Antivirals for Pandemic Influenza: Guidance on Developing a Distribution and Dispensing Program (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4 Who Should Get Antivirals and Where? I n this chapter, the committee examines the two main components of plans for antiviral distribution and dispensing: decision making about the groups that will receive antiviral medications for treatment or pro- phylaxis (consistent with the ultimate goals of pandemic influenza plans) and about the locations where dispensing would take place. DIAGNOSIS AND TREATMENT The committee������������������������������������������������ ’s discussion about diagnosis and treatment was informed by the knowledge that the ����������������������������������� federal���������������������������� government and states soon may reach the target of 75����������������������������������������������� ���������������������������������������������� (of a total of 81) ��������������������������� million course������������� s, an amount intended to be used for treatment (see discussion in Chapter 2). However, the committee notes that if program goals for the current antiviral sup- ply include both�������������������������������������������������������� treatment and some level of p�������������������������� ����������������������������������������� rophylaxis���������������� ,��������������� it may become �������������� necessary to p������������������������������������������������������������� rioritiz����������������������������������������������������� e���������������������������������������������������� groups for treatment������������������������������� ��������������������������������������������������� �������������������������������������������� in addition to prophylaxis on the basis of the national ethical framework and the recommendations of the advisory body described in Chapter 3, as well as the characteristics of the pandemic�� .� A number of tests are available to diagnose seasonal influenza but few are sufficiently accurate (i.e., high sensitivity and specificity) for use in a pandemic in the context of antiviral scarcity and need for targeted use. Also, there are some tests available for use at the point-of-care, but they have significant limitations (WHO, 2005; CDC, 2006c). The optimal diagnostic tool would be an accurate test that could be 53 

54 ANTIVIRALS FOR PANDEMIC INFLUENZA used on large numbers of patients at the point-of-care, would provide rapid results (in 30 minutes or less), and would require minimal skill to administer and interpret (Griffin, 2007; Grijalva et al., 2007). Current diagnostic tools include the following: • Culture (of the virus) is traditionally considered the gold stan- dard, with 100 percent specificity, but a culture also requires 48 hours or longer, so it cannot be used to determine treatment with antivirals that require prompt use for maximum efficacy. • Real-time polymerase chain reaction (RT-PCR) is at least as, and perhaps more, sensitive than culture. It currently is expensive, of limited availability, and requires somewhat sophisticated labora- tory capabilities. • Direct immunofluorescence (or direct fluorescent antibody assay, DFA) provides results in approximately 1–6 hours and has 80–95 percent sensitivity. • Enzyme immunoassay (EIA) rapid test gives results in 30 min- utes, and has 50–90 percent sensitivity (median 70 percent) and 90–95 percent specificity. • Serology is mostly useful for epidemiologic studies and takes more than 10 days. Most tests available do not provide information about influenza A subtypes (CDC, 2006a; FDA, 2007b). Only PCR and culture currently can detect the H5N1 strain. Even fewer tools may be available if a different strain with pandemic potential emerges. The performance of laboratory tests varies by the type of sample (swab versus aspirate or wash; nasal versus nasopharyngeal versus throat), the age of the patient (generally more sensitive in children because of higher viral loads), and the time since onset of symptoms (CDC, 2006b; FDA, 2007b). In addition, the operating characteristics (e.g., the sensitivity, specificity, likelihood ratio) of any laboratory or clinical diagnostic tool depend on the prevalence of influenza in the population, which influences the predictive value. For example, the positive predictive value of rapid tests is fairly good (i.e., most positives are true positives) when disease is highly prevalent, helping rule in true influenza. However, negative pre- dictive value at the peak of a pandemic will be poor, so rapid tests would not be a good tool for denying treatment. The clinical diagnosis of influenza is difficult because the symptoms overlap with those of other respiratory viral infections. A number of studies have examined the ability of clinical symptoms to predict which patients with respiratory infections have influenza (Call et al., 2005). The results are influenced by study design, the age of the patients, the 

WHO SHOULD GET ANTIVIRALS AND WHERE? 55 presence of co-morbid states and the prevalence of influenza. One of the largest studies examined 3,744 participants in clinical trials of zanamivir, who were pre-selected as having a high pre-test probability of influenza (Monto et al., 2000). The combination of fever and cough had a sensitiv- ity of 64 percent and a specificity of 67 percent. In this setting where the prevalence of influenza was 66 percent, the positive predictive value was 79 percent and the negative predictive value was 49 percent. Clinical diagnosis in children may be even more difficult because the symptoms vary by age, and many respiratory viral infections mimic influenza. In one study of 128 children, the triad of cough and headache by history and clinical finding of pharyngitis had a sensitivity of 80 per- cent (95 percent CI [confidence interval], 69–91 percent), specificity of 78 percent (95 percent CI, 67–89 percent), positive predictive value of 77 percent (95 percent CI, 61–88 percent), and a negative predictive value of 81 percent (95 percent CI, 70–92 percent) (Friedman and Attia, 2004). In another study of children ages 5–12 years old participating in a zanamivir trial, cough and fever had a positive predictive value of 83 percent (Ohmit and Monto, 2006). Thus, clinical diagnosis in children and adults has a modest positive predictive value in high prevalence settings, but with low specificity and negative predictive value. In other settings, such as the beginning of a pandemic when disease is still rare, the positive predictive value may be considerably lower. In most settings, clinical diagnosis does not perform as well as rapid tests, which have modest sensitivity. In a pandemic, symptom frequency in a naïve population could potentially be different than during seasonal influenza, and it is possible that severe influenza could facilitate clinical diagnosis (unusually severe and distinctive symptoms would help to differentiate it from seasonal influenza and other common respiratory illnesses). However, public per- ception, public health concerns, patient pressure, and ethical concerns may lead to pressure to overdiagnose (M’Ikanatha et al., 2005). The implication of a modest positive predictive value of diagnostic tests is that a greater number of treatment courses will be needed. If the positive predictive value is 50 percent, half of those diagnosed will not have influenza and twice as many doses will be needed. Clearly, more accurate, rapid, and simple-to-use diagnostic tools would improve the efficiency of antiviral use during a pandemic. Cost would undoubtedly be an additional consideration, so inexpensive tests that meet all criteria described above would be ideal. Many questions arise when considering the processes to be used for diagnosis and treatment. What are the criteria for prescribing? A prescrib- ing algorithm? Should these systems be tested during seasonal influenza? Other outbreaks? What about concerns about attempts by individuals not 

56 ANTIVIRALS FOR PANDEMIC INFLUENZA in a priority group to obtain drugs, or attempts to obtain or sell drugs unlawfully outside established distribution systems? In regard to the final question, the committee believes that it will be impossible to keep people from gaming the system, and that it would be a better use of time and resources to ensure that the greatest proportion of people who need antivirals actually get them. Several options (or a combination) could be used to diagnose cases of influenza and, where applicable, link them with a dispensing site or dispensing mechanism (discussed later in this chapter). Decision mak- ing about the appropriate site for diagnosis will be influenced by several considerations, including concern about mixing infected and uninfected persons, point in time during the pandemic wave (for example, diagnosis could occur in the primary care setting early in the pandemic), ability and feasibility of dispensing at the same site, and implications of referring to another location. Two potential sites for diagnosis are described below. Later in this chapter, dispensing sites are discussed, some of which could function as sites for diagnosis. Remote Diagnosis: Telephone- or Website-Based Triage The National Health System in the United Kingdom is planning to use the National FluLine as its main mechanism for providing information, and for diagnosing, triaging, and referring individuals to dispensing loca- tions (Alcock, 2007; UK Department of Health, 2007). In the United States there are multiple locally or regionally based call centers, such as poison control centers, county extension offices, nurse advice lines, and tobacco cessation hotlines. Some jurisdictions, such as the state of Iowa, have con- sidered or are considering use of such call centers to support response to a pandemic. A study funded by the Agency for Healthcare Research and Quality (AHRQ) and conducted by Denver Health (2007) has explored the use of call centers to support the following activities in a crisis: provi- sion of health information, disease surveillance, triage and decision sup- port, quarantine and isolation support and monitoring, outpatient drug information and adverse event reporting, and mental health support and referral. All or most of these functions would apply to antiviral distribu- tion and dispensing in an influenza pandemic. Use of web-based support systems in a pandemic could be of great help, but would require careful planning including the development of an effective algorithm and clear, easy-to-understand language and format. On a more basic level, such sys- tems would depend on functioning technology (e.g., telecommunication hubs) and a stable electrical power supply, both of which could be affected in a severe pandemic due to worker illness and absenteeism. It has been 

WHO SHOULD GET ANTIVIRALS AND WHERE? 57 suggested that the federal government could make advance arrangements to use the telephone- and/or web-based customer service system of a large business with an online retail presence (B. Wolcott, personal com- munication, December 17, 2007). However, this would require rapid and effective training (if at all feasible) of phone personnel and contingency planning for pandemic-related attrition in regional or international call centers. Existing state- or local-level call centers may be a more realistic mechanism for diagnosis and referral, although they would experience human resource demands similar to those of dispensing sites (perhaps minus the risk of infection). An algorithm for diagnosis and triage would be needed for these systems. Clinician Examination Diagnosis by one’s own health care provider may not be possible during a pandemic, especially after the pandemic strain is known to have entered a community. Clinicians in the primary care setting could be overwhelmed (by both legitimate demand and perhaps by pressure from the worried well) during a severe or even a moderate outbreak, and there are concerns about infection control, storing and securing antivirals, linking with a data-gathering system, and other potential problems; thus it is highly unlikely that dispensing antivirals could occur in this setting (see discussion below). Furthermore, as noted earlier, clinical diagnosis would have somewhat limited predictive value, and in a severe pandemic with insufficient antiviral stockpiles, an accurate and rapid point-of-care test would be necessary (this would be most useful early and late in the pandemic). PROPHYLAXIS The committee believes current antiviral supplies will not be ade- quate for all of the uses considered. If expanding antiviral stockpiles is not feasible or desirable, a prioritization scheme developed by a transparent and inclusive national process will be needed for decision making (see recommendations in Chapters 2 and 3). The vaccine prioritization strate- gies (see Table 4-1) that have been developed by the federal government are the result of a process that considered different scenarios of pandemic severity, made explicit the values that guided decision making, and used input from the public in determining basic values. Recommendation 4-1: The committee recommends that in the pre- pandemic period, the Department of Health and Human Services undertake an effort similar to that for influenza vaccine priorities— 

58 ANTIVIRALS FOR PANDEMIC INFLUENZA TABLE 4-1 Groups to Receive Influenza Vaccine (from Draft Guidance on Allocating and Targeting Pandemic Influenza Vaccine, October 2007) Four categories of target groups: • Homeland and National Security • Health Care and Community Support Services • Critical Infrastructures (CI) • General Population Each category includes multiple target groups. The tiers cut across categories: Tier 1 Tier 2 Tier 3 Tier 4 Critical occupations Critical occupations Critical occupations High-risk - Deployed forces - Military support - Other active duty population - Critical health care - Border protection - Other health care - High-risk - EMS - National Guard - Other CI sectors adults - Fire - Intelligence services - Other government - Elderly - Police - Other national people - Government leaders security High-risk population - Community services - Healthy children High-risk population - Utilities - Pregnant women - Communications - Infants - Critical government - Toddlers High-risk population - Infant contacts - High-risk children Tier 5 includes the remaining population not included in the target groups listed above. national in scope, inclusive of diverse populations and viewpoints, and in keeping with a shared ethical framework—to discuss and develop a prioritization scheme for antiviral treatment and prophy- laxis that is capable of adjustments in real-time in response to the influenza pandemic. The national dialogue and public engagement activities will ideally include a discussion of the goals of an antiviral dispensing program, for example, maintaining the functioning of society or mitigating death and hospitalizations, and whether the program can address some or all poten- tial goals. As noted in Chapter 2, the amount of antivirals stockpiled will need to be commensurate with antiviral use goals. There are two main types of prophylaxis that could be undertaken as part of an antiviral dispensing program for pandemic influenza: post-   See Chapter 3. 

WHO SHOULD GET ANTIVIRALS AND WHERE? 59 exposure prophylaxis for those who come in close contact with infected individuals (either in an occupational setting or through household con- tact), and outbreak or seasonal prophylaxis that would be prolonged (i.e., for the duration of the first wave of the pandemic) and provided to groups with occupational exposure. There are at least three factors that warrant consideration in identi- fying priority groups. First, there seems to be widespread consensus at the national and state levels that ethical principles are needed to inform policy and program decisions on the allocation of antiviral drugs, particu- larly if supplies are very limited (see discussion in Chapter 3). The epide- miologic profile of the disease is a second factor in decision making, but it will emerge from the data gathered in the early weeks of the pandemic (e.g., age groups and types of health status, including pregnancy, linked with greatest likelihood of death or severe disease requiring hospitaliza- tion). A third factor in decision making is based on both practical and ethical issues: occupations in general and work duties in particular that (1) pose greater likelihood, intensity, and frequency of exposure; and (2) are associated with situations where personal protective equipment may not be effective, usable, or available. The most basic level of prioritization requires deciding whether a limited antiviral supply (beyond that needed for treatment) would be pro- vided to household contacts of infected individuals, or to certain groups of front-line workers, either as prolonged or post-exposure prophylaxis, depending on the amount of drugs available. Despite the theoretical and ethical arguments for post-exposure prophylaxis among household members who participate in voluntary quarantine, there is little evidence to support the feasibility or efficacy of this approach. Moreover, this approach requires a large stockpile (more than 88 million courses accord- ing to the estimate provided in the Department of Health and Human Services [DHHS] draft proposed guidance dated November 6, 2007) and may lead to unanticipated consequences. Recommendation 4-2: The committee recommends that pandemic influenza planners at all levels make outbreak prophylaxis for health care and emergency personnel who are in short supply and will have repeated and difficult-to-control exposure a first priority for prophylactic antiviral use. Post-exposure prophylaxis for other health care personnel and emergency responders should be a sec- ond priority. Post-exposure prophylaxis of household contacts of infected individuals should be a third priority if stockpiled antivi- rals are insufficient to meet all prophylaxis objectives. In the absence of pharmacologic means of protection, alternate meth- ods for containing disease spread in and beyond the household would be 

60 ANTIVIRALS FOR PANDEMIC INFLUENZA applied, including quarantine and isolation (voluntary or not, depending on circumstances) and advance instructions for implementing household infection control. ��������������������������������������������������������������������� F�������������������������������������������������������������������� or a 1918-type (i.e., severe) �������������������������������������� ����������������������������������������������������� pandemic, a��������������������������� ge, occupation, and health status would be������������������������������������������������������� ��������������������������������������������������������������� among the characteristics of note for prioritization. The differential age-based mortality risk seen in 1918�������������������� , as one of several potential epidemiologic features of a pandemic,�������������������������� is an ethically relevant consideration for ����������������������������������������������������������� prioritization��������������������������������������������� , especially if, as in 1918, healthy persons� who are���������������������������������������������������������������������� in the prime of life and key to infrastructure stability are at high risk of mortality due�������������������������������������������������� most likely�������������������������������������� to their robust immune response (���� ���������������������������� the cytokine storm phenomenon������������������������������������������ )����������������������������������������� . Further�������������������������������� more���������������������������� , people who put themselves in harm�������������������������������������������������������������������� ’������������������������������������������������������������������� s way for the sake of all in a pandemic should be given protection not only because of the utility of doing so (lessening mortality and strengthening������������������������������������������������������������������ ����������������������������������������������������������������� infrastructure��������������������������������������������������� ), but also the fairness (reciprocity) of it. Some groups may be at���������������������������������������������������������� a�������������������������������������������������������� ������������������������������������������������������� disproportionate��������������������������������������� ly high ������������������������������� risk��������������������������� of mortality, as pregnant� ��������� women were in the 1918 pandemic due to their pregnancy-suppressed immune systems����������������������������������������������������������� . I�������������������������������������������������������� f supplies are adequate and antivirals are deemed ������ ������������� safe, these groups ����������������������������������������������������������� would be a likely high������������������������������������� -������������������������������������ priority group���������������������� for occupational and household post-exposure prophylaxis����������������������������������� . Likelihood, frequency, and inten- ��������������������������������� sity of e������������������������������������������������������������������� xposure to the in�������������������������������������������������� ���������������������������������������������������� flu����������������������������������������������� enza������������������������������������������� virus are��������������������������������� ������������������������������������ other �������������������������� important����������������� consideration��� s�� . In a public health crisis when supplies are short, ���������������������� those who voluntar���� ������������ ily� assume risk on behalf of others������������������������������������������ ������������������������������������������������������������� both need and deserve first-priority sta- tus������������������������������������������������������������������������ . With an indeterminate supply of antivirals, primarily����������������� �������������������������� prioritized for treatment in the first wave, ������������������������������������������� offering prophylaxis from a limited supply to household contacts of ill individuals and to family members of those ��������������������������������������������������������������������� ����������������������������������������������� with occupational���������������������������������������������������� risk would likely undermine capacity for treatment and protection of �������������������������������������������������� health care workers and emergency response person- nel��������������������������������������������������������������������� . ������������������������������������������������������������������� Because������������������������������������������������������������ severe pandemic by definition involves both high mortality and infrastructure degradation, the group characteristics ����������������� to��������������� be considered for prioritization can certainly be named and considered in advance, and tentative ����������������������������������������������������������������� prioritizations�������������������������������������������������� per particular supply���������������������������� ���������������������������������� levels (as in scenarios A, B, and C) can be developed���������������������������������������������� on the basis of national dialogue and public engagement process���������������������������������������������������� , subject to adjustment during the actual pandemic.� Outbreak Prophylaxis The severity and characteristics�������������������������������� ����������������������������������������������� of the pandemic itself and the available supply of antivirals are key considerations in developing and adjusting a prioritization strategy. Because pandemic-specific ������������ information� is����������������������������������������������������������������������������� so crucial to selecting group prioritization characteristics, the committee   Outbreak prophylaxis refers to prophylaxis offered for the duration of an outbreak (i.e., the first pandemic wave), and is also referred to as seasonal prophylaxis, estimated by fed- eral planners to last approximately 12 weeks. 

WHO SHOULD GET ANTIVIRALS AND WHERE? 61 only������������������������������������������������������������������������� indicate���������������������������������������������������������������� s characteristics to be considered.����������������������������� �������������������������������������������������������������� As described above, factors to be considered in deciding what groups will receive outbreak prophy- laxis include epidemiologic information available early in the pandemic, practical and ethical considerations about occupations and duties, and availability of drugs. The following characteristics will need to be considered in determin- ing prioritization of occupational groups for outbreak prophylaxis: • Societal functio�n • I������������������������������������������������������������ rreplaceable societal function������������������������������ (this will be challenging to define) • Risk level for ������������������������������������� in����������������������������������� flu�������������������������������� enza���������������������������� -related mortality/morbidity • Risk level of exposure ���������������������������������������� (most current plans seem to assume that a large proportion of health care workers will experience certain exposure, without any effort to stratify them further, for example, by those who will be exposed and those who may be exposed) • Risk of transmi����� ssion • Age • Assumption��������������������������������������������� of risk for others and society (reciprocity) The committee affirms that notwithstanding limited supplies of antiviral stockpiles, some antiviral medication will need to be used for prophylaxis to protect essential health care workers who are providing care directly to severely ill patients in both the inpatient and outpatient settings. Antiviral use in this situation will best be done as an adjunct to other infection control measures in hospitals and clinics and within the broader community. There are groups and individuals in certain occupations, in settings, and with duties that will have the greatest risk of infection and the least ability to control exposure (e.g., through the use of infection control and personal protective equipment [PPE]). These groups include, but are not limited to emergency department and triage personnel; direct care nurses and nurses’ aides; emergency medical technicians and others who may conduct early diagnosis; respiratory therapists and other hospital per- sonnel assigned to care for patients with influenza-like illness; critical operating room personnel; and public health workers with laboratory, epidemiology, and antiviral distribution-related responsibilities. Depending on the pandemic’s severity, effect on the health care and   For a complete ethical framework including strategies for allocation of antivirals, see Vawter, D. E., J. E. Garrett, A. W. Prehn, D. A. DeBruin, C. A. Tauer, E. Parilla, J. Liaschenko, M. F. Marshall, and K. G. Gervais. For the good of us all: Ethically rationing health resources in Minnesota in a severe influenza pandemic. Minnesota Pandemic Project, Minnesota Department of Health, 2008. 

62 ANTIVIRALS FOR PANDEMIC INFLUENZA emergency services workforce, and availability of antivirals, pandemic planners could consider strategies to minimize the need for prolonged prophylaxis (and thus decrease the amount of antiviral regimens needed, as well as minimize safety concerns about antiviral use longer than 6 weeks) by limiting health worker exposure to infected individuals. This could perhaps be done by organizing front-line workers into subsets or cohorts and deploying them in turn to care for cases of influenza. The use of personal protective equipment would constitute an additional strategy, assuming some level of availability at least during part of the pandemic. Although the committee was not charged with examining dimen- sions of pandemic influenza planning outside antiviral distribution and dispensing, the committee notes that given the likely use of antiviral pro- phylaxis in conjunction with other strategies to reduce the risk of health care personnel and emergency responders, greater clarity is needed on official recommendations for the use of masks and respirators in health care settings, workplaces, and homes. To the extent that the lack of clarity relates to inadequate scientific evidence, it will be important to conduct studies of the efficacy of infection control methods that could be used with antiviral prophylaxis and in order to decrease the need for prophylaxis in groups with occupational exposure. Further, the committee notes that the cost and opportunity costs associated with personal protective equip- ment will be a consideration. Also, the recent IOM report Preparing for an Influenza Pandemic: Personal Protective Equipment for Healthcare Workers described maintenance and reusability as two priority areas for research given concerns that personal protective equipment supplies at the state, local, and hospital level are limited and will be depleted rapidly in an influenza pandemic. The need for the strategies described above may vary depending on local circumstances. A small emergency department in a rural area, for example, may require continuous prophylaxis for all staff members dur- ing the entire first phase of a pandemic; that may or may not be the case for a large, well-staffed, urban emergency department. Earlier in this chapter, the committee noted the limitations of existing rapid diagnostic tests and the great need for accurate, rapid, easy-to- use point-of-care tests to ensure the most judicious use of antivirals in a scenario of scarcity. An additional argument for the development and deployment of accurate diagnostics is the ability to adequately identify health care personnel and other “front-line” workers who develop influ-   “Careful consideration should be given to the trade-offs between disposable and reusable PPE [personal protective equipment], particularly given the extreme demands that would be placed on a disposable PPE supply in an influenza pandemic. Maintenance and reuse are key factors for consideration in developing performance requirements” (IOM, 2008). 

WHO SHOULD GET ANTIVIRALS AND WHERE? 63 enza and recover. They will be able to work safely without prophylaxis or personal protective equipment. Identifying priority groups among emergency responders may be made more complex by the fact that other personnel and trained volun- teers may play emergency response roles in a pandemic. Thus, groups other than the obvious emergency responders (law enforcement, emer- gency medical services, and fire department personnel) may need protec- tion to do work that exposes them to unusually high risk of transmission. It will be necessary to identify the likely levels of exposure among all tra- ditional emergency response workers, in addition to including other types of personnel or volunteers. Further, to target prophylaxis more accurately, it may be important to make an advance distinction between those who will undoubtedly be exposed and those who may be exposed. The committee asserts that final determination of priority groups for outbreak prophylaxis cannot be made before the pandemic because the epidemiology of the disease is not known (e.g., who will be most likely to get sick, and who will be most likely to die). To use antivirals spar- ingly and strategically based on available epidemiologic data and local circumstances, Recommendation 4-3: The committee recommends that efforts be made to minimize the need for outbreak prophylaxis among health care and emergency responders, and efficiently allocate scarce health resources. Necessary measures include proper and consistent use of personal protective equipment and grouping of workers in subsets to stagger their exposure to infected patients, thus reducing the numbers who need prophylaxis at any given time and shortening the duration of needed prophylaxis. From a communication and public relations standpoint, pre- identifying a given group or groups as priority targets for prophylaxis could have unintended consequences, if it later became necessary to change their status as more information (e.g., epidemiologic pattern) becomes available or circumstances change. Labor-related concerns will also arise because protecting all personnel in the same occupational cat- egory may be impossible. This area undoubtedly will require advance dialogue with relevant labor organizations and other stakeholders. Not identifying priority groups in advance could be disruptive if likely groups have no advance warning. Also, although a great deal of attention is given to the potential benefits of antivirals in decreasing severe morbidity and mortality, antivirals also pose risks, and their risk–benefit profile may change if emerging data in a pandemic indicates that the pandemic strain is less susceptible to antivirals than anticipated in planning. Although the exact groups that will be prioritized to receive antivirals 

64 ANTIVIRALS FOR PANDEMIC INFLUENZA may not be known until a pandemic has begun, it may be helpful to iden- tify in advance the sites best suited to dispense antiviral drugs to various potential occupational groups, which may include workplaces. Also, it may be helpful to plan basic communication tools, including a focus on supporting prophylaxis adherence. Federal planners anticipate a 12-week duration of prophylaxis for certain health care and emergency services workers, on the assumption that non-pharmaceutical interventions will lengthen the duration of the first pandemic wave. When antiviral use guidelines are determined, it would be helpful to have national guidance (provided by the Centers for Disease Control and Prevention [CDC]) to ensure a level of consistency across state and territorial borders. From past experience (the 1998 meningitis outbreak, the 2004–2005 influenza vaccine shortage), it is known that differences among jurisdictions may cause confusion, distrust of the guidance pro- vided, movement of citizens to areas where they think they can improve their chances of getting the drug or vaccine, and complaints to the public health authorities (GAO, 2000; Hannan, 2008) (see Appendix A). Thus, consistency on some aspects of the program across the United States (and possibly with neighboring nations), will be needed to preserve the credibility of the guidelines and of public health officials, to underscore the scientific, ethical, and practical basis for decision making, and to help encourage the greatest possible public compliance with the guide- lines. However, the committee recognizes that in some circumstances, the guidelines may need to be clarified or modified somewhat by state and local authorities to address local realities. For example, there could be variation in the manifestations of the pandemic and variations in the populations served, which might potentially lead to minor differences in public health response. The committee believes that provision of prophylaxis to health care personnel and other relevant workers during a pandemic would rely on their health care facilities’ plans and existing information systems, distri- bution mechanisms, and procedures. However, it is important to note that some emergency response workers may not have occupational “bases” that lend themselves equally well to providing a site for dispensing medi- cations, or the education and record keeping activities needed for the duration of the pandemic or until a well-matched vaccine is available. A further factor in decision making relates to the assumption��������� of risk for others and society (reciprocity)����������������������������������������� . It has been suggested that health care workers’ absenteeism could be motivated by concern about their families, both about infecting members of their household, and about needing to care for stricken family members (Wynia and Gostin, 2004). Some present- ers at the committee’s information-gathering meetings also suggested that some health care workers might take the prophylaxis intended to 

WHO SHOULD GET ANTIVIRALS AND WHERE? 65 them and give it to family members, thus leaving themselves vulner- able to infection. For these reasons, the possibility of providing some type of prophylaxis (or perhaps priority treatment in the event of very scarce antiviral supplies?) to family members of health care personnel has been raised. There are ethical issues of fair treatment for families of those who assume a risk for the public good and in doing so may increase their families’ risk of contracting influenza (as suggested by experience during the Severe Acute Respiratory Syndrome outbreak of 2003) (Tufts, 2003; Wynia and Gostin, 2004). However, one public engagement activity revealed that although the public agrees with prioritization of workers who risk their own health and life to care for those who are stricken, the view on protecting the families of such workers with prophylaxis is less favorable (The Keystone Center, 2007). This clearly is one of many issues that warrant a broad public dialogue about antiviral use in a pandemic. There may be other means of preventing exposure of family members of health care workers, including meticulous and comprehensive infection control measures, but in-depth examination of these issues is beyond the scope of this report. Post-Exposure Prophylaxis for Household Contacts The ability to provide post-exposure prophylaxis for household con- tacts of cases is dependent on decisions made about the size of stockpiles (see scenario C, described in Chapter 2). The committee has recommended that this prophylaxis goal should be a lower priority after prophylaxis of certain types of health care workers and emergency responders. More- over, providing this type of prophylaxis will likely be more difficult than offering prophylaxis to the occupational-exposure groups who are easier to gather, communicate to and with, dispense to, and monitor for drug adverse events and adherence. The committee does not propose a prioritization scheme for house- hold contacts for the same reasons outlined above. However, the commit- tee lists some of the key considerations below: • What groups experience the highest attack rates? • What groups are most likely to transmit the disease (i.e., schoolchildren)? • Who is getting sickest (requiring hospitalization)? • Who is most likely to die? It is possible that the same group(s) will be the answer to all these epi- demiologic questions, or different groups may fit each of these dimensions of the epidemiologic profile that are expected to become evident early in 

66 ANTIVIRALS FOR PANDEMIC INFLUENZA a pandemic. Additional consideration may be needed for d������������� isadvantaged or special population groups w������������������������������������������� ho experience b���������������������������� arriers��������������������� in accessing health care services in general, and to groups in�������������������������������� stitutionalized with congregate living�������������������������������������������������������������������������� (e.g., long-term care facilities, correctional facilities). National pan- demic influenza strategy in certain circumstances could involve asking members of households to care for ill individuals in the home in the event that health care inpatient facilities have exceeded all surge capacity. In that case, consideration will be needed for providing post-exposure pro- phylaxis to home-based caregivers who are being asked to remain in the home caring for a family member. In addition to ethical and other considerations that will be needed in prioritizing groups to receive post-exposure prophylaxis, there will be logistical challenges. It will be necessary to enumerate household members of infected cases, to dispense drugs to them perhaps without a clinician actually examining them, and to provide a diagnosis if it appears they are developing symptoms (depending on the timing of the presenta- tion of the infected household member). Legal Considerations During an influenza pandemic, declarations of emergency, disaster, or public health emergency at federal, state, or local levels alter the existing legal environment to allow government public health and safety officials and others sufficient flexibility and powers to respond. Though essen- tial to emergency responses, such flexibility may also allow decisions or actions in real-time that may not be consistent with prior planning for the distribution of antivirals (Hodge, 2006b). In essence, government offi- cials may be empowered during declared states of emergency to deviate from pre-emergency plans for the distribution of antivirals. For example, they may require household prophylaxis for specific persons, even if supplies are scarce. Public health or emergency management authorities may seek shifts in existing prioritization plans for many reasons, but ideally decisions should be grounded in what are viewed as the most effective strategies to protect the public’s health. Household prophylaxis, for example, may be viewed as critical to garner the efforts of essential health care personnel or others to treat patients with influenza, or others suffering from other life-threatening conditions. Regardless of the justi- fication, deviations from existing distribution strategies are predictable during emergencies as the extent of the impact of pandemic influenza is measured within populations and available supplies are assessed. As a result, although it may not be possible to specify legal responses to emer- gency circumstances arising from pandemic influenza, advance planning and real-time assessments of public health needs should heavily influence communal actions. 

WHO SHOULD GET ANTIVIRALS AND WHERE? 67 Enumerating and Dispensing to Household Contacts Diagnosis of infected individuals will take place in person (depend- ing on severity of symptoms, either in the primary care setting or in the emergency department) very early before the pandemic strain is known to have entered a community. The first cases in the United States and in a given community will be confirmed by current (or better) diagnostic tests. Thus, the availability of rapid, accurate point-of-care testing will be very important during that early phase. In the middle to latter phases of the outbreak, diagnosis (or rather identification of cases) is more likely to be based on syndromic diagnosis. Thus, it may be more feasible to perform diagnosis and prescribing for many patients over the telephone or via a web-based interface in order to help prevent transmission and to lessen the likely effect on the health care delivery system. Thus a triage function would need to be built into the telephone or web algorithm so those who are acutely ill are chan- neled in a different direction (hospital, clinic, etc.). If sufficient antiviral courses are available to provide prophylaxis to household contacts, once the pandemic strain is in the community, this would trigger the points-of- dispensing (PODs) or other dispensing sites to begin functioning. In the context of diagnosis for initiating empiric treatment, voluntary isolation of cases and household prophylaxis, there are tradeoffs to remote diagnosis (e.g., via a telephone or web-based algorithm, see discussion earlier in this chapter). There are tradeoffs between the risk of having transmission at the site of outpatient care and the potential increased accuracy of diagnosis. In the absence of abundant and accurate point-of-care diagnostics where influenza would be diagnosed based on simple clinical rules, there may be advantages to making the presumptive diagnosis remotely in persons not requiring acute care. However, if the goal is to quarantine the household and initiate prophylaxis, there is a high cost that must be acknowledged: due to the modest predictive value of clinical diagnosis, up to three times as many people will be quarantined and up to three times more drug will be used than is necessary. Thus, algorithm- or telephone-based diagnosis is very costly if voluntary quarantine and post-exposure prophylaxis of household contacts are planned. Strategies for post-exposure prophylaxis of household contacts will need to be evaluated and reevaluated in the course of program implemen- tation based on the following interrelated factors:   Point-of-dispensing and point-of-distribution are sometimes used interchangeably in literature about state and local planning. 

68 ANTIVIRALS FOR PANDEMIC INFLUENZA • The availability, accessibility, and cost of a rapid, accurate, point- of-care test to establish that a person with influenza-like illness has the pandemic strain and how readily it can realistically be deployed • Availability of antivirals, since much of the decision making about household contact prophylaxis may occur on the basis of remote diagnosis versus in-person diagnosis • Weighing the risk of exposing people to an infected person in the course of in-person diagnosis against the risk of misusing antivirals by dispensing them for treatment and prophylaxis of someone without confirmed pandemic influenza infection or exposure There are a number of tasks related to household post-exposure pro- phylaxis that have to be resolved after diagnosis. These include enumera- tion; reasonable identification and screening of household members to receive prophylaxis; identification of the appropriate dispensing site for that household, whether pharmacy or POD or other site (or a delivery method); issuing a valid set of prescriptions; and recording the process. The details might vary in each jurisdiction based on the dispensing site, but the tasks will remain constant. For example, if a person is identified as being a candidate for treatment, and the strategy at the time is to offer prophylaxis to household contacts, the clinician on the telephone or inter- acting on the Web could have a household contact form that requests the number of household contacts, names, ages (noting children, the elderly), and known contraindications (including pregnancy), if any. Depending on the dispensing sites used in a given jurisdiction, there may be plans to have the clinician call in (or communicate in some way) the prescription for treatment of the infected individual and for post-exposure prophylaxis for household contacts to the POD, pharmacy, or other dispensing site that will be accessible to the respective household. This will require care- ful planning and decision making by each jurisdiction to ensure optimal placement of dispensing sites to ensure the most rapid and equitable pro- cess for giving antiviral courses to those who meet dispensing criteria. DISPENSING SITES Allocation and Distribution to Dispensing Sites The importance of timely and efficient distribution of potentially life- saving drugs in responding to a pandemic cannot be overstated. All or most states have planned and tested their mechanism for taking the mate- riel obtained from their portion of the Strategic National Stockpile (SNS) 

WHO SHOULD GET ANTIVIRALS AND WHERE? 69 and placing it at predetermined locations around the state. If distribution mechanisms other than the SNS mechanism are considered to disperse some or all of a jurisdiction’s antiviral supply, the respective public health agencies and their partners would benefit from considering private sector and Department of Defense expertise in supply chain science: efficiently moving product or material from point A to point B. After the declara- tion of a pandemic, drugs need to be moved to dispensing sites smoothly and quickly to address treatment needs and, if appropriate, prophylaxis needs. To facilitate prompt dispensing, the CDC supplemental guidance for pandemic influenza asked grantees to identify sites where antivirals may be pre-positioned when a pandemic is judged to be imminent (DHHS and CDC, 2006). A recent survey from the Association of State and Territorial Health Officials found that “21 state public health agencies plan to or have already prepositioned antivirals at hospitals, local health agencies and other locations around the state, and 26 do not plan to do so” (ASTHO, 2008). Selecting Sites of Dispensing Assuming adequate antiviral stockpiles are available to undertake post-exposure prophylaxis of contacts of infected individuals, the most challenging aspect of any dispensing program would be determining how to quickly and efficiently dispense antivirals on a large scale in circumstances where the pandemic influenza virus has entered a com- munity and disease prevalence is increasing rapidly. There may be good reasons to consider using existing SNS distribution mechanisms (even for sources of medication other than the SNS, e.g., state stockpile). The vast majority of state and local public health agencies in the United States have developed considerable experience with PODs, which are the final destination of SNS supplies and are intended for the distribution of medi- cations or vaccines rapidly and efficiently to large numbers of people. Despite potential drawbacks that may include the strict nature of the dis- tribution channels and potentially expired supplies within the SNS, most jurisdictions are familiar with this distribution system, have invested resources, including funds and staff time for exercises, and have identified points-of-dispensing. Once the SNS or state antiviral supplies have been delivered to the local level, there are a number of options for dispensing sites. The com- mittee has learned about a wide array of locations (or mobile dispensing,   some states, the plan is that the state-purchased supply of antivirals will be sent out to In the local agencies, freeing up the state facilities to receive the state’s portion of the SNS. 

70 ANTIVIRALS FOR PANDEMIC INFLUENZA in the case of visiting nurses, etc.) that could be used. Sites that could be considered, and many of which are already considered by jurisdictions around the country, include the following: • Public health agency POD • Hospitals • Physicians’ offices or other primary care settings • Pharmacies • Businesses • Schools • Drive-through pharmacies or other stores • Parking lots • Parish nurses or visiting nurses • Homes, through mail or parcel delivery service, either pre- distributed or just-in-time • Other private-sector delivery mechanisms employing a mail or parcel delivery service (e.g., Netflix-like system) • Health maintenance organization systems • Nursing homes or other long-term care facilities Other dispensing sites could be used depending on local circum- stances and needs. No one type of site will be suitable in all cases, and it is likely that if prophylaxis on a medium-to-large scale is the objective, many jurisdictions may need to plan for a combination of dispensing sites to reach all groups identified to receive antivirals. Site selection will ideally build on what has been learned about PODs and consider demographics and population density. Dispensing sites also may be characterized by a range of advantages and disadvantages, strengths and weaknesses. These include, but are not limited to the following: • Integration of dispensing site features with other interventions in a jurisdiction’s pandemic plan, such as social distancing (the SNS guidance makes this a condition for selection of points of dispensing) • Capacity for infection control in general, including size of facil- ity and efficiency of processing, availability of hand sanitizer and posted educational material on hygienic practices, and focus on other avenues for transmission such as sharing pens or clip- boards, if applicable • Accessibility and convenience • Capacity and/or throughput (ability to process large numbers of people quickly and efficiently) 

WHO SHOULD GET ANTIVIRALS AND WHERE? 71 • Suitability for facilitating early treatment (within 48 hours of dis- ease onset) • Ability to implement crowd control and ensure security of drug supply (required by the SNS guidance) • Ability to collect and report important data, including tracking who gets the drugs and a contingency pen-and-paper option in the event of technology or power failure, to (other) public health agencies • Appropriateness for diagnosis if (1) decision is to use some/all sites for treatment in addition to prophylaxis, and if (2) exposed contacts provide information that suggests they may be infected, should there be a need to provide diagnosis at prophylaxis-only sites • Implications for communication and education (e.g., drive- through sites offer little opportunity to educate, but how much education can be given if the goal is to process people efficiently and minimize contact? Is it useful to couple dispensing with an education component? Or should education take place through the mass media and new media [e.g., text messaging] instead?) • Issues related to using private distributors to deliver antiviral medications to private homes A major consideration in selecting dispensing sites is suitability for vulnerable or special-needs populations (e.g., ease of access, linguistic and cultural competence, geographic proximity so transportation does not become a barrier, perceived safety of the site, hours of operation). The choice of well-trusted community sites using personnel who have the greatest potential to deliver public health information and antiviral medications in a linguistically and culturally competent manner will be important. Examples of these sites may include community clinics, houses of worship, neighborhood (senior) centers, and other trusted community organizations. If security and other logistical concerns can be addressed, antiviral dispersal at these nontraditional sites will help to ensure that the service delivery is linguistically competent and that services are delivered by trusted agencies within communities across the nation. The provision of linguistically and culturally competent information and distribution services can help to facilitate the elimination of disparities in knowledge and avoidance of disparities in pandemic outcomes. Below, the committee discusses the features of some of the dispensing site options, including public health agency points-of-dispensing, hospi- tals, homes, pharmacies, and physicians’ offices. Public health agency PODs have been exercised in most jurisdic- tions and are well-established mechanism for mass dispensing. A poten- 

72 ANTIVIRALS FOR PANDEMIC INFLUENZA tial disadvantage is that they may not be sufficiently dispersed in some jurisdictions and therefore not accessible to all residents in a community. PODs also may include a one-size-fits-all approach that may not be opti- mal for hard to reach or vulnerable populations. PODs could, however, be expanded to include a wider range of setting and greater breadth of geographic distribution. Several jurisdictions have explored the use of a head-of-household model to facilitate rapid and efficient dispensing of antibiotics. For example, the Seattle and King County health departments collaborated with the University of Washington to exercise their pub- lic health POD and distribute mock medication to heads of households (Stergachis et al., 2007). In this exercise, which lasted 68 minutes, more than 600 10-day courses of “antibiotics” were dispensed to 254 heads of household in a process that included a triage element (including by age and need for pharmacist consultation). Philadelphia employed a similar process to test antibiotic dispensing and found that express dispensing to heads of household was a viable option (Agócs et al., 2007). In this exercise, 717 heads of household picked up drugs for a total of 2,120 household members in a period of 2 hours. The patients were directed to one of two lines, one of which was intended for those requiring additional screening (pediatric dosage and other issues). One noteworthy feature of this activity was that 42 of 50 point-of-dispensing staff were trained in the hour before the exercise began. The head-of-household dispensing model could perhaps be used to dispense antivirals for households with a known case of influenza. Something similar could occur in cases where individuals diagnosed with pandemic influenza are well enough to give information about their household to the health care provider who would be treating them and to return home with courses dispensed for all house- hold members. The major advantage of the head-of-household dispensing model is that it involves only one point of contact for each household with an infected member, but that may also be a potential disadvantage if some households do not have a single member who is well enough to go and retrieve medication for the rest of the household. Although hospitals are a typical site of diagnosis and dispensing, they are less than ideal for timely dispensing of antivirals (within the 48-hour window from symptom onset), because people may go to the hospital only when their condition is worsening. Also, hospitals may rapidly become overwhelmed in a pandemic, making them undesirable for further activities that would place strain on available human and other resources. In a pandemic, many people may try to avoid the hos- pital believing it to be a locus of infection. Hospitals may have security problems—including securing the antiviral supply, other hospital assets, and patient information—if they were used as community dispensing sites with large numbers of non-patients streaming through. 

WHO SHOULD GET ANTIVIRALS AND WHERE? 73 Pharmacies have well-established channels for acquiring, dispensing, tracking inventory, and screening for contraindications and proper dos- age. However, most pharmacies sell a wide range of products that will continue to be in demand in a pandemic, making pharmacy facilities sites where the ill and the well would come in contact. Also, there may be good reasons to offer diagnosis and dispensing at the same site, and it is unclear whether pharmacies are appropriate sites for diagnosis. The primary care setting has at least in principle the advantages of familiarity, ability to screen for contraindications and prescribe the proper dose for each individual patient, ability and duty to follow patients and refer in the event of adverse events, and so on. However, in a moderate- to-severe pandemic, primary care providers may be rapidly overwhelmed with demand. Also, providers do not typically dispense drugs, and locat- ing antivirals in a primary care setting may present a logistic and security challenge. If tracking antiviral use is an objective, using primary care sites to dispense would greatly multiply the sites of dispensing, with implica- tions for information technology resources. Finally, having individuals with influenza-like illness travel and congregate with others in a mul- titude of primary care waiting rooms may not support infection control objectives that may call for immediate self-isolation of sick persons and the avoidance of public settings. If the home were the site or focus of dispensing, there are several models that have been explored by the Department of Health and Human Services (DHHS) and its state and local partners. One model is found in the MedKit project, which involved pre-positioning mock drugs with households and instructing household members to wait until further notice to use the contents of their kit. A second model has been tested through the Cities Readiness Initiative distribution of identical just-in- time packages containing bottles of “antibiotics” to homes in two ZIP Codes each in Boston, Philadelphia, and Seattle (Koh et al., 2008). This activity used U.S. Postal Service workers paired with local police person- nel to deliver packages to homes. A third model that has been suggested is to encourage households to purchase antivirals in advance personal stockpiles. In the past, limited manufacturing capacity meant that per- sonal stockpiling threatened the availability of antivirals for seasonal use and creating central stockpiles. This is no longer an issue. Personal stockpiling, however, is likely to be inconsistent and heterogeneous. It is unlikely to be an effective resource for public health use. Pre-positioning of antiviral kits in homes or delivering them “just in time” to households would be a considerably more difficult endeavor than delivery of antibiotics, because given the gradual spread of the pan- demic, the timing of the initiation of treatment or prophylaxis is complex. Common misconceptions about what constitutes “flu,” anxiety, rumors, 

74 ANTIVIRALS FOR PANDEMIC INFLUENZA and the difficulty of accurately identifying influenza using symptoms will make it difficult to ensure that antivirals are deployed appropriately. (The communication challenges are formidable, but must be addressed in advance, as noted in Chapter 3.) Some data exist for the MedKit proj- ect, but one cannot easily extrapolate from the ability to keep a blister pack of ciprofloxacin in reserve for a declared anthrax attack to the more complex requirements of antiviral use. Concerns have been raised about the risk of increased resistance from inappropriate use, but there are few data available to address this or other concerns. If pre-positioning is to be considered, the committee believes that careful exploration of these issues through extensive feasibility studies will be needed. Legal Barriers Related to Dispensing Sites The use of some sites for dispensing may present legal impediments. During declared states of emergency, governments may determine that public properties need to be used as dispensing sites, but interjurisdic- tional or other legal disagreements may interfere. For example, while the federal government may lawfully seek to use state or local property to facilitate the dispensation of antivirals, legal disputes about the selection of specific sites or other grounds can result in delays. Legal issues may include assessing responsibility for maintaining the premises, determin- ing compensation for its use, providing adequate security to ensure public safety, and resolving liability concerns. On the other hand, state and local governments may not typically use federal properties for dispensing sites without federal approval, even during emergencies. Though the Pan- demic and All-Hazards Preparedness Act (PAHPA) (PL 109-417, § 101 et seq. [2006]) and other federal laws have streamlined federal organization and responses to public health emergencies through DHHS, obtaining advance authorization to use federal sites for state or local dispensing programs will be helpful. Political or other legal factors may interfere with dispensing site selec- tion. Some potential dispensing sites may not be able to fully accom- modate persons with disabilities, or pose risks to the health or safety of others, thus raising concerns about equal protection, disability rights, liability, and other issues. For example, local elementary schools may be deemed by public health officials in some communities as suitable sites for distributions of antivirals. However, such plans to use these facilities may be mired in legal concerns over the short- and long-term conse- quences of using these facilities for these purposes (e.g., potential for contamination to delay the reopening as school facilities). As discussed above, emergency laws are designed to allow government authorities to quickly resolve these types of legal issues (Hodge, 2006b). In reality, 

WHO SHOULD GET ANTIVIRALS AND WHERE? 75 however, competing legal norms may still interfere with plans for the use of public properties as dispensing sites. On a practical level, some of the planning issues described above may be addressed by taking steps such as engaging in dialogue with community entities about expectations and roles and drafting memorandums of understanding. Emergency laws may also allow government to use private-sector facilities as dispensing sites (Gostin et al., 2002). In many instances, private-sector entities may collaborate with government agencies to avoid potential legal concerns while protecting the public’s health. However, resistance of private-sector entities to the use of their property as dis- pensing sites may lead to legal disputes. Government is positioned dur- ing formal declarations of emergency to commandeer the use of private property for public health purposes. Government acquisition of private property for this purpose would likely be viewed as a constitutional tak- ing, requiring compensation for the temporary use of the facilities and potential restoration after the emergency. Even if private-sector entities offer their properties for use as dispensing sites, they may seek assurances from government that the entity will not face claims of civil liability or increases in property insurance premiums for any harms that arise on the property for acts or ordinary negligence. Such protections are not automatically guaranteed under emergency laws. To the contrary, most emergency liability protections are designed to protect individuals, not corporate or other entities. As a result, although health care volunteers providing assistance at dispensing sites may be protected from liability claims, a private-sector entity that hosts the site or is responsible for its operation may not be entitled to similar liability protection (Gostin et al., 2002). This legal gap has been addressed legislatively in a few states, but is largely unresolved in most states and at the federal level (University of North Carolina Institute for Public Health, 2008). Staffing A number of jurisdictions have information available on the human and other resource requirements for setting up points of dispensing. In terms of the types of personnel that will be used, 36 states intend to use hospital staff for dispensing, 35 plan to use public health staff, 17 intend to use medical reserve corps, 15 will use other emergency response per- sonnel, 12 will use emergency systems for the advance registration of volunteer health professionals (ESAR-VHPs), 14 intend to use pharmacy personnel, and 15 plan to dispense via local health departments, primary care providers, and community health centers (ASTHO, 2007). It cannot be assumed that adequate personnel will be available to roll this out in the time needed. Plans ideally will consider the effect of absenteeism 

76 ANTIVIRALS FOR PANDEMIC INFLUENZA due to pandemic-related morbidity in personnel or family members, the dual roles of many responders, and the human resources available. The actual attack rates and epidemiologic features of the pandemic will help determine staffing decisions. One model that is used by many local juris- dictions to plan the staffing and other aspects of their point-of-dispensing is the Agency for Healthcare Research and Quality–supported Weill/ Cornell Bioterrorism and Epidemic Response Model, which was designed for mass dispensing planning by hospital and health system officials.  Potential Legal Issues Related to Staffing Human resources are essential to the operation of a POD, but the participation of various persons may require the resolution of potential, significant, legal issues in advance. Government and private-sector enti- ties responsible for their operation may seek to staff PODs with suitable, trained employees to provide essential services. The use of employees to staff nontraditional operations during public health emergencies impli- cates an array of legal concerns for the employees and their employers, as discussed below. As prior emergencies have demonstrated, emergency planners must also be prepared for significant staff shortages during pandemic influenza. Some employees will not be comfortable working in direct contact with the public because of the deadly threat of influ- enza, others will simply not report to work, and regrettably, some will be stricken with the condition they are being asked to help address. Meeting surge capacity of persons seeking antivirals through PODs will likely necessitate the deployment and use of volunteers, specifically health practitioners who are pre-vetted, trained, and organized to pro- vide essential public health or health care services (unlike spontaneous volunteers who may simply show up at emergency sites) (Hodge et al., 2005). To discourage spontaneity and promote organization of volun- teer health practitioners during emergencies, Congress (PL 107-188, 42 U.S.C.A. § 247d-7b [United States code annotated, 2005]) directed DHHS to fund states and territories to develop interoperable emergency systems for the advance registration of volunteer health professionals (HRSA, 2005). Virtually all states and territories have created (or begun to cre-   The model may be found at http://www.ahrq.gov/research/biomodel3/toc.asp.   Public health emergencies, such as pandemic influenza, have consistently featured sup- port from volunteer health professionals (VHPs) (e.g., physicians, nurses, public health workers, lab technicians, emergency medical responders, psychologists). Emergency re- sponse planners count on VHPs to fill surge capacity and provide needed medical expertise and related support functions. Some volunteers are organized, trained, and directed to respond through governmental programs (e.g., Disaster Medical Assistance Teams [DMATs], Medical Reserve Corps [MRC]) and private-sector efforts (e.g., American Red Cross). Others 

WHO SHOULD GET ANTIVIRALS AND WHERE? 77 ate) these and other registration systems (e.g., Medical Reserve Corps at the local level) (Hoard and Tosatto, 2005) to organize and register skilled health practitioners who are willing to volunteer their services during emergencies. While the need for volunteers during emergencies is essential to staff PODs or other public health or health care sites, volunteers, employees, and the entities that host or support them face an array of critical legal issues, including the following: Liability  When and under what circumstances may persons staffing PODs be personally liable for their actions during declared states of emer- gency? This question is critical to employees and volunteers alike, espe- cially during emergencies when standards of care in phases of medical triage may rapidly change. However, their exposure to personal liability is highly variable. For example, some actors, such as governmental workers, may be largely immune from liability claims for acts of negligence during emergencies (HRSA, 2006). Many volunteers (who do not fit the defini- tion of uncompensated volunteers under the federal Volunteer Protection Act [PL 105-119 (1997); 42 U.S.C.S. § 14501 et seq. (2004)] or similar state laws) may be potentially liable. Federal legislative proposals following Hurricane Katrina to better protect volunteers from liability have been introduced, but not passed (Hodge et al., 2006a). Some states’ existing emergency laws may provide some liability protections for volunteers (HRSA, 2006). Other states have recently passed or are currently consid- ering passage of a 2007 model law, the Uniform Emergency Volunteer Health Practitioners Act (UEVHPA) (National Conference of Commission- simply show up at the site of a disaster or nearby health care facilities. These “spontaneous volunteers” are ready to help, but lack organization, identification, credentials, and, ulti- mately, utility. Their presence can actually impede effective emergency responses. Prior experiences concerning complications in the deployment and use of VHPs led Congress to authorize DHHS to fund and assist states and territories to develop emergency systems for the advance registration of VHPs (ESAR-VHPs). These systems are designed to recruit and register prospective VHPs within each jurisdiction to help ensure a ready supply of trained, vetted volunteers during actual emergencies. As currently organized at the state and territorial levels, ESAR-VHP systems typically include verifiable, current information regarding a volunteer’s identity, licensing, credentialing, accreditation, and privileging in hospitals or other health care facilities that might need volunteers. With the passage of the Pandemic and All-Hazards Preparedness Act in December 2006, DHHS is authorized to link ESAR-VHP systems and comparable volunteer registries organized via the MRC at the local level into a single, national verification system to better organize volunteers for fed- eral emergency response efforts. The establishment of interoperable, state- and local-based registration systems will help federal, state, and local emergency response coordinators and others to quickly identify and better utilize VHPs during public health emergencies. However, a series of legal and regulatory questions impact their use and participation, most notably liability concerns. 

78 ANTIVIRALS FOR PANDEMIC INFLUENZA ers on Uniform State Laws, 2006), to provide greater liability protection for volunteers (Hodge et al., 2007). It should be noted that no laws protect persons staffing PODs from liability for their wanton, willful, or criminal acts (Hodge et al., 2006). Workers’ compensation Beyond issues of personal liability for actions causing harm to others, when may staffers be responsible for the personal injuries that they may incur during emergencies? In non-emergencies, employees who are injured or killed at work are covered through work- ers’ compensation programs, which provide compensation regardless of fault. These same benefits should continue during emergencies for employees, but not necessarily for volunteers. By definition, volunteers are not employees, and thus may not be entitled to typical workers’ compensation benefits while serving as volunteers (HRSA, 2006). Some employers may deem volunteers as covered through their workers’ com-   To address specific legal concerns underlying the deployment and use of volunteer health practitioners during declared states of emergency, the National Conference of Commission- ers on Uniform State Laws (NCCUSL) has approved the Uniform Emergency Volunteer Health Practitioners Act (UEVHPA) as of August 2007. Among its key provisions, the Act (1) establishes a system for the use of volunteer health practitioners that is capable of func- tioning autonomously even when routine methods of communication are disrupted; (2) defines “volunteers” to include compensated and uncompensated individuals; (3) requires pre-deployment registration in a recognized system to facilitate subsequent deployment and streamlining of volunteers to a disaster site; (4) provides reasonable safeguards to assure that volunteer health practitioners are appropriately licensed and regulated to protect the public’s health; and (5) allows states to regulate, direct, and restrict the scope and extent of services provided by volunteer health practitioners to promote disaster recovery opera- tions. The UEVHPA also provides immunity against civil claims for negligence or other acts to volunteers enrolled in a registration system who provide services through a local host agency in cooperation with local emergency management requirements and adhere to scope of practice limitations imposed by their licensing state and host state. The Act offers two leg- islative options for state legislatures to determine the level of liability protection to provide volunteer health practitioners. Alternative A provides strong, comprehensive liability pro- tections for the negligent acts of volunteer health practitioners during emergencies. Under Alternative A “a volunteer health practitioner who provides health or veterinary services pursuant to this [act] is not liable for damages for an act or omission of the practitioner in providing those services.” Volunteer health practitioners are not protected against willful misconduct, or wanton, grossly negligent, reckless, or criminal conduct, intentional torts, breach of contract, or an act or omission relating to the operation of a motor vehicle, vessel, aircraft, or other vehicle. Alternative A also provides some liability protections for entities that host volunteer health practitioners, such as hospitals, clinics, or disaster response agen- cies. The liability protections of Alternative B, on the other hand, are more comparable to existing liability protections found in the federal Volunteer Protection Act, which provides liability protections to largely uncompensated volunteer health practitioners. As of January 2008, four states (Colorado, Kentucky, New Mexico, and Tennessee) have enacted versions of the UEVHPA. Additional legislative enactment efforts are underway in multiple, additional jurisdictions in 2008. 

WHO SHOULD GET ANTIVIRALS AND WHERE? 79 pensation plans, but this is atypical. As a result, unless legal protections are provided, volunteers may provide services in risky environments dur- ing emergencies without any guarantee of protections from harms except from their own personal health insurance (which may not provide simi- lar coverage as workers’ compensation plans). In response, some states’ emergency laws extend workers’ compensation protections to volunteers (largely those volunteers providing direct services to government). The Uniform Emergency Volunteer Health Practitioners Act, which is under consideration in multiple jurisdictions, seeks to provide workers’ com- pensation benefits to registered volunteers as if they were state employees (Carpenter et al., 2008). Insurance coverage limits Related to the two major issues discussed above are the limits of health, medical malpractice, workers’ compensa- tion, disability, or life insurance coverage during emergencies. Staffers, whether employees or volunteers, may have to deal with varying limits of coverage under these and potentially other types of insurance during declared states of emergency. For example, some life insurance policies may not cover individuals for deaths resulting from pandemic influenza or other public health emergencies. Medical malpractice coverage for health care practitioners may contain express limits as to liability expo- sures during emergencies. Since staffers may make decisions based on their perceptions of existing insurance protections, any legal change that tends to downgrade the scope or benefits of this coverage may be inimical to the staffers’ participation. Unauthorized use of personnel Staffing responsibilities for various work- ers or volunteers at PODs may involve legal issues inherent in the delivery of medical services or products to individuals and populations. During non-emergencies, non-licensed personnel are not authorized to distrib- ute prescription drugs (like antivirals) to individuals in the interests of protecting public health and safety. Staffing shortages or other exigencies during a pandemic influenza may nevertheless require the use of unli- censed personnel to assure ready distribution of antivirals to persons in need. The potential illegality of this practice may be waived under federal or state emergency laws (HRSA, 2006). Federal authorities may even con- sider re-characterizing antivirals as non-prescription medications during declared emergencies. Absent direct waivers or federal interventions, however, non-licensed personnel and the entities organizing their service may be concerned about the legality of their actions and refuse to allow non-licensed personnel to perform certain actions. Interstate volunteer health care practitioners (assuming their availability during pandemic influenza) face a different, potential legal restraint. Since they are not 

80 ANTIVIRALS FOR PANDEMIC INFLUENZA licensed in the jurisdictions in which they seek to volunteer, any medical service they may provide may be considered unlawful (Hodge, 2006a). To remedy this issue, many existing state emergency laws and the Uni- form Emergency Volunteer Health Practitioners Act allow states to view out-of-state volunteer health practitioners as licensed in the jurisdiction for which they provide services during the duration of the emergency (Hodge et al., 2007). Site Set-Up, Flow, and Infection Control States and local jurisdictions have already done a great deal of work developing and exercising (sometimes in live events) point-of-dispensing set-up and flow. If jurisdictions select a variety of dispensing sites for post-exposure prophylaxis, they may benefit from including in their plan- ning early communication directing people to the appropriate site (e.g., individuals without insurance, employment, or with disability, or in cases where there are multiple sites of dispensing in one’s neighborhood). Effec- tive management of flow through a dispensing site and crowd control have been studied, planned for, and exercised by most state and local jurisdictions, as part of public health preparedness exercises and drills. Separating the well from the ill may be an objective of pandemic influ- enza response, or if social distancing is one of the non-pharmaceutical interventions a jurisdiction or the nation as a whole undertakes, but the benefits of doing so are unclear, given uncertainty about the rate of transmission and the limited evidence about the efficacy of various non- pharmaceutical interventions (Aledort et al., 2007). Instituting other infec- tion control measures may be useful, for example, encouraging and sup- porting use of cough etiquette and hand sanitizers at dispensing sites. INFORMATION SYSTEMS FOR MONITORING DRUG USE AND SAFETY The committee is aware of the long-standing efforts in the govern- mental public health infrastructure to develop comprehensive informa- tion systems that are compatible with one another and can communicate and exchange data. Preparing for an influenza pandemic presents addi- tional needs for information systems or functionalities that can be used to track antivirals and other countermeasures, such as vaccine. Most states are required through their CDC grants for public health emergency preparedness to develop compatible public health information networks (PHINs). The 2002 CDC guidance noted that CDC and its state partners had the shared goal of adopting and implementing “standards- based, integrated, and interoperable information technology (IT) systems” 

WHO SHOULD GET ANTIVIRALS AND WHERE? 81 to support public health activities. Any information system that would be used in an influenza pandemic would ideally be integrated with or the same as a system that supports other preparedness activities. The 2008 Pandemic Influenza Funding Announcement for Competitive Proposals (Activi- ties) (CDC, 2008) asks grantees to develop demonstration projects that explore ways to integrate existing state-based immunization information systems (IISs) with National Countermeasure and Response Administra- tion (CRA) systems to track doses of pandemic influenza vaccine, facilitate electronic laboratory data exchange supporting pandemic influenza sur- veillance, develop statewide PHIN-compliant electronic mortality report- ing systems and explore ways to distribute and dispense drugs to isolated or quarantined persons in a pandemic influenza event. Also the language of the guidance does not specifically suggest use of immunization regis- tries or similar systems to track or gather any other information related to antiviral drugs. The committee has learned that some localities intend to use or are considering the use of immunization registries to track both influenza vaccine and antiviral use (Biedrzycki, 2007; Zucker, 2007). Also, some states intend to use immunization registries to track use of pandemic influenza vaccine (IOM Meeting Two Transcript [Williamson, 2008]). For example, Michigan has enacted legislative changes to expand the use of its immunization registry to include antivirals (American Immunization Registry Association, 2006). For a variety of reasons having to do with efficiency (cost, staff time, training needs, etc.), it is reasonable to suggest that information systems for antivirals be the same as those used for influenza vaccines. It is unclear to what extent and what proportion of state immunization registries have the capacity and functionality to be used for tracking antiviral use. Immunization registries may not be easily adapted to include antivirals. Such registries are used largely in pediatric clinical settings, so few or none of the sites that will be used for antiviral dispensing have a link to an immunization registry, and potentially not even to a clinical setting through with data about dispensing could be linked to subsequent clinical information or health care databases. (The committee is aware that there are jurisdictions that are exploring the addition of adult immunizations to registries, perhaps beginning with health care worker immunizations.) Further, some immunization registries are still paper-based. The volume of data entered in a pandemic mass dispensing setting would be likely to place unprecedented demands on most existing public health informa- tion systems, such as registries. Finally, some jurisdictions have extremely limited information technology resources that do not meet existing public health practice needs, let alone those of an emergent pandemic response. Use of immunization registries simply may be unfeasible with existing personnel, technology, and systems. Despite all this, systems for tracking 

82 ANTIVIRALS FOR PANDEMIC INFLUENZA who gets antivirals will be needed, especially in the context of a severe pandemic and with limited supplies, although it is unclear how quickly a system could be ready in cases where a usable system is not already available. Even where a usable information system is in place, a backup system that is technology-independent would be helpful to include in planning exercises and, if possible, in real-life use.10 (Preventing “gam- ing” the system—i.e., fraudulent attempts to secure antivirals—would not be an objective of a tracking system, but rather the objective would be to prevent the inadvertent duplication of dispensing.) Further, expansion of existing systems to address pandemic response functions may constitute the best use of resources in some cases. Ideally, this work will be completed before the pandemic begins. Recommendation 4-4: The committee recommends that the Depart- ment of Health and Human Services support and fund public health agencies to develop or expand information systems for tracking dispensed antivirals. The development or expansion of these sys- tems should make use of existing information resources or systems, consider information technology needs for other dimensions of pandemic influenza response, comply with the Centers for Dis- ease Control and Prevention standards, and be interoperable and robust. Any proposed system would need to be tested. One context for address- ing this important area of planning may be the Cities Readiness Initiative (a CDC pilot project to enhance the capacity of cities to deliver medica- tions and medical supplies in a public health or similar emergency). 11 For reporting adverse events related to antiviral use, the proposed DHHS draft guidance states that the Food and Drug Administration Adverse Event Reporting System (AERS)/Medwatch should be used. This system is passive and not ideally suited to rapidly capture, inter- pret, and convey information needed to evaluate a course of action. It is also possible that providers will be unable to recognize and/or report adverse events in an environment of extreme surge and human resource shortages. Also, the public may not be concerned enough in the face of a life-threatening disease outbreak to report medication-related adverse events. AERS does not address the need of state and local jurisdictions to monitor and respond to adverse events, so public health agencies and their partners may need some additional measures to prepare and plan 10  This would be needed in the event of power failure and other critical infrastructure failures. 11  Information about the initiative is available at http://www.bt.cdc.gov/cri/. 

WHO SHOULD GET ANTIVIRALS AND WHERE? 83 to respond to safety signals, whether real or perceived. It also is unlikely that existing systems, for jurisdictions that have something reasonably capable in place, are sufficiently robust to be repositories of dispensing records that link data to a visit or diagnosis around an adverse event. Some health care organizations may have such a database capability, but if antiviral distribution occurs outside the medical home (i.e., people may not receive antivirals and other care from their primary care provider), health care organization and health plan data would have little to no data related to antiviral distribution. Finally, unlike the need for complete antiviral tracking data, it is not necessary or realistic to attempt to gather comprehensive information about each antiviral drug-related adverse event. The use of sentinel sites, such as those used to conduct epidemiologic surveillance of seasonal influenza could be one solution to get some statistically meaningful data. CDC (with FDA input) could solicit specific types of adverse event report- ing, or initiate a type of active surveillance based on individual reports and hypotheses about emerging safety signals. Special attention should be given to adverse events in certain vulnerable populations such as pregnant women and young children, for whom safety data is limited or non-existent. Also, planning could focus on systems to capture only seri- ous and unexpected adverse events. Recommendation 4-5: The committee recommends that the Depart- ment of Health and Human Services consider options in addition to the Food and Drug Administration Adverse Event Reporting System to capture adverse events resulting from use of antiviral drugs to ensure active and timely reporting. One option is a network of sentinel sites that can collect data that are representative of antiviral use nationally. (Data gathered and compiled by such a system would be provided to the advisory body described in Chapter 3 for analysis and determination of whether changes are needed in national recom- mendations on the use of antivirals.) Legal Issues Related to Information Systems What legal norms exist to protect the privacy of individuals receiving medical services or products during emergencies? In non-emergencies, a panoply of health information privacy laws at the federal, state, and local levels apply to the provision of health services. Highlighted by the Health Insurance Portability and Accountability Act (HIPPA) Privacy Rule (45 C.F.R. § 160.100 et seq. [2004]), these laws seek to assure patients that their health data are entitled to reasonable protections against unwarranted acquisition, use, disclosure, and storage (Hodge, 2004). Adhering to legal 

84 ANTIVIRALS FOR PANDEMIC INFLUENZA privacy protections during emergencies may be viewed as impractical or impossible. In part, this led DHHS to briefly suspend adherence to the HIPAA Privacy Rule for patients affected by Hurricane Katrina in 2005 (Hodge et al., 2007). Still, personnel at dispensing sites may need to be sensitive to the privacy expectations of individuals seeking antivirals dur- ing pandemic influenza. Community perceptions that privacy norms are largely being ignored or intentionally breached may lead some to avoid accessing medications through PODs. Privacy and security implications also underlie public health surveil- lance systems. These issues depend on whether surveillance systems tracking the use of antivirals are distinct systems or tied into existing fed- eral, state, or local databases. Federal and state laws regulate public health surveillance systems to assure that the acquisition and use of identifiable health information are purposeful and the data are held confidential. Dis- closures outside public health or other governmental authorities are often restricted, absent specific written authorization of individuals who are the subjects of the data. Security protections may require these databases to protect against unwarranted access by non-approved users. If the data to be collected through this type of public health surveillance are truly non-identifiable, privacy or security issues are not relevant, but of course non-identifiable data are of limited utility to public health practitioners seeking to intervene to protect communal health. EXERCISES AND DRILLS Reviews from drills and exercises find few that are specific to antiviral distribution; the bulk of the drills and exercises that have been conducted have been in the context of the Cities Readiness Initiative and dispensing materiel from the SNS (e.g., ASTHO, 2006). Most of these activities do not involve the complex assumptions and considerations required for treatment and prophylaxis. However, they do provide an experience base from which to draw. Unfortunately, many of the preparedness drills and exercises that have been conducted throughout the country do not have objective, numerical measures or outcomes associated with them. RAND Corporation’s work on countermeasure distribution through the SNS has focused on the development of drills and exercises that break the com- plex tasks down into discreet components, and identify clear metrics for each component. Once a health department can successfully complete the key component tasks, it may conduct a more comprehensive functional exercise. It is extremely difficult to assess and improve the capability to distribute countermeasures without a robust and quantitatively focused exercise program. A primary resource is the Department of Homeland 

WHO SHOULD GET ANTIVIRALS AND WHERE? 85 Security Homeland Security Exercise and Evaluation Program toolkit, which includes an Exercise Evaluation Guide that can be customized. Some jurisdictions have used actual disease outbreaks or other emer- gencies as opportunities to practice their pandemic influenza or public health emergency response plans. A number of local public health agencies have used seasonal influenza immunization clinics to test their prepared- ness for mass dispensing in the event of an influenza pandemic. During the mumps epidemic in 2006, Iowa dispensed measles–mumps–rubella vaccine to all counties and provided some special vaccination clinics to college-age adults, utilizing the state’s public health emergency prepared- ness plans (ASTHO, 2006; McCormick, 2006). There are many other states, counties, and cities that have had similar experiences. The CDC public health emergency preparedness and Strategic National Stockpile guidance to grantees is largely oriented toward table- top exercises and full-scale exercises (CDC, 2006b, 2007a, 2008). It is unclear whether jurisdictions are considered to have met funding require- ments when they implement plans to respond to actual events (such as an outbreak of measles or meningitis, or seasonal influenza immunization clinics), develop after action reports, and implement changes to address identified shortcomings. Some real-life activities may present more realis- tic challenges and constitute better tests of how well pandemic influenza plans will function in a pandemic. Also, well-planned efforts to measure performance in real-life events may contribute to an evidence base that demonstrates what works. Recommendation 4-6: The committee recommends that federal pandemic influenza grant guidance explicitly state that jurisdic- tions receiving federal funding may fulfill the exercise requirement through the implementation of response to actual biologic emer- gency situations or similar events, if the appropriate benchmarks are used, performance is evaluated, and necessary corrective action is taken. CLOSING OBSERVATIONS Implementation of an antivirals program for an influenza pandemic, whether it occurs in the near or distant future will need to take into account multiple factors, many of which are evolving or cannot be known in advance, including supply of antivirals, shelf-life, resistance, and vaccine technology. The epidemiologic characteristics of the pandemic strain— e.g., age of greatest impact and/or mortality, mode of spread, rapidity of development of resistance—constitute large unknowns that will affect when, how, and which individuals are treated with antivirals. Regardless 

86 ANTIVIRALS FOR PANDEMIC INFLUENZA of the final shape of the pandemic, it is clear to the committee that many of these issues need to be addressed in advance and provide a foundation for later decision making. Several overarching goals need to be kept in the forefront: developing in advance an ethical framework, communication and education of the public with clear and consistent messages, the need to reconcile actual supply and antiviral program goals, and flexibility on the one hand to react to the changes in the course of the pandemic and on the other hand to address the diverse needs of localities.