Admitted patients experience unexpected adverse events in about 10% of cases, 7.3% of which could be fatal. Although not all unexpected adverse events are preventable or predictable, 30% to 83% are so [1,12,13]. Approximately 80% of in-hospital cardiorespiratory arrest showed at least one abnormal sign such as blood pressure, heart rate, respiratory rate, body temperature, or change in consciousness within 8 hours before the event [14-16]. This suggests that there is an opportunity for intervention before patient condition worsens. However, these antecedents can go unrecognized in general wards. In most hospitals, continuous monitoring of vital signs is usually available only in the intensive care unit (ICU). The number of physician rounds for ward-patients is one to two times a day. The interval between measurements of vital signs is usually 8 hours or longer. Respiratory rate and mental state should be measured under direct observation, which is prone to error [15,17]. Even if there is a warning sign, it may not be recognized or properly alerted depending on personal experience, attitude, work environment, and position of the responding nurse or the doctor. As there is a long chain of command to activate—from nurse to intern or resident, from resident to fellow, fellow to attending physician—an alert may be delayed in each step [2].

In the surgical ward, doctors are not readily available because they might be in the operating room. Even if the patient’s warning signs are recognized and reported in a timely manner, it can be difficult to provide appropriate monitoring, intervention, or treatment because of the limited medical resources available in general wards. The Surviving Sepsis Campaign emphasizes early recognition and treatment. Accordingly, the current guideline combines 3- and 6-hour bundles into a “1-hour bundle” to emphasize immediate resuscitation and management [18]. Considering this, the current conventional medical system might be inadequate for proper early response.

An RRS is composed of doctors and nurses who specialize in intensive care medicine to cope with patients at risk. An RRS consists of an afferent limb, efferent limb, patient safety, quality improvement with feedback from data analysis, and administrative components including education of staff [3], as shown in Figure 1. The afferent limb is a way to detect patient deterioration [19]. A nurse or a ward-physician can contact the responding team according to “calling criteria.” Alternatively, using scores such as the modified Early Warning Score (MEWS), patients at risk can be actively screened [20,21]. As under-recognition of monitored abnormal values leads to delay in activation, a surveillance system using a scoring system like MEWS (Table 1) [20] or the National Early Warning Score (NEWS) (Figure 2) [22] can be useful for detecting unexpected adverse events.

Once deterioration has been detected, the responding team works as an efferent limb. The forms may vary—for example, doctor and nurse can work as a team, or the nurse can respond first, and the doctor can cooperate when necessary. According to the composition and character of the responding team, it can be referred to as a rapid response team (RRT), medical emergency team (MET), critical care outreach (CCO) program, or critical care transit program. An RRT is often a nurse-led team. An MET is usually physician-led, offering several critical care interventions such as airway management, central vascular line insertion, and resuscitation. A CCO program adds active surveillance for patients at risk to RRT, managing high-risk patients in advance. The responding team should be equipped with real-time monitoring devices, vascular access devices, oxygen delivery devices, and intervention medications [3].

Continuous quality improvement including self-assessment and data analysis with feedback is another important part in RRS. This quality improvement limb is essential to prevent or prepare for future events. To determine event rates, resource requirements and outcomes, indicators such as number of calls, reason for call, unit where the call started, time of each call, number of transfers to ICU, and the number of calls converted to cardiopulmonary resuscitation (CPR) are collected. Monthly meeting for review of potentially avoidable cases would be an effective strategy for quality improvement.

The Institute for Healthcare Improvement started the “100,000 Lives Campaign” in 2004 to engage U.S. hospitals to prevent 100,000 needless inpatient deaths. To reach this goal, the Institute for Healthcare Improvement suggested six clinical interventions including deployment of an RRT. About 3,100 hospitals (about 3/4 of all hospitals in the United States) were enrolled in the campaign and prevented an estimated 122,300 needless deaths [23]. Sixty percent of hospitals enrolled in the campaign implemented an RRS to detect and intervene in those in crises. As a result, cardiac arrest rate in the general ward was reduced by 50%, transfer from ward to ICU decreased by 58%, and overall mortality decreased by 37%. The Institute for Healthcare Improvement followed with the 5 Million Lives Campaign, and related campaigns were launched in Canada, Australia, Sweden, Denmark, and the UK.

This system is widely supported by accreditors and quality improvement organizations such as the Joint Commission in the USA. The 2009 Joint Commission’s National Patient Safety Goal recommended implementation of an RRS [24]. The 2015 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care also recommended it to reduce the incidence of cardiac arrest, particularly in the general care ward [25]. As a result, an RRS is now implemented in more than 3,700 hospitals throughout the United States and has spread to other continents including Europe and Asia [2,4,26-28].

The preferred composition of the responding team differs among countries and institutions. In the UK, the RRS team may be nurse-led [29], and in the USA, nurses or respiratory therapists may lead [27]. In Australia, New Zealand, and Scandinavia, a physician-led RRS is favored [26,30].

The Medical Early Response Intervention and Therapy (MERIT) trial [4] was a multicenter, cluster-randomized controlled trial of MET, which failed to demonstrate benefit regarding composite endpoint of death, unexpected cardiac arrest, and unplanned ICU admission. However, in the MERIT trial, the cardiac arrest response team was utilized in the control arm about 50% of the time, even before the actual event of cardiac arrest. This may have decreased the gap between the control and the intervention arm with RRS. In the past, the effects of RRS were controversial, but in recent years, there have been solid results supporting this system. Many observational before-and-after studies showed that RRS resulted in decreased in-hospital mortality and cardiac arrest [6-11,30]. The study by Salvatierra et al. [11] was an observational cohort study including 471,062 adult patients from 2001 to 2009. It revealed that RRS improved in-hospital mortality by 24% compared to the pre-RRT time period. Another recent retrospective study found a significant decrease in unexpected and overall mortality [10]. Table 2 is a summary of studies reporting in-hospital mortality [4,7,11,26,27,29-32]. Although the studies mentioned above show considerable heterogeneity, a meta-analysis by Maharaj et al. [33] has consistently shown similar effectiveness of RRS. Implementation of RRS in the adult population was associated with a reduction of overall hospital mortality and cardiorespiratory arrest. In a tertiary hospital in Korea (Asan Medical Center), we found that the number of inhospital cardiac arrests decreased with increase of RRS calls (Figure 3). Based on these results, implementation of RRSs and annual RRT calls are increasing [34].

Despite the lack of accurate data related to patient safety events in the Korean medical system, one study found that about 7% of patients experienced at least one adverse event, and that 61% of events were preventable. In Korea, use of an RRS started in 2008, and the number of hospitals operating an RRS has since been increasing. Asan Medical Center first started an RRS in March 2008; this response team was called the medical alert team (MAT). At first, this system was applied to limited patients as a pilot project. It then was expanded to every patient in the hospital in May 2009. In addition to a calling system by attending physician or nurse in the general ward, the MAT in Asan Medical Center is based on a surveillance program using an electronic medical recording (EMR) screening system [35]. Since MAT was implemented, several hospitals started to prepare for RRS. Samsung Medical Center also prepared an RRS and started pilot operation in January 2009, followed by Hanyang University Hospital in 2011 [36], Seoul National University Bundang Hospital in 2012 [37], and Seoul St. Mary’s Hospital in 2013. Since then, awareness of RRS has been heightened due to promotions in several critical care-related conferences and academic meetings.

To implement RRS successfully, several barriers should be overcome, especially the culture. Korean culture is characterized by a somewhat shy and hierarchical relationship between doctors and nurses. Thus, the amount of direct calling for RRS is small at the beginning. Many Korean RRS overcame low calling rate by introducing EMR screening method including vital signs, labs, and proactive rounds (Figure 4).

Korean RRS was the topic of several papers during this period [35-42]. A single-center before-and after-study of RRS by Kwak et al. [36] and a Korean multicenter study [39] showed that RRS decreased in-hospital cardiac arrest. Kim et al. [38] compared in-hospital cardiac arrest rates by RRS working hours and showed that the incidence of in-hospital cardiac arrest was decreased only during RRS operating hours. RRS not only decreased in-hospital cardiac arrest rates, but also improved mortality from sepsis and first-attempt success rate of intubation (Figures 5, 6) [43,44]. However, many hospitals do not operate 24 hours a day, and physicians dedicated to such a team are rare. Most medical staff are working as intensivists and as members of the RRS team at the same time and are vulnerable to mental and physical exhaustion.

Leaders in hospitals know that an RRS is a good system. There is particular value in Korea where there are not enough medical personnel and resources. However, implementation of an RRS is potentially expensive, and there was no RRS management fee until now. Thus, investment in RRS has been challenging. In 2018, as the Korea Institute for Healthcare Accreditation mandated items for patient safety improvement, more hospitals are adopting an RRS.

2019 is a milestone in the history of RRS in Korea. In May 2019, the Korean Health Insurance Review and Assessment Service and Ministry of Health and Welfare started an RRS pilot program, expanding RRSs further. The third edition revised accreditation standards for acute care hospital accreditation included implementation of an RRS. These changes are expected to make RRS an essential part of improving patient safety. Through introduction of the RRS, hospital mortality rate, number of cardiac arrests, length of ICU stay, and rate of readmission to ICU may be reduced. This might lead to a reduction in overall medical costs by improving the quality of medical care.

In 2019, more than 50 hospitals in Korea will run an RRS. Many hospitals are just beginning to overcome obstacles in implementing an RRS such as lack of awareness of the system, activation criteria, response protocol, organizational culture, resource scarcity, and fear of appearing inappropriate [45]. Constant, repeated education, feedback for quality improvement, and administrative support should be used to move forward. The optimal composition of RRS is not certain. It should be decided according to available institutional resources, preferences, and goals. Simulation training for RRS responders and hospital staffs can result in better management and communication. Appropriate calling and screening criteria as an afferent limb are essential. A novel informatics approach might be an option in the near future [46]. As wearable devices are more widely used, data from these may become available, leading to more accurate and timely assessment. As more hospitals adopt the RRS, associated Korean multicenter data will accumulate and will enable future development.