**1. Introduction**

"The 2018 wildfire season was the deadliest and most destructive wildfire season on record in California with a total of 8,527 fires" [1]. These wildfires led to various evacuations in healthcare settings including neonatal intensive care units (NICUs), and evacuation protocol reevaluation. Perinatal patients may be critically ill and highly dependent on medical staff and technology for care [2]. Babies cared for in NICUs are most likely to be impacted by acute evacuation from their units [3].

In addition to wildfires, it is inevitable that other types of disasters will also impact California healthcare systems in the future. The California Association of Neonatologists (CAN) resource, the "Neonatal Disaster Preparedness Toolkit," addresses a plethora of forms of disaster preparedness from bioterrorism to an active shooter [2]. With the purpose of "provid(ing) guidance to NICU leadership in developing comprehensive disaster response plans that are in compliance with Joint Commission Standards and based on community, best-practice models" [2], the toolkit delves into the command center, six critical elements of disaster response [4] and the TRAIN™ tool (triage by resource allocation for in-patients).

The TRAIN™ tool is a way that NICU leaders categorize the infants under their care every day to prepare for evacuation if needed [5]. TRAIN™ tool assigns ambulance asset needs but not NICU level [2]. The assignments for transport can be car, BLS (basic life support), ALS (advanced life support), CCT (critical care transport), or Specialized with the categories being life support, mobility (car/car seat, stretcher, incubator, immobile), nutrition, and pharmacy [2]. The TRAIN™ tool provides a more subjective and efficient way to transport inpatients during a disaster evacuation [5]. We utilized the

TRAIN™ tool because it has been studied and researched in its original form, as well as modified to be applicable to all neonatal and pediatric inpatients [5]. It is also a triage tool that meets the three needs of evacuation, surge capacity, and communication [5] and is endorsed by the CAN and the District IX AAP (American Academy of Pediatrics) section on Perinatal Pediatrics. Unlike the TRAIN™ tool, other available triage tools do not meet the three needs [5]. For this reason, our study exclusively used and focused on the TRAIN™ tool (Table 1).

**Table 1.** Triage by resource allocation for inpatients (TRAIN) tool. (Adapted from Dr. Lin's *Triage by Resource Allocation for Inpatients: A Novel Disaster Triage Tool for Hospitalized Pediatric Patients*, 2018) [5].


PO: Per os (taken orally); TPN: Total parenteral nutrition; IV: Intravenous medication.

Although guidelines have been published both in and out of California for various disasters, there is a gap in our knowledge in how these guidelines are implemented in actual disaster settings, particularly in the context of the recent wildfires in California. Additionally, we wanted to study the experiences and the practices of NICU personnel during evacuation or acceptance of babies. What are some of the barriers in healthcare delivery for NICUs, especially if needing evacuation, that we may have learned from recent California wildfire experiences? What have healthcare team members and leaders learned about how to improve patient safety during a crisis like a wildfire that is encroaching on the hospital's door? What new ways have healthcare workers discovered and are enacting to effectively and efficiently evacuate NICUs? We answered these questions in an open-ended fashion as we wished to learn what health care team members were doing. These questions do not have absolute requirements based on certain guidelines.

In order to answer these questions, we researched NICUs that evacuated or accepted babies due to a California wildfire, contacted key members of the healthcare team (NICU medical directors, neonatologists, nurse managers/directors, neonatal clinical nurse specialists, NICU nurses, patient care manager, and NICU department managers), and conducted interviews from April 2019 to May 2020. The purpose of this study was to learn more about the issues faced by NICU healthcare providers during wildfire disasters, with the longer-term goal of improving NICU patient safety. We focused on NICUs that evacuated, as they made more of our sample pool of interviews.

### **2. Materials and Methods**

We researched California fires using CALFire, newspaper articles, and internet search engines. Through newspaper articles and internet search engines, fifty-nine NICUs were investigated and a total of seven hospitals' health care worker(s) were interviewed. The NICUs were spread out geographically and included parts of Northern and Southern California where the wildfires happened. Two interviewee's NICUs are located in Southern California. Eight interviewee's NICUs are located in Northern California, with six in the same city of Northern California. Of the six in the same city, hospital A had two interviewees, hospital B had one interviewee, and hospital C had two interviewees/three interviews, as one health care team member was interviewed twice due to two separate wildfires in different years affecting the same hospital.

We obtained contact information for the health care team members at these NICUs through contacts from the California Perinatal Quality Care Collaborative (CPQCC), referrals from previous participants that were interviewed, and other relevant sources. CPQCC is a California network of NICUs and HRIF (high risk infant follow-up) clinics, whose goal is to improve care for California's mothers and most vulnerable infants [6]. We did not obtain contact information from HRIF clinic sources.

Once the informant was identified, we audio recorded approximately 60-min interviews over phone or remote conferencing software. We also gained additional information through online surveys. Of the ten interviews conducted from April 2019 to May 2020, three were obtained only through online survey. The online survey questions on Google Form were the same as the questions asked by phone or remote conference software and were reserved for participants with scheduling conflicts.

This qualitative study consisted of ten questions. By interviewing key members of the health care team, we were able to assess key components of patient safety in the hospital related to wildfire disaster response, specifically in regard to improving newborn care. A semi-structured interview process was used (Appendix A—interview guide). The interview questions contained four major sections: Participant's Background, Institutional Perspective, Evacuation Experience, and Lessons and Insights. Please see Appendix A for the interview protocol.

Data were analyzed using qualitative research methods. The audio recordings were transcribed and read iteratively to find common themes in the interviews. We used a grounded theory approach based in "data systematically gathered and analyzed" but primarily focused on practical aspects that could be potentially generalizable to further work in the area of NICU preparedness for disaster experiences [7].

Before interviewing participants, interviewees filled out a consent and demographic form. All questions on the demographic form were optional. After the interviews, we emailed each participant a USD 25 gift card in appreciation for their time and effort. This study was approved by the Stanford University Institutional Review Board.
