Canada.³  The higher rate may have occurred because not all unplanned transfers in our institution are preceded by RRT activation and the acuity was probably higher in the children who required RRT activation.

Our results are similar to those of Wang et al,⁸  who found that emergency response teams for resuscitation were called more often during the daytime hours. In a study in adult patients, Jones et al,⁵  found that more calls were made during the change of nursing shifts. Although the time of transfer to the PICU did not affect length of stay or mortality in our study, more RRT calls were made during the daytime and more children were transferred to the PICU after RRT calls during the daytime. The reason for more calls for rescue during the daytime is not clear. Possibly, critical worsening of clinical status is more often detected during daytime hours because of the availability of a higher number of health care staff than during the nighttime. In addition, the change of shifts from day to night is another time when systematic review of patients’ condition is done and may provide an opportunity to escalate care. Jones et al⁵  have speculated that less patient observation is performed at nighttime and at the change in shifts when patients are reevaluated and worsening of clinical status is detected.

Of note, we detected an increase in mortality for patients transferred to the PICU from acute care units compared with the mortality of patients transferred from other sources. In another single-center study in the United States, Odetola et al¹⁰  found a higher risk of mortality in patients transferred from acute care units than in patients transferred from the emergency department. In a tertiary care PICU in Brazil, mortality was twice as high in patients admitted from acute care units compared with patients admitted from the emergency department.¹¹  Although unplanned read-missions are infrequent, the outcomes are worse than those of planned readmissions.^12,13  In a study by Jäderling et al,⁹  adult patients admitted to the ICU from general medical units after RRT initiation had higher mortality than did patients admitted after direct physician contact and admission with no activation of the RRT. In our study, we were unable to compare mortality rate between RRT and non-RRT transfers to the PICU because the data in our RRT database was delinked from patient identifiers such as name and medical record number.

Our findings of increased RRT use during daytime hours and increased mortality among patients with unplanned transfers from an acute care unit to the PICU are important for planning resource allocations and improving quality and safety measures. With increased use of outpatient and ambulatory services, sicker patients are being admitted to children’s hospitals. Children admitted to acute care units are at risk for further worsening in clinical status; they are less frequently monitored than are children in other units and are at risk for unsuccessful resuscitation if their condition worsens further. Several system-based quality improvement methods have been advocated for an early recognition and treatment of children with critical changes in clinical status. Various pediatric early warning systems have been studied.^14–18  The effect of combined use of pediatric early warning systems for early detection of critical worsening in clinical status and rapid response systems for early treatment is yet to be determined. In addition, mandatory RRT activation determined by using a pediatric early warning scoring system algorithm may further enhance the effectiveness of RRT use. In a study by Panesar et al,¹⁹  mandatory RRT activation based on a pediatric early warning scoring system increased the rate of nighttime activation of RRT.

Our study has several limitations. The documentation of disposition of all RRT calls was incomplete. Our RRT database was delinked from patient identifiers; hence, we could not compare the outcome among RRT transfers and non-RRT transfers. Because our data were collected from a single institution, generalization of the findings must be done with caution.

Our results suggest several additional areas of investigation. Detailed evaluation of causes for higher rate of RRT activation and transfer to a PICU during the daytime is warranted. Day-night variation in outcomes differs according to the population studied. Outcomes are worse for out-of-hospital and in-hospital cardiac arrests,^20–24  trauma admissions,²⁵  unplanned cesarean deliveries,²⁶  and emergency percutaneous coronary intervention²⁷  if these events occur during the nighttime. However, this diurnal variation is not present in all hospital admissions,²⁸  in patients with community-onset bloodstream infections,²⁹  and in patients with severe head injuries admitted to a neurological ICU.³⁰  Diurnal variation may not be detected in a single-center study for infrequent outcomes such as cardiac arrests and mortality in a children’s hospital. Infrequent or less optimal monitoring and evaluation of children admitted to acute care units during the nighttime is a potential reason for the diurnal variation in RRT activation and needs to be further evaluated.

Only 25% of transfers from acute care units to the PICU occurred after RRT activation. The majority of rapid responses were called during daytime hours (7 am –7 pm), and most patients (63%) were transferred to the PICU. Mortality was significantly higher in patients transferred from acute care units than in other PICU admissions. Because a significantly higher number of RRTs are called during the day shift and mortality is higher among children transferred from acute care units, further evaluation is needed to pinpoint the reasons. The finding that only 25% of patient transfers to the PICU occurred after RRTs suggests that RRTs may be underused and that education of the staff as well as patients’ families about RRT may need to be increased.