Rapid Response Systems

Christopher P. Bonafide

Richard J. Brilli

James Tibballs

Christopher S. Parshuram

Patrick W. Brady

Derek Wheeler

KEY POINTS

Rapid response systems (RRSs) aim to identify hospitalized general ward children who exhibit early signs of clinical deterioration and to intervene before respiratory or cardiac arrest occurs.

RRSs operate on the assumptions that early, reversible clinical deterioration can be identified and that assistance from a team of critical care experts can improve patients’ outcomes.

RRSs include two clinical components (afferent and efferent limbs) and two organizational components (process improvement and administrative limbs).

The role of the afferent limb of RRSs is to identify patients at risk of deterioration and trigger an appropriate response based on the level of risk.

The role of the efferent limb of RRSs is to deploy specialized teams of skilled personnel to hospital wards to address urgent care needs.

The effectiveness of medical emergency teams in reducing hospital arrest rates and mortality is controversial; no cluster-randomized trial shows benefit, but several pediatric before-and-after studies show improvements in outcomes.

The role of the process improvement limb of RRSs is to assess the overall success, assess opportunities to optimize the system, and design tailored improvement interventions.

The role of the administrative limb of RRSs is to manage each of the RRS components, focus on implementing the system, and support its ongoing operation.

The rate of urgent requests to the pediatric critical care team to provide expert advice and management for patients on the wards exhibiting early signs of clinical deterioration continues to increase. The systems that focus on the prediction, detection, and management of clinical deterioration in non-intensive care areas are known as rapid response systems (RRSs) (1). The hallmark of RRSs is their focus on identifying and mitigating reversible early signs of clinical deterioration in ward

settings to prevent respiratory and cardiac arrest. They operate on the assumptions that, in at least a subset of patients, (a) early, reversible clinical deterioration can be identified using tools that facilitate detection and standardize escalation of care on the wards, and (b) consulting a multidisciplinary team of critical care experts with the capability to rapidly intervene

at the bedside can improve patients’ outcomes. The organizational structure of RRSs includes two clinical components (afferent and efferent limbs) and two organizational components

(process improvement and administrative limbs).

The objectives of this chapter are (a) to provide the pediatric intensivist with an overview of RRSs and their components, (b) to review early warning scores (EWSs) and the calling criteria that comprise the afferent limb, (c) to summarize the response mechanisms that comprise the efferent limb, (d) to provide a set of process improvement measures that can be used to evaluate RRS effectiveness, and (e) to discuss the administrative issues associated with implementing and managing an RRS.

ORIGINS, DISSEMINATION, AND PREVALENCE OF PEDIATRIC RAPID RESPONSE SYSTEMS

Origins

Over the past two decades, RRSs that are focused on identifying and managing pre-arrest status have been implemented in thousands of hospitals throughout the world. The medical emergency team (MET) concept was first reported in an adult hospital in Australia in 1995 (2). The team was developed to rapidly detect and correct vital sign abnormalities that represent early disturbances in cardiorespiratory function that precede arrest in patients with severe trauma. Team members included medical and nursing staff with training in resuscitation. The team could be activated when urgent help was required or when calling criteria based on specified vital sign parameters or conditions were attained (2). Definitions for MET, rapid response team, and critical care outreach team are provided in the section on The Efferent Limb.

Ten years later, Tibballs, Kinney, and colleagues reported the first pediatric RRS implementation at Royal Children’s Hospital in Australia (3). The team was composed of physicians and nurses from the ICU and emergency department, as well as a medical registrar (analogous to a physician fellow in the US system). Like the adult system, the team could be
activated at any time by a concerned nurse or physician or when calling criteria were attained. An important innovation from this pediatric implementation was the inclusion criteria based on age-specific parameters for heart rate, blood pressure, and respiratory rate.

Dissemination

In the years that followed, hospitals around the world began describing their results after implementing RRSs, and patient safety organizations took notice. In 2005, the Institute for Healthcare Improvement launched the 100,000 Lives Campaign, a nationwide initiative to reduce morbidity and mortality in the American healthcare system (4). The campaign expanded in 2006 as the 5 Million Lives Campaign and included over 4000 participating hospitals (5). One of the key components of these campaigns was the implementation of RRSs as a means of rapidly responding to early signs of patient deterioration.

In 2008, the Joint Commission, an organization that accredits and certifies more than 19,000 healthcare organizations and programs in the United States, set a National Patient Safety Goal requiring that hospitals seeking accreditation “empower staff, patients, and/or families to request additional assistance when they have a concern about the patient’s condition” (6). While compliance with this safety goal did not require implementation of a formal RRS, the issuance of this goal likely contributed to further uptake of RRSs.

Organizational bodies outside the United States (in Canada, the United Kingdom, and Australia) supported RRS implementation to varying degrees and with differing models (7,8,9).

Prevalence

Estimates of the prevalence of pediatric RRSs began to emerge in the mid-2000s. A 2005 survey of 181 children’s hospitals in the United States and Canada showed that 100% had an immediate-response code blue team that responded for cardiopulmonary arrest, and that 17% had an MET that responded to children clinically deteriorating but not at risk of imminent cardiopulmonary arrest. In 21% of the hospitals with METs, discrete calling criteria were used to determine when to activate the team (10). A 2010 survey to estimate the prevalence and characteristics of pediatric RRSs among 130 US children’s hospitals with PICUs found that 79% had an MET that “quickly responds to patients on the general wards at early stages of instability” (11). They also found that 34% used automatic triggers, defined as predetermined changes in the patient’s vital signs or overall clinical status, to activate the MET.

THE AFFERENT LIMB

The role of the afferent limb of RRSs is to identify, or track, patients at risk of deterioration and trigger an appropriate response

based on the level of risk. In a consensus statement on the afferent limb of RRSs, systems for prediction of deterioration were distinguished from systems for detection of deterioration (12). Predictive tools focus on “traits” (such as a diagnosis of epilepsy) rather than “states” (such as a heart rate of 200) and do not require continuous data collection. Detective tools, in contrast, focus on identifying states consistent with critical illness by recognizing signs of deterioration using highly time-varying data like vital signs. Detective systems require frequent intermittent measurements or continuous data collection to provide early identification of departures from clinical stability and prevent progressive deterioration.

Predicting Deterioration

In comparison to detective tools, little work has been performed in the area of developing tools to predict clinical deterioration in hospitalized children using patient characteristics. The first detailed case series of hospitalized children provided with urgent assistance from an MET described the clinical characteristics of the patients: 44% had surgery during the hospital admission, 36% had an ICU admission, and 20% had a diagnosis of chronic encephalopathy (13). Additional chronic conditions that occurred frequently included congenital syndromes, chronic lung disease, and abnormal upper airways. This case series provided a snapshot of the populations at risk of deterioration; however, estimates of association between these conditions and pediatric deterioration could not be determined.

More recently, a predictive model for clinical deterioration using non-vital sign patient characteristics was developed using a case-control design (14). The predictive model resulted in a 7-item weighted score that included age under 1 year, epilepsy, congenital/genetic conditions, history of transplant, presence of an enteral tube, hemoglobin less than 10 g/dL, and blood culture drawn in the preceding 72 hours. Patients were grouped into risk strata based on their scores. The very low-risk group’s probability of deterioration was less than half of baseline risk. The high-risk group’s probability of deterioration was more than 80-fold higher than the baseline risk. Predictive tools like this have the potential to assist in identifying and triaging a subset of high-risk children who should be intensively monitored for early signs of deterioration at the time of admission. The converse may also be helpful; the predictive tool may help identify very low-risk children who, in the absence of other clinical concerns, can be monitored less intensively.

Detecting Deterioration

Single-Parameter Calling Criteria

The simplest and most widely used form of a detective tool is a set of single-parameter calling criteria. They are easy for bedside clinicians to use; if any one of the criteria is met, the efferent limb should be activated. While the parameters are commonly objective clinical findings such as vital signs, they can also include diagnoses (such as suspected shock), events (such as seizures), subjective observations (such as increased work of breathing), and intuitive concerns (such as worried about the patient). Parameters for heart rate, respiratory rate, and blood pressure are usually presented within distinct groups to account for variability by age. The differences in parameter cut points across studies reflect that evidence supporting age-based vital sign parameters is very limited. A list of single-parameter calling criteria is given in Table 29.1.

Multiparameter Early Warning Scores

Multiparameter tools combine several of the core components of single-parameter calling criteria into EWSs. The first EWS was developed to detect deterioration among hospitalized adults in 1997 (15). These scores trade increased complexity for potentially better accuracy in identifying deterioration. The scores may be either weighted or unweighted; weighted scores allocate a variable number of points based on the degree to which patients’ vital signs deviate from a (usually arbitrarily developed) “normal” or “expected” range. The scores are periodically calculated either by hand or within the electronic health record, and the sum total score is used to trigger the efferent limb.

TABLE 29.1 PRE-ARREST SINGLE-PARAMETER CALLING CRITERIA USED IN RECENT MET STUDIES

	▪ BRILLI ET AL. (51)	▪ HANSON ET AL. (48)	▪ HUNT ET AL. (52)	▪ KOTSAKIS ET AL. (7)	▪ SHAREK ET AL. (47)	▪ TIBBALLS ET AL. (50)	▪ ZENKER ET AL. (49)
Heart rate		•		•	•	•	•
Respiratory rate		•		•	•	•	•
Blood pressure		•		•	•	•	•
Oxygen saturation	•	•	•	•	•	•	•
Respiratory distress	•	•	•	•		•	•
Airway threat		•		•		•
Circulatory compromise/shock			•				•
Cardiac dysrhythmias			•
Mental/neurologic status	•	•	•	•	•	•	•
Seizures		•	•			•
Staff concern	•	•	•	•	•	•	•
Family concern	•	•	•			•	•

An ideal score would balance high sensitivity (minimizing the number of children who deteriorate without being identified by the score) with high specificity (minimizing the number of children who trigger the MET but are not deteriorating, unnecessarily consuming pediatric intensive care resources). The achievement of high sensitivity and specificity when identifying a broadly defined condition such as clinical deterioration is challenging.

Only gold members can continue reading. Log In or Register to continue