Emergency Medical Response for Non-Hospitalized Person Events in a Children’s Hospital Free

Our institution is a 404-bed, freestanding, pediatric, academic, urban, quaternary care center with approximately 25 000 annual inpatient admissions. The campus includes a complex of buildings that contain inpatient pavilions, outpatient clinics, a psychiatric inpatient facility, administrative and support services, and research facilities. It neighbors a medical school and 2 major adult medical centers. Within the campus, most buildings are contiguous at 1 or more levels, some buildings are connected by bridges, and some clinical and support programs are located across a major street. Our institution primarily offers services to children from birth to age 21 years; however, at times, adults with pediatric-onset chronic conditions are also treated at our care center. There are ∼2000 outpatient visits each weekday, and >16 000 clinical and nonclinical employees and staff.

The hospital’s code blue (CB) team is composed of clinicians and staff who respond from several locations, including the emergency department and ICUs. In addition to physicians and nurses, there are team members from respiratory care, pharmacy, social work, and security. The team responds to all clinical and nonclinical areas of the campus for CB events, except those involving an HP in the ICUs, operating rooms, or emergency department. Security helps with crowd control and stopping traffic when the CB team needs to respond to events in atypical locations (eg, the clinic across the street or the parking lot). The typical response time to a CB event is approximately 3 to 5 minutes. All code team clinicians are trained in basic life support, pediatric advanced life support, and advanced cardiac life support. Any person on the hospital campus can activate an emergency response by calling the operator. For HP, a rapid response or CB can be requested; however, for NHP needing assistance, at the time of this study, only a CB could be activated.

Our hospital’s institutional review board approved this study. We performed a retrospective review of all CB activations for a 1-year period from January to December 2017. An NHP event was defined as any CB activation for an individual who was not currently admitted to the hospital under inpatient or observation status. An adult was defined as any person aged 18 years or older.

Subjects were identified from the institution’s resuscitation program’s quality improvement (QI) response and transfer database, which tracks all levels of emergency response. Any activation of our emergency response system results in a page that the database manager receives, like the code team. This manager enters and reviews all activations into the database in real time or retrospectively (eg, for activations on weekends). Data from the QI record and the associated code team documentation were reviewed. Data were collected and managed using REDCap electronic data capture tools hosted on our hospital’s network.⁹  If a medical record number was unavailable, reasonable efforts were made to search by other means to identify missing data.

Data collected included demographics, patient type (eg, inpatient, outpatient, visitor, or staff), event location and timing (eg, time of day and day of the week), interventions performed by the CB team, and patient disposition. We used a priori defined categories to group the reason for CB activation as follows: neurologic, cardiovascular, trauma, respiratory, gastrointestinal, miscellaneous, and other or unknown. These categories were defined by the research team and align with categories of emergency response activations previously published in adults.^6,7  Some patients had multiple triggers for a single CB event. Missing data were ignored.

We performed a descriptive analysis of the epidemiology, features and outcomes of HP, and NHP CB events. Categorical data were reported as percentages and continuous data as median (25th–75th percentile). Comparisons between groups for categorical variables were made by using the χ² test. Age was compared between 2 groups by Wilcoxon rank sum test because of its abnormal distribution. A 2-sided P value ≤.05 was considered statistically significant.

A total of 170 CB team activations occurred during the study period. Two were excluded because records indicated the activation was called in error, making the final study cohort 168 CB events involving 164 individuals. Medical records were available for all 33 (100%) HP events and 127 (94.1%) NHP events, whereas specific CB paper documentation was available for 21 (63.6%) of HP and 73 (54.1%) of NHP events.

The majority of CB events involved NHPs compared with HPs (135 of 168, 80% vs 33 of 168, 20%, P < .0001). NHP activations were more often for adults (n = 91, 67.4%) with a median age of 25.2 years (range 0.2–86.6), of which 35 (25.9%) were male. Only 6 (18.2%) of the HP responses were for patients ≥18 years of age. Figure 1 depicts the age distribution of CB activations for HP, and NHP, and demographic details are summarized in Table 1.

For both HP and NHP groups, CB activations were more common during the day shift (7:00 am–7:00 pm) than the night shift (Table 1). Significantly more NHP than HP CB events occurred on weekdays (93.3% vs 60.6%, P < .0001). For HP, the CB activation represented an escalation in response from a previous consultation or rapid response in 5 of 33 (13.2%) cases.

CB responses for NHP most often involved outpatients (43.7%), followed by visitors (including parents) (25.2%), and staff (14.8%). In Tables 1–4, we include NHP subgroup analysis of those classified as outpatient NHP (OP) versus all other NHP (AOP) (eg, visitors and staff) at the time of CB activation. In addition to the differences between HP and NHP, we report the differences between OP and AOP, HP and OP, and HP and AOP.

Events most commonly occurred in clinical areas (eg, inpatient ward or outpatient clinic) for both groups. CB activations for HP occurred more commonly on inpatient units than they did for NHP (75.8% vs 25.9%, P < .0001). Five (15.2%) HP and 45 (33.3%) NHP events occurred in an ambulatory area, whereas 3 (9.1%) HP and 28 (20.7%) NHP events occurred in a public area, such as the cafeteria.

The most common reasons for CB activation among HPs were neurologic emergencies (45.5%), namely seizures (30.3%), and respiratory decompensation with desaturation or hypoxia (27.3%). For NHP, the most common trigger was cardiovascular (57%), specifically syncope (42.2%). The other reasons for activation and their frequencies are outlined in Table 2.

Hospitalized patients were more likely than NHPs to require an intervention from the CB team (29 of 33 HP [87.9%] vs 83 of 135 [61.4%], P = .0039). Critical interventions were more common in the HP group compared with the NHP group (Table 3). The most common critical interventions were bag-mask ventilation (2.2% NHP vs 39.4% HP, P < .0001) and medication administration (14.1% NHP vs 42.4% HP, P = .0003). Events requiring cardiopulmonary resuscitation (CPR) were rare in both groups (0.7% NHP and 9.1% HP, P = .0044).

After a CB activation, the majority of HPs were transferred to the ICU (60.6%) for additional monitoring and care (Table 4). The majority of NHP (82.2%) were transferred to an emergency department. Our institution is connected by a bridge to an adult hospital, so 55 of 91 (60%) adult NHPs were transferred directly there by members of the code team. For those NHP seen in our institution’s pediatric emergency department, 15 (11.1%) were admitted to an inpatient ward, 3 (0.7%) to an ICU, 7 (5.2%) were transferred to another facility (eg adult ED), and 44 (32.6%) were discharged from the hospital.

Our study represents the first report of emergency responses for NHP in a freestanding, large academic children’s hospital. In contrast to other reports in the literature regarding emergency responses for NHP,^6–8  this study also compares the characteristics of those responses to events involving HP. A single previous study by Sahin described CB activations in a children’s hospital but focused on the response pattern of the code team, whereas ours describes the characteristics of the patients requiring response.¹⁰  Other reviews of rapid response events in pediatric hospitals, such as the one by Wang et al in 2011, often exclude the NHP population.¹¹  Interestingly, in that cohort of 1537 team activations over 13 years, only 26 (1.7%) were excluded for being an adult visitor or adult employee.¹¹ 

As expected, some patient characteristics were similar for both groups and consistent with previous reports; for example, most CB activations occur on weekdays and during the day shift (7:00 am–7:00 pm) for both HP and NHP groups.^12,13  We suspect that our finding of significantly more NHP than HP codes occurring during this time likely correlates with the expected high-volume time for outpatient and visitor traffic on a medical campus. We also observed that the rate of escalation to CB for HP was low if the patient had already been evaluated by the ICU team. This is similar to findings in a report by Jayaram et al in 2017 that revealed events requiring CPR in a pediatric hospital were infrequently preceded by activation of the hospital rapid response system.¹⁴  Another important finding is that the majority of our CB activations were for adults, especially in the NHP group. Although pediatric providers in quaternary care centers commonly care for patients over the age of 18 years, those adult-aged patients infrequently have the type of medical emergency that we observed in the NHP group. Interestingly, Adams has published several reports relating to the greater-than-expected frequency of in-hospital arrest for hospital staff and visitors, which represent approximately 1% of all in-hospital cardiac arrests.^15–17  He also describes the importance of public access to an automated external defibrillator (AED) in a hospital setting, given that bystander response with an AED results in a faster time to defibrillation compared with a code team response.¹⁶  The 1 NHP in our cohort who suffered a cardiac arrest on campus received defibrillation by one of our publicly available AEDs before code team arrival.

Our institution’s code team requires advanced cardiac life support training for all attending- and fellow-level ICU and ED physicians because we care for many patients in the adult age range (>18 years of age). Other institutions use the approach that the provision of basic life support is sufficient for the adult population if there can be rapid transfer to an adult critical care setting such as an emergency department or ICU. Federal Emergency Medical Transportation and Active Labor Act regulations address the transfer of patients with an emergency medical condition in any hospital receiving Medicare funding.¹  Hospitals are required to provide stabilizing care by using their available resources before transfer to another facility, but the expectations for the care of adults in a pediatric facility are not specified. The Joint Commission requires that “an evidence-based training program is used to train appropriate staff to recognize the need for and use of designated equipment and techniques in resuscitation efforts”¹⁸ . The interpretation of this standard may vary, but the implication for adult patients in pediatric institutions is unclear.

Another key point from these data relates to the utilization of critical care resources. Although code team members are trained for both adult and pediatric medical emergencies, the majority of CB activations for NHP at our institution do not require critical interventions. Most NHPs were triaged to either an adult or pediatric emergency department for additional evaluation.

Although the code team response at the time of this study involved critical care physicians and nurses, our data suggest that it may be unnecessary to draw resources from critical care areas for NHP events. Although difficult to accurately assess, there is a potential risk for harm to existing patients if ICU-based code team members are responding to an emergency in another area.¹⁹  Our data support the recommendations by Dechert and Nett that a team for NHP emergencies could be nurse- or paramedic-led.^8,10  As Dechert notes, this does not preclude the option to “ramp up” to a higher level of response, should the specific situation warrant.⁸ 

As a result of preliminary analysis of these data, our institution piloted a medical assist team (MAT) in October 2018 as an intermediate level of response for medical emergencies in NHP. Specific MAT activation criteria include verbally responsive, nonhospitalized persons, with nonlife-threatening emergencies while on campus. MAT activations exclude patients with chest pain, breathing difficulties, or decreased level of consciousness. The team consists of a senior pediatric resident, emergency department nurse, nurse coordinator of patient placement, and security staff. The team rapidly assesses NHP to determine if transfer to an adult or pediatric emergency department is indicated and has the option to escalate to a CB if needed. Since implementation, our experience indicates that the frequency of CB calls for NHP has decreased, and these MAT calls are infrequently escalated to a CB.

This study, although the fourth published regarding NHP emergencies, is the first to compare the responses for HP versus NHP by the same emergency response team and the first based in a freestanding children’s hospital. It is important to recognize the limitations of a single-institution report and our results may not necessarily be generalizable to other systems. The retrospective nature of the review and the use of a database designed for quality improvement purposes raises the risk of missing or inaccurate data elements. Although some handwritten code sheets were missing for both HP and NHP, the data were often available in the QI database and/or other aspects of the medical record. This finding also suggests that our system and training for documenting interventions at CB events needs to be improved.

At our pediatric institution, NHPs frequently experience medical events while on the children’s hospital campus and use the code team more often than HP. The majority of NHP events in our children’s hospital are for adult persons, who are frequently triaged to an emergency department and do not require emergent interventions delivered by a critical care-based code team. A separate emergency response team for medical emergencies outside of inpatient areas may more efficiently use hospital resources to manage these events. We intend to use these findings to revise the system of care for NHP experiencing medical emergencies in our pediatric institution. The high frequency of adult events within a pediatric hospital also has implications regarding the appropriate training for emergency responders. Future studies are warranted to evaluate the most effective pediatric CB team structure, training, and certification for NHP emergency responses in other children’s hospitals.