|Home | About | Journals | Submit | Contact Us | Français|
The federal government states that local communities are primarily responsible for public health planning and implementation during a severe pandemic. Accordingly, an assessment of the current healthcare capabilities in these communities and planning for deficiencies is required. This article assesses the impact and healthcare capabilities of a specific model local community in a mid-Atlantic state. Two statistical models demonstrate the likely impact of both mild and severe pandemics on local healthcare resources. Both models reveal significant deficiencies that local communities may face. In the event of a severe 1918-type pandemic influenza or a mild influenza pandemic, many local community healthcare systems will likely have inadequate resources to respond to the crisis; such a healthcare emergency would likely overwhelm local community resources and current public health practices. Proper planning at the community level is critical for being truly prepared for such a public health emergency.
Although the exact moment of a calamity may not be foreseen, disasters, in general, are predictable events that occur with some periodicity. Pandemic influenza is such a menace on the horizon. A worldwide outbreak of influenza will occur1; the main questions are when and how severe. Many important stakeholders will be involved in the response to a massive biological threat such as a severe influenza outbreak. The local community healthcare system will be required to play a pivotal role. The purpose of this article is to estimate the demanding effects of a pandemic influenza outbreak on a local community healthcare system. The current federal plan in the United States stresses that local communities should not expect outside assistance during a pandemic; therefore, such an assessment is necessary to realistically highlight the challenges many communities would face during such a healthcare emergency.2
This article is divided into five sections. Following this introduction, “Current threat of an influenza pandemic” section briefly describes the current threat of pandemic influenza and discusses a likely causative organism, the virulent H5N1 avian influenza virus. “Model county statistics” section presents the statistics that are relevant for planning for a potential pandemic in a sample suburban county in a Mid-Atlantic state. “Effects of a pandemic” section estimates the number of hospitalizations, deaths, outpatient visits, and the healthcare capacity that could be expected in the sample community if a pandemic were to occur. Finally, in the “Discussion” section, we make several recommendations that may mitigate the effects of a pandemic on local healthcare resources. On the basis of the model presented, it appears that many local community populations and healthcare systems would be overwhelmed in the event of either a severe 1918-type or mild influenza pandemic. To avoid such an eventuality and mitigate such effects, planning for this potential healthcare crisis needs to occur now.
Influenza is a virus from the family Orthomyx-oviridae.3 Within this family, there are three types of influenza: A, B, and C.3,4 Influenza A has been the causative agent in several pandemics and poses the gravest public health risk.3,4 Of particular importance are influenza A’s membrane glycoproteins that mediate entrance into host cells.4,5 The hemagglutinin (HA) and neuraminidase (NA) glycoproteins bind to the sialic acid glycoproteins on host cells and promote envelope-to-host cell fusion to cause entrance of the virus into a host cell.6 Influenza A is unique among the subtypes due to the fact that it maintains a large reservoir of genetic diversity in zoonotic hosts.4 As a result, an animal may be infected with multiple strains of influenza A concurrently. For example, both a strain prone to infect humans and a zoonotic strain may infect the same host. This can result in reassortment (or exchange of genetic material between the two strains) and can produce a novel virus transmissible to immunologically naïve humans.3,4 This process, called antigenic shift, can result in a strain of influenza virus that contains a new HA antigen and is capable of infecting humans. Such a new strain may cause a pandemic.3,4 Exposure of humans to a virus with a novel HA glycoprotein antigen has resulted in numerous pandemics in the past.5
Seasonal influenza A causes 30,000–50,000 deaths in the United States each year.1 Epidemic outbreaks tend to occur every 2–3 years and are likely the result of point mutations in the HA glycoprotein of influenza A, which slightly alter this antigen and allow infections.3 As noted earlier, a reassortment of influenza A in an animal may result in a strain of virus that is predisposed to infect human hosts and contains a completely novel HA antigen. Such reassortments are believed to have resulted in three major pandemics in that last century: the 1918 H1N1 Spanish Influenza, the 1957 H2N2 Asian Influenza, and the 1968 H3N2 Hong Kong Influenza.7 Pandemics tend to appear in waves and last 12–36 months.1 Their virulence varies. The Asian Influenza and the Hong Kong Influenza pandemics were mild, but the 1918 Spanish Influenza pandemic was much more pathogenic, resulting in a mortality 8.22 times higher than that of the 1968 Hong Kong Influenza.8 In addition, the 1957 and 1968 pandemics both followed a mild pattern of disease and most deaths occurred in high-risk individuals.9 In contrast, the 1918 Influenza pandemic caused significant mortality in young, healthy patients.4
Even a mild influenza pandemic like the 1968 Hong Kong Influenza would likely significantly tax a local healthcare system in the United States. It would probably cause between 89,000 and 207,000 deaths, 314,000–734,000 hospitalizations, and 18–42 million outpatient visits, with 20–47 million people being made ill.10 A severe pandemic like the one that occurred in 1918 would likely devastate the economy and stretch our already-strained healthcare system to the breaking point. Such a pandemic could result in 1.7 million deaths in the United States and 180–360 million worldwide.1 Clinical, epidemiologic, and laboratory evidences suggest that the current H5N1 virus, if were able to rapidly reproduce in humans, could cause a pandemic akin to the 1918 strain rather than a mild 1957- or 1968-type event.1,11
Currently, H5N1 exists primarily in duck, geese, and shorebird reservoirs.5 This virus, however, can cross over to infect humans and has done so, resulting in high mortality rates.12–14 There is evidence that isolated, nonsustained human-to-human transmission has occurred; however, it does not appear to spread like human influenza, ie, via inhalation of infectious droplets, direct contact, indirect contact, and self-inoculation.12,13 The primary source of this infection appears to be direct contact with infected poultry.12,13 However, the virus’s ability to spread between humans has the potential to change. Accordingly, H5N1 is a worrisome candidate for the pathogenic cause of the next pandemic. However, it should be noted that H5N1 is not the only candidate pathogen that could be the cause for the next pandemic, as clearly shown by the recent global spread of H1N1 influenza. Other causes of significant concern include coronaviruses, such as the one that caused Severe Acute Respiratory Syndrome (SARS), as well as other unknown viruses that could cause acute respiratory infections.
Of particular concern is the severity of H5N1 infection in humans. As of June 2009,60.5 percent, or 262 of 433, confirmed H5N1 infected patients were reported to have died as result of their disease.14 This infection appears to cause a viral pneumonia and an increased inflammatory response that is reminiscent of the pathology seen with the 1918 influenza pandemic.1,4,11–13 H5N1 data itself represents only laboratory-confirmed cases of infections and fails to include subclinical infections with benign courses.12 Accordingly, the virulence may be less than it appears. Regardless, the high disease mortality and severity of laboratory confirmed H5N1 infections are of great concern.12–14
The most common cause of death from H5N1 infection was respiratory failure.12 Most patients required intubation, ventilator support, and intensive care unit (ICU) care within 48 hours after admission.12 From the perspective of resource-utilization, this requirement is frightening when one considers that the United States only has 105,000 ventilators and that 75,000–80,000 of them are in use at any given time.1 Furthermore, the number of ventilators in use increases to 100,000 during a typical yearly human influenza season.1 Obviously, the healthcare system does not have the capacity to properly handle a major pandemic in which the majority of admitted patients require ventilatory support within 48 hours after admission.1 In addition, the healthcare system likely lacks adequate bed capacity. The American Hospital Association estimates that the average proportion of open hospital beds is around 4–6 percent.11 As a result, the surge capacity for most hospitals is limited.11
To best illustrate the impact of an influenza pandemic, we selected a specific county in the mid-Atlantic region of the United States. This county was evaluated for its ability to address the health impacts of an influenza pandemic. This model suburban county had a population of 271,118 in a 2005 census. Thirty percent of residents (n = 81,335) were between 1 and 19 years of age, 60 percent (n = 162,671) were between 20 and 59 years of age, and roughly 10 percent (n = 27,112) were between 60 and 74 years of age.
The only healthcare facility in the county is a full-service hospital with 204 licensed beds in 2006. The hospital has 16 ICU beds and 16 ventilators available for use. This compares with an average number of licensed hospital beds of 286 and an average number of ICU beds of 32 in Central MD, which includes Baltimore City and its two major academic centers.15 The hospital has no memoranda of understanding with any other healthcare facilities, but it is part of a large university system in the state. A number of other hospitals and public health entities are located in nearby counties, but they would likely be overwhelmed during a severe pandemic and therefore not able to offer substantial assistance to the county’s primary healthcare facility. The closest urban environment, Baltimore, MD, has the potential to act as a major medical resource. Specifically it could provide hospital beds, critical care beds and ventilators, alternative care facilities, and additional personnel.
In the event that our model county experienced a pandemic, it is likely that it would not be able to rely on federal or state assistance. As noted in the White House’s Implementation Plan for the National Strategy for Pandemic Influenza, “local communities will have to address the medical and nonmedical effects of the pandemic with available resources.”2 The report also advises that “the center of gravity of the pandemic response, however, will be in the communities …” and that “communities should anticipate that in the event of multiple simultaneous outbreaks, there may be insufficient medical resources or personnel to augment local capabilities.”2 Accordingly, it is imperative that every community takes stock of its resources and plan to face this crisis.
To estimate the potential effects of a pandemic on the county, we used the FluAid 2.0 and FluSurge 2.0 programs created by Dr. Martin Meltzer at the Centers for Disease Control and Prevention (CDC).16,17 These programs estimate the results of a pandemic on a region based on population and medical resource availability. Both programs have a default setting for a mild pandemic consistent with the 1968 Hong Kong flu.810 For this analysis, both models were used to estimate the numbers of deaths, inpatient admissions, outpatient visits, and ICU admissions as well as ventilator needs and hospital capacity in the county in worst-case and best-case scenarios.
The default Hong Kong Influenza virulence factors of mortality and admission rates were the basis of our best-case scenario. By multiplying admissions and mortality by a factor of 8.22, we created a worst-case scenario simulating the 1918 Spanish Influenza pandemic.10 Further, because the federal government assumes an attack rate of 30 percent, we used this estimate as a future Influenza attack rate.2,11(“Attack rate” is defined as the incidence of symptomatic patients in a community.) Finally, we estimated the number of hospital unit beds, ventilators, and ICU beds available in the county, based on the capacity of an actual suburban general hospital in the model county. The resources provided by the urban environment were not included in the analysis.
If the county were to experience a pandemic similar to the 1918 Spanish Influenza, with a 30 percent attack rate, total mortality in the county would range from 841 to 2,767, with a most likely estimate of 1,589 deaths over a 8-week period. The 8-week period is derived from the mean duration of time for a pandemic to build, peak, and defervesce in a community11 Total hospitalizations would range from 2,492 to 9,431 admissions, with a most likely estimate of 7,267 patients admitted. At 5 weeks, at the peak of the local epidemic, it is estimated that 1,051 patients would need to be admitted to a non-ICU bed and that 304 patients would require ICU care. These requirements translated to 515 percent of the inpatient non-ICU bed capacity and 1,903 percent of the ICU bed capacity of the county general hospital (Table 1).
Further, we made the unrealistic assumption that all the 16 ventilators in the county hospital would be available. At 5 weeks, an estimated 152 ventilators— or 952 percent of the assumed capacity—would be required (Table 1).
The number of outpatient visits could not be accurately predicted because the CDC model’s conversion factor of 8.22 did not allow for this estimation, since multiplying by this factor would result in a number of outpatients in excess of symptomatic individuals.10 It is reasonable to expect the number of outpatient visits would at least equal the numbers seen during a mild pandemic, ie, 36,508–51,380 visits (Table 2). In light of the fact that the total number of outpatient visits at our model county general hospital is 114,148 per year, such a huge number of patients presenting over 8 weeks could not be treated effectively without the implementation of an established surge capacity program.18
The ventilatory support requirements and mortality associated with the current H5N1 influenza A strain exceed those of the 1918 pandemic.2,10–14 Therefore, this worst-case scenario probably is an underestimate. However, as noted earlier, the H5N1 virulence is likely overestimated. In addition, it is thought that H5N1 would become less virulent if it mutated to become a pandemic strain. Accordingly, the 1918 Influenza pandemic, with a severely decreased virulence compared to H5N1, is a fair estimate of a potential severe pan-demic.11 Regardless, without additional surge capacity resources, it is unlikely that the county healthcare facility or the county could alone meet the needs of such an infectious disease disaster. As a result, a 1918-type pandemic would overwhelm the local healthcare system.
Finally, on a national level, Dr. Thomas Inglesby of the University of Pittsburgh Center for Biosecurity has estimated that our national surge capacity would also be inadequate during such an event.11 At the peak of a 1918-type Influenza pandemic, Dr. Inglesby estimates that 197 percent of all hospital beds, 198 percent of all ventilators, and 461 percent of all ICU bed capacity would be required.11 These numbers, though better than the above local worst case scenario, appear similarly bleak and support a finding that the country would not be able to manage the surge in patient volume that would result from a severe 1918-type pandemic.
If the county were to experience an epidemic similar to the Hong Kong Influenza over a 8-week period with a 30 percent attack rate, total mortality would range from 102 to 336, with a most-likely estimate of 194 deaths. Additionally, total hospitalizations would range from 302 to 1,148 admissions, with 884 admissions being likely over 8 weeks. At 5 weeks, during the peak of the pandemic, 128 patients would need to be admitted to non-ICU beds. That is, 128 of the 204 available beds, or 63 percent of the hospital’s bed capacity, would be filled by these patients. Also at 5 weeks, 37 patients would need to be admitted to the ICU. As the county hospital has only 16 ICU beds, this represents 231 percent of capacity. Nineteen ventilators would be required during the fifth week of an epidemic, or 119 percent of the hospital’s capacity (Table 2).
Further, outpatient visits would range from 36,508 to 51,380 during this short period. As in the worst-case scenario, it would be difficult to manage this volume of outpatients, considering that the model hospital’s annual volume is 114,148. Ventilators and antibiotics would be needed to treat secondary bacterial pneumonia and respiratory failure. Although the financial and personnel demands of such an epidemic would be difficult for the community to meet, it would probably not be completely devastating to the local healthcare system (Table 2).
In the event of an outbreak of a 1918-type pandemic with catastrophic results or even a mild influenza pandemic with a smaller but significant increase over baseline levels, many local community healthcare systems will likely have inadequate resources to respond to the crisis. Moreover, such a healthcare emergency would likely overwhelm local community resources and current public health practices. However, a response does not take place in a vacuum, and there are a number of additional surge capacity strategies that may lessen the local community’s demands. To mitigate the deleterious effects of an incident of pandemic proportions, we recommend the following actions to local community healthcare system administrators.
Proper planning for surge capacity is necessary to be prepared to respond to the demand for ward beds that will accompany a pandemic. Local community planners should make efforts to designate surge capacity areas. These activities need to be done in a coordinated manner with all stakeholders, particularly medical care providers and public health authorities. Local Emergency Operations Centers (EOC) and State emergency management officials can likely provide guidance regarding the location of sites such as gymnasiums, schools, or arenas that could be or are designated to be used if increased patient care capacity is needed. In addition, standardized guidelines for mobile triage, nursing and physician care, and resource allocation at such sites need to be developed and reviewed.
In addition to physical space, healthcare providers should work with public health authorities to develop plans that modify standard operating procedures in ways that decrease daily nonpandemic demand. For example, hospitals should develop means of canceling elective surgeries and of quickly discharging appropriate patients so as to rapidly increase admitting capacity.19 These procedures would have the additional benefit of assisting at other times of increased inpatient demand, such as during periods of emergency department crowding. Of note, there have been local, state, and federal efforts to develop a mechanism for changing treatment standards during surge events to provide care for a larger number of patients. This strategy has been termed “altered standards of care” and involves a shift in the practice of medicine to maximize benefit for the greatest number of patients. The Department of Health and Human Services (DHHS), including its Agency for Healthcare Research and Quality (AHRQ), have specifically advocated for such treatment shifts as a potential mechanism for increasing patient care capacity during such incidents.20 Essentially, “altered standards of care” are changes in medical treatment from focusing on maximizing care for individual patients to maximizing benefits for the greatest number of patients.20 This can involve modified triage and allocation of resources based on potential patient outcome. Currently, states are considering guidelines for such altered standards as a mechanism for increasing hospital capacity during overwhelming surge. Planners should contact their states regarding such discussions and review any relevant protocols.
A pandemic will increase the demand for intensive care unit beds and ventilators. This is true even if the severe scenario overestimates the needs considering that there are now potentially effective medications that were not available at the time of previous pandemics. Creative planning to meet this demand must commence now. Makeshift intensive care beds can be created with the use of portable monitors, pulse oximetry, and electrocardiogram monitors. Additional ventilators and other appropriate equipment should be secured and stored now in anticipation for the significant demand that will accompany a pandemic. Alternatively, less-expensive portable ventilators and Ambu bags can be stockpiled in the event that resources are inadequate to purchase or rent ventilators. Also, specific guidelines and recommendations regarding the expansion of ICU capacity within hospitals and the allocation of ICU beds have been created by groups such as the Task Force for Mass Critical Care.21,22 Such guidelines should be reviewed by local community planners in advance of a potential surge event.21,22
The issue of funding has been addressed in the March 2008 Pandemic Planning Update V, published by the US DHHS, which summarizes the progress of the domestic preparedness efforts thus far, and outlines the US government’s current initiatives and financial com-mitments.23 The Update details that Congress has allocated approximately $2.3 billion in funding for pandemic preparedness since 2006, which includes NIH and HHS grants to private and public sectors for vaccine and antiviral development, and an additional $2 billion allocated to state and local governments for public health preparedness and emergency response.23
Healthcare personnel need to be updated, trained, and organized according to their local community’s disaster action plan. The increased volume of patients during a pandemic will require an organized response involving supplemental nurses and physicians. A designated core of leadership in the local government is necessary to assume the organizational abilities of contacting which type of personnel are needed, of processing the volunteer physicians and healthcare workers, and of drafting and enforcing the most ethical and fair guidelines for overriding standards of confidentiality and consent, and directing workers to appropriate sites.24
Within the healthcare community, this level of communication and coordination can be accomplished through interaction and action plans drafted between the local government and hospital CEO’s, provider network CEO’s, or even coordinated by the AMA, providing immediate notifications and alerts with already-established plans for response.24 The National Incident Management System already has an organizational model established termed the Incident Command System, which is a “standardized, on-scene, all-hazards incident management approach that allows [for such pandemic-type event coordination].” (www.fema.gov/emergency/nims/)
This level of healthcare provider coverage can be difficult to achieve because many personnel have commitments to other healthcare organizations. Potential conflicts need to be considered in advance to ensure adequate staffing. In addition, steps need to be taken to ensure the welfare of staff members and their families to minimize absenteeism during a disaster. Also, the likely disproportionate effects of a pandemic on healthcare professionals (as seen in the SARS epidemic and the 1918 pandemic) are decimation of the healthcare workforce and likely absenteeism; this requires planning for the recruitment of community volunteers to assist with subsequent waves of the pandemic.11,25–27 Precertification of appropriate personnel can assist in the credentialing process and lessen the challenges of this critical step in an emergency.
Pharmaceuticals such as antibiotics for treatment of secondary bacterial pneumonia and resuscitation fluids should be stockpiled in anticipation of a pandemic. Of note, a recent study published in The Journal of Infectious Diseases has indicated that secondary bacterial infections caused a significant mortality during the 1918 pandemic. As a result, efforts should be made to stockpile antibiotics and bacterial vaccines as well.28 Additionally, protective gear such as face shields, gowns, gloves, and N-95 respirator masks should be secured now, before a pandemic emerges. Community administrators must continue to monitor influenza vaccine development as well as communicate with local state EOCs and public health officials about stockpiles of antivirals such as oseltamivir (Tamiflu) and procedures for medication distribution to local hospitals. A large amount of such medication is held in the federal Strategic National Stockpile (50 million courses) and must first be allocated, proportionally, to the state by the federal government and then sent out to local communities.29 Suffice it to say, that the logistics of dispersing these medications may be challenging. Of note, with respect to H5N1 infections, the use of oseltamivir appears to have a survival benefit in uncontrolled trials and may be of use in a severe pandemic.13 As a result, gaining clarity and insight into the mechanism of distributing such medication may be of significant importance for community planning.
Finally, community interventions such as school closures and banning of public gatherings appear to have correlated with decreases in peak and total mortality during the 1918 pandemic.30 In a recent retrospective analysis, researchers found that cities that initiated such measures faster and for a longer duration appear to have had decreased overall mortality and total mortality during the 1918 pandemic.25 In light of this, local public health authorities should plan for the early and sustained institution of such measures in the event of such a severe pandemic. By doing this, local communities may decrease their patient burdens and thus mitigate surge demands on their local healthcare systems. Guidance for such nonpharmaceutical interventions is available through the CDC at www.pandemicflu.gov/plan/community/commitigation.html#XVI.
Only through proper planning at the community level will our country be truly prepared for the next public health emergency. On June 11, 2009, the WHO raised its pandemic alert level to 6, indicating that an influenza pandemic is under way. The causative organism is the novel H1N1 virus that was initially isolated in Mexico.31 Currently, the virulence of this influenza virus is moderate, but this may change.31 As a result, it is not clear what local communities will need to address during this pandemic. This article provides a spectrum of potential local healthcare demands based on virulence and makes some suggestions for preparation. In light of this current pandemic, the assessment and strategies reviewed by this article may be of significant importance to community planners.
We would like to thank Linda J. Kesselring, MS, ELS, for her thoughtful manuscript review. Dr. Hirshon was supported by NHLBI grant 5K08HL073849. Dr. Mareiniss received support from The Horizon Foundation.
Darren P. Mareiniss, Senior Consultant for Medical and Legal Policy, University of Maryland Center for Health and Homeland Security, Baltimore, Maryland.
Jon Mark Hirshon, Associate Professor, Department of Emergency Medicine, University of Maryland School of Medicine, Baltimore, Maryland; Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore, Maryland.
Bryan C. Thibodeau, Department of Emergency Medicine, University of Maryland School of Medicine, Baltimore, Maryland; Department of Pediatrics, University of Maryland School of Medicine, Baltimore, Maryland.