The use of prescription antiviral drugs to treat and prevent infection will be an important component of a pandemic influenza response. While current antiviral drug use strategies and publicly maintained stockpiles are targeted primarily for treatment of persons with pandemic illness, expanded antiviral drug production has allowed additional strategies to be considered. An interagency working group, with input from representatives of State, local and tribal public health agencies, considered scientific issues, ethics and values, and perspectives of stakeholders in developing draft guidance on antiviral use strategies and stockpiling. The antiviral drug use guidance in this document replaces the recommendations developed in 2005 which are published as part of the Department of Health and Human Service’s (HHS) pandemic preparedness and response plan.
The proposed guidance on antiviral use is based on the national pandemic response goals of slowing the spread of pandemic disease, reducing impacts on health, and minimizing societal and economic disruption. The working group recommends the following strategies and settings for antiviral use to meet these goals:
Antiviral drugs are being stockpiled by HHS as part of the Strategic National Stockpile, (SNS) and by States. The current public sector stockpile target is 81 million drug regimens: 6 million regimens for containment and for slowing the entry of pandemic disease into the United States, and 75 million regimens for treatment. Implementation of recommendations for prophylaxis of healthcare and emergency services workers who have high-risk exposures and for PEP in recommended settings will depend largely on private sector organizations and businesses purchasing and stockpiling antiviral drugs for their employees. The working group encourages governments, healthcare organizations and other employers, and families and individuals as appropriate, to purchase and stockpile sufficient antiviral drug supply to support recommended antiviral drug use strategies and to plan for effective implementation at the time of a pandemic as part of comprehensive pandemic planning and preparedness.
In addition to the proposed national recommendations on treatment and prophylaxis, businesses that provide goods or services essential to community health, safety, or well-being (“critical infrastructure” sectors) should strongly consider antiviral prophylaxis for critical workers as part of comprehensive pandemic preparedness planning, especially those workers who are individually critical and whose absence would jeopardize provision of essential services. Other employers may consider antiviral prophylaxis for workers to maintain business continuity or protect employees.[2] PEP for household contacts of persons with pandemic illness will reduce their risk of infection and may decrease overall rates of pandemic disease in communities. Despite these potential benefits, however, further work is needed to assess the feasibility of this strategy and identify approaches for purchasing and stockpiling the antiviral drugs to support its implementation. Therefore, the working group makes no recommendation for household antiviral PEP at this time.
Antiviral medications from the SNS, other than those targeted for containment and use at U.S. borders, will be allocated pro rata and delivered to Public Health Emergency Preparedness Project Areas (includes 50 States, 4 major metropolitan areas, and 8 U.S. territories) when a pandemic occurs. The working group recommends that public sector antiviral drug supply be prioritized for treatment of all persons who may benefit from therapy based on assessment of medical need. Treatment is preferred to prophylaxis in settings of limited antiviral drug supply; targeting some antiviral drug supply for prophylaxis and prioritizing treatment for certain groups would raise significant ethical and logistical challenges. Effective implementation of community mitigation strategies to reduce rates of illness and greater accuracy in diagnosing pandemic influenza illness would reduce antiviral drug needs, potentially leading to an ability to provide some prophylaxis while maintaining a treatment policy. Among prophylactic antiviral drug uses, protecting front-line healthcare and emergency services personnel is the top priority.
Ongoing discussions with stakeholders and the public are important as part of a transparent process and to move forward in addressing implementation issues. Rapid implementation of these strategies during a pandemic will pose substantial challenges. Periodic reassessment of national antiviral drug guidance will be important based on scientific and technological advances. Strategies also should be reassessed when a pandemic occurs to take into account the characteristics of the virus, epidemiology of disease, and impacts on society.
The use of influenza antiviral medications (specifically, prescription medications that have activity against the influenza virus, also referred to as antiviral drugs or antivirals) will be an important component of a multi-faceted response to an influenza pandemic. Other key response measures include non-pharmaceutical approaches such as social distancing (with multiple strategies described in the guidance on community mitigation[3]), improved hygiene, and use of respiratory protection measures such as facemasks and respirators in appropriate situations. Vaccination against the pandemic influenza virus is likely to provide the most durable protection against pandemic illness but pandemic influenza vaccine only can be developed once the pandemic virus is identified; no vaccine or only limited quantities of stockpiled pre-pandemic vaccine[4] may be available when the first U.S. pandemic wave starts. Once pandemic vaccine becomes available, it will be administered according to a prioritization strategy that targets high risk and critical occupational groups and vulnerable populations first.[5]
By contrast with pandemic influenza vaccine where availability and supply at the onset of U.S. pandemic outbreaks cannot be predicted, antiviral medications can be stockpiled and availability assured at the onset of the pandemic. Plans for stockpiling antiviral drugs should be based on strategies for their use so that sufficient quantities are on hand to support recommended interventions. The current national target for Federal and State antiviral drug stockpiles is 81 million regimens. This includes 6 million regimens to contain or suppress initial pandemic outbreaks overseas and in the United States, and 75 million regimens targeted for treatment of ill persons. Of the 81 million regimens to be stockpiled, 50 million have been purchased by the Federal Government and 31 million are allocated for State purchase proportional to population, with a 25% Federal cost share. Antiviral agents recommended for inclusion in the stockpile based on a treatment strategy include the neuraminidase inhibitors, oseltamivir (Tamiflu®) and zanamivir (Relenza®), with about 80% and 20% of the Federal stockpile component made up of the respective agents.[6] In addition, several million regimens of rimantadine, purchased in a season of influenza vaccine shortage, are still held in the Strategic National Stockpile (SNS). Additional stockpiling of M2-inhibitor antiviral drugs (amantadine and rimantadine) has not been recommended because resistance to these agents among circulating influenza A viruses is frequent and, among susceptible viruses, develops rapidly when they are used to treat influenza A virus infections.
The 2005 HHS Pandemic Influenza Plan defines treatment as the primary strategy for antiviral drug use. [7] Treatment was recommended for several reasons:
At the time the 2005 recommendation was made, global production of the neuraminidase inhibitors was limited. In the context of requirements for seasonal influenza use and pandemic stockpiling by the United States and other countries, sufficient drug was not available to support more expanded recommendations and stockpiling. Recent expansion of antiviral drug production capacity – including a U.S. based supply chain for oseltamivir – has made increased stockpiling possible. This, in combination with several recent analyses suggesting substantial potential benefits of prophylactic antiviral drug use in a pandemic, prompted a re-assessment of pandemic antiviral drug use strategies and potential stockpiling targets.
This document updates the 2005 recommendations and provides guidance to Federal, State, local, and tribal planners on antiviral drug use strategies and the number of antiviral regimens that would be needed to support implementation.[14] This guidance supports national pandemic response goals to: 1) stop, slow, or otherwise limit the spread of a pandemic to the United States; 2) limit the domestic spread of a pandemic, and mitigate disease, suffering, and death; and 3) sustain infrastructure and mitigate impact to the economy and the functioning of society. The guidance should be considered “interim”; recommendations should be reassessed as new scientific and technological advances are made, and at the time of the pandemic when the characteristics of the pandemic virus and epidemiology of disease are known.
HHS convened a Federal interagency group that included representatives from Federal agencies[15] and obtained input from State and local public health experts. Tribal health was represented by participants from the Indian Health Service.
In its deliberations, the working group considered the national goals of a pandemic response; results of scientific studies on the effectiveness of antiviral treatment and prophylaxis for seasonal influenza infections; treatment for H5N1 avian influenza infections; surveillance data and studies of antiviral resistance; results of mathematical modeling of antiviral drug and non-pharmaceutical interventions; perspectives of State, local, and tribal health officials; and public values and ethical principles. The potential impacts of antiviral treatment and PEP in households of persons with pandemic influenza disease were estimated using a mathematical model.
Key assumptions underlying estimates of antiviral impacts and the number of regimens needed to support proposed strategies include the following:
Quantitative estimates of the number of antiviral drug regimens[19] needed to completely implement recommended strategies are based on these assumptions and therefore are subject to substantial uncertainty. Estimates should be reconsidered as additional scientific data become available and as new technologies (for example, sensitive and specific point-of-care diagnostic tests) are developed.
Prophylaxis strategies considered by the Working Group include PEP and outbreak prophylaxis. PEP is given within 48 hours following close contact with a person who has pandemic influenza illness, for example those living in the same household, and requires one regimen per person to implement. Outbreak (“pre-exposure”) prophylaxis is given before exposure occurs at the onset of a community pandemic outbreak and is continued for the entire period of the local outbreak, potentially requiring up to eight antiviral drug regimens per person.[20]
The working group considered prophylactic antiviral drug use strategies in various populations and settings. Among occupational groups, a key consideration was the risk of occupational exposure as defined by the Occupational Safety and Health Administration (OSHA) risk pyramid.[21] Workers with exposures to persons known to be infected with pandemic influenza are defined as being at very high and high risk; those who have unavoidable and frequent close contact with persons not known to be infected are defined as being at medium risk; and those without frequent close contacts are defined as being at low risk. A second consideration was the worker’s role in maintaining essential community services or continuity of business operations that may be threatened in a severe pandemic. For non-occupational settings, people’s health status (e.g., the presence of immunocompromising illness or therapy) and living situation (e.g., residence in a group setting such as a nursing home or prison) were considered. A description of specific groups considered for antiviral drug prophylaxis is provided in the Appendix.
Guidance on the use of antiviral drugs in a severe influenza pandemic should be linked with the available supply. U.S. antiviral drug strategies and stockpile targets were included in the 2005 HHS pandemic influenza plan and were established when production capacity and supply were limited. Expanded capacity now allows expanded recommendations. However, these recommendations do not obligate the Federal Government or public sector to purchase or stockpile antiviral drugs for all recommended uses and implementation will depend on the actions of other sectors of society. Pandemic preparedness is a shared responsibility of all levels of government, businesses, families, and individuals. Responsibility for purchasing and stockpiling antiviral drugs for prophylaxis of occupationally defined groups will largely be in the private sector. Federal efforts to reduce barriers to purchasing and stockpiling antiviral drugs are ongoing. Federal guidance for employers is available at [website TBD].
Recognizing that expanding antiviral drug stockpiles will take time and that some States are unlikely to purchase their full allocation to support treatment of those with pandemic illness, prioritization may be needed until supply is sufficient to support all recommended uses. Public sector stockpiles should be prioritized for treatment because it represents the most efficient use of a limited drug supply, because prophylaxis for some while others are denied treatment would not be perceived as equitable, and because other measures can be implemented to protect workers and reduce the risk of exposure and infection.
In addition to the antiviral drugs that are stockpiled before the pandemic, there is a possibility that oseltamivir production in the United States may provide an opportunity to acquire limited additional supply when a pandemic is imminent and before (or during) U.S. disease outbreaks. Any additional antiviral drugs that may be produced would provide additional flexibility. Because this supply likely will be limited and can not be predicted in advance, planners should not decrease stockpile targets based on an assumption that ongoing production will fill supply gaps. As all 20th century pandemics have occurred in both fall and winter disease waves (and during the 1918 pandemic, a first spring wave also occurred), this capacity may contribute to preparedness for later waves. Guidance for use of antiviral drugs may be different during a second pandemic wave when many people already would be immune because of prior disease or vaccination and when impacts on critical infrastructures may differ.
Allocation of antiviral drugs that are in limited supply raises several ethical issues. The following principles were considered by the working group in developing antiviral drug use recommendations:
These principles are consistent with the values expressed in several public engagement and stakeholder meetings that were held to consider pandemic influenza vaccination prioritization. When asked to rate the importance of potential program goals, participants rated most highly the goals of protecting those who contribute to a pandemic response, who provide care for people with pandemic illness, who maintain essential community services, and who are at increased risk of infection because of their job.
Consistent with ethical principles of transparency and inclusiveness and recognizing the value of public input in developing policy, representatives from the working group met with stakeholders in State and local government, public health, healthcare and emergency services, businesses, organized labor, and the public. Objectives of these meetings were to obtain input on the proposed recommendations; identify potential barriers to implementation; and define interventions that would facilitate implementation. Overall, 13 meetings were held with representatives from about 400 organizations who participated either in-person or by telephone. From this process, there was general agreement with expanding the antiviral drug guidance to include prophylactic uses, including consensus support for a recommendation of outbreak prophylaxis for healthcare and emergency services personnel who have frequent exposure to persons with pandemic influenza infection. Additional findings from the stakeholder engagement activities are included in the section on Implementation later in this document.
The working group recommends that antiviral drugs should be used in the following settings (see Table).[23]
The total number of antiviral drug regimens needed to fully implement the working group recommendations substantially exceeds current public sector stockpiling targets. Because of this, implementation will require governments, healthcare organizations and other employers, and families and individuals as appropriate, to purchase and stockpile sufficient antiviral drug supply to support these strategies, and plan for effective implementation at the time of a pandemic as part of comprehensive pandemic planning and preparedness. Public sector stockpiles should be targeted for containment, delaying a U.S. pandemic, and treatment. Guidance from the Strategic National Stockpile recommends that antiviral drugs distributed from the SNS to the States be used for treatment. Implementing prophylactic antiviral drug strategies, therefore, will require the establishment of stockpiles in the private sector. This approach is consistent with the role employers play in protecting their workers and operations against other types of risk.
In addition to these recommended uses of antiviral drugs, the working group recognizes that antiviral prophylaxis also may be beneficial in other settings. Businesses that provide services essential to the health, safety, and well-being of communities have a special responsibility to plan to maintain those services in an influenza pandemic.[24] Antiviral prophylaxis is one intervention that can protect the health of workers, decrease absenteeism, and help preserve the ability to deliver essential goods and services. Because non-pharmaceutical measures also can provide substantial protection and not all companies have the capability to manage an antiviral drug program that includes stockpiling and dispensing a prescription medication, the working group does not recommend antiviral prophylaxis for essential workers in critical infrastructure businesses. However, these businesses should strongly consider whether antiviral prophylaxis (post-exposure or outbreak) should be included as part of a comprehensive pandemic preparedness and response plan. Outbreak prophylaxis may be particularly important for workers who are individually critical to the provision of essential services, where their absence during a pandemic would jeopardize the delivery of those services. In addition, employers in all sectors may consider antiviral prophylaxis as a part of business continuity planning for a pandemic.
Post-exposure prophylaxis for household contacts of persons with pandemic influenza illness has been proposed to reduce secondary transmission to close contacts where risk is high and to reduce overall spread of the pandemic in communities. Clinical studies in seasonal influenza show PEP to be very effective in preventing illness among household contacts and mathematical modeling results suggest that this intervention, if widely applied, may reduce pandemic disease in communities. Despite these potential benefits, however, further work is needed to assess the feasibility of this strategy and identify approaches for purchasing and stockpiling the antiviral drugs to support its implementation. Therefore, the working group makes no recommendation for household antiviral PEP at this time.
| Condition | Setting and target population | Antiviral strategy | Rationale | Pandemic response goals addressed by antiviral strategy | Estimated number of regimens (million) |
|---|---|---|---|---|---|
| Recommendations for antiviral drugs in public sector stockpiles | Initial pandemic outbreaks overseas and in the U.S. | Treatment; PEP, targeted prophylaxis |
|
Slow pandemic spread | 6 |
| Exposed travelers entering the U.S. early in a pandemic | PEP | Contributes to a risk-based policy to reduce the entry of infected persons and delay U.S. outbreaks | Slow pandemic spread | ||
| Persons with pandemic influenza illness | Treatment |
|
Reduce health impacts | 79 Note | |
| Recommendations for antiviral drugs in private sector stockpiles and if public sector supplies are sufficient | Healthcare and emergency service workers | Outbreak (pre-exposure) prophylaxis for workers with high-risk exposure |
|
|
103 |
| PEP given exposure of other workers | |||||
| Outbreak control in closed settings (e.g., nursing homes, prisons) | PEP |
|
Reduce health impacts | 5 | |
| Persons who are severely immuno- compromised | PEP |
|
Reduce health impacts | 2 | |
| TOTAL NUMBER OF ANTIVIRAL DRUG REGIMENS FOR FULL IMPLEMENTATION | 195 | ||||
Note: A prior estimate of 75 million regimens has served as the basis for public sector stockpiling. The 79 million regimen estimate was calculated using updated planning assumptions
Antiviral medications from the Strategic National Stockpile will be allocated to States pro rata when a pandemic occurs. Whereas the Federal share combined with State-purchased antiviral drugs was estimated to be sufficient to treat all persons who present for care and may benefit from therapy in a pandemic, not all States have purchased their full allocation. Despite potential shortfalls when antiviral stockpiles are limited, treating all persons based on assessment of medical need is considered preferable to targeting certain priority groups for treatment. This approach better meets the ethical principle of fairness and recognizes the significant uncertainty in estimating stockpiling requirements. Effective implementation of community mitigation strategies to reduce rates of illness and greater care in diagnosing influenza-like illness may help reduce antiviral needs. Monitoring supply and antiviral use rates during a pandemic is important so that adjustments could be made in policy, if needed.
Treatment is preferred to prophylaxis in settings of limited antiviral drug supply as the need is clear and benefits likely to accrue for those who are treated. By contrast with the single regimen needed for treatment, if community mitigation measures are effective at reducing the illness rate in a pandemic to 15% and outbreak prophylaxis is given as eight antiviral drug regimens (to protect the recipient for a 12 week period), it would take about 53 antiviral drug regimens to prevent one infection. Diligent application of non-pharmaceutical measures can offer substantial protection for high-risk and critical workers, and information and education on the effectiveness of these measures can help reduce absenteeism due to fear of infection in the workplace. If needed to maintain critical functions, PEP after unprotected exposures (e.g., in a healthcare setting where respiratory protection with a respirator was not used) may have similar effectiveness and would be more efficient than outbreak prophylaxis when antiviral drug stockpiles are limited.
Reserving some antiviral drugs at the State level to use for outbreak control in closed and high-risk settings is recommended as an effective and efficient use of antiviral drug supply. PEP is recommended as a component of outbreak control only in settings defined by a high risk of transmission that may result in an explosive outbreak with high attack rates; a high risk of severe complications and death among those infected; and a lack of other effective control measures. Examples of such settings include long-term care facilities for persons who are elderly or have underlying illnesses or hospital bone marrow transplant units. Use of antiviral drugs in these settings should be guided by results of an epidemiological investigation and recommendations by public health personnel.
Establishment of antiviral stockpiles by healthcare organizations, long-term care facilities (e.g., nursing homes), businesses, and families that include a person who is immunocompromised will expand antiviral drug availability and allow implementation of the antiviral prophylaxis recommendations without limiting the ability to provide treatment to persons who may benefit. Because separate caches of antiviral drugs will be procured and stockpiled for these purposes, no prioritization strategy can be proposed as organizations or employers will stockpile to address their specific needs.
Although this document does not focus on implementation, the working group has considered feasibility in making its recommendations. Rapid diagnosis and dispensing are critical for optimal treatment and PEP. While studies with seasonal influenza show benefit if started within 48 hours, more rapid administration of antiviral drugs has been shown to result in greater treatment impact. Pandemic planners are encouraged to develop strategies that lead to more rapid implementation, if possible. This requires education for ill persons to seek medical consultation or care soon after symptom onset, a healthcare system capable of rapidly diagnosing influenza (potentially including telephone triage and diagnosis), an ability to identify who qualifies for treatment or PEP, and rapid dispensing of antiviral drugs. In healthcare and emergency service sectors, workers who provide direct patient care and have high-risk exposure must be identified, antiviral drug stockpiles must be established, and plans must be made for dispensing at the time of a pandemic. Monitoring systems must be established to track public sector antiviral drug supply and utilization to assess whether supplies are likely to be adequate for recommended uses, to assess rates and impacts of antiviral drug resistance, and to identify adverse events associated with antiviral drug use.
Because implementation of this guidance will rely, in part, on private sector antiviral drug stockpiling, stakeholder engagement activities identified potential barriers to establishing such stockpiles. Prominent concerns included cost, shelf-life, regulatory and logistical issues, potential liability, and the possibility that private stockpiles could be appropriated by the government. Each of these issues is being considered further and approaches are being considered to decrease potential barriers.
This draft guidance on antiviral drug treatment and prophylaxis is based on consideration of scientific, behavioral and logistical issues, as well as societal values. Ongoing discussions with stakeholders and the public will be useful as part of a transparent process and to move forward in addressing implementation issues. Antiviral prophylaxis guidance should be reassessed periodically as additional information becomes available and new infrastructures, capabilities, and materials are developed. Logistical issues surrounding stockpiling (e.g., warehouse requirements for an expanded stockpile) and implementation also should be considered.
The working group recommends the following actions to improve the scientific basis for guidance and the ability to effectively implement an antiviral strategy during an influenza pandemic:
The working group considered antiviral drug use in a variety of settings. Following is a summary of working group deliberations leading to its recommendations. Note that not all of the antiviral drug use settings and strategies discussed by the working group were recommended in the proposed national guidance.
Treatment with neuraminidase inhibitors has been shown to shorten the duration of seasonal influenza illness by about 1 – 1.5 days when begun within the first 48 hours after symptom onset.[25] A Canadian observational study conducted after regulatory approval suggested that the magnitude of benefit increases with a shorter interval from symptom onset to treatment: among persons presenting for treatment within the first 12 hours of illness, the duration of symptoms was shorter by 3 to 4 days.[26] Significant benefit has not been documented when treatment is begun more than 48 hours after symptom onset.
Post-hoc combined analyses of data from randomized controlled trials suggest that early neuraminidase inhibitor treatment of patients with uncomplicated influenza significantly reduces rates of lower respiratory tract complications requiring antibiotic treatment (bronchitis and pneumonia) and hospitalization. For oseltamivir, reductions of 55% and 59% occurred for the respective outcomes. However, the number of persons experiencing these outcomes was limited, confidence limits around these estimates were wide, relatively few persons at high risk of influenza complications were included in the study groups, and the analysis combined subjects from different trials and was not specified in the original protocols.[27] Epidemiological studies comparing outcomes of patients who did or did not receive oseltamivir treatment also found significant reductions in pneumonia and hospital admission.[28] Clinical trials were not powered to assess a potential impact of treatment on influenza mortality because of their limited sample size. Such an impact would be predicted based on the reduction in illness severity, defined by lower respiratory infection or a need for hospitalization.
An epidemiological study in Toronto, Canada, assessed the impacts of oseltamivir treatment at hospital admission for persons with confirmed influenza infection compared with an untreated control group. Most of the treated patients were elderly (mean age 79 years old) and many had underlying health conditions. Mortality in the treated and untreated groups was 3.9% and 10%, respectively. Treatment was associated with a significant reduction in mortality (Odds Ratio 0.21, p = 0.03).[29] In addition, a study of Canadian nursing home outbreaks showed significantly fewer deaths in nursing homes where oseltamivir was used for treatment compared with nursing homes where no treatment or amantadine treatment was given.[30]
Treatment effects for pandemic influenza could differ from those for seasonal influenza illness. Oseltamivir and zanamivir were not available at the time of previous pandemics; amantadine which was used during the 1968 pandemic was effective as treatment. Although not a pandemic virus, a case series of patients with avian H5N1 infection reported by WHO found similarly severe outcomes among oseltamivir-treated and untreated persons. Treatment was begun, however, between days 4 and 22 of the illness and many patients had complications including pneumonia at the time therapy was begun.[31] More recently, a WHO review suggested that early oseltamivir treatment may reduce H5N1 associated mortality and, because of prolonged replication of this virus, that treatment is warranted even with late presentation.[32] In vitro and animal model data show that oseltamivir can be effective as treatment of H5N1 avian influenza, though one study suggested that optimal effectiveness may require a longer duration of therapy.[33] An expert panel convened by WHO made a strong recommendation for oseltamivir treatment of persons with avian H5N1 influenza infection.[34]
In addition to its direct effects on duration and severity of illness, treatment may also have indirect benefits. Results of mathematical models suggest that treatment early in the course of illness can reduce transmission to close contacts. The magnitude of this effect is limited because the period of peak infectiousness occurs shortly after symptom onset, generally before treatment begins. Results of mathematical models should be interpreted with caution as they depend on assumptions and parameters which often are based on limited data. Another indirect effect of treatment, associated with shorter illness duration and decreased worker absenteeism, would be a better ability to maintain effective healthcare, emergency services, and other essential community functions. The potential magnitude of this impact cannot be quantified.
In estimating the direct health benefits of treatment and the number of antiviral drug regimens needed to support a treatment strategy, key parameters include the attack rate of pandemic disease, the proportion of persons with pandemic illness who are treated, the responsiveness of the pandemic virus to treatment and the potential need to adjust the dose or duration of treatment, and the ability to target treatment to persons with pandemic illness in the absence of a sensitive and specific point-of-care diagnostic test. We assume an attack rate of 15% for pandemic illness in the context of effective community mitigation and a positive predictive value of a clinical diagnosis of influenza-like illness (ILI) of 35%, as described above. We assume that 60% of persons with pandemic illness will be treated within 48 hours of illness onset; this estimate is slightly more than the approximately 50% of persons with seasonal influenza who seek care at any point during their infection. Health benefits are calculated assuming a 59% reduction in hospitalizations consistent with clinical trials for seasonal influenza, and an assumed 25% reduction in pandemic mortality.
Given these assumptions, for the first wave of a 1918-like pandemic, antiviral treatment could prevent about 144,000 deaths and about 1.85 million hospitalizations. A total of about 79 million antiviral regimens would be needed to support a treatment strategy. This suggests that one death would be prevented for about 550 antiviral regimens and one hospitalization would be prevented for about 40 antiviral regimens. At current Federal contract costs, this translates into a cost of about $11,200 per death prevented and $1,000 per hospitalization prevented – a highly cost effective intervention.
For seasonal influenza, PEP among household contacts of persons with influenza infection has been shown to be very effective in preventing illness. The effectiveness of PEP using a neuraminidase inhibitor to prevent seasonal influenza illness has been 70% to 90% in household settings when started within 48 hours of the case’s illness onset.[35],[36],[37] There have been no studies of neuraminidase inhibitor PEP for pandemic influenza or for H5N1 avian influenza where secondary infections in close contacts have been extremely rare; studies in animal models show pre-exposure prophylaxis to be effective in preventing H5N1 infection.
PEP of household contacts will have the direct benefit of preventing infection, illness, and its consequences within the household – a setting where about one-third of all influenza transmission is estimated to occur. In addition, because persons in the house of a case-patient are less likely to become infected when PEP is given, they will not transmit infection to others in the community, reducing the overall spread and burden of influenza disease. Antiviral treatment is likely to have little impact on the overall rate of illness in communities. By contrast, household PEP may be more effective because it is given before people become ill and are infectious to others. The magnitude of this “indirect” benefit depends on the proportion of households that receive prophylaxis, how soon they receive the antiviral medication, and the effectiveness of prophylaxis. Mathematical models have assessed the potential impact of household PEP when applied in combination with non-pharmaceutical interventions including isolation of cases, voluntary quarantine of household members, dismissing children from schools and preventing them from re-congregating elsewhere, and reducing close contacts in communities and workplaces (“social distancing”). Under one scenario where antiviral PEP was implemented in 60% of case households at 24 hours after illness onset in the case-patient, one model predicts an 8% relative reduction in illness attack rate compared with a scenario where antiviral PEP was not included.[38] Other models, using the same pandemic scenario, suggested smaller incremental benefits of adding antiviral PEP.[39],[40]
Household antiviral PEP also may contribute to the success of community mitigation strategies by improving compliance with recommendations for voluntary household quarantine. Well family members may be more willing to stay home with an ill person if they are protected with antiviral PEP. Some planners have framed this as an ethical issue: if persons are asked to assume greater risk by remaining at home with an ill family member, one should provide protection through household PEP.
In estimating the potential impacts and number of antiviral drug regimens needed to support a household PEP strategy, we assumed that 60% of case households would be included, prophylaxis would be started at 24 hours after illness onset in the case-patient, and that all household contacts would receive the medication. As described above, we assumed the positive predictive value of a clinical diagnosis of influenza is 35%. The impact of PEP on reducing illness attack rates was extrapolated from Ferguson’s mathematical model.
Assuming a 15% attack rate with community mitigation and antiviral treatment, the model predicts a reduction in attack rate to 12.5% with the addition of antiviral PEP. This corresponds to about 155,000 fewer deaths and about 838,000 fewer hospitalizations. Overall, the combination of treatment and household PEP is estimated to reduce pandemic deaths by about 288,000 and hospitalizations by about 2.4 million. To achieve these outcomes, based on the assumptions used in the model, would require a total of about 167 million antiviral regimens. The incremental antiviral requirement for PEP compared with a treatment strategy alone is 88 million regimens. Overall, this strategy could lead to the prevention of one death for about 580 antiviral regimens and prevention of one hospitalization for about 70 antiviral regimens. This represents a cost of about $11,800 per death prevented and about $1,400 per hospitalization prevented when treatment and household PEP are combined. Cost per death and hospitalization prevented by the PEP component alone would be about $14,000 and $2,600. Again, these estimates represents cost efficient outcomes based on commonly accepted standards and when compared with routinely recommended medical interventions.
Prevention and cost-effectiveness estimates for antiviral treatment and household post-exposure prophylaxis strategies alone and in combination are shown in the Table.
| Parameter | Treatment alone | Household PEP alone | Treatment and Household PEP |
|---|---|---|---|
| Number of deaths prevented | 144,000 | 155,000 | 288,000 |
| Number of hospitalizations prevented | 1.845 million | 838,000 | 2.427 million |
| Number of antiviral regimens | 79.4 million | 106.4 million | 167.1 million |
| Cost per death prevented Note1 | $11,200 | $14,000 | $11,800 |
| Cost per hospitalization prevented Note1 | $900 | $2,600 | $1,400 |
Note: Average cost per regimen based on Federal contract price for oseltamivir and zanamivir, the relative proportions of each agent targeted for acquisition for the national stockpile, rounded to the nearest $100
Maintaining effective healthcare and emergency response
services (includes Emergency
Medical Services, fire, and law enforcement personnel) will be essential in
preventing adverse health outcomes and protecting public safety in a pandemic.
The healthcare sector will face a massively increased burden while coping with
a workforce diminished by illness and possibly other causes of absenteeism –
for example, caring for an ill family member or due to fear of becoming
infected in the workplace. In a survey of public health personnel in three Maryland county health departments,[41] only 54% of respondents indicated that they would likely report to work during
a pandemic. In a multivariable analysis, one factor significantly associated
with the likelihood of reporting was confidence in one’s personal safety. Respondents
were not directly asked about antiviral drug treatment or prophylaxis and
responses to a hypothetical scenario must be interpreted with caution. Limited
information from the 1918 pandemic and experience in Toronto, Canada, during the recent SARS outbreak suggest much lower rates of absenteeism among healthcare
workers. Nevertheless, the Maryland findings raise the possibility that absenteeism
could be substantial and that antiviral prophylaxis may reduce absenteeism both
by preventing illness and by improving perceptions of safety in the workplace.
Several potential strategies for prophylaxis in healthcare and emergency service settings could be considered. Because exposure to ill persons during a pandemic outbreak will be frequent for healthcare workers and emergency service personnel with direct patient contact, post-exposure prophylaxis would be essentially equivalent to outbreak prophylaxis – as soon as one 10-day course of PEP ended, another would likely begin. A modification of the PEP strategy may be to dispense PEP only when “unprotected” exposure occurred. Potential concerns with this approach for those with frequent high-risk exposures include whether it would be sufficient to reduce absenteeism that may occur due to fear of occupational infection, whether unprotected exposures could be accurately identified and how frequently they would occur in a heavily exposed population. In addition, there is a lack of data on the effectiveness of personal protective equipment measures in preventing influenza transmission. A hybrid strategy that includes outbreak prophylaxis for workers with frequent high-risk exposures and post-exposure prophylaxis when unprotected exposure occurs for those who have less frequent or intensive patient contact tailors the intervention to the level of risk and is the preference of the working group.
Estimating the number of antiviral drug regimens needed to support prophylaxis for healthcare and emergency service workers using this strategy requires defining populations of workers with more frequent higher risk exposures and those at lower risk. Of the approximate 13 million workers in the healthcare sector as defined by the Bureau of Labor Statistics, we estimate that two-thirds of healthcare workers, or about 8.7 million, may have frequent high-risk exposures along with 2 million persons in emergency services sectors, encompassing Emergency Medical Services, fire service and law enforcement personnel. The remaining 4.3 million healthcare sector workers would receive post-exposure prophylaxis when unprotected exposure occurs, estimated as 4 times during a 12 week community outbreak. Based on these estimates, a total of 102.8 million antiviral regimens would be needed. Additional work to define specific groups at higher and lower risk and their respective numbers is needed.
The health benefits of this prophylactic strategy cannot be easily quantified. Several studies suggest that healthcare workers who have patient exposure have increased rates of seasonal influenza infections.[42],[43] In addition to the direct effect of reducing pandemic influenza illness and its consequences, prophylaxis also would reduce the risk of transmission to family members, co-workers, and to patients. Influenza prevention by vaccination of healthcare workers has been shown to reduce nosocomial infection in acute care hospitals[44] and mortality in long-term care facilities for the elderly.[45] An additional impact would be to reduce absenteeism among workers in these critical sectors, improving the quality of healthcare and public safety. Many studies have shown improved health outcomes with a greater staff-to-patient ratio. During an influenza pandemic when healthcare burden is markedly increased, this effect may be even greater.
Maintaining community services during an influenza pandemic is essential to achieving national pandemic response goals of mitigating adverse health consequences and reducing societal and economic disruption. In addition to healthcare and emergency services, critical infrastructure sectors defined by the Department of Homeland Security include utilities (electricity, natural gas, water), communications and information technology, transportation, food and agriculture, banking and finance, pharmaceutical, chemical, oil, and postal and shipping. Overall, about 70 million persons are employed in these sectors excluding healthcare and emergency services. A study done by the National Infrastructure Advisory Council (NIAC) developed a preliminary estimate of between 4 million and 8 million persons in these sectors as critical to maintenance of essential functions during a pandemic. By contrast with healthcare and emergency services, these critical workers, in general, will not have high-risk occupational exposure to pandemic infection, as defined by the Occupational Safety and Health Administration’s risk pyramid.[46]
Various measures are available to protect workers and maintain essential functions. Workplace exposures can be reduced by changing practices to decrease close contact between workers including having conference calls instead of meetings, and promoting teleworking and flexible scheduling. Educating workers not to report to work if ill with influenza-like symptoms and allowing leave when a household member is sick also will reduce workplace exposures. Providing education and materials to promote hygiene and the use of facemasks and other personal protective equipment, where appropriate under OSHA recommendations, may reduce the risk exposure to influenza if workplace contacts occur. Guidance for pandemic planning by businesses has been provided.[47]
Antiviral post-exposure prophylaxis may provide an incremental benefit in preventing disease and could reduce absenteeism. The magnitude of this increment depends largely on what other antiviral and non-pharmaceutical strategies are implemented. Assuming that workers who are ill or exposed to an ill person in their household would receive treatment or PEP under the strategies described above, fewer workers are likely to be exposed in the workplace. In a setting of effective planning and implementation of non-pharmaceutical workplace interventions to reduce close contacts and the risk of infection if contacts occur, workers and essential functions can be protected, and the additional benefit of antiviral prophylaxis following a workplace exposure likely would be small.
The number of antiviral drug regimens needed to support prophylaxis for critical infrastructure workers outside of the healthcare and emergency services sectors would depend on the prophylaxis strategy and the target population. Two potential target populations may be considered:
If outbreak prophylaxis was provided to all workers considered by NIAC to be critical, 32 – 64 million antiviral drug regimens would be needed. While fewer regimens would be required for a PEP strategy, it may not be feasible or acceptable to the employer or workforce to target only the subset of workers NIAC designates as “critical”: in the setting of a workplace exposure, providing prophylaxis to some exposed workers but not others may not be possible. Therefore, a broad PEP strategy for critical infrastructure, assuming an average of 1 exposure per worker during the course of a community outbreak, would require 70 million regimens based on targeting all workers in these sectors.
During seasonal influenza outbreaks, clusters of influenza infections often occur among elderly persons living in long-term care facilities (LTCFs). Factors promoting transmission of infection in LTCFs include frequent close contact between staff and residents and among residents, high patient density, and reduced immune response to vaccination among LTCF residents. Antiviral post-exposure prophylaxis has been shown to be effective in controlling influenza outbreaks in LTCFs and represents a standard public health practice.
Another setting where explosive influenza outbreaks may occur is aboard ships. Cruise ship outbreaks frequently are reported; risk factors include crowding and low vaccination rates among passengers and crews. Many elderly persons are passengers on these ships which increases the number of influenza complications that may be seen. Outbreaks also have occurred on Navy vessels although vaccination of crew members has substantially decreased this risk with seasonal influenza. Jails and prisons represent another high-risk setting due to crowding and a limited ability to apply other measures to reduce transmission of infection.
If one exposure per person is assumed for outbreaks in closed and high-risk settings, about 5 million antiviral drug regimens would be needed for PEP of persons in LTCFs and jails/prisons (estimated populations of 3 million and 2 million persons, respectively[48]). The number of regimens that may be needed to respond to ship-board outbreaks is uncertain. Recommendations against non-essential travel are likely to be issued during a pandemic and a substantial reduction in the number of persons who choose to take a cruise during a pandemic is likely.
Persons whose immunity is suppressed by disease (e.g., resulting from HIV, leukemia, congenital immunodeficiencies) or due to medical therapies (e.g., receipt of a hematopoietic stem cell transplant [HSCT] or cancer chemotherapy) are at high-risk for influenza complications and mortality and are less likely to be protected by influenza vaccination. Antiviral drug prophylaxis has been shown to be effective in reducing risk of infection and its consequences in immunocompromised groups. Although PEP following household exposure for this high-risk group would be provided under a broad household PEP strategy, PEP also may be given for community exposures. Alternately, protection may be provided through prophylaxis for the duration of a community pandemic influenza outbreak.
Among immunocompromised persons, the largest population is those who have invasive cancer and are receiving chemotherapy. Assuming that the number of immunosuppressed cancer patients is equal to the annual incidence, 1.4 million persons would be included.[49] Estimating a total of 1.7 million persons who are severely immunocompromised and an average of non-household exposure per person, 1.7 million antiviral regimens would be required to support PEP. If outbreak prophylaxis were provided for 12 weeks, the total number regimens needed would be 13.6 million.
Policies at U.S. borders to reduce entry of persons with pandemic influenza infection may delay the occurrence of the pandemic wave. Risk-based policies are based on assessing whether incoming travelers may be infected based on their exposure history and whether they have symptoms of influenza illness. Implementing quarantine has been considered for persons exposed during travel to someone who is potentially infected, for example passengers on the same airplane as a possible case. Because facilities for quarantine are limited and this strategy could impose a substantial burden on border enforcement and public health personnel, an alternative may be to provide antiviral prophylaxis and to instruct recipients to report to their local public health authorities if symptoms develop. Because rates of travel may be reduced substantially during a pandemic and border policies to reduce entry of pandemic infection would be implemented only for a short period before the U.S. pandemic wave began, the number of antiviral regimens needed to support this strategy would be limited.
Post-exposure prophylaxis could be provided broadly in workplaces in addition to healthcare, emergency service, and critical infrastructure sectors. The approach of providing PEP in households, workplaces, and schools has sometimes been referred to as “targeted antiviral prophylaxis” (TAP). A mathematical model of community mitigation strategies with TAP suggests that the combination of interventions may reduce the rate of illness to less than 1% of the U.S. population.[50] If this outcome occurred, the number of antiviral regimens needed and the logistical challenges with implementation would be limited as PEP rarely would be needed. Despite the optimistic modeling results, however, there is concern about the assumptions on which these projections are based and the ability to achieve the assumed levels of compliance. If effectiveness was less, implementation challenges for TAP and the number of antiviral drug regimens required could be much greater.
Outbreak prophylaxis for household members of critical workers has been proposed as a strategy to decrease absenteeism and improve maintenance of healthcare and other essential services. Protecting workers’ family members would reduce work loss occasioned by caring for an ill family member and, by decreasing concern among workers that they may infect family members based on their occupational exposures. The psychological and behavioral effects of protecting family members on the willingness of critical workers to remain on the job cannot be estimated. Based on an average household size of 2.6 persons, outbreak prophylaxis of critical healthcare and emergency service worker households for 12 weeks would require about 116 million antiviral drug regimens.
[1] Policies to reduce the introduction of pandemic infection into the United States and the specific strategies for antiviral drug use in support of this objective are still being developed.
[2] See Guidance for Employers of Antiviral Drug Prophylaxis at [website TBD]
[3] See http://www.pandemicflu.gov/plan/community/commitigation.html
[4] Pre-pandemic vaccine is vaccine made and stockpiled against novel influenza A virus subtypes. Pre-pandemic vaccine, if available against the influenza A virus subtype that causes the pandemic, is likely to only be partially protective against a pandemic virus because the antigenic match between the vaccine and pandemic viruses likely will be imperfect.
[5] See http://www.pandemicflu.gov/vaccine/prioritization.html
[6] See Working Group report for Pandemic Influenza National Implementation Plan action 6.1.6.1.
[7] U.S. Department of Health and Human Services Pandemic Influenza Plan, Appendix D, www.hhs.gov/pandemicflu/plan/appendixd.html
[8] Aoki FY, Macleod MD, Paggiaro P, et al. Early administration of oral oseltamivir increases the benefits of influenza treatment. Journal of Antimicrobial Chemotherapy 2003;51:123-9.
[9] Kaiser L, Wat C, Mills T, Mahoney P, Ward P, Hayden F. Impact of oseltamivir treatment on influenza-related lower respiratory tract complications and hospitalizations. Archives of Internal Medicine 2003;163:1667-72.
[10] Nordstrom BL, Sung I, Suter P, Szneke P. Risk of pneumonia and other complications of influenza-like illness in patients treated with oseltamivir. Current Medical Research & Opinion 2005;2:761-8.
[11] Bowles SK, Lee W, Simor AE, et al. Use of oseltamivir during influenza outbreaks in Ontario nursing homes, 1999-2000. J Am Geriatrics Soc. 2002;50:608-16
[12] McGeer A, Green KA, Plevneshi A, Shigayeva A, Siddiqi N, Raboud J, Low DE, Toronto Invasive Bacterial Diseases Network. Antiviral therapy and outcomes of influenza requiring hospitalization in Ontario, Canada. Clin Infect Dis 2007;45:1568-75.
[13] Lee N, Chan PK, Choi KW, et al. Factors associated with early hospital discharge of adult influenza patients. Antivir Ther 2007;12:501-08.
[14] Note that statements about antiviral drugs in this document and guidance for their use may not represent FDA-approved uses or FDA policy.
[15] Participating Federal agencies included HHS (with representation from CDC, NIH, FDA, IHS, and the National Vaccine Program Office), Department of Homeland Security, Department of Veterans Affairs, Department of Defense, Department of State, and Department of the Treasury
[16]See http://www.pandemicflu.gov/plan/community/commitigation.html.
[17]Although an animal model study showed that a longer duration of treatment increased survival among mice infected with avian H5N1 virus, an expert panel convened by the World Health Organization (WHO) has not recommended changes in dose or duration when treating people with H5N1 infection.
[18] Current rapid diagnostic tests for seasonal influenza are estimated as 70% sensitive and >90% specific but their accuracy for pandemic influenza is not known. The assumed positive predictive value (PPV) of 35% for a clinical influenza diagnosis is extrapolated from seasonal influenza outbreaks. Higher PPVs have been obtained in clinical investigation settings and among adults in some studies. Although PPV may be higher during a pandemic, especially if the attack rate of influenza is higher, increased care seeking for non-influenza respiratory disease and by the “worried well” also could decrease the value.
[19]A regimen is defined as 10 drug doses: treatment, provided twice daily for 5 days, and post-exposure prophylaxis, provided once daily for 10 days both require a single regimen. By contrast, prophylaxis for the duration of a community outbreak may require up to 8 regimens for a 12-week outbreak.
[20] The package inserts for each antiviral drug should be consulted for the duration of prophylaxis for which efficacy and safety data are available. Any additional safety issues that might arise with increased duration of exposure cannot be predicted with confidence in the absence of longer-term data. If changes in drug regimens are considered that might warrant use of an Investigational New Drug application or an Emergency Use Authorization, FDA should be contacted as far in advance as possible to discuss regulatory needs and recommendations.
[21] See http://www.osha.gov/publications/influenza_pandemic.html
[22] See http://www.cdc.gov/od/science/phec/guidelinesPanFlu.htm
[23] Recommendations do not include antiviral drug use strategies or requirements to protect homeland and national security. The Department of Defense (DOD) currently maintains an antiviral drug stockpile and has developed strategies for antiviral drug use among essential personnel. Other Federal agencies also are considering their specific needs. The DOD general beneficiary population (e.g., dependents of active duty personnel) is included under the civilian guidance.
[24] See http://www.pandemicflu.gov/plan/pdf/cikrpandemicinfluenzaguide.pdf.
[25] See package labels, Oseltamivir: http://www.fda.gov/cder/foi/label/2002/21087S10lbl_update.pdf ; Zanamivir: http://www.fda.gov/cder/foi/label/2001/21036s4lbl.pdf
[26] Aoki FY, Macleod MD, Paggiaro P, et al. Early administration of oral oseltamivir increases the benefits of influenza treatment. Journal of Antimicrobial Chemotherapy 2003;51:123-9.
[27] Kaiser L, Wat C, Mills T, Mahoney P, Ward P, Hayden F. Impact of oseltamivir treatment on influenza-related lower respiratory tract complications and hospitalizations. Archives of Internal Medicine 2003;163:1667-72.
[28] Nordstrom BL, Sung I, Suter P, Szneke P. Risk of pneumonia and other complications of influenza-like illness in patients treated with oseltamivir. Current Medical Research & Opinion 2005;2:761-8.
[29] McGeer A, Green KA, Plevneshi A, Shigayeva A, Siddiqi N, Raboud J, Low DE, Toronto Invasive Bacterial Diseases Network. Antiviral therapy and outcomes of influenza requiring hospitalization in Ontario, Canada. Clin Infect Dis 2007;45:1568-75.
[30] Bowles SK, Lee W, Simor AE, et al. Use of oseltamivir during influenza outbreaks in Ontario nursing homes, 1999-2000. J Am Geriatrics Soc. 2002;50:608-16
[31] Beigel JH, Ferrar J, Han HM, et al. Avian influenza A (H5N1) infection in humans. N Engl J Med 2005;353:374-85.
[32] Writing committee of the second WHO consultation on clinical aspects of human infection with avian influenza A (H5N1) virus. Update on avian influenza A (H5N1) virus infection in humans. N Engl J Med 2008;358:261-73.
[33] Yen HL. Monto AS. Webster RG. Govorkova EA. Virulence may determine the necessary duration and dosage of oseltamivir treatment for highly pathogenic A/Vietnam/1203/04 influenza virus in mice. J Infect Dis 2005;192:665-72.
[34] Schunemann HJ, Hill SR, Kakad M, et al. WHO rapid advice guidelines for pharmacological management of sporadic human infection with avian influenza A (H5N1) virus. Lancet 2007;7:21-31.
[35] Welliver R, Monto AS, Carewicz O, et al. Effectiveness of oseltamivir in preventing influenza in household contacts: a randomized controlled trial. JAMA 2001;285:748-74.
[36] Monto AS, Pichichero ME, Blanckenberg SJ, et al. Zanamivir prophylaxis: an effective strategy for the prevention of influenza types A and B within households. J Infect Dis 2002;186:1582-8.
[37] Hayden FG, Belshe R, Villanueva C, et al. Management of influenza in households: a prospective, randomized comparison of oseltamivir treatment with or without postexposure prophylaxis. J Infect Dis 2004;189:440-49.
[38] Neil Ferguson, Imperial College, London, UK, unpublished data.
[39] Ira Longini, University of Washington, Seattle, WA, unpublished data.
[40] Steven Eubank, Virginia Tech, Blacksburg, VA, unpublished data
[41] Balicer RD, Omer SB, Varnett DJ, Everly GS Jr. Local public health workers’ perceptions toward responding to an influenza pandemic. BMC Public Health 2006;6:99. www.biomedcentral.com/1471-2458/6/99
[42] Elder AG, O’Donnell B, McCruden EAB, et al. Incidence and recall of influenza in a cohort of Glasgow healthcare workers during the 1993-4 epidemic: results of serum testing and questionnaire. BMJ 1996;313:1241-2.
[43]Kawana A, Teruya K, Kirikae T, et al. Syndromic surveillance within a hospital for the early detection of a nosocomial outbreak of acute respiratory illness. Jpn J Infect Dis 2006;59:377-9
[44]Salgado CD, Giannetta ET, Hayden FG, Farr BM. Preventing nosocomial influenza by improving the vaccine acceptance rate of clinicians. Inf Cont Hosp Eipdemiol 2004;25:923-8.
[45] Carman WF, Elder AG, Wallace LA, et al. Effects of influenza vaccination of healthcare workers on mortality of elderly people in long-term care: a randomized controlled trial. Lancet 2000;355:93-7.
[46] See http://www.osha.gov/Publications/OSHA3327pandemic.pdf
[47] See http://www.pandemicflu.gov/plan/pdf/cikrpandemicinfluenzaguide.pdf
[48]LTCF totals include nursing facilities (about 1.8 million beds) and other residential facilities for elderly persons. Prison and jail estimate from the U.S. Department of Justice (http://www.ojp.usdoj.gov/bjs/prisons.htm)
[49]See http://www.cdc.gov/cancer/npcr/npcrpdfs/US_Cancer_Statistics_2003_Incidence_and_Mortality.pdf
[50] Germann TC, et al. Mitigation strategies for pandemic influenza in the United States. Proceedings of the National Academy of Sciences 2006.