The province of Manitoba is situated centrally in Canada with a population of roughly 1.2 million people living in an area of 650,000 km. The overall population density is approximately three people per square kilometer. The province has one of the highest proportions of First Nation (aboriginal/indigenous) peoples in Canada (approximately 13%). The two largest cities, Winnipeg and Brandon, account for 75% of the population of the province. The hospital catchment area approaches 1.5 million.
Critical care beds are concentrated in Winnipeg and Brandon, where there are two adult teaching hospitals and five community hospitals. One teaching hospital provides ten beds for both medical and surgical critically ill patients and a second 10-bed intensive care unit (ICU) exclusively for postcardiac surgical patients. The other teaching hospital provides 12 medical and 11 surgical ICU beds with a six-bed intermediate ICU for patients who require long-term ventilator weaning; this hospital is the trauma center for the province and also treats all patients with burns and neurosurgical issues. The five community hospitals provide 30 additional ICU beds with a limit of 17 ventilated patients. Any patient requiring dialysis while in the ICU must be admitted to one of the teaching hospitals.
With regionalization of services, ICU patients are transferred to the appropriate hospital for medical procedures not offered in the admitting hospital and also for bed management purposes. As a result, the network of ICU beds in Winnipeg functions as one unit. Under normal circumstances, the entire ventilator-capable ICU bed capacity for the province is 63, although less than half of those beds carry ventilated patients at any given time on average. Capacity runs at greater than 90% on most occasions. All ICUs run on a closed-unit model with 24/7 coverage by housestaff or house medical officers.
Salt Lake City, the capital of Utah, resides within Salt Lake County, a metropolitan area containing roughly one million people or 37% of the state's population. The county's hospital catchment area exceeds two million. Critical care beds are located throughout Salt Lake County in over ten hospitals, only four of which are academic teaching hospitals. There are over 100 ICU beds in the county's academic hospitals, although ventilated patients generally occupy <50% of beds and capacity runs at approximately 85% on most occasions. ICUs function primarily in a closed-unit model and provide 24/7 intensivist coverage with housestaff and/or nurse practitioners also providing clinical care at each of the three hospitals.
During April 2009, as pandemic influenza spread to several continents, ICU physicians in Winnipeg and other affected cities noted a rise in the number of patients with rapidly progressive atypical pneumonia requiring intubation and prolonged mechanical ventilation. Initially, the standard treatment of community-acquired pneumonia (antibiotics and supportive care) was instituted. Within a few days, it became clear that these patients were different in that they quickly progressed to acute respiratory distress syndrome. The typical patient was young and relatively healthy with a remarkable oxygenation defect. Many of the Manitoba patients were of First Nation origin.
The first Manitoba patient was a 50-yr-old woman undergoing chemotherapy; the second was a 29-yr-old previously healthy man. Initial respiratory cultures were negative, but Streptococcus pneumoniae was isolated from both several days after presentation. After approximately 1 wk, the attending physician became concerned with the atypical clinical course and lack of response to standard antimicrobials, electing to send samples for viral detection.
As a result of a lack of familiarity with the appropriate sampling methods and viral sample transport kit, initial efforts at providing material for testing were rejected by the clinical laboratory. It took several attempts (and approximately 10 days) before a nasopharyngeal swab was accepted by the laboratory for influenza A testing. It was soon found that these patients with progressive pneumonitis were in fact the leading cases in a wave representing the largest localized outbreak of pandemic 2009 influenza virus (pH1N1) infection in North America.
As the number of ICU patients with suspected pH1N1 started to rise, several diagnostic and management challenges presented themselves to the ICU team.
In retrospect, the cases seen early in the epidemic represented classic viral pneumonia with bacterial superinfection, but significant difficulties in recognizing the epidemic were apparent. Despite the knowledge that pandemic influenza requiring ICU care had been seen recently in Mexico City and an awareness that cases were being recorded on several continents indicating pandemic spread, there was a level of indecision and skepticism during the first week when intensivists were attempting to engage the public health authorities.
The initial outbreak was seen in underserviced First Nation communities. It was only when significant numbers of ICU cases were identified that public health agencies were alerted to the problem. The initial response to the local epidemic was delayed because of the incongruity of the number of cases being admitted to the ICUs while disease activity was overlooked within isolated northern communities. In addition, remarkably few adult patients required ward care compared to ICU care (initial ratio of 1:4 in adults), which caused further delay in perceiving the nature of the epidemic.
Although the initial cases presented with findings suggestive of bacterial pneumonia, several different presentations were noted. The most common (75% to 80% of cases) was that of a rapidly progressive bilateral interstitial pneumonitis typically occurring 3–4 days after onset of a viral prodrome associated with severe refractory hypoxemia and requiring intubation within 12 to 24 hrs of presentation to hospital. Interestingly, hypoxemia associated with diffuse pneumonitis appeared to be resistant to high levels (>20 cm H20) of positive end-expiratory pressure; however, several institutions independently reported that patients exhibited a consistent sensitivity to aggressive diuretic therapy during the first week of ventilatory support.
Once recognized, patients presenting in this manner were often pre-emptively intubated while still on moderate levels (15 L/min) of oxygen. The need for early intubation seemed to wane once a strategy of early antiviral therapy was introduced into at-risk populations. Other clinical presentations included: exacerbation of preexisting asthma/chronic obstructive pulmonary disease; destabilization of chronic disease; and secondary bacterial pneumonia (S. pneumoniae and methicillin-sensitive Staphylococcus aureus being the most common).
Unlike the seasonal variety, the 2009 pH1N1 influenza seemed to attack primarily young, relatively healthy adults with an average age of 32 yrs in the Manitoba cohort. Approximately 70% of patients did not have serious comorbidities as defined by chronic organ failure (2). In addition, pregnant women seemed disproportionately affected with a high rate of fetal loss in some sites. The epidemiologic profile of patients in Salt Lake City, UT, Orlando, FL, and Mexico City was similar. At one point, a single hospital complex in Orlando housed >20 pregnant women in the ICU. Interestingly, in the Manitoba experience, delivery of the fetus was not associated with improvement in the mother's clinical status.
As a result of the high number of affected pregnant patients and the rate of spontaneous preterm precipitous labor, a fully equipped neonatal ICU bed occupied one of the Winnipeg adult ICU rooms. The availability of resources for management of precipitous preterm delivery should be strongly considered if pregnant patients are in the ICU, because the Winnipeg anecdotal experience demonstrated significant delays in obtaining the necessary pharmacologic agents and equipment for a critically ill neonate.
The Winnipeg experience was also notable for a disease predilection among the aboriginal population with this group accounting for 50% of our cases of severe respiratory illness, most of which were 2009 pH1N1 influenza cases.
Septic shock and multiorgan failure were relatively uncommon in these patients, but we did note a high occurrence of rhabdomyolysis (20%), transaminitis, and altered level of consciousness (15%), similar to findings reported from Mexico and Salt Lake City.
After intubation, most patients required extremely high levels of sedation. It was not uncommon to have patients sedated with simultaneous infusions of fentanyl at >400 μg/hr, midazolam at >30–40 mg/hr, and propofol at 3–5 mg/kg per hr. Several patients required the addition of ketamine. Also notable was the extreme drive to breathe with minute ventilations upwards of 20 L/min in such patients, often despite substantial levels of sedation. One consequence was that patients often exhibited extremely large tidal volumes, far in excess of those recommended using a pressure-limited low tidal volume strategy (1). This finding may be a consequence of the relative youth of patients and the very severe degree of hypoxemia noted. A significant subset of patients were unable to sustain arterial saturations of >85% without rescue therapies for days (or weeks) at a time.
Once it was discovered that the initial patients had influenza pneumonitis, the challenge became the timely diagnosis of this disease in patients presenting with severe respiratory illness.
Because of the rapid inflection in the number of cases seen across the province of Manitoba, the Provincial Laboratory was unable to process the large number of samples requested during the early phase of the epidemic. Samples from community clinics, peripheral hospitals, and wards in the main teaching hospitals were processed at the same priority level as samples collected from critically ill ICU patients. Over the first 2 wks, with the support of the Provincial Laboratory, the decision was made to prioritize samples from the ICU and acute care units. The laboratory responsible for performing the test first typed viral samples with a real-time polymerase chain reaction for influenza A and, if negative, no further testing was performed. If the sample contained influenza A, the laboratory attempted to match the type to the usual seasonal variants. If influenza A was not typable, it was presumed to be 2009 pH1N1, and the patients were labeled as pH1N1-positive.
This initial multistep approach took up to 4 days to perform. Given the time required for definitive results, the standard initial approach of requiring confirmation of influenza before initiation of antiviral therapy was quickly abandoned in favor of early empiric therapy as soon as the clinical suspicion of 2009 pH1N1 influenza arose.
The experience in Salt Lake City was similar to that in Winnipeg with the laboratory being overwhelmed with the number of samples. Unlike Winnipeg, Salt Lake City did not prioritize laboratory samples, instead choosing to restrict testing to hospitalized patients in an attempt to reduce the workload to a manageable level.
Several specific diagnosis-related learning points became apparent during the initial wave of 2009 pH1N1 severe respiratory illness cases. First, the majority of patients admitted to the ICU had provided their viral specimens before transfer to ICU; thus, the laboratory did not identify these as ICU cases and samples were not prioritized for testing. A hotline was set up between the ICU and the diagnostic laboratory. Each morning, the ICU attending would call the hotline with a list of new admissions suspected to have 2009 pH1N1 illness. The laboratory then would prioritize these specimens and attempt to have results available by the end of the day. Despite initial discussions regarding 7-day-a-week testing, the hotline system was maintained as a Monday through Friday process. This led to extended isolation for some patients but was a compromise with the laboratory to accommodate the increased workload. A dedicated intensivist in Salt Lake City was also assigned the task of determining results of the 2009 pH1N1 influenza test. This seems to be the best method to ensure proper sample handling and result-reporting.
Second, in a significant number of cases, nasopharyngeal swabs were negative but tracheal aspirate specimens were positive. It became routine to send paired nasopharyngeal and tracheal specimens on all intubated patients with suspected 2009 pH1N1; subsequent surveillance was done using whichever specimen had yielded a positive result. This duplicate testing resulted in an increased workload for the laboratory staff, but it was agreed that this was a necessary step to ensure a diagnosis of influenza was being made appropriately. This process was incorporated into a city-wide order set that has been approved for all pH1N1-suspect severe respiratory illness cases (Appendix 1).
Isolation of Patients With Severe Respiratory Illness
Patients who presented with a severe respiratory illness consistent with 2009 pH1N1 influenza were immediately isolated in negative-pressure rooms while such rooms were available. This involved droplet precautions including N95 masks, gowns, and gloves; eye protection was also required for any aerosol-generating medical procedure. Thus, the hospital supply of these items was quickly strained. Gown use at one of the teaching hospitals went from a baseline of 50,000 per month to 100,000 per month during the peak of the outbreak. Furthermore, the limited city-wide availability of negative-pressure ICU rooms was rapidly exhausted.
A universal finding in all the centers that experienced an early outbreak was the demand for N95 masks. In Mexico City, hospital staff, including police, dietary workers, and cleaning staff, demanded N95 masks. All three centers soon ran out and had to come up with novel solutions to this problem. Education about risk of viral transmission seemed to help somewhat in Mexico City. In Salt Lake City, surgical masks were used with patients on closed circuit ventilation, and N95 use was restricted to aerosol-generating medical procedures.
To decrease resource use, a discussion was undertaken regarding the necessity for full personal protective equipment in the setting of closed circuit ventilation. It was decided to maintain full personal protective equipment with all ventilated patients. This was partly based on fear, on the part of healthcare staff, but also based on the real possibility of endotracheal tube disconnection resulting in a break in the circuit, particularly during the acute phase of illness. It quickly became apparent that published guidelines for isolation had not fully considered the risks of ventilator circuit breaches, lack of ventilator expiration filters in older ventilators, and the use of certain advanced ventilatory techniques (e.g., intentional creation of an endotracheal cuff leak for high-frequency oscillation ventilation) in the formulation of isolation policies. We found that application of many such policies without consideration of the realities of patient management issues in the acute and intensive care environments could lead to significant exposure risks for healthcare workers.
The other issue quickly encountered was a lack of negative-pressure isolation rooms. Although these rooms were used whenever available, most patients were managed in nonpressurized rooms. Priority for negative-pressure rooms was given to non-2009 pH1N1 cases as required (for example, isolation for tuberculosis). One solution to this problem was cohorting of pH1N1-infected patients in an emergently refurbished, recently unused, open-room ICU within the hospital.
Cohorting was also used when the medical ICU capacity in the main teaching hospital in Winnipeg became strained. Under normal circumstances, medical patients in excess of medical ICU beds within the primary teaching hospitals are cared for in the surgical ICU under the care of the medical ICU attendings. However, during the epidemic, an attempt was made to concentrate patients with 2009 pH1N1 in the medical units, placing the non-2009 pH1N1-infected patients in the surgical ICU. This was partially the result of a lack of isolation rooms in the surgical units, but also in an attempt to decrease dissemination of 2009 pH1N1 throughout other ICUs.
After confirmation of the diagnosis, the major issue became determination of the duration of isolation. We noted that full isolation of ICU patients represented a major drain on nursing staff. Initial infection control orders mandated that all ICU patients remained isolated until asymptomatic, but many patients had prolonged (2–5 wks) ventilation, which was interpreted by some infection control consultants as ongoing symptoms of disease. During the height of the epidemic, this created an intolerable manpower strain. To minimize this stress (and with the agreement of infection control and microbiology laboratory staffs), weekly polymerase chain reaction of previously positive sites (tracheal aspirates if pneumonitis; nasopharyngeal swab if nonpneumonitis infection) was initiated. Release from isolation was considered with either a negative polymerase chain reaction (of a previously positive site) or positive polymerase chain reaction but negative viral culture (Appendix 1).
In Salt Lake City, in contrast, isolation guidelines from the Centers for Disease Control and Prevention were followed (isolation for 7 days) as a rule. However, some patients remained critically ill after 7 days, and the decision to retain or remove isolation precautions was determined on an individual basis.
Isolation on the medical wards also was difficult as a result of few available isolation rooms. Patients with confirmed 2009 pH1N1 often shared two-bed hospital rooms. Isolation was maintained until resolution of symptoms. Initially, if these patients decompensated and required intubation, this was done on the ward. This was problematic in that those called to intubate the patients were often junior housestaff who lacked experience in this procedure. Furthermore, the personal protective equipment required for this procedure was sometimes not used as a result of the urgency of the intervention. This led to early, elective intubation of these patients in the isolation rooms of the ICUs. Although considered, there was no formal plan to have anesthesiologists intubate these patients as there had been with the severe acute respiratory syndrome epidemic in Toronto, Canada (3).
Experience at several sites demonstrated that patients who recovered from the initial illness were at high risk of decompensation after being transferred to the ward. They needed close monitoring that the wards were not able to provide, and a stepdown unit for these medical patients was not available at our hospitals. To provide these recovering patients with 2009 pH1N1 with appropriate care, a recently vacated intensive care unit in the main Winnipeg teaching hospital was converted into a medicine stepdown unit where isolation precautions could be maintained and a higher level of monitoring (including continuous oxygen saturation monitoring and lower a nurse:patient ratio) could be offered.
Although isolation was technically easy to maintain for staff and patients, the concern soon focused on visiting families. Because many of these patients came from small centers in the northern parts of Manitoba, family members would travel with the patients in close quarters. Many times, three or four relatives of the index case would present with evidence of 2009 pH1N1 influenza within a short period. Because of the large number of patients, the concern for the family members being exposed and subsequent transmission to others in the waiting room and hospital common areas, the medical team (in collaboration with the infection control department and hospital administration) initiated a policy in which family members were not allowed into the patient's room. Family members were encouraged to delay visits if they were feeling ill and were strongly encouraged to seek urgent care if they began showing signs or symptoms of a severe respiratory illness.
To compound the problem, some hotels near the Winnipeg hospitals refused lodging if they found out that families had loved ones admitted to the hospital with influenza. Some families took to sleeping in a local park near the hospital because they were unable to obtain adequate lodging. With the help of the regional health authority, they were placed into hotels to ensure their safety and their health.
One particular northern community had an inordinately high number of 2009 pH1N1 cases. Soon after the recognition of the outbreak, some in the public health sector recommended quarantine with cancellation of all nonessential travel. This recommendation was not enacted, leaving a question as to whether it may have decreased the number of cases among the First Nations peoples. Internationally, the Chinese government quarantined those with suspected pH1N1 and limited travel to combat the epidemic; reported numbers of cases and deaths seemed lower than neighboring countries, suggesting that this policy may have some merit.
Pharmacologic Therapy and Pharmacy Supplies
If 2009 pH1N1 influenza was suspected, patients in Winnipeg were started on both broad-spectrum antibiotics for community-acquired pneumonia and oseltamivir. In consultation with members of the infectious disease department, 75 mg oseltamivir twice daily was deemed sufficient for treatment of pH1N1 influenza. However, the early recommendation that antiviral therapy be started only if symptoms had existed for fewer than 48 hrs was quickly ignored given the severity of disease in patients presenting to the ICU.
Initially, antiviral therapy (oseltamivir) in Winnipeg was terminated after 5 days of treatment according to the standing recommendations in ambulatory patients; however, several intensivists noted a worsening of respiratory parameters 2 to 3 days after stopping oseltamivir, and intensivists began to continue it for a minimum 10-day course. Because of the early perceived need to continue isolation for the duration of ventilation, consideration was also given to continuation of oseltamivir for the same period (often 3–5 wks) or until viral shedding had stopped (up to 3–4 wks). A significant problem was the uncritical application of “evidence-based” guidelines developed on ambulatory patients with influenza infection without regard to the variant pathogenesis, presentation, disease progression, and risk in critically ill patients. This problem spanned clinical care issues from diagnostics to therapy and was the source of recurrent disagreements between intensivists and various consultants.
As noted earlier, most patients required extremely high levels of sedation to tolerate mechanical ventilation while in the ICU. This is likely related to the young age of the patients, the lack of other organ failure, and the profound hypoxia that led to an increased respiratory drive. Data collected by the pharmacy at one of the teaching centers revealed a substantial increase in the amount of antibiotics and sedatives required during the outbreak (4).
The pharmacy looked at antibiotic and sedative use during the fiscal quarter when the outbreak occurred (April–June 2009). The use of piperacillin–tazobactam, levofloxacin, and vancomycin increased by 43%, 131%, and 56%, respectively, compared with the same quarter of the previous year. The use of fentanyl and midazolam increased by 166% and 333%, respectively. To put this sedative use into perspective, the amounts of midazolam and fentanyl required during this time represented 104% and 62%, respectively, of the previous year's total for the entire hospital. This level of sedative use was not anticipated at the beginning of the outbreak and required emergency shipments from suppliers. Physicians began using alternative sedative agents that did not act on opiate or γ-aminobutyric acid receptors, because patients seemed refractory to these agonists. An important lesson from these problems was the tenuousness of “just-in-time” supply-chain dynamics during times of acute stresses on the system.
Treatment of the first few cases also included the use of late-acute respiratory distress syndrome doses of steroids. The Winnipeg anecdotal experience, mirroring that of the ARDSNet late steroid trial, was that of a high rate of critical illness myopathy, resulting in functional quadriplegia requiring prolonged ventilation (5). Subsequent use of steroids was low and case-specific.
Mechanical Ventilation and Rescue Therapy
As might be expected, the number of patients requiring mechanical ventilation placed a great strain on our respiratory therapy (RT) department. At the start of the epidemic in Winnipeg, 35 adult ICU ventilators were available in the main teaching hospital, although less than half were in use at any given time. At the peak of the epidemic, the city transiently exceeded normal full capacity and required emergency mobilization of recently and routinely acquired “replacement” ventilators and the shipment of federal emergency ventilators. Ventilator demand for confirmed or suspected pH1N1 influenza cases across the city of Winnipeg during the outbreak is shown in Figure 1. Since the spring epidemic, the city has increased hospital capacity to 66 ventilators.
The increase in the number of ventilated patients caused a strain on RT staffing. The typical 1:6 ratio of RTs to patients at our institution was stretched to 1:12 at the peak of the epidemic and was resolved by increasing the complement of RTs by one to 1.5 staff per shift. Another issue related to staffing was the significant number of ventilator changes these patients required as a result of frequent desaturation. RTs often would spend upwards of 1 hr adjusting the ventilator or manually ventilating patients to maintain adequate oxygenation.
Similar to our nursing staff, the RTs experienced stress, fatigue, and burnout when having to spend several days in a row looking after these high-risk patients. One of the consistent issues that the RTs described as a source of stress was the sheer number of patient encounters during any given shift and the infection control precautions required before making contact with the patients.
Various forms of rescue therapy were required in very high volumes during the epidemic, including high-frequency oscillation, nitric oxide, and extracorporeal membrane oxygenation. The requirement for each was limited to some extent by local availability, and efforts have been implemented to augment supplies for any potential recurrence of the epidemic. In addition, supplementary training for nurses has been required to support increased use of extracorporeal membrane oxygenation in the event of insufficient perfusionist manpower.
As the clinical management of these patients was being discussed, a surge of patients began arriving from throughout Manitoba. In the initial weeks of the local epidemic, two to three patients were admitted per week; by early June, two to six patients were being admitted each day to ICUs across the city (Fig. 2).
The concern about resources had existed before this patient surge, but support by the hospital administration had been limited. Lack of material supplies was the first hurdle. This included not only personal protective equipment and sedatives, but also ventilators. Fortuitously, a supply of replacement ventilators was delivered soon after the initial outbreak. They were rapidly deployed not as replacements, but as additions to the ventilators already in use.
As the number of cases rapidly increased, the surge overwhelmed the ICUs. The administrative plan was enacted and, within hours, more resources were mobilized and the ICU capacity increased by 10% to 20%. The first phase involved opening unstaffed ICU beds. The nurses required for this were made available through the cancellation of vacation and the use of nurses willing to work above their clinical requirement. In several cases, mandatory overtime was required of the ICU nurses to meet demands.
As the pandemic evolved and existing ICU beds were fully occupied, there was expansion into adjacent physical spaces (i.e., recovery rooms and coronary care units). Increased nursing staff was secured largely by assigning nurses with previous critical care experience to the ICU (i.e., temporarily assigning them from a different patient care area) after activation of emergency staffing articles of the nursing contract. A notable effect was the cessation of most ICU-oriented research at the institutions because most ICU research coordinators were former ICU nurses. Another domino result was nursing shortages in other clinical areas such as coronary angiogram suites and hemodialysis treatment areas.
Notably, there was no substantial increase in sick leave despite the high level of concern regarding vulnerability to infection. This may have been a consequence of educational efforts that emphasized the high prevalence of community infection (i.e., this was not dominantly a nosocomial transmission risk), the low risk of individual progression to severe disease, and the availability of early antiviral therapy to all healthcare workers exhibiting potential symptoms. Along with the issue of personal vulnerability to infection, childcare responsibilities had a large impact on staffing decisions and should be considered in planning for staffing in the future.
In terms of physician support, there was an increase in the nighttime support for the residents with two residents, instead of one, scheduled in the large medical ICUs at the teaching hospitals. The rotating residents in the ICU were asked to increase their number of overnight call to support each other, which led to less time between shifts and more fatigue. Residents were extremely anxious about the severity of illness in the ICU and their ability to make treatment decisions for their patients. They were concerned about the increased number of acute deteriorations in oxygenation and ventilation in these patients. An effort to publicize the availability of psychologic counseling for staff was initiated.
Critical care fellows in our institution were called on to aid the attending physicians in covering the increased patient load and to support the rotating residents. This meant that most fellows were on call every second night, leading to fatigue, stress, and an inability to perform additional nonclinical work.
The attending intensivists did not have a formal plan to deal with the increased workload but asked for additional support on a case-by-case basis. As a result of our staffing model of a 24/7 call from home intensivist attending for 1 wk at a time, physician burnout was always a concern as more in-house time was demanded of the attending physicians. All physicians in our ICU group have primary appointments and clinical responsibilities in other departments, so it was difficult to obtain additional help for the attending physicians. In the community hospitals without resident coverage at night or during the day, there was no change in the coverage model, although attempts were made to send all patients with pH1N1 to the tertiary care ICUs to reduce the spread of infection to other hospital sites.
Other staffing issues involved pregnant medical staff. Given the high rate of severe illness noted in pregnant women, pregnant nurses were not allowed to care for any patient with confirmed or suspected 2009 pH1N1. Pregnant housestaff were strongly encouraged to avoid contact with potential 2009 pH1N1 cases and were allowed to change scheduled rotations to facilitate this.
Burnout was a continuous problem throughout this period. The nursing staff was required to work significant amounts of overtime and became highly fatigued and stressed with substantial increases in voiced levels of dissatisfaction. Some nurses who had not worked in the critical care environment for several years were concerned that they no longer had the skills to care for critically ill patients. Fear of infection, stressful work conditions with difficult-to-manage patients, and a changing work environment contributed to significant nursing staff concerns. However, their number of because of the transient nature of the high workload, there were few voluntary separations from ICU employment. In anticipation of a recurrence of the epidemic in the Fall, remedial training programs for all former ICU nurses who may be asked to support clinical ICU operations are recommended.
By the beginning of July, the number of cases admitted to the ICU dropped significantly. The problem became the ongoing management of these patients with significant ventilator requirements. As patients slowly improved and were extubated, the ICU became less busy, so that by the beginning of August, the steps taken to increase ICU capacity were reversed. Reassigned nurses returned to their usual positions and all staff vacations were honored.
Planning for epidemics like the 2009 (pH1N1) influenza outbreak requires the coordination of all members of the healthcare team. Significant resources and the coordination of government, hospital administration, and clinician actions are needed to adequately care for such patients.
At the hospital level, all members of the healthcare team (cleaning staff, pharmacy, respiratory therapy, laboratory services, nursing, and physicians) need to collaborate on solutions to the myriad of logistic issues that arise.
In our cities, epidemic plans were in place, but unanticipated logistic issues were encountered causing significant management problems. It is hoped that by sharing our collective experience, other medical centers will be better prepared to deal with this and future epidemics.