In recent years, multiple disasters, both natural and man-made—from forest fires to terrorist attacks—have required response from health care providers, including resident physicians. Fortunately, these types of disasters are still infrequent; however, infrequency leads to an absence of training opportunities1 and leaves medical professionals unprepared to handle the burden of a disaster when it strikes.
Every year, in late June, incoming medical residents across the United States undergo orientation to their respective training programs. At our institution, Corpus Christi Medical Center (CCMC), emergency preparedness is addressed during this orientation and the program director is tasked with finding volunteer residents for the disaster team, which is activated only in times of emergency. In the Coastal Bend of Texas, the primary emergencies of concern are tropical storms and hurricanes.
On August 13, 2017, one such emergency began to take shape as a tropical wave off the west coast of Africa. The disturbance organized into a tropical cyclone four days later, before again becoming a disorganized tropical wave, which moved into the Yucatán Peninsula on August 22. U.S. and European weather models disagreed on the trajectory and potential severity of the storm. It was not until the storm system rapidly intensified in the warm waters of the Gulf of Mexico, slightly more than 24 hours before landfall,2 that the clear and present danger to the Texas coast became apparent. Hurricane Harvey made landfall as a Category 4 storm at approximately 10:00 pm CDT on August 25, a mere 30 miles northeast of Corpus Christi, Texas. It was the first hurricane to hit the middle Texas coast since Celia in 1970, 47 years previously.
The Storm Story
The decision to activate the disaster team at CCMC was made as the storm approached and a state of disbelief hung in the air. Coastal Texas counties began issuing evacuation orders for citizens. Because of the recommended evacuation, only chief residents and a junior administrator remained available to staff the graduate medical education (GME) department at CCMC, adding to the challenges of preparing for a looming disaster. The disaster plan regarding GME remained simple. Residents would be divided into two teams: Team A, preferably those without pets, significant others, or dependents; and Team B. Team A would stay behind to man the hospital, and Team B would evacuate with instructions to return and relieve Team A within 48 hours after the storm passed.
Our institution has multiple campuses within the city, the two main hospitals being Corpus Christi Medical Center–Bay Area (BA) and Corpus Christi Medical Center–Doctors Regional (DR), which are located approximately seven miles apart. Residents’ primary task as the storm approached was to empty BA of all patients, be it by discharging them home, evacuating them to another city, or transferring them to DR, which had the advantage of being nearly twice the elevation above sea level as BA. Transferring patients represented an enormous challenge, as the hospitals in the area typically operate close to their maximum capacities. With a collaborative effort from residents, hospitalists, nurses, and the patients themselves, BA was entirely evacuated only 24 hours before the eye of Hurricane Harvey began to encroach on the Texas coast. In total, two full workdays were required to get everyone discharged, transferred, or evacuated.
With BA closed, Team B residents began evacuating the city, and Team A residents relocated to DR. Twelve Team A internal medicine residents in their first to third year of training gathered extra food and water for their group. Residents who had been working at the DR campus the night before had begun acquiring extra blankets and mattresses from storage in the hospital. As it turned out, mattresses were a premium resource. The team of 12 had to make do with only seven mattresses, which were divided between two call rooms.
In disaster planning, having as many physicians as possible on hand may seem like an advantage, but being overstaffed in tight quarters was almost as bad as being understaffed. In addition, one-quarter of the residents on Team A were in their intern year and required direct supervision at all times. Arguably, the 1.5 months of residency training the interns had experienced at that time is simply not sufficient to amass the requisite skills for disaster relief.
After Team A became settled at DR it became clear that there was no plan in place for the residents’ duties, nor was the program director or designated institutional official (DIO) on-site to clarify these duties. The junior administrator of the GME program assumed the responsibility of liaison between residents and hospital administration. The chief residents made assignments for patient care, dividing coverage into 12-hour shifts, with six residents on each shift. During each shift, two residents were stationed in the 21-bed intensive care unit and the other four were to be present on their respective wings of the hospital, maintaining an even presence throughout the hospital.
During their shifts, residents encountered a shortage of hospitalist physicians and nurses. The attending physicians decided that residents would best be used to round on patients and write progress notes. Hence, the 12-hour shifts were abandoned, as six residents would not be enough to write all the progress notes for the near-capacity, high-acuity patient census. Residents began to question whether their presence was really needed if they were going to be relegated to clerical tasks. To make matters worse, the cafeteria was operating during limited hours, and it was overcrowded when it was open, making it nearly impossible to obtain food. With tensions running high, it was implied that we should go home. However, it was already past the time for safe evacuation.
As residents endured conflicting messages between all parties involved, a feeling of helplessness, and waiting for the unknown, the hospitalists proceeded with a new plan. After 48 intense hours, residents were tasked with running the temporary discharge lounge in human resources. This task would normally be handled by ancillary nursing staff, but because there was a shortage of nurses, residents were forced to fill that need.
The storm passed, and on August 29, Team B was beginning to arrive back in the city. Exhausted members of Team A returned to their homes to assess the damage and sleep. Several residents on Team B were not able to get back to the city because they evacuated north to areas that were impacted by the aftermath of Hurricane Harvey and its flooding. The BA campus remained closed for minor flood damage and infection control for several days after the city evacuation order was lifted. The residents who were not on inpatient assignment organized themselves into teams that were deployed to help fellow residents handle damage to their property, and to help with the overall cleanup and food bank in the city.
As mentioned above, the rarity of disasters means a lack of opportunity to learn how to respond to them effectively. We offer the following lessons learned from our experience during Hurricane Harvey. We hope these are valuable for other GME programs’ disaster planning efforts.
We recommend that the hospital evaluate the minimum number of staff needed for each department to maintain quality and safety. Using overqualified personnel for clerical tasks is not a good allocation of resources, nor is asking improperly trained staff to complete the task at hand.
Having weathered a disaster with minimal supplies, we recommend that nonperishable food, walkie-talkies, and extra mattresses be kept on the designated campus so that there is no lack of resources when a major disaster happens. We were fortunate that cell service and electricity were not disrupted; without these resources, we would have been critically deficient in means of communication because of a lack of battery-powered handheld radios. We recommend including radio communication devices in future disaster plans. The limited cafeteria hours also posed a threat, making it difficult to obtain food. A few dangerous runs to the nearby grocery store and a few crockpots of soup kept residents fed. A stash of nonperishable food would have been a welcome resource when food was scarce.
Optimal team structure
Because of their limited training, interns should not be included on the team of residents who remain on duty during the disaster. In addition, the team on duty should be limited to six residents per campus. A smaller team of more experienced residents would adequately staff each campus and eliminate the need for direct supervision.3
We experienced a lack of communications on all fronts. We received mixed messages from hospital administration and attending physicians. Communication to the GME program representatives was lagging in the absence of the DIO. Secondhand information, mixed messages, and guesswork all contributed to the challenges we encountered. The communication errors coupled with high stress were the main breakdown in the disaster plan.4
Many Team B residents evacuated to Houston. In retrospect, however, Houston was a poor choice for evacuation because it fell within the projected path of the storm and experienced severe flooding, preventing the timely return of five residents on Team B. Residents who evacuated also encountered difficulty locating vacancies at hotels. Because residents were delayed in leaving the city relative to the rest of the population, they were forced to drive much farther outside the city to find hotels with vacancies.
Our initial disaster plan contained no directives, and the DIO was not present at DR to advise us. In the future, the existing disaster plan should not only include specific numbers of residents to ensure adequate staffing and optimal team structure but also define duties and expectations for residents during the incident. Decision making should not fall to a single individual.
More than a year after Hurricane Harvey, retrospective root cause analyses of our experience point to a main breakdown in communication on all fronts: A clear, descriptive plan of expectations was not laid out; the administrator had blind, sporadic communication; and there were too many residents involved. The new emergency preparedness plan should include having an administrator or physician leader from GME on-site—primarily for communication, setting expectations, and leadership but also so residents will not feel abandoned.
Much of the conflict and many of the challenges could have been avoided if a properly developed disaster plan had been in place and available to all members of the staff. Our GME program’s disaster plan, which was woefully inadequate, consisted of dividing residents into teams and leaving the rest to the DIO to address as needed. Our perfect storm—riding out a Category 4 Hurricane with no direct deaths while the main administrator was on leave and discovering that we were underprepared—has allowed us to critically analyze and correct our pitfalls. Our mistakes have led us to better insight and given us the opportunity to convey these lessons to others.